THE KAYAK

A STUDY IN TYPOLOGY AND CULTURE HISTORY

1
By Jarmo Kankaanp
INTRODUCTION
Although ethnologists and archaeologists studying the Eskimo deal with largely similar
manifestations of material culture, interdisciplinary studies combining materials from both fields
are surprisingly rare among both groups and are usually limited to a fairly low level of
sophistication. An ethnologist studying a certain implement type may trace its linear formal
evolution through the archaeological record, and an archaeologist describing his finds may point
to similarities with historical types, but neither will usually draw any broad conclusions from
these data. An ethnologist comparing different variants of the same basic implement will rarely
try to explain the observed structural differences by referring to relationships between
archaeological cultures, nor is the archaeologist studying the movements of prehistoric cultures
particularly interested in the areal grouping of historical implement types; both generally prefer
to stick to their own respective disciplines and the material and methods they offer.
This study is an attempt to combine ethnographic and archaeological data pertaining to a
particular implement type and a number of other, closely related cultural features, in order to
connect the evolution and spread of regional variants with recognied prehistoric cultures and
events, and to throw light on certain typological and historical !uestions that have proven
difficult to answer from the material and analytical premises of one discipline alone.
The study begins with a description and structural analysis of historical kayak types used by the
Eskimo, Aleut, "hukchi, and #oryak. $ased on the analysis, the types are classified into three
main groups according to certain basic structural !ualities. The typological and possible
historical relationships between these groups are first assessed purely on the basis of structural
criteria, after which each group is traced as far as possible in the archaeological record. The
structural patterns characteriing the main groups are connected with recognied archaeological
cultures or traditions, and the structural and archaeological data are combined to present a
hypothetical outline of the evolution of the historic types.
1
This paper was originally presented as a %.A. thesis at the &niversity of 'elsinki in 1()). *t
has been published in +innish as ,#a-akki. Typologinen -a etnohistoriallinen tutkielma. Helsingin
yliopiston kansatieteen laitoksen tutkimuksia 1../ 'elsinki0 1liopistopaino 1()(. The
archaeological !uestions touched upon in this study have since been discussed more thoroughly in
the author2s 3h.4. thesis ,Thule subsistence/ 5$rown &niversity, 1((6, available through &.%.*.7.
DISTRIBUTION AND STRUCTURAL FEATURES
Di!ri"#!ion
The geographical distribution of native kayak types extends from the 8ulf of 3enshina 59iberia7
in the west to Ammassalik 5East 8reenland7 in the east. The most southerly areas where the
kayak has been used during historical times are the southern :abrador coast and the Aleutian
*slands; the northern border in the ;ew <orld coincides with that of human habitation. *n many
areas, the kayak is only present on a relatively narrow strip ad-acent to the coast, but especially
in ;orth Alaska and central ;orth "anada it has also been used extensively on inland waterways.
Apart from the Eskimo, the kayak has also been used by the Aleuts and the Tanaina *ndians of
southern Alaska, and by the "hukchi and #oryak of eastern 9iberia. *n "anada and ;orth
Alaska, however, the kayak is exclusively an Eskimo implement and has not been adopted by the
neighboring *ndian groups. The kayak is still used as a hunting craft in 8reenland, :abrador, and
southwestern Alaska, but the decline of aboriginal hunting methods has led to its total disap=
pearance from many other areas.
To a European, the term >Eskimo kayak> usually brings to mind a 8reenland=style low=decked,
sharp=ended, single=seat craft propelled with a double=bladed paddle and used mainly for sea
mammal hunting. %any kayak types fit this description in a general way, being dedicated sea
boats used for hunting seals, walrus, and small whales as well as for fishing and fowling.
'owever, in the central "anadian arctic and parts of ;orth Alaska the kayak has been used above
all for hunting swimming caribou at crossing places on inland waterways, and the kayaks of
these areas have conse!uently developed into craft !uite different from the pelagic eastern types.
;ot even all sea kayaks resemble the 8reenlandic models. Especially the <est and 9outh
Alaskan types with their vertical stern, ridged deck, and curious bow structures are very different
from their eastern counterparts. 4ouble and triple=hatch kayaks have also been common in Alas=
ka, as has the single bladed paddle; the #oryak of 9iberia have even used short single paddles in
pairs, one for each hand.
S!r#$!#ra% Par!
The hull of a kayak consists of two separate assemblies, the wooden frame and the skin covering.
Aside from materials and methods of manufacture, these assemblies also differ from each other
with respect to procuring group, manufacturing timetable, use life, and method of storage.
& T'( )ram(
The frame forms the actual stress=bearing structure of the kayak and gives the hull its typical
shape. The frame of a kayak is self=supporting, i.e., it is built as a unit that retains its integrity
and shape even when the covering is detached. *n this sense, it differs from the frame of the birch
bark canoe where the individual slats and ribs are kept in place only by being compressed
between the gunwales and the bark cover 5e.g., Adney ? "hapelle 1(6@01@=1., 1(A; Taylor
1()B7.
C
9ince wood 5as opposed to skin7 does not re!uire any special treatment and associated critical
timing prior to working, most parts of the frame may be manufactured beforehand in free order
and without a strict timetable 5Arima 1()D016; Eimmerly 1(D(0D17. $uilding a kayak frame may
take several weeks, since it is often not possible to complete the work at a sitting due to other
duties.
A hunter usually builds the frame for his kayak himself, either alone or with the help of one or
several male companions. A father may build a kayak frame for his adolescent son, but otherwise
there are few examples of having the frame F or any other hunting implement for that matter F
manufactured to order 5one notable exception being kayaks built for museum collections, e.g.,
Arima 1(D. or Eimmerly 1(D(7. This is at least in part due to the fact that all the measurements
used are anthropometric and should ideally be those of the user himself. Ergonomy is in fact
considered essential, since controlling the kayak in a heavy sea is difficult if the various points of
support do not fit the anatomy of the kayaker 5"urtis 1(AB01A; 3etersen 1()6b0D, 1.7.
9ince most groups using kayaks live far from the tree line, all larger structural members of the
frame are generally made from driftwood. The traditional method of manufacture consists of
splitting a driftwood log with small wedges into beams of suitable sie and trimming these down
with a small ade and crooked knife. *n some areas even driftwood is so scarce that large parts
such as the gunwales have to be built up from several pieces 5e.g., Arima 1(6@0CC1=CCC;
1()D0@C7. This method is still popular because store=bought !uality lumber is expensive and hard
to get; a beam split along the grain is also more durable than a sawn plank 5e.g., +reeman
1(6@0DC; Eimmerly 1(D(0)=1B, C@=C); 3etersen 1()6a01(; Arima 1()D01.7.
$uilding a kayak frame from scratch in the traditional manner is a time=consuming enterprise,
and the owner thus has a good reason for trying to keep his frame in usable condition as long as
possible. &se life may be extended through careful storage on a special wood or stone rack
during the winter and annual renewal of the more delicate parts. <ith good care, the frame has a
comparatively long use life; on the $elcher *slands of eastern 'udson $ay the thicker members
such as the gunwales are estimated to last about 1B years and the thinner ones such as the ribs
and stringers about five to six years 5+reeman 1(6@06)7.
= T'( $o*(r
The cover has traditionally been fashioned from the skins of medium=sie sea mammals such as
seals or sea lions; some inland groups have also used caribou skin. The use life of the cover is
much shorter than that of the frame since even the best sealskin rots and wears out faster than
wood. <ith care, a skin cover may last three years, but under normal circumstances a new cover
is fitted every year, usually in the spring before the beginning of the open water season 5+reeman
1(6@06.; Eimmerly 1(D)0AC; 3etersen 1()6a0A17.
9ewing the cover is often a communal enterprise since the cover must be finished before the
skins 5which have been soaked for stretching7 dry out. "overing the kayak, like all chores related
to sewing and dressing skins, is traditionally performed by the women. The men lend a hand only
towards the end, when the cover is stretched over the frame 5'awkes 1(160DC; $irket=9mith
1(C(G*01)6; Hasmussen 1(A101DC; +reeman 1(6@0D@; Arima 1(6@0CAD, CA(=C@B; 8uemple
1(6D01.C=1.D; 3etersen 1()6a0A17.
A
B#i%+in, a Kayak
The simplest type of kayak frame consists of a pair of gunwales and a series of deck beams, ribs,
and deck and bottom stringers. Assembly of the frame starts with morticing the deck beams
between the gunwales. The middle beams are put in first, and the ends of the gunwales are
gradually pulled together as more beams are added, until they are finally -oined when all beams
are in place. The deck stringers may now be fastened and the frame turned over. <ith the frame
upside down, the ribs are morticed into the lower edge of the gunwales; the -oin is sometimes
locked with a dowel 5cf. Arima 1()D01(7. After the bottom stringers have been fastened to the
ribs with lashings, the frame is ready for covering 5cf., e.g., %athiassen 1(C)a0(@; +reeman
1(6@0DA=D@; 'oltved 1(6D0D); Arima 1(D.011@=1C17.
$esides the parts mentioned above, some kayak types have special -oint blocks or form giving
members in the bow and stern. At which stage of assembly these are added depends on their
method of fastening and their structural function, and conse!uently varies widely between types.
*n some kayaks the wooden cockpit rim is also fastened permanently to the frame while in others
it is only fastened to the cover after this has been completed.
"overing the frame begins with sewing the skins into a long mat with pockets at the extremes.
After these pockets have been fitted over the bow and stern, a thin line is run igag fashion back
and forth over the deck and through small cuts made only through the underside of the skin close
to F but not !uite at F the edges. This line is gradually tightened, until the edges of the cover are
pulled together 5or as close together as they will go7 over the deck. The edges are then sewn
together and any remaining holes are patched with separate pieces, leaving only the cockpit hole
in the middle of the deck. The stretching line disappears under the seams but is not removed.
+inally, the cockpit hole is trimmed and the edges fastened to the wooden hoop=like rim 5cf. $oas
1(B1=BD011; $irket=9mith 1(C@0C6@; %athiassen 1(C)a0(.; 'oltved 1(6D0D)=)1; Arima
1(D.01CD=1C); 3etersen 1()6a0AC=AA; 1()6b0.(=6B, 6.=667.
According to Iohn 'eath 51(D)0CA7 one of the differences between the <est and 9outh Alaskan
ridged=decked kayaks and the northern and eastern flat=decked models is that on the former, the
skins are not sewn together prior to covering but are -oined on the frame itself. Iudging by
published accounts of covering ridged=decked kayaks, however, the operation seems to
commence !uite similarly to that performed with flat=decked types 5e.g., "urtis 1(AB01.; $irket=
9mith 1(.A0@D; :aughlin 1()B0AD; Hobert=:amblin 1()B01B7. A double and a single hatched
ridged=decked Aleut kayak 5belonging respectively to the Turku 'istorical %useum and the
;ational %useum of +inland7 which the present writer has had the chance to examine both have
the same type of igag stretching line under their deck seam as the flat=decked types, indicating
that at least the final stage of cover fitting was essentially similar in both areas.
@
PRE-IOUS RESEARCH IN KAYAK TYPOLOGY
The structural properties of various regional kayak types have been treated to some extent in the
>classical> ethnographies of the late nineteenth and early twentieth centuries 5e.g., $oas 1))) and
1(B1=BD, %urdoch 1)(C, Turner 1)(@, ;elson 1)((, Iochelson 1(B., 3orsild 1(1., 'awkes
1(16, $irket=9mith 1(C@ and 1(C(, %athiassen 1(C)a7, but these presentations are mainly
descriptive and comparisons with other types F when present F are fairly superficial. Typological
descriptions focus mainly on outward appearance; the presentation of the actual structural
features is usually !uite limited, and exact measured plans or drawings are almost nonexistent.
The first attempt to classify all historical kayak types into morphological groups appeared in
Iames 'ornell2s comprehensive study on watercraft from 1(@6. 'ornell listed six main type
groups0
*. "aribou Eskimo
**. $eaufort 9ea
***. $affin :and
*J. 8reenland
J. Alaska
J*. 9iberian #oryak
According to 'ornell, the "aribou Eskimo type is found only in the $arren 8rounds tundra area
west of 'udson $ay, the $eaufort 9ea type on the arctic coast from 3t. $arrow to the #eewatin
4istrict in northeastern "anada, the $affin :and type on $affin :and 5$affin *sland7 and also on
the :abrador coast and at 9mith 9ound 5northwestern 8reenland7, the 8reenland type elsewhere
in 8reenland, the Alaskan type in <est and 9outh Alaska, and the #oryak type in eastern 9iberia
on the northern coast of the 9ea of Kkhotsk and on the $ering 9ea coast between the 8ulf of
Anadyr and #amchatka. $esides these main types, 'ornell also mentions the river kayak of the
5Heindeer7 "hukchi and the sea kayak of the 9iberian Eskimo and coastal "hukchi; however, he
does not classify these types under any of his typological groups 5'ornell 1(@60166=1D@7.
'oward *. "hapelle2s study of arctic 5mainly American7 skin boats from 1(6@ remains the basic
source of kayak typology because it is the only work to date which presents detailed line
drawings with measurements of representatives of almost all known regional types. "hapelle did
not present an actual typological analysis or define typological groups, but he did note structural
similarities between the kayak types of certain areas 5e.g., 9outh and 9outhwest Alaska, ;orth
Alaska and central ;orth "anada, ;ortheast "anada and ;orthwest 8reenland7 5Adney ?
"hapelle 1(6@01(D, CB@7.
*n a short, unpublished paper on kayak typology from 1(DB, Iohn 'eath divided kayaks into
eight groups0
51.7 9iberian kayaks
.
5C.7 3acific kayaks
5A.7 $ering 9ea kayaks
5@.7 Arctic Alaskan kayaks
5..7 %ackenie 4elta kayaks
56.7 "aribou Eskimo kayaks
5D.7 East "anadian kayaks
5).7 8reenland kayaks
'eath noted that groups 5@.7 and 56.7
1
were related, and that groups 5C.7 and 5A.7 differed in many
ways from the other groups. Kf the 9iberian kayaks, 'eath only mentions the #oryak kayak
5'ornell2s type J*.7, which he considers to be genetically related to the 8reenland kayaks. Kf the
others, group 56.7 is clearly the same as 'ornell2s "aribou Eskimo type, group 5D.7 is 'ornell2s
$affin :and type, and number 5).7 is 'ornell2s 8reenland type. 'ornell2s Alaskan type is divided
by 'eath into two groups 5numbers C. and A.7, and the ;orth Alaskan kayaks 5@.7 and %ackenie
kayaks 5..7 are split off 'ornell2s $eaufort type, while the remaining "opper Eskimo and
;etsilik kayaks are left unclassified, although 'eath does mention them in connection with both
the ;orth Alaskan and the "aribou Eskimo kayaks.
According to 'eath, the ob-ective of his study was to present the ,genetical relationships/
between kayak types in order to help eskimologists unravel their common origin. 'e did not
himself analye the results from an evolutionary or cultural historical perspective, and was
content to present some unrelated comments on the areal distribution of certain structural
features.
The latest and most comprehensive general typology of kayaks has been presented by Eugene 1.
Arima in his monograph on the "aribou Eskimo kayak from 1(D.. Arima lists the following
groups and variants 5pp. 6D=)670
1. #oryak
C. Heindeer "hukchi
A. 9iberian Eskimo F "oastal "hukchi
@. $ering 9ea
= $ering 9trait
= ;orton 9ound
= ;univak *sland
= $ristol $ay
.. The Aleutians and 9outh Alaska
= Aleut
= #onyag
= "hugach
6. ;orth Alaska, "entral Arctic "anada
= ;orth Alaska
= "opper Eskimo
= ;etsilik
= *glulik
= "aribou Eskimo
D. %ackenie
). Eastern "anada
1
'eath seems to include also the other "entral "anadian caribou kayaks in group 56.7,
although this is not explicit from the text.
6
= Atlantic :abrador
= 'udson and 4avis 9traits
= eastern 'udson $ay
= +oxe $asin
= 3olar Eskimo 5;orthwest 8reenland7
(. 8reenland
= &pernavik 4istrict
= 4isko $ay
= southern central <est 8reenland
= 9outh 8reenland
= East 8reenland
*n this list, Arima brings the number of 9iberian groups up to three by adding the unconnected
types mentioned by 'ornell, and combines 'eath2s ;orth Alaskan group and "aribou Eskimo
group with the "opper Eskimo and ;etsilik kayaks to form one group; in other essentials the
division is similar to 'eath2s. &nfortunately, Arima2s descriptions of the various subgroups are
fairly superficial; in many cases the author merely lists the subgroups without giving any details
of their distinctive features. The 8reenlandic subgroups may be identified from the two
references Arima gives 5$irket=9mith 1(C@ and 3orsild 1(1.7; most of the ;orth Alaskan F
"entral "anadian and $ering 9ea subgroups are to be found in Adney ? "hapelle, though Arima
does not in this case always give references. 'owever, the $ristol $ay subgroup and all East
"anadian subgroups remain uncertain as Arima does not describe their appearance or give their
exact geographical boundaries, nor does he explain his grouping criteria or give any references.
Arima also mentions a few archaeological finds F mainly kayak models F from which he seeks to
draw lines of development to historical types. 'owever, since the focus of the study is the
"aribou Eskimo kayak, the archaeological section is mainly concerned with finding forbears for
this type; connections between other historical types and prehistoric finds are only discussed
very briefly.
After Arima2s monograph, only one study on general kayak typology has appeared. This is a
short article by 'eath dating from 1(D), in which the author develops his idea of dividing all
kayak types into two primary groups, the first comprising all the ridged=decked Aleut, 3acific
Eskimo, and $ering 9ea types and the second containing all other 5flat=decked7 types. 'eath
connects these two groups with the two main linguistic subgroups of the Eskimo language, 1upik
and *nupik, noting that their boundary coincides closely with that of the kayak types 5pp. CC=C@7.
The structural similarities between the #oryak kayak and the 8reenland types are also
considered, and the conclusion is drawn that these may derive from a common ,archetype/, the
grandfather of all kayaks 5pp. C1=CC, C@=C.7.
D
TYPES AND SUBTYPES
The following division is based primarily on Arima2s scheme, although the order is reversed and
is now from east to west. The reason for this change in orientation is that the gradually growing
complexity of the frame structure that one encounters when moving from eastern "anada to
9outh Alaska is more easily grasped if the presentation starts with the more simple forms. The
subgroups are not all similar to Arima2s subgroups because his criteria and areal distributions are
not always clear; especially as concerns 8reenland, the change is also partly due to the
availability of newer data.
$ecause the focus of this study is on evolution and history rather than on pure taxonomy, the
goal of the following presentation of types is not so much the clarification of minute differences
between subtypes as the discussion of the structural features that are typical and common to all
types within a group. +or this purpose, differences between subgroups are important mainly as
criteria for classifying the features themselves. $y studying the differences and similarities
between, on the one hand, subtypes of the same main group used in different environments and,
on the other hand, subtypes of different main groups used in a similar environment, we may
attempt to define which structural features are environmentally determined
1
and thus useless as
clues of spatial historical events. This analytical perspective could in a sense be called anti=
functionalistic, since the successful application of a diachronic, evolutionary historical approach
essentially depends on the ability to assess, which features specifically ar( no! products of their
environment but rather reflect historical events and processes.
T'( Gr((n%an+ Kayak
Typical structural features of the true 8reenlandic types 5+igs. 1a and Aa7 are the triangular form
blocks attached to the ends of the frame and the use of multiple, short deck stringers. There are
only three bottom stringers0 a keelson and two side stringers. All three are !uite sturdy 5c. CB x
C. mm > . mm7. The keelson is furnished at both ends with a flat, vertical pro-ection to which
the ends of the side stringers are fastened; the upper edge of this pro-ection is itself lashed to the
-oined ends of the gunwales and to the front edge to the triangular block which carries the lines
of the keelson and the gunwales to the end of the bow and stern.
The ribs were formerly made from fresh willow or -uniper; during the colonial period wooden
barrel hoops obtained from traders became popular 53etersen 1()6a0C@7. The form of the ribs
varies; in some kayaks the ribs are evenly curved, in others the middle portion is flattened. %ost
of the deck beams are straight, only those above the kayaker2s feet curve up slightly higher than
the gunwales. The most curved deck beam is the masik, which supports the fore end of the
cockpit coaming and also acts as a carrying handle. The fore deck stringers do not come back all
the way to the masik itself as they do in other kayak types, instead, they are fastened to another
deck beam directly in front of the masik.
The cockpit coaming itself is roundish = usually egg shaped = and only slightly elevated at the
front edge. The coaming is made from a thin wood slat by bending; during the sailing ship era
ready=made wooden mast rings were also often used 53etersen 1()6a01(7. The cover is still
preferably made from harp seal, bladdernose, or ringed seal skin, but the decline of the seal
1
>Environment> refers in this case to all non=cultural exterior factors, not -ust natural
surroundings.
)
population has brought about an increased use of sailcloth impregnated with oil=based paint
53etersen 1()6a0AB7.
There have been changes in the form and distribution of 8reenlandic kayaks even in the current
century, so typological and regional classification is somewhat problematic. The data offered by
sources of varying age do not necessarily coincide, and it is difficult to define a point in history
when each area was using the ,original/ local type. The number of discrete types defined by
various authorities on the basis of differing criteria is large, but the actual descriptions are often
so vague that it may be impossible to identify a kayak in an old photograph without knowing,
when and where the photograph was taken.
The first to classify 8reenlandic kayaks was ".<. 9chult=:orenten 51(B@0A1B=A1C7, who
postulated two structural groups0
a7 9outh 8reenland kayaks, identified by the use of morticed -oints, a straight and low general
shape, and a cover made from light 5scraped7 skin.
b7 "entral and ;orth 8reenland kayaks, identified by the use of lashed -oints, a high and
curved general shape, and a cover made from dark 5unscraped7 skin.
A few years later, %orten 3. 3orsild divided the 8reenlandic kayaks into six groups, found in the
following areas 5names of communities have been added70
1. East 8reenland 5Ammassalik7
C. <est 8reenland 6BL F 61L ; 5#ap +arvel F Arsuk7
A. > 61L F 6DL ; 5+redrikshMb F *tille!7
@. > 6DL F D1L ; 5'olsteinsborg F &ummanna!7
.. > D1L F DAL ; 5*llorsuit F &pernavik7
6. 9mith 9ound 5Thule 4istrict7

Kf these groups, C. and A. are identical with 9chult=:orenten2s 9outh 8reenland and "entral=
;orth 8reenland types, @. and .. are high=sterned types from northern <est 8reenland, 1. is an
extinct type with a high stern, and 6. is actually an East "anadian type introduced by immigrants
from $affin *sland. According to 3orsild, types 1., C., and A. are interrelated, as are also types @.
and .. 53orsild 1(1C061)=61(; 1(1.01C1=1CC7.
The next classification was introduced by #a- $irket=9mith, who considered those of 9chult=
:orenten and 3orsild too simple and also criticied their structural criteria. According to $irket=
9mith, the use of morticed -oints had been adopted from Europeans, the high stern was not
originally limited to northern <est 8reenland, and both light and dark covers were used in some
areas 5$irket=9mith 1(C@0C6)=CDB7. $irket=9mith2s own classificatory system 5see %ap **a7 ran
as follows 5op. cit.0CD17; note that both 4anish and 8reenlandic names follow the old
orthography0
1. J#%ian('aa" . Fr(+(rik'aa"/ 'ardly any sheer Nflex of gunwale lineO. :ong stems.
Almost only light skin. +lat NdeckO seams. . *ntroduced by the immigrated Nsic!O
Eskimos from +rederik J*2s #yst Nthe southern part of the east coastO0 similar type,
though with a nearly flat bottom and vertical sides.
C. Go+!'aa"/ 4istinct sheer. 9horter stems. 4ark skin with flat seams.
(
A. S#kk(r!opp(n . Ho%!(in"or,/ 9heer. Jery short stems, the stern formerly curving
upwards, as is still done in #angPmiut. 4ark skin. +lat seams.
@. E,(+(min+( . Diko Bay/ 9heer; in out=of=the=way places still of the shape of a tall,
sharply built stem with a deck inclining strongly backwards. :ong stem, shorter stern
which until a short time ago was curved upwards. 4ark skin. Haised seams Nto keep
water from collecting at the seams and seeping insideO.
.. 0m1na2 . S3n+r( Up(rni*ik4 :ike the preceding type, but with a somewhat shorter
stem and stern. The latter still curved upwards.
6. Up(rni*ik Di!ri$!5 north of 9Qndre &pernivik0 ;o sheer. 9hort stem and !uite a short
stern, the former slightly, the latter very strongly curved. 4ark skin. Haised seams. The
two straps immediately in front of the manhole as a rule obli!ue.
$esides these <est 8reenlandic types, $irket=9mith also mentions the ,#ing +rederik J*2s #yst/
5#ing +rederick J* "oast7 type, the Ammassalik type, and the Thule 4istrict type, but gives no
description of the latter two.
There are several differences between the regional distributions of $irket=9mith2s and 3orsild2s
types. $irket=9mith divides 3orsild2s A. and .. groups in two, and his #ing +rederik J*2s #yst
type does not appear on 3orsild2s list although it may actually form a part of 3orsild2s group 1.
since 3orsild does not specify whether the degrees he gives apply only to the coast west of #ap
+arvel or whether they also refer to the southern part of the east coast. The borders between the
groups are also slightly different; for instance, $irket=9mith seemingly assigns the communities
of +rederikshMb and 'olsteinsborg to a more southerly and &ummanna! to a more northerly
group than 3orsild, although 5again7 this apparent discrepancy may actually be due to the
inexactness of the rounded=off degrees used by the latter.
The regional groups defined by Arima are almost exactly similar to those suggested by 3orsild.
<hy Arima nevertheless also mentions $irket=9mith as a reference is unclear, since he does not
choose to adopt $irket=9mith2s division of the A. and .. groups, nor does he mention the #ing
+rederik J*2s #yst group at all.
The latest typological classification of 8reenlandic kayaks has been presented by a 8reenlandic
authority, '.". 3etersen 51()6a0@)=@(7. According to 3etersen, 8reenland kayaks may be
divided into four groups distributed as follows 5%ap **b70
1. The flat type0 ;arssa!, IulianehMb, ;anortalik, East 8reenland.
C. The curved type0 +redrikshMb F 'olsteinsborg.
A. The avasisaartoq type0 fell out of use around the turn of the century, features still found
in the northwestern part of &ummanna! +iord and *llorsuit.
@. The ;orth 8reenland type0 #angaatsiak F &pernavik, lately also the Thule 4istrict.
The flat type 5+ig. .a7 is clearly e!uivalent to $irket=9mith2s type 1. According to 3etersen,
kayaks of the flat type are very long, narrow, and low=sided. The gunwales are almost vertical,
1B
and the sheer line is straight or only slightly negatively curved in the middle. The stem and stern
are very long, and since the keel line is straight and the bottom flat, the craft has a very shallow
draught.
The longest and narrowest kayaks of this group are found on the central east coast at
Ammassalik, where the flat type replaced the former curved type toward the end of the 1(th
century 5cf. 3etersen 1()6a0.C7. The old curved=stern model was still the dominant type when
8ustav 'olm2s expedition visited Ammassalik in 1))@, but had been completely replaced by the
southern type by 1)(@ 5Thalbiter 1(1C0A)@7. A flat type with vertical gunwales 5$irket=9mith2s
,#ing +rederik J*2s #yst/ type7 seems to have been formerly used on the southern east coast,
but this model was discontinued after the population of that area moved to the ;anortalik 4istrict
in the early 1(BB2s 53etersen 1()6a0@)7.
3etersen2s curved type 5+ig. .b7 appears to be similar to $irket=9mith2s type C and is also found
in part of the area which $irket=9mith defines for type A; the distribution is in fact the same as
that of 3orsild2s and Arima2s group A. *n the curved type, the middle part of the sheer line is
almost straight, but the gunwales curve up slightly toward the ends. The upper edges of the
gunwales also lean slightly outward, so there is no chine in the cover at the lower edge. The stem
and stern are shorter than in the flat type and rise at a sharper angle; the curved kayaks also have
higher sides than the flat types. The keel line is curved; in the 'olsteinsborg and 9ukkertoppen
areas some kayaks also have a concave curve toward the rear, forming a rudder=like pro-ection at
the -unction of the keel and the stern. The hull profile varies according to regional and personal
preference, but the J=angle of the bottom appears to be generally sharper than in the flat types.
The avasisaartoq type 5+ig. .c7 seems to be similar to $irket=9mith2s description of the old
kayak types found in his A. and @. group areas. According to 3etersen, the avasisaartoq type has
a high stern and is more strongly curved F though with lower sides F than the curved type.
3etersen2s ;orth 8reenland type is the shortest of the 8reenlandic kayaks. The stem and stern
are very short and the gunwales have a slight negative 5concave7 sheer. A variant from Ja-gat has
a rising stern which 3etersen considers a carryover from the avasisaartoq type 53etersen
1()6a0@(7. *n outward appearance and distribution, this type comes closest to $irket=9mith2s ..
and 6. groups, although the latter is defined as having a straight sheer line. 9ince neither author
provides a line drawing of his type, the actual degree of similitude is difficult to assess.
$irket=9mith2s description of the 6. group is in fact fairly close to a line drawing published by
"hapelle of an eastern kayak of unknown provenance in the 9mithsonian collection 5cf. Adney ?
"hapelle 1(6@ +ig. 1(17; however, the construction of the frame, particularly the stem and stern,
is not apparent from "hapelle2s drawing. "hapelle also presents a drawing of a kayak
purportedly from ;orthwest 8reenland 5Adney ? "hapelle 1(6@ +ig. CB.7, but this probably
refers to the Thule 4istrict 5;orth 8reenland7 since the kayak in !uestion is clearly of the East
"anadian type.
The $affin *sland=derived kayak type formerly used in the Thule 4istrict was replaced by the
1(@B2s by a type originating from northern <est 8reenland, so 3etersen considers the current
Thule kayak to represent the ;orth 8reenland type 53etersen 1()6a0@(, .67. 'owever, -udging
by a drawing published by 'oltved 5+ig. .d7, the new type is not purely 8reenlandic.
8reenlandic features in the new Thule type are, e.g., the narrow and high stern with concave
lines, the double foredeck stringers, and the absence of rear deck stringers. Kn the other hand,
11
the tripartite, angular ribs are similar to those used on the old $affin *sland model. This may of
course be an environmentally determined feature since green wood suitable for making curved
ribs is not found in the Thule 4istrict 5cf. 3etersen 1()6a0.@7; it might be noted that according to
'oltved, there is currently a preference for imported green willow stems, which can be bent into
shape 5'oltved 1(6D0D.7.
Another feature that distinguishes the new type from both the other 8reenlandic kayaks and the
East "anadian types is the construction of the stem and stern. The new Thule kayak does not
have the widening keelson and triangular end blocks typical of other 8reenlandic types, instead,
it has straight stem and stern posts which connect the ends of the keelson to the ends of the
gunwales. There is no mention in the literature of this construction being found in the kayaks of
the northern <est "oast, and it is clearly exceptional in the eastern Eskimo area.
$esides these main types, 3etersen also introduces three specialied types0 a transportable
lightweight kayak, a storm kayak with very wide gunwales but no side stringers, and a >cult
kayak> with small stem and stern >horns> reminiscent of those found on %ackenie 4elta kayaks
53etersen 1()6a0.B=.17. These types are not described in detail. Toward the end, 3etersen also
mentions three >local types>0 a type from *kerasaarsuk with a long rear >horn> to facilitate lifting
the kayak on the ice, the old Thule model derived from $affin *sland mentioned above, and a
type from Ammassalik similar to the East 8reenland type of 3orsild and $irket=9mith 53etersen
1()6a0.C=.67.
According to a drawing published by 'olm, the old Ammassalik kayak had an upcurved stern
which, however, did not rise as steeply and was longer than that of the avasisaartoq type 5cf.
'olm 1(11 +ig. AA7. The keel line was straight in the middle and curved up evenly at the ends
with no noticeable break; 'olm2s drawing also shows a straight gunwale although informants
stated that the deck curved upwards toward the stem as well 5cf. Thalbiter 1(1C0A)@; +ig. (B7.
The gunwales in the drawing lean outward and the ribs are evenly curved, but tradition has it that
the ribs used to be made in the same manner as in the old Thule kayaks, i.e., by -oining three
straight pieces 5cf. Thalbiter 1(1C0A)@7. The kayak thus had a flat bottom and, according to
Thalbiter, was conse!uently classified by the natives as an umiak 5loc. cit.7. 'olm2s drawing
shows three fore deck stringers but no aft deck stringers. 9ome kayaks of the flat type also have
three fore deck stringers 5cf. 3etersen 1()6a +ig. @@7, others are similar to the curved types and
have only two.
T'( Ea! Cana+ian Kayak
The East "anadian kayaks 5+igs. 1b ? Ab7 are recogniable by their long and very narrow bow
and wide, flat stern. The deepest part of the hull is at the foot of the bow and the widest part is
behind the cockpit. $ecause the volume and thus also the buoyancy is greater toward the stern,
the cockpit is situated clearly abaft the midpoint. The sheer line has a moderately negative curve.
These kayaks do not have the 8reenland=style triangular blocks at the ends of the frame, but the
-oint of the gunwales and keelson is sometimes strengthened with an internal support and the
stern may have a protruding end block.
The thick driftwood ribs are straight in the middle and usually either bent very acutely toward
the ends or partially broken to produce an angle. The bottom is conse!uently !uite flat, though in
some models the high keelson gives the hull profile a shallow J shape. As in 8reenland, there
are usually three wide, boardlike bottom stringers which, however, are laid on the flat; some
1C
models have a pair of additional short filler stringers between the side stringers and the gunwales
and rarely also between the side stringers and the keelson 5e.g., Arima 1()D011@=11.7. The fore
edge of the cockpit is raised somewhat higher than in the 8reenland kayaks, and the coaming is
often 4=shaped and furnished at the fore end with a flat plate acting as a paddle rest.
:ike all other American kayaks, the East "anadian models have only one central deck stringer
fore and aft; these stretch all the way from the cockpit to the end of the prow and stern. The
masik deck beam, which supports the fore edge of the cockpit coaming, is higher and also
sturdier than those found on 8reenland kayaks because it must support not only the coaming
itself but also the long paddle which is rested on the coaming or the rest plate. The deck beams
ahead of the masik are slightly curved 5as in 8reenland models7, but the rise of the deck in front
of the cockpit is nevertheless sharper than on the 8reenland types.
*t is difficult to divide the East "anadian kayaks into distinct regional groups because diagnostic
features in the structure and form often merge without clear=cut areal boundaries. The outside
measurements may change according to use 5Arima 1()D0()7; historical accounts show that there
have been changes even in this century due to, e.g., the introduction of firearms 5Arima 1()D0(D
1B., 11B7. The most diagnostic regional differences relate to the form of the prow, the stern, and
the ribs, and to the angle of the gunwales. Kn the basis of these features the following regional
groups may be distinguished 5%ap ***70
1. ;orthwest 'udson $ay, %elville 3eninsula, $affin *sland 5cf. Arima2s +oxe $asin group7
5+ig. 6a70 9hort and high prow with a fairly steeply rising keel line forming a concave curve
at the upper end 5the so=called >clipper prow>7. 8unwales vertical, ribs curved at the ends.
5Cf. $oas 1)))0@)D, +ig. @1.; $oas 1(B1=BD01B, +ig. 1e; DD, +ig. 1B6b; %athiassen
1(C)a0(C, +ig. .Ca; Adney ? "hapelle 1(6@ +ig. 1(6; Arima 1()D +igs. AB=AC7. At least the
following types of bow and stern supporting structures are found0
a7 3ond *nlet 5northern $affin *sland70 Triangular interior support block in the bow;
short, protruding block in the stern 5%athiassen 1(C)a +ig. ..7.
b7 "umberland 9ound 5southern $affin *sland70 +lat connecting piece between the upper
edges of the gunwales in the bow; no separate support block in the stern 5$oas 1(B1=
BD01B, +igs. 1d, 1f7.
C. The northern coast of the :abrador 3eninsula opposite 'udson 9trait 5cf. Arima2s 'udson
and 4avis 9traits group7 5+ig. 6b70 Jery long hull and very long and low prow; angle of keel
line rise at the bow fairly low. Jertical gunwales, acutely bent or cracked ribs, flat bottom.
5Turner 1)(@0CAD=CA); Arima 1(6@; Adney ? "hapelle 1(6@0CB.=CB6, CBD +ig. 1(D; Arima
1()D +igs. CD=C), A)7.
A. ;ortheast coast of :abrador 5cf. Arima2s Atlantic :abrador group7 5+ig. 6c70 Jery close to
the previous type, but higher prow and somewhat more curved keel line toward the stern.
Jertical gunwales, acutely bent ribs, low J=angle in bottom profile. 5Adney ? "hapelle
1(6@0CB6, +ig. 1(); Arima 1()D0()=((, +igs. A@ ? A.NRO7.
@. East coast of 'udson $ay 5cf. Arima2s Eastern 'udson $ay group7 5+ig. 6d70 Jery short and
wide hull, full and steeply rising prow, short, protruding pro-ection in the stern. 8unwales
lean markedly outward, ribs are bent at the ends in a roundish curve. 5Cf. Arima 1()D +igs.
@B=@17.
1A
.. $elcher *slands 5eastern 'udson $ay7 5+ig. Da70 9hort, straight, and steeply angled prow.
Jertical gunwales, flat angular bottom, strait keel line. 3row may have interior supporting
block; ends of gunwales and keelson form a short pro-ection in the stern. Also found in
double=hatch form 5the kayakpaakalik or pacalik7, which is otherwise unknown in the East.
This group is not mentioned separately by Arima and seems to be included in the previous
group. The general form, the angle of the gunwales, and the structure of the bottom,
however, are so different that treating these kayaks as a separate group is warranted. 5Cf.
+reeman 1(6@06) ? plates; 8uemple 1(6D7.
6. Thule 4istrict, ;orth 8reenland 5+ig. Db70 +lat bottom, flat sides due to outward leaning
gunwales; profile trapeoidal. 3row short and steep, often similar in shape to group 1. Kld
examples often have three=part angular ribs, newer kayaks also have curved ribs. The
cockpit coaming, which is fre!uently furnished with a paddle rest plate, is often four sided
with the front board shorter than the back board. 1ounger versions may have a rounded
coaming and a separate paddle rest on the foredeck 5cf. 'oltved 1(6D0D6, +ig. .A7. At least
the protruding stern block is used 5cf. 3etersen 1()6a0.., +ig. ..a7. This is the type that
came to the district with immigrants from northern $affin *sland during the 1)6B2s 5the
3olar Eskimo did not use kayaks in the early 1)BB2s7 and was replaced by a <est 8reenland
type in the 1(@B2s. 5Cf. 9teensby 1(1B +ig. A); 'oltved 1(6D0D@=)C; Adney ? "hapelle
1(6@0CB6, CBD, +igs. 1((=CBB; 3etersen 1()6a0.A=.67.
*n addition to these types, there are also specialied versions for ice edge and inland hunting
which may differ substantially in construction and measurements from the typical sea kayaks 5cf.
Arima 1()D +igs. AA ? AD7.
T'( Ar$!i$ Kayak
The kayak types found in the central "anadian Arctic and ;orth Alaska are long, narrow, and
lightly constructed and have a strongly tilted cockpit and straight or positive 5convex7 sheer
5+igs. Ca and @a=b7. The gunwales are fairly broad in the middle but narrow toward the ends. The
thin, evenly curved or slightly flattened ribs are made from fresh wood. There are usually five or
seven thin bottom stringers. The keelson is generally no thicker than the rest of the stringers, so
the bottom profile is roundish and no actual keel is discernable. $oth the single deck stringers
and all the bottom stringers reach all the way to the bow and stern.
Jiewed from above, the line of the gunwales is convex both at the bow and the stern, and the
keel line also curves smoothly up to the ends of the frame. The bow and stern are actually often
so similar in shape that only the position of the cockpit coaming shows which end is which.
9ome types have hornlike pro-ections on the bow and stern; the sie and shape of these differs
according to the region. The cover is usually made from ringed seal skin, but inland groups also
use caribou skin which is lighter though not as impermeable as sealskin.
The arctic kayaks are found in two geographically separate areas, an eastern province consisting
of the coastal area between 'udson $ay and "oronation 8ulf and the inland tundra of the $arren
8rounds in northern "anada, and a western province comprising the area between the "anadian
border and #otebue 9ound in ;orth Alaska. These two areas are not connected, for between
them lies the %ackenie 4elta which has its own kayak type.
1@
The structure of the actual weight=bearing frame is very similar in all arctic kayaks; however, in
the Alaskan types the foredeck rises in a convex curve starting in front of the paddler2s feet while
in the "anadian types it rises in a sharp concave curve directly in front of the cockpit. %any
"anadian types also have a bow and stern ,horn/.
The Arctic kayaks may be divided into the following groups 5%ap *J70
1. The $arren 8rounds and the west coast of 'udson $ay 5the "aribou and Aivilik Eskimo7
5+ig. (a70 9even rectangular=profile bottom stringers, fastened on edge to give rigidity. Two=
piece J=shaped masik. :ong and thin bow and stern ,horn/, the former horiontal and the
latter angled upward. 9mall step at the root of the bow horn. "over often caribou skin. 5$oas
1(B1=BD0D6=D), +igs. 1B., 1B6a ? 1B6c; $irket=9mith 1(C(G*01).=1)(; Adney ? "hapelle
1(6@0CB@, +ig. 1(A; 'eath 1(DB016=1D; Arima 1(D.0((=1AB; Arima 1()D06B=6D, +igs. 1A=
C17.
C. $oothia 3eninsula, #ing <illiam *sland, and the $ack Hiver area 5the ;etsilik Eskimo7 5+ig.
(b70 +ive or seven bottom stringers grouped closely around the keelson. 9hort horiontal
bow ,horn/, almost nonexistent stern horn. 9mall connecting block between keelson and
gunwales in the bow. $esides the usual oval cockpit coaming also a 4=shaped coaming is
known, and although Arima considers this an intrusive feature only adopted in the present
century 5Arima 1(D.0D67 the 4=coaming is already present in the ;etsilik kayaks collected
by Amundsen during his navigation of the ;orthwest 3assage in 1(BA=B. 5cf. Taylor
1(D@01167. The masik is in two parts and shaped like a sharp or rounded J 5cf. Adney ?
"hapelle 1(6@ +ig. 1(@; Arima 1()D +ig. 1C7; the cockpit coaming is sometimes fastened to
the frame, which is exceptional east of Alaska 5Arima 1()D0@C=@), +ig. 1C7. 4ifferences in
the lines of the frame allow the definition of two separate subgroups0
a7 "learly positive sheer, outward=leaning gunwales, ribs almost straight at sides and
bottom. Jery short, ,stepped/ end horns. "ockpit noticeably aft. 9tringers flat, fastened
on their side. 'ull profile straight=sided, bottom almost flat.
b7 ;o sheer, vertical gunwales. Houndly curved ribs which meet the gunwales at an
angle. ,'orns/ longer than on the preceding type; as in the "aribou Eskimo model, the
fore horn is horiontal and the aft horn rises at an angle. Thin, cylindrical bottom
stringers, hull profile rounded. 5Cf. Adney ? "hapelle 1(6@ +ig. 1(@7. Arima considers
the example presented by "hapelle to be a hybrid of the original ;etsilik type 5subtype
a7 and the "aribou Eskimo type 51()D0@C7.
A. "oronation 8ulf 5the "opper Eskimo7 5+ig. (c70 Three or five bottom stringers, very small
or nonexistent end horns. <eakly positive sheer. 'igh and curved but slender masik, cockpit
coaming often notably high. Jertical gunwales, bottom profile flattish curve. 5Ienness
1(@601A(=1@1; Adney ? "hapelle 1(6@0CB@, +ig. 1(C; Arima 1()D0C)=A1, +ig. 117.
@. ;orth Alaska 5the ,tareumiut/ and ,nunamiut/ Eskimo7 5+ig. (d70 9even slatlike bottom
stringers, hull lines very similar to "opper Eskimo kayak except for the convex foredeck
and also the upturned ends of the ;unamiut kayaks of the inland tundra 5cf. *ngstad
1(.@0)B=)1; $inford 1(D)0A@) +ig. D..7. The masik is curved on the northern coast and J=
shaped in #otebue 9ound and inland; according to "hapelle2s plan drawings also the
foredeck bulge itself would seem to be higher and longer in #otebue 9ound than on the
;orth "oast 5cf. %urdoch 1)(C0AC(; Adney ? "hapelle 1(6@0CBB, CB1 +igs. 1)6 ? 1)D7. *n
1.
;orth Alaskan kayaks the cockpit coaming is often furnished with short supports connecting
the sides of the coaming with the gunwales as in the ridged=decked $ering 9ea types. 5Cf.
%urdoch 1)(C0AAB7.
T'( 6a$k(n7i( Kayak
The kayaks used in the %ackenie 4elta of northern "anada are short and high=sided when
compared with the arctic types 5+igs. 1c ? )7. The cockpit is round or egg=shaped; the coaming
often sags and conse!uently the front part curves up while the aft part is horiontal. The ,horns/
of %ackenie kayaks are symmetrical, either short and ,plug=like/ or longer and sharply pointed;
the latter are usually somewhat curved and almost vertical. The sheer is straight or slightly
negative, the keel line likewise. The keel line rises in a sharp curve at both ends. Kut of the
water, the %ackenie kayak seems to have high sides, but due to the narrow hull the actual
freeboard is !uite small when the kayak is afloat. 5Cf. Adney ? "hapelle 1(6@ +ig. 1)); Arima
1(D. +igs. 1@a ? 1@b, +igs. 1=)7.
The deck beams are often slightly curved and the deck stringers are fairly high F particularly so
close to the cockpit F so the deck has a noticeable median ridge similar to that found on <est and
9outh Alaskan kayaks. There are usually three, sometimes five, bottom stringers; both the
keelson and the side stringers are very thin and wide in the middle but grow narrower toward the
ends. The keelson is not continuous but furnished fore and aft with separate extensions which
form the ,horns/. The thin willow ribs have a round or elliptical curve which is sharpest in the
middle though there may be a short flat portion in the centre. 5Adney ? "hapelle 1(6@0CBB=CBC;
'eath 1(DB01A=1@; Arima 1(D.0)B=)C; Arima 1()D01.=C17. According to some sources, the cover
was made from beluga skin 5e.g., 3etitot 1))D01D(; <hittaker 1(AD01DA=1DD; cf. Arima 1(D.0)C;
1()D01A=1@7; however, kayaks in museum collections seem to be covered with sealskin 5Arima
1()D01@D7.
The %ackenie 4elta is inhabited by several Eskimo groups whose kayaks probably differed
somewhat from each other at least as to the shape of the ,horns/; however, the published
ethnographic material is not sufficient for a more comprehensive description of variants 5cf.
Arima 1()D01@7.
T'( Koryak Kayak
The kayak used by the #oryak of the 8ulf of 3enshina on the northeastern coast of the 9ea of
Kkhotsk represents an extreme example among kayak types at least as concerns measurements
5+igs. Ac ? 1A7. The #oryak kayak is not over A m long and only CB=C. cm high but over 6B cm
wide 5Adney ? "hapelle 1(6@01(.7. The widthGlength ratio of a #oryak kayak illustrated in
Iochelson2s monograph 51(B.=B) +ig. )A7 is actually no more than 10A.6. The hull has a very
shallow draught and a slight J=profile. There is a carrying handle at the bow and stern. The
weakly curved flat ribs support three similarly flat bottom stringers; of these, however, only the
keelson is full=length. There are three deck stringers fore and aft, but only two deck beams.
A #oryak kayak in the American %useum of ;atural 'istory apparently originally had vertical
supports connecting the outer deck stringers to the ribs 5Adney ? "hapelle 1(6@01(.7; a kayak in
16
the %useum of Ethnography in :eningrad has a similar support between the central deck stringer
and the keelson 5$irket=9mith 1(C(0**,D)7. The deck beams are slightly curved, so the deck is
somewhat arched both fore and aft 5cf. Arima 1(D.06(7. The cockpit is oval and very large, being
!uite as wide as the frame itself. *t is also placed forward of the centre, which is a uni!ue feature;
however, since the occupant sits at the rear of the opening 5as shown by the flat ,seat/7 his torso
is actually behind the centre line 5Arima 1(D.06D7. The example illustrated in Iochelson2s study
has the aft part of the cover made from hairy skin, but no explanation or mention of the
fre!uency of this trait is given.
Two published descriptions of the specimen in the American %useum are in conflict as regards
the structure of the ribs. According to Adney ? "hapelle the ribs are made from one piece
51(6@01(.7; according to 'eath they are bipartite, the two halves being connected under the
keelson 51(DB0A7. $ipartite ribs are not found in any other kayak type, however, and it should be
noted that Iochelson 5who collected this particular example while employed by the Iesup ;orth
3acific Expedition7 makes no remark as to the structure of the ribs in his description 5cf. 1(B.=
B)0.@B7.
T'( R(in+((r C'#k$'i Kayak
A drawing of a 9iberian Heindeer "hukchi kayak published by $ogoras shows a fairly short,
wide, and high craft characteried by a teardrop=shaped cockpit and very steeply rising prow and
stern which meet the keelson at an acute angle 5+ig. 1@7. The gunwales seem !uite stout, the ribs
and stringers fairly thin. The number of bottom stringers is not evident from the drawing but the
chines on the cover suggest that three were used.
The deck is flat and there seems to be almost no sheer; the shape of the keel line is not
discernable, nor is the number of deck stringers. According to the text, the cover is of reindeer
skin. The stem and stern are oddly flattened; however, the drawing is not based on a full=sie
kayak but on a model, so the structural details and proportions may be faulty. 5$ogoras 1(B@=
B(01A.; Arima 1(D.06(7.
T'( B(rin, S(a Kayak
Typical features of the kayaks used in the $ering 9ea area are the short, dumpy form, the high
deck ridge, and the fastening of the cockpit coaming to the gunwales with slablike vertical
supporters 5+igs. Cb ? @c7. The prow is constructed using a large separate former block which
anchors the deck 5ridge7 stringer, the gunwales, and the keelson together; the upper part of the
stern is often also a separate block. %any $ering 9ea types have a carrying handle or hole in the
bow and stern and a noticeably large cockpit which can, when necessary, accommodate two
occupants sitting back to back. The keelson, deck stringers, and gunwales are stout and almost
s!uare in profile while the side stringers are !uite thin and often round. There are usually seven,
nine, or eleven bottom stringers, keelson included. The stern is flattened and rudder=like, and
some models have a very pronounced positive sheer. The cover was formerly made from bearded
seal skin; currently sailcloth is popular.
The kayak types used in the $ering 9ea area do not differ very much in frame structure, but the
differences in outward appearance F especially the form of the bow and stern F are notable. The
1D
geographical distribution and distinctive features of the five main types may be summed up as
follows 5%ap J70
1. The 9eward type 5Arima2s $ering 9trait subgroup, cf. 1(D.0DA7 5+ig. 1Ba70 9eward 3eninsula
and #ing *sland. 9traight or slightly positive sheer, with the sheer line curving up in the bow
following the keel line. 9harp, upcurved prow. The ridge of the aft deck curves slightly
down toward the stern. The cockpit is oval and fairly small. The thin ribs are bent
elliptically, so the bottom profile is !uite narrow. There are nine flat bottom stringers, the
keelson being somewhat thicker than the rest. The #ing *sland variant has a small hole
above the prow and sometimes a thin, recessed handle above the stern. 5;elson 1)((0CCB,
3l. :SS*S0@=.; Adney ? "hapelle 1(6@01((=CBB, 1() +ig. 1)17. #ayaks were not used in
historical times on 9t. :awrence *sland in the $ering 9ea, but a kayak model collected there
by ;elson 51)((0A@6 +ig. 1A@7 may indicate that the former kayak type resembled the
9eward type. The model has a small transverse hole in the bow and stern, but it is difficult to
say if these were also found on the full=sie version or if they were only meant for hanging
the model inside a real kayak as an amulet.
C. The ;orton type 5Arima2s ;orton 9ound subgroup, cf. 1(D.0DA7 5+ig. 1Bb70 ;orton $ay and
the south coast of ;orton 9ound to the 1ukon 4elta. 4eck ridge, sheer, and keel line nearly
straight, making the kayak look rectangular when viewed from the side. The prow curves up
!uite sharply and ends in a hooklike beak surmounted by a thin lifting handle. The stern is =
flattened and almost perpendicular; the aft end of the deck ridge forms a recessed lifting
handle above the stern similar to that found on #ing *sland kayaks. The cockpit is slightly
oval and somewhat wider than the hull. There are seven bottom stringers. The sides of
;orton kayaks lean outward like those of the 9eward type, but the bottom of the hull is flat=
ter. 5;elson 1)((0CCB, 3l. :SS*S0A; Adney ? "hapelle 1(6@0CBB, +ig. 1)C7.
A. The ;univak type 5+ig. 1Bc70 ;univak *sland and the mainland coast opposite to it. :arge,
round cockpit, clearly >humped> gunwale line 5positive sheer7. +lattened stern furnished
with a wide, protruding lifting handle. %ainland variants usually have a rounded, paddle=
like stem with a large hole; ;univak *sland variants may also have a stem with one or two
hornlike pro-ections or a straight stem with one or two sharp >bumps> on top of the deck
ridge 5cf. "urtis 1(AB01@7. *n the straight=stemmed variants the stem block is not a flat board
but a curved s!uare beam with a wider portion at the bottom to support the hull stringers 5cf.
Adney ? "hapelle 1(6@, +ig. 1D@7. The stern lifting handle on some ;univak *sland types
droops downward and may have a fist=sie hole at the root. There are seven or nine bottom
stringers, and the hull is very wide and flat. 5;elson 1)((0C1(=CCB, 3l. :SS*S01=C; Adney
? "hapelle 1(6@01(D=1((, +ig. 1)B; Eimmerly 1(D)7.
@. The $ristol type 5Arima2s $ristol $ay subgroup7 5+ig. 1Bd70 $ristol $ay. According to
Arima, $ristol $ay kayaks have eleven bottom stringers and are heavily built 5Arima
1(D.0DA; cf. Adney ? "hapelle 1(6@01(6=1(D, 1((7. &nfortunately, he does not present a
more detailed description, drawing, or photograph of the type he is referring to, for it does
not appear in any other source under this name. The Alaskan collection of the +innish rear
admiral Adolf Arvid EtholTn
1
51D((=1)D6, "hief %anager of the Hussian American
1
According to prevailing custom, EtholTnUs name was >Hussianied> in Hussian documents to
Adolf #arlovich Etolin, a form which unfortunately persists in American literature and in the
spelling of the various topographical features in Alaska which were named after him 5 e.g., Etolin
*sland south of <rangell in 9< Alaska, "ape Etolin on ;univak, and Etolin 9trait between
;univak and the mainland7. Another +inn to suffer the same fate was Iohan 'ampus +uruh-elm,
1)
"ompany 1)@B=@.7 in the +innish ;ational %useum contains several walrus ivory kayak
models from 9outhwest Alaska, some of which clearly represent the ;univak type with its
large stem hole; however, there are also a number of models that closely resemble a kayak
illustrated by "hapelle, which has been dubiously designated as obtaining from #odiak
*sland 5Adney ? "hapelle 1(6@01(6, +ig. 1DD7. +eatures common both to the models and
"hapelle2s illustration are a flat bottom, a sharply rising hornlike prow separated from the
hull by a very narrow curved slit or seam 5portrayed on the models by a scored line7, and a
sector=shaped cutout at the base of the stern lifting handle which gives the kayaks a
semblance of the ;orton type. $oth the models and the illustrated kayak also exhibit a very
clear downward slope of the deck ridge from the cockpit toward both ends. The kayak
illustrated by "hapelle has eleven bottom stringers. 9ince the models actually seem to
originate from $ristol $ay
C
, there is a fair probability that the type in !uestion is in fact
identical to Arima2s $ristol $ay group.
.. The 9iberian type 5+ig. 1C70 The coast of the easternmost part of 9iberia. According to
"hapelle the kayak used by the 9iberian Eskimo and the "oastal "hukchi resembles the
;orton 9ound type but has no lifting handles in the bow or stern 51(6@01(.7. Iudging from
an illustration published by $ogoras 51(B@=B(01A. +ig. @Dc7, the 9iberian sea kayaks had a
nearly perpendicular bow and stern and a large, round cockpit which was situated uncom=
monly far astern. 4etails of frame construction are not discernible in the drawing, but the
deck ridge seems to be lower than on other $ering 9ea types 5cf. Arima 1(D.0D1, (A7. +or
some reason or other, Arima does not include this type in the $ering 9ea group but sets it
apart as a separate group 5cf. 1(D.0D17.
T'( So#!' A%akan Kayak
"ompared with the $ering 9ea kayaks, the 9outh Alaskan kayaks or baidarkas 5a name adopted
from the Hussians during the colonial period7 are long and graceful 5+igs. Cc and @d7. These
types also come in two and three=hatch versions, which are similar in basic construction to the
single=hatch models. Hepresentative features are a bifid bow, slightly concave 5negative7 sheer,
and a shallowly ridged deck; the gunwales do not curve down toward the ends as in many $ering
9ea types but, instead, coalesce level with the ends of the deck stringers. The keelson is much
thicker and wider than the side stringers and forms a clear keel ridge in the bottom profile. The
height of the keelson in a double=hatch Aleut baidarka belonging to the Turku 'istorical
%useum, as measured by the author, was @C mm., only C mm. less than that of the gunwales.
A two=piece stem block connects the keelson to the deck stringer and gives the stem its typical
form; the block consists of a paddle=like lower fork and a flattened upper fork and has been
assumed to have prevented the bow from >digging in> into the bottom of a wave 5cf., e.g., 'eath
1(DB0.7. Two horiontal triangular supports are fastened to the sides of the stem block at the
level of the deck; these serve as fastening plates for the gunwales and give the stem block lateral
rigidity.
the penultimate "hief %anager, who opposed the sale of Alaska to the &nited 9tates. +uruh-elm is
usually referred to in Hussian and American texts as >*van +uruhelm>.
C
3ers. comm., %s. 3ir-o Jar-ola, curator of the Exotica collections of the +innish ;ational
%useum. *t might be mentioned in passing that walrus ivory, the material used for these kayak
models, is very typical of the $ering 9ea coast where walrus hunting has considerable economic
importance whereas east of the Alaska 3eninsula, e.g., on #odiak *sland, walrus hunting is almost
unknown and walrus ivory is conse!uently used much less 5cf. Hay 1()B0AD, +ig 17.
1(
At the stern the keelson, deck stringer, and gunwales are connected to a vertical plank=like block.
4epending on the shape of the gunwales, the stern is either laterally flattened or 5when viewed
from astern7 T=shaped; in the latter case the aft deck remains wide all the way to the fore edge of
the stern block. The keelson is in two or three parts; this has been thought to improve the
flexibility of the hull and allow it to bend with the waves 5cf. :aughlin 1()B0A@=A6; Hobert=
:amblin 1()B0(7. The sides of the cockpit coaming are connected either with longitudinal
supports to the deck beams or with vertical supports to the gunwales.
The common covering material is the skin of the 9teller sea lion; single=hatch models could also
be covered with harbor seal skin 5Hobert=:amblin 1()B0)7. There is a fair amount of variation in
the bottom lashing. The two=hatch Aleut model in the Turku 'istorical %useum has each rib
fastened to the keelson with a separate lashing and also a separate continuous lashing on each
side to attach the rib to the stringers; the lashings are made not with rawhide line but with pleated
sinew and baleen strands. Kn the other hand, many of the wooden models in the EtholTn
collection have a continuous longitudinal lashing fastening the keelson to the ribs and a igag
transverse lashing 5going up one rib and down the next7 on both sides fastening the side stringers
to the ribs. There are also models with transverse lashings going all the way from gunwale to
gunwale on one rib as they do in the $ering 9ea types, as well as models with longitudinal
lashings following the stringers instead of the ribs as in the flat=decked types; however, this last
system is atypical in ridge=decked kayaks.
The original builders of 9outh Alaskan kayaks were the Aleuts, the #oniag and "hugach
Eskimo, and later the Tanaina and 1akutat *ndians. 4uring Hussian colonial times, however,
multiple=hatch 9outh Alaskan types could be found as far as #otebue 9ound, "alifornia, and
9iberia, for the triple=hatch version was the preferred conveyance of colonial officials. Aleuts,
with their double=hatch hunting baidarkas, were also transferred by force to other areas,
including the formerly uninhabited 3ribilof and "ommander *slands, to hunt fur seals and sea
otters which were so important to the Hussian fur trade.
The original subtypes were distributed as follows 5%ap J70
1. The Aleut type 5+ig. 11a70 The Aleutian *slands and the southern part of the Alaska
3eninsula up to 3ort %Vller; during the colonial period also in the 3ribilof and "ommander
*slands. :ow and narrow hull, T=shaped stern formed by :=shaped gunwales which turned
sharply inward at the fore edge of the stern block. Jery thin 5c. ) x 1C mm7. oval=profile ribs
and side stringers, keelson almost as thick as the gunwales 5in the example illustrated by
"hapelle c. C( x 1( mm., in a single=hatch model belonging to the EtholTn collection no less
than @@ x C. mm7. Kval cockpit coaming fastened to the gunwales with short and narrow
vertical supports. 9everal regionally diverging prow styles; both ,horns/ usually low and
blunt=ended. An old model which evidently disappeared already during the Hussian period
had nearly horiontal bow horns. 5:-apunova 1(6@0CCA=C@C; Adney ? "hapelle 1(6@01(6=
1(D; :aughlin 1()B0A@=AD; Hobert=:amblin 1()B0.=CB7.
C. The 3acific Eskimo type 5+ig. 11b70 The #odiak *slands and the south coast of Alaska from
"hignik to 1akutat $ay. :aterally flattened stern similar to the $ering 9ea types. 'ull clearly
higher and wider than in the Aleut type 5Iacobi 1(AD01C(7, for both the #oniag and "hugach
Eskimo and the *ndians paddled on their knees 5:aughlin 1(6C01C1; Iacobi 1(AD01C(7.
"ockpit usually round. 4ue to the width of the frame, the central cockpit in a three=hatch
model 5and at least in the "hugach type also the other two7 was supported at the sides by
CB
longitudinal bars running from the deck beam behind the opening to the one in front. As
opposed to the Aleut type, the upper bow ,horn/ of the 3acific Eskimo kayak ended in a
sharp vertical pro-ection which rose higher than the wide, paddle=like and blunt lower horn.
The "hugach often used harbor seal skin also for multiple=hatch baidarka covers 5cf. $irket=
9mith 1(.A0@.=@(; de :aguna 1(.6, 3l. .D067. Tanaina and 1akutat baidarkas are poorly
documented, but they probably resembled the "hugach type 5cf. Ksgood 1(AD06D=6(7.
C1
STRUCTURAL ANALYSIS
T'(or(!i$a% Coni+(ra!ion
As mentioned in the section on research history, several authors have noted the existence of
structural similarities between kayak types and have used these to propose ,genetic/
relationships. 'owever, not one of these writers has explained why the proposed similarities
should relate specifically to ethnocultural traditions instead of, for example, the effects of the
physical environment. Wuite often the selection of features used in formulating ,genetic/
groupings does not seem to have been based on a systematic analysis of kayak structural design
itself, but rather on intuition or on the chosen features2 compliance with the previously known
distribution of various unrelated phenomena, such as linguistic groups or geographical
boundaries.
The formulation of the evolutionary theory used in this study is based on the assumption that the
structural features found in kayaks may be divided into three functional groups F demonstrative,
adaptive, and configurative F on the basis of their primary ,raison d2Xtre/. The definition is
purposely in the present tense, for it is usually impossible to deduce with any degree of certainty
the actual mechanism through which a feature originally came into being. 'owever, from an
evolutionary historical point of view it is more important to be able to evaluate its relative
permanence and typological representativeness, and to this end it is important to know, which
class of phenomena it is dependent upon.
4emonstrative features are mainly social signaling systems which function through displayable
symbols denoting, e.g., tribal identity, social ranking, or clan affinity. $ecause the function of
these features is informative, they are usually prominently displayed and easily read by any
person familiar with the particular symbolic system. 'owever, to retain their informative value
demonstrative features must be adaptable to variation and cannot therefore be functionally
related to the environment, nor can they be expected to remain unchanged for any greater length
of time. Although demonstrative features thus are clearly culturally specific, they usually reflect
phenomena of chronologically and spatially limited scope which have little effect on the main
course of developmental history.
Adaptive features are connected with cultural ecological phenomena such as the availability of
raw materials, the demands of the physical use environment, and exploitative specialiation.
%any adaptive features are answers to specific practical problems relating to factors such as the
immediate use environment F the waves, the ice, the wind F and various other phenomena
affecting use, such as the behaviour of important prey species.
*t is fairly common for the kayak builderGuser to be conscious of the functional value of a
problem=specific adaptive feature and to be able to explain why the feature appears in a specific
form. $ecause adaptive features are products of their historical users and areas of distribution,
they reflect the adaptation of the users to their present environment but reveal very little about
the prehistory of the kayak types or their users, i.e., about events preceding the development of
the historical situation.
"onfigurative features are usually not connected with recognied problems but rather form an a
priori basis F a configurative mental image 5a >template>, if you will7 of the kayak as a >thing> F
to which the demonstrative and adaptive features are applied; in other words, the function of
CC
configurative features is to form a basic blueprint from which various specialied types may be
developed.
9uch features can often be connected with long=term ethnocultural traditions, i.e., their
distribution reflects macro=level ethnocultural events such as the spread of peoples, cultures, and
technocomplexes. "onse!uently, they may have originated in areas far removed from where they
were found in historical times. To be able to cross environmental borders and to remain
unchanged through local environmental fluctuations such as climatical shifts, these features must
be relatively independent of natural phenomena, so their distribution in principle cannot directly
reflect either current or former adaptive processes.
The most widespread configurative features of kayaks are those which are universal and thus
form the basic definition of a kayak0 the covered deck, the self=supporting frame, one pair of
gunwale strakes, etc.. 'owever, the configurative features more interesting from an analytical
point of view are those which represent alternative solutions to structural problems which are
common to all kayaks, such as giving rigidity to the frame and supporting the bottom cover. *t is
these features which make it possible to trace lines of historical development by tying together
types 5which may have diverged under the influence of environmental and other external factors7
into configurational groups in the same way one can use primary physical features to tie together
divergent species of specialied animals into genera deriving from a common ancestor.
The spatial behaviour of the phenomena affecting the occurrence of the different classes of
features varies notably, and because it is also probable that the distribution of the affecting
phenomena is reflected in the distribution of the features themselves, the functional classification
of features may be presumably determined by analying their distribution and especially the
coincidence of structural borders with changes in various environmental and social phenomena.
"onsidering the spatial behaviour of the effecting phenomena, we may expect the features
belonging to the main functional groups to occur as follows0
4emonstrative features act as emblems of units with fairly limited areal ranges, such as
social or political groups. They may thus be expected to vary within culturally and
environmentally homogeneous areas, i.e., they may be termed regionally heterogeneous. Kn
the other hand, the features are usually also limited to that area, i.e., they are regionally
uni!ue.
Adaptive features represent adaptations to specific surroundings and tasks, and may
conse!uently be expected to react to changes in the environment, subsistence technology, or
subsistence economy, but not necessarily to ethnic, social, or political boundaries. Adaptive
features can be regionally heterogeneous like demonstrative features because they may
relate, e.g., to different types of specialied use within the same area. Kn the other hand,
they may also appear in similar form in geographically separate areas due to similar
environmental conditions and use needs. Thus, they are not regionally uni!ue.
"onfigurative features are not dependent on social or environmental factors and thus usually
have distributional areas which are large and continuous but may display a wide range of
environmental variation. The distributional limits may be connected with present or former
cultural or ethnic boundaries. Kccurrence is usually regionally homogeneous and exclusive,
i.e., mutually exclusive alternatives do not occur contemporaneously in the same area. :ike
adaptive features, configurative features are not regionally uni!ue; however, their
appearance in two separate areas is not tied to environmental or other external similarities.
CA
These formulae are hypothetical and derive F as stated F from the assumed spatial behaviour of
the influencing factors, i.e., from the contexts of occurrence of the features themselves. <hether
or not they actually work may be best ascertained by looking at a few features which fulfill the
conditions set forth in the formulae but whose context may also be determined through ethno=
graphical and historical data.
= D(mon!ra!i*( )(a!#r(
Kne group of demonstrative features that is fairly well documented ethnographically comprises
the various types of bow forms of the $ering 9ea and 9outh Alaskan kayaks. Kn politically
unified ;univak *sland the different bow forms serve to denote, for example, the owner2s age
and matrimonial status 5cf. :antis 1(@6016), C@(7, while on the mainland, where tribal
boundaries are pronounced, they function as regional emblems. That the outward shape of the
bow in this case is clearly independent of external physical factors is evident from the fact that
the different models are all used for identical purposes in identical circumstances; also, the range
and scale of variation is so large that it is impossible to define an ,original/ type.
9imilar regional variation may be noted in the ,horns/ of the Arctic types, which are also found
in an area where the environment, language, and culture are all very homogeneous but where the
regional ,tribal/ groups are nevertheless clearly defined. Kn the other hand, the variation in the
bow profiles of the East "anadian kayaks is not necessarily purely demonstrative, for this area
also exhibits a fairly large amount of variation in environmental and use contexts.
= A+ap!i*( )(a!#r(
;ext to the range of raw materials and the thickness of the structural parts, one of the most
important complexes of adaptive features comprises the exterior measurements of the kayak hull.
As mentioned in the typological section, hull width in, e.g., 8reenland and East "anada is
consciously modified according to the proposed use context, and it is thus !uite probable that
also the differences between main regional groups in hull proportions are at least partially due to
practical considerations related to use and environment.
The effect of use conditions and hunting techni!ues on hull proportions can be seen in the length
and width measurements of the regional groups which display several environment and use
related tendencies 5cf. scatter diagrams a ? b and appendix70
#ayaks used in scattered floe ice 5e.g., eastern 'udson $ay, $ering 9ea, #oryak7 are usually
short.
#ayaks used mainly in open water 5e.g., :abrador, $affin *sland, 9outh 8reenland, 9outh
Alaska7 are often long.
#ayaks used in seal and walrus hunting where the hunter waits for the !uarry to approach or
tries to sneak up on it unobserved 5e.g., 8reenland, East "anada, $ering 9ea, #oryak7 tend
to be fairly wide.
C@
#ayaks used in caribou, beluga, and sea otter hunting where the !uarry is pursued before
being struck 5e.g., the arctic caribou kayaks, %ackenie, Aleut7 are usually !uite narrow.
The reasons for variation in hull length are clear0 the long hull is faster and easier to hold on
course and is thus better in open waters, while the short hull is structurally sturdier and more
maneuverable and thus more useful in the ice leads, especially as it is also easier to lift over the
floes when necessary. 'ull width, on the other hand, does not seem to depend on the environ=
ment but rather on the hunting method. Kpen water sealing often means waiting or approaching
by stealth, because a !uickly swimming and diving seal is very difficult to chase; the catch must
also fre!uently be transported a long distance either on deck or in tow 5$irket=9mith 1(C@0A1)=
ACB; 8ad 1()@011B=1117. The stability provided by a wide hull is useful when the hunter must try
to stay immobile for long periods of time, and both the stability and the greater buoyancy are
assets in transporting the catch. The success of a chase, on the other hand, depends on speed, and
as there is furthermore no need to transport the catch on deck in either the reindeer hunt or the
beluga hunt, no exceptional stability or buoyancy is re!uired and the hull may conse!uently be
made as narrow as possible to cut down drag and weight 5cf. Arima 1(D.0((=1BB; 1()D06B7.
= Con)i,#ra!i*( )(a!#r(
A good example of a configurative feature F or rather a complex of features F is the occurrence
of a kayak type belonging to the East "anadian group in ;orth 8reenland. *n this case, the
reason for the occurrence is an historically known case of culture contact, i.e., the migration of a
group of $affin *sland Eskimos to 8reenland in the 1)6B2s. *t should be noted that the old
"anadian=type ;orth 8reenland kayaks were on average clearly shorter than the actual $affin
*sland kayaks because the ice conditions were different; also, the structure of the ribs was
different, owing at least in part to a dearth of suitable raw material. *n spite of these adaptive
modifications, however, the ;orth 8reenland kayak retained all the typical features of the East
"anadian kayaks F the single deck stringer, the flat bottom structure, the simple end construction,
the 4=shaped cockpit coaming, and the wedge=like hull form F up to the time when a new
contact phase with <est 8reenland introduced a new type of basic frame structure.
T'( Con)i,#ra!i*( Gro#p
As may be concluded from the above, many of the diagnostic features of the various regional
kayak type groups F such as special structural details and measurements F have demonstrative or
adaptive functions and conse!uently do not reflect the developmental history of the type as a
cultural trait but rather the adaptation of the kayak to changes in various social and
environmental phenomena. The definition of a permanent configurative feature is, in fact, best
met by certain compound structural features of the frame, such as the deck and bottom structure,
which due to their primary nature most easily become unconscious ,idTes fixes/, established
configurative assumptions which can only be changed through strong intrusive impulses.
*f it is possible to divide the different kayak types into ,genetic/ groups which reflect their true
historical evolutionary relationships and the ethnocultural events which caused the spread and
development of the types, this is best done specifically on the basis of the configurative
compound structural features. <e must, however, take into account the fact that each kayak type
contains several such compound features and that not all of these are e!ually old. *t would thus
C.
be a mistake to assume that the whole history of development could be reflected by the complex
of features found in any existing individual type.
Iohn 'eath has divided all kayak types into two families according to the following criteria0
= #ayaks of the ridged=decked family 5the $ering 9ea and 9outh Alaska types7 have the
cockpit coaming permanently attached to the frame, J=shaped deck beams, narrow
gunwales, transverse bottom lashings, and a cover which is sewn together on the frame and
goes over the cockpit coaming.
= #ayaks of the flat=decked family 5all other types7 have a separate ,floating/ cockpit
coaming, straight deck beams, wide gunwales, longitudinal bottom lashings, and a cover
which is sewn together beforehand and fastened to the cockpit coaming from underneath.
5'eath 1(D)0CC=C@7.
&nfortunately, 'eath does not explain in his paper why specifically these features and these
alone are crucial, and why, e.g., several structural features which are found in all ridged=decked
types and in many but not all flat=decked types are not. The systematic weakness with 'eath2s
families lies in the fact that the division is not based on an impartial analysis of all features but
rather on a kind of in=groupGout=group classification dominated by one specific feature complex.
'eath appears to have begun his classification by forming one exclusive family 5the ridged=
decked kayaks7 and defining as its features all those which separate the ridged=decked types
from all other kayaks; most of these features are naturally connected specifically with the deck
structure. The other family, the flat=decked kayaks, remains in effect an out=group, the inner
variations of which 'eath does not consider. This is demonstrated, e.g., by the fact that while
'eath notices differences in bottom lashing direction, he does not consider the differences in
bottom stringer construction. This does not vary between the ridged=decked kayaks and their
closest flat=decked neighbors, but does present substantial variation within the flat=decked
family.
*n ridged=decked kayaks, the structural function of the deck structure is totally different from that
of the flat=decked kayaks. *n the flat=decked types, the deck beams give transverse support to the
gunwales, but the deck stringers which run over them only support the deck cover while the
weight of the hull rests mainly on the wide gunwales. $ecause of their thinness and geometrical
position, the actual load=bearing function of the deck stringers is very limited, a fact amply
demonstrated by the 8reenlandic custom of deleting the aft deck stringers completely. As
structural concerns thus do not limit variation in the deck structure, several different patterns are
to be found in the flat=decked types.
*n the ridged=decked types, on the other hand, the deck or ridge stringer, which is supported by
the J=shaped deck beams, forms a structural bridge which supports the ends of the frame and
prevents the hull from sagging. $ecause the rigidity of the hull is thus achieved through
structural geometry, the gunwales are stressed less than in the flat=decked types and may conse=
!uently be made thinner and lighter without affecting the structural strength of the frame.
The use of this bridge construction on the kayak has, however, one drawback0 in order to be
rigid, the bridge should form a continuous span from bow to stern, but since the occupant sits on
the centerline the deck stringer must be cut at the cockpit. This problem has been solved by using
C6
a very sturdy cockpit coaming which is fastened permanently to the frame between the ends of
the fore and aft deck stringers and thus forms a structural part of the span.
The fastening of the coaming to the deck stringers unfortunately also has the effect of lifting the
coaming above the gunwales and rendering it prone to wobbling; this is countered by adding
supports to the sides of the cockpit. 9ince the coaming is now permanently lashed to the frame at
four points, it is obviously more practical to draw the cover over the coaming since drawing it
under the coaming would re!uire cutting holes at the connection points.
+rom the above it is evident that four of the six features used by 'eath for classifying kayak
families actually form F specifically in the ridged=decked in=group F a functional complex where
the various sub=features are all dependent on one another and which should thus in fact be
considered only as one feature. The mechanical integrity of the ridge=type structure is also
reflected by the limited amount of variation0 the steepness of the ridge changes somewhat from
type to type, but the structural form itself is almost identical in all ridged=decked models.
The two other features used to define the ridged=decked family are much more suspect. As
mentioned in the introductory notes on structure, 'eath2s assertion that the cover was sewn
together on the frame is not supported by published accounts, and the presence of the hidden
tightening lashing in the Aleut kayaks proves that the cover was sewn together first and only then
stretched onto the frame. The !uestion of the transverse bottom lashings is also not all that
straightforward in 9outh=Alaskan types0 both models and full=sie examples exhibit longitudinal
keel lashings and the transverse side lashings are either separate for each end of every rib or form
a continuous igag pattern, running up one rib and down the next but stopping short of the
keelson. *t seems that 'eath formed his view of the bottom lashing of ridged=decked kayaks
solely on the basis of $ering 9ea types and assumed that 9outh Alaskan kayaks are similar
because they also have a similar deck. Thus, the >ridged=decked family> is actually based on a
single structural feature, i.e., the ridged deck structure; the other proposed diagnostic features are
either direct corollaries of the first, false generaliations, or unfounded assertions.
The exclusive nature of 'eath2s family division is also evident from a more thorough analysis of
the flat=decked out=group. This family is in fact not at all as uniform in structure as the ridged=
decked in=group, and the definition of, e.g., the diagnostic features of the deck structure solely
with a mind to demonstrating the exclusivity of the ridged pattern leaves unconsidered several
features which distinguish flat=decked types from each other but either are not uni!ue to the
ridged=decked types or do not appear in them at all.
'eath has also proposed that the #oryak and 8reenland kayaks might be ,genetic/ relatives
because both have multiple deck stringers and three bottom stringers. *n addition, the peripheral
distribution of these types has prompted 'eath to assume that these features are especially old
and represent an ,archetype/ of the kayak. This hypothesis also has various problems which
derive on the one hand from the unnecessarily vague classification of deck structure types and on
the other from the definition of the ,archetype/ specifically as a combination of two structural
patterns.
The proposition that 8reenland and #oryak kayaks are genetically similar because both have
multiple deck stringers does not seem convincing because the actual deck structure in these types
is very different. *n 8reenland kayaks the deck stringers F be there two or three F are e!ually
long; they are also truncated and fastened only to the deck beams, not to the gunwales. *n the
#oryak kayak the long central deck stringer extends all the way to the bow and stern
CD
respectively, while the shorter side deck stringers are fastened at one end to a deck beam, at the
other to the gunwale.
The essential configurational difference between these two patterns is that the 8reenland kayak
retains no trace of the full=length median deck stringer typical of all other types while the #oryak
kayak is basically a median=stringer kayak with additional side deck stringers. $oth types also
present possible external reasons for using a multiple=stringer pattern0 the 8reenland kayak
complex includes an exceptionally large selection of complimentary e!uipment 5weapons, line
rack, ice scrapers, etc.7 most of which is carried within easy reach of the hunter on the portion of
the deck supported by the stringers. The #oryak kayak, on the other hand, is exceptionally wide
and the deck cover also needs more support from the stringers because there are only two deck
beams.
'eath2s other ,archaic/ feature, the triple=stringer hull, is not found only in #oryak and
8reenland kayaks but also in all East "anadian types, and its distribution in the east is thus
actually !uite wide and homogeneous. This fact, however, is not mentioned by 'eath, apparently
because the East "anadian types have a single deck stringer and do not thus conform to his idea
of the ,archaic/ type. The three=stringer construction, at least in the eastern area, is clearly not a
demonstrative or adaptive feature since it is regionally homogeneous and appears in both sea and
river kayaks, in open water and among the floes, in both seal and caribou kayaks, and even in a
double=hatch traveling kayak. Iudging by its distribution as well as its basic structural nature, the
three=stringer construction clearly answers to the description of a configurative feature.
Another feature connected with hull construction that 'eath does not consider at all is the
multiple=stringer construction found in the $ering 9ea, 9outh Alaska, and Arctic types. This
feature is also clearly independent of external factors as it is found both in inland caribou kayaks
and open sea kayaks; however, it is of no importance to 'eath2s preconceived division since its
distributional borders do not coincide with those of the ridged and the flat deck.
The hull structure is clearly not affected by the deck structure, for the three=stringer hull is found
with both single and multiple stringer decks and the multiple stringer hull with both the ridged
and the flat deck. The three=stringer hull is not found combined with the ridged deck nor the
multiple=stringer hull with a multiple=stringer deck, but this is probably due rather to the
separateness of their respective distribution areas than to any structural reason.
Iust as there is no change in the hull structure at the distributional border between the ridged and
the flat deck, there is also no substantial change in the deck structure at the border between the
hull structure types. The Arctic types do exhibit a clearly less robust masik and a steeper cockpit
angle than do the East "anadian types, but these are clearly adaptive features0 the tilted coaming
allows a necessarily !uick entrance into the cockpit of the fairly cramped caribou kayak, and the
use of a heavy paddle supported on the coaming is not necessary or even advantageous in inland
waters. "onsidering the fact that neither type of hull structure seems to be of itself connected
with external factors, the distributional border between them is all the more interesting because it
coincides with a clear environmental boundary.
As mentioned above, the multiple=stringer hull is found in Alaska both in sea kayaks and in
caribou kayaks. *n "anada, however, it is only found in caribou kayaks while all "anadian sea
kayaks are of the three=stringer pattern. *t is difficult to offer any functional explanation for this
phenomenon since the three=stringer structure is not necessarily mechanically superior to the
multiple=stringer hull but rather !uite the opposite. $eing structurally more simple, the three=
C)
stringer pattern also gives the impression of being older. Therefore, it would seem that the
,division of labour/ between the three=stringer and the multiple=stringer hull in the "anadian
Arctic is more likely the conse!uence of F rather than the reason for F the patterns2 geographical
distribution.
9ince the supplanting of an existing multiple=stringer sea kayak with a structurally inferior three=
stringer design seems improbable, it would seem more logical to assume that the multiple=
stringer type was the later to arrive, probably from the west 5-udging by its historical
distribution7, and specifically in the guise of a caribou kayak. This new type may have
supplanted any previous three=stringer caribou kayak types west of 'udson $ay, if any such
types indeed existed, but the three=stringer construction nevertheless remained in use in sea
kayaks in areas east of the eastern limit of the multiple=stringer type. This hypothesis, however,
does not yet explain why the spread of the multiple=stringer pattern should have stopped
specifically at an ecological border, nor why the feature was not incorporated into the eastern sea
kayak types although there should have been no cultural or technical restraints.
T'( R(%a!i*( A,( o) !'( D($k an+ H#%% Pa!!(rn
*n flat=decked kayaks there is no structural difference between fastening the cover to the cockpit
coaming from below or above, for the coaming is not an integral part of the frame and the
fastening methods themselves do not differ in complexity to any significant extent. Kn the other
hand, the internal method 5bringing the cover over the coaming7 is connected specifically with
the integral coaming found on ridged=deck kayaks. 9ince flat=decked kayaks almost universally
employ the external method 5bringing the cover up inside the coaming7, the ridged deck would
seem to be a later development; otherwise we would expect the internal fastening to be
commonly found also in flat=decked kayaks from areas not ad-acent to the deck pattern border.
Another feature pointing to the relative lateness of the ridged structure is the pattern2s
complexity. The ridged structure itself, the corollary thin gunwales, integral coaming, and sepa=
rate bow and stern blocks form an integrated feature complex whose structural parts are
dependent on each other; nothing like this construction is found on any other type of vessel. Kn
the other hand, the only essential parts in the flat deck structure are the straight deck beams
which function as spreaders of the gunwales and are also found in many open boat types; the
deck stringers fastened on top of the deck beams are basically nonessential. Their sole function is
to support the deck cover in situations of exceptional stress. The frame structure of the flat deck
pattern with its wide gunwales is thus probably a carryover from a preceding undecked boat type
while the ridged deck structure must clearly have come into use only after the invention of the
fully covered deck as it is improbable that the essential integral coaming should have been used
in an open boat.
The ridged deck pattern may also be younger than the multiple=stringer hull, for its area of
distribution is much more limited. 9ince F with the possible exception of the bottom lashings F
there are no differences between the Arctic types and the ridged=decked types that cannot be
attributed to the ridged deck complex itself, it is conceivable that the ridged deck originated and
spread as an independent feature complex within the formerly homogeneous flat=decked,
multiple=stringered group. The Arctic types would thus represent a pattern which was formerly
common also in <est and 9outh Alaska and which formed the basis from which the ridged=
decked pattern was later developed.
C(
The hull structure would thus belong at least in ridged=decked kayaks to a stratum of structural
features older than the deck pattern, and as the three=stringer hull pattern seems older than the
multiple=stringer pattern we might discern at least three basic structural types of different age.
The oldest stratum would be represented by the flat=decked, three=stringered pattern which is
found in the East "anadian and 8reenlandic types and the #oryak kayak, possibly also in the
Heindeer "hukchi kayak. The next oldest pattern would be the flat=decked, multiple=stringered
type represented by the Arctic types. The youngest structural type would be the ridged=decked,
multiple=stringered pattern found in the $ering 9ea and 9outh Alaskan kayaks.
AB
KAYAK TYPES AND ARCHAEOLOGY
T'( 6#%!ip%(&!rin,(r(+ Typ(
= T'( Ar$!i$ ,ro#p
The round=bottomed, flat=decked structure of the historical Arctic types has a long history in the
$ering 9ea area and in ;orth Alaska. The oldest currently known kayak models are from the
%iyowagh site on 9t. :awrence *sland 5"ollins 1(AD01.(, 3l..(017 and the Ekven cemetery on
the "hukchi 3eninsula 5Arutyunov et al. 1(6@0A@@ +ig.D7, both sites belonging to the Kld $ering
9ea culture which dominated the $ering 9trait area c. B=)BBA.4.. These models are round=
bottomed and flat=decked, as are also models both from the 3unuk culture which developed from
Kld $ering 9ea c. )BB A.4. and from the $irnirk culture which arose in ;orth Alaska c. .BB
A.4. 5+igs. 1Da ? b; cf. also "ollins 1(AD0C@., 3l.)A0.7. The humped foredeck typical of
historical ;orth Alaskan kayaks 5any more than the actual ridged deck7 does not yet seem to
have been in use in the area at this time.
The historical Eskimo cultures of ;orth Alaska, "anada, and 8reenland are currently considered
descendants of the Thule culture, which developed in the 1Bth century in ;orth Alaska from
$irnirk through 3unuk influence and spread east through a rapid migration. :ike 3unuk, the
Thule culture specialied in hunting large baleen whales with the umiak, a large open skin boat
which, due to its large cargo capacity, could also function as an efficient means of transportation
and speed the spread of the culture. Kn the north coast of Alaska the whaling=based <estern
Thule culture continued until the early 1(BB2s, but on the arctic coast of "anada contacts started
to decline soon after the expansion as whaling lost import and the umiak went out of use. After
settling down, the ancestors of the present "opper, ;etsilik, and *glulik Eskimos gradually
formed distinct cultural groups; however, their common heritage was still evident from many
implement types which still followed the Thule patterns common to all Eskimo groups in ;orth
Alaska and "anada.
The kayak type used by the Thule people seems to have been closely related to the earlier types
of the $irnirk and Kld $ering 9ea F 3unuk cultures. 4rawings on a Thule drill bow found in
Arctic $ay, $affin *sland, portray several flat=decked kayaks with bow and stern >horns>
apparently pursuing a swimming caribou 5+ig. 16a7; a similar kayak type is represented by a
model found in ;orth 8reenland 5'oltved 1(@@0CAB=CAC, 3l.@(0C.7, possibly also a drawing on a
bone implement from "ape #rusenstern, Alaska 58iddings 1(6)0(1 +ig.C@d7. ,'orns/ may
already have been used on $irnirk kayaks as their wooden formers have been found at the
$irnirk type site in ;orth Alaska 5+ig. 1.a7.
Arima has traced the form and structure of the "aribou Eskimo kayak back through Thule to
$irnirk and Kld $ering 9ea 5Arima 1(D.0.A, )D=(B7, and it is !uite evident that in fact the whole
Arctic group with its multiple stringer hull structure and symmetrical form derives from the
Thule type as the group is structurally so homogeneous that the various types must have a
common origin. The geographical distribution of the group also points toward a Thule origin as it
comprises both the ,original Thule homeland/ in ;orth Alaska and "entral Arctic "anada, where
the Thule culture spread into a practically empty land and where outside influences were thus at
a minimum.
A1
= T'( ri+,(+&+($k(+ ,ro#p
*t is not possible to follow the 9outh and <est Alaskan ridged=decked types very far back in
time, for there is a dearth of archaeological material. The oldest currently known kayak find of
the ridged=decked type is from continental Alaska south of 9eward 3eninsula, more specifically
from the ;ukleet site in ;orton $ay. A portion of the dwelling site dated to the 1@BB2s produced,
among other things, a bifurcate bow piece and two J=shaped deck beams 5+ig. 1.b7. The kayak
from which these parts derived was very close to the historical ;orton type, although the shape
of the stern is unknown as no stern parts were found 5cf. 8iddings 1(6@0)A7.
The oldest kayak finds from 9outh Alaska are somewhat younger. The pre=contact site #ar=D16
on #odiak *sland has produced two gunwale fragments. These are fairly narrow and thin, and
both have a series of elongate holes placed c. @> apart on one side, apparently for morticing in
the ribs 5"lark 1(D@0(A=(@, 3l.1)0"=47.
A prehistoric kayak find from 3alutat "ave in the northern part of 3rince <illiam 9ound
comprises a gunwale broken into three pieces, a wooden bar thought to be a ridge stringer, 11
possible ribs, a single=bladed paddle, and several paddle fragments. The gunwale is somewhat
thinner than those from #ar=D16 but the distance between the rib mortices is similar. The find
also contained a thin, triangular wooden beam c. )2 (> long, which was thought to be part of the
keelson of a two=hatch baidarka 5de :aguna 1(.60C@.=C@(7. The 3alutat find presumably dates to
the later prehistoric or protohistoric period 5de :aguna 1(.606.7, i.e., to the 1Dth or 1)th century.
The length of these two kayaks cannot be determined because the gunwales were fragmentary,
but the relative thinness of the members indicates that both boats were fairly small and probably
single=hatch. The thinness and fragility of the gunwales also suggests that the boats had ridged
decks since such flimsy gunwales would not have given a flat=decked kayak sufficient rigidity;
however, we have no decisive proof of this feature since no deck beams were found.
&nfortunately, neither find contained bow or stern parts, making comparisons with historical
9outh Alaskan types impossible.
Although the ;ukleet find shows that the ridged deck and the associated bow block reached the
northern limit of their distribution already by the 1.th century, the ridged deck has been assumed
to have spread into western Alaska at a fairly late date 5Arima 1(D.0(A7. There is no direct evi=
dence of kayak use in the ;orton culture 5c. .BB $.". F 1BBB A.4.7 nor in the preceding "horis
culture 5c. 1BBB $.". F .BB $."., beginning possibly as early as 1.BB $.".7 but -udging from
dwelling site finds, the ;orton people seem to have practiced open=water sea mammal hunting
!uite actively. The ;orton populace could thus be assumed to have used at least some type of
seaworthy hunting boat, most likely of the kayak type since the umiak was not used for hunting
in <est Alaska even in the historic period 5cf., e.g., :antis 1(DC0@.7 and other sea boat types are
not known from the area F in fact, even in ;orth Alaska, the use of the umiak as a hunting boat
does not seem to have begun before the 3unuk culture started to use it for whaling in the (th
century A.4..
Kld $ering 9ea, as a coastal and insular culture, was totally dependent on sea mammal hunting,
and the kayak must conse!uently have figured in its repertoire of hunting gear right from the
start. The origin of the Kld $ering 9ea culture is still unclear, but it has been seen as developing
from ;orton 5<. <orkman 1()C011.=1167. *f this is true, we may assume that ;orton used a flat=
decked kayak type at least up to the beginning of the "hristian era 5the divergence of K$97,
AC
since Kld $ering 9ea would probably have used the ridged=decked pattern if it had been previ=
ously present in ;orton.
"ertain younger kayak models seem to suggest that the ridged deck was adopted on 9t.
:awrence before the use of the kayak ceased on the island during the late 3unuk phase 5cf. 8eist
? Hainey 1(A601CA, 3l.AC +ig.@; E.<. ;elson 1)((0A@6 +ig. 1A@7. 'owever, if it actually did get
to 9t. :awrence, the ridged deck must first have spread over the $ering 9trait to 9iberia since the
permeability of skin boat covers probably made sailing directly from 9t. :awrence to the <est
Alaskan coast impossible 5cf. Hainey 1(@10@6A7 and there were conse!uently no direct contacts
between these two areas before the historical period. *t might be noted in this connection that
though the kayak shown in ;elson2s drawing is, according to the text, ,from 9t. :awrence
*sland, and is representative of the boats used at that place/ 5;elson 1)((0A@D7, according to
8eist and Hainey 51(A601C1, 1CA7 the 9t. :awrence Eskimo in the 1(CB2s and AB2s no longer
used kayaks, nor did they even remember them having been used, which seems odd if they still
were using them when ;elson visited the island some @B years earlier. ;elson seems to have
stayed on 9t. :awrence only for a short time collecting specimens and did not study the living
culture to any notable extent 5cf. ;elson 1)((0C17, so the absence of the kayak may have escaped
him. ;elson2s model kayak represents a typical 9eward kayak with the sharp, upturned bow and
the kinked, sloping stern ridge, and the model seems to be new and in good condition. *t is not
impossible to think that the 9t. :awrence Eskimo, who derived an important part of their trade
income from selling walrus ivory carvings to passing ships, could have borrowed the pattern for
their ,tourist carvings/ of kayaks from the 9eward 3eninsula Eskimo if they did not have a type
of their own. The use of foreign patterns in carvings meant for the tourist trade is well
documented especially from the 9eward 3eninsula area 5cf. Hay 1()B7, and -udging from their
large number in the EtholTn collection, kayak carvings seem to have been popular items already
in Hussian America in the mid=1)BB2s.
Iudging from its historical distribution, the ridged deck probably spread to <est Alaska from the
south, i.e., from the Aleutian *slands or the #odiak F Alaska 3eninsula area. 9ince the structural
function of the ridged deck is to prevent the frame from sagging in a swell, it may be assumed to
have developed specifically in the sea kayak, in which it is still solely found. 9ea mammal
hunting held a central position in 9outh Alaskan coastal cultures very early on F in the Kcean
$ay * culture of #odiak and the east coast of Alaska 3eninsula c. @BBB $.". and in the 3aleo=
Aleut culture as early as 6DBB $.". F while the younger <est Alaskan paleoeskimoid cultures
such as 4enbigh 5from c. CCBB $.".7 and "horis still represented a dualistic coastal=inland
foraging strategy. ;orton is the first culture in this line to be considered a developed maritime
hunting culture, and the ;orton expansion to, e.g., ;univak *sland and ;orth Alaska has been
seen as a conse!uence of the adoption of efficient maritime hunting techni!ues and technology,
possibly from the contemporaneous 9outh Alaskan cultures 5cf. 4umond 1(DC0@17.
Although there is no archaeological kayak material older than the previously mentioned finds
from #ar=D16 and 3alutat, ethnographic information suggests that sea kayaks have been used in
9outh Alaska for a very long time. Especially in the Aleutians, institutions such as open sea
navigation, kayak hunting techni!ues and technology, magical rights and taboos related to kayak
hunting, and of course the construction of the kayak itself were highly developed. 8eological
studies have prompted the opinion that the 3aleo=Aleutian culture must have used seaworthy
boats, since the original spread of the culture into the archipelago would otherwise have been
impossible 5$lack 1(D@01C6, 1A(; Aigner 1(D60@C7. +aunal remains from prey species living
only in the open sea suggest that the Aleuts were already practicing open sea kayak hunting
during the earliest Anangula phase 5:aughlin 1()B0A@7. Kn the Alaska 3eninsula, boats are
AA
presumed to have been used for open sea hunting during the Takli Alder phase 5c. @BBB=C.BB
$.".7 because faunal remains include such species as cod, halibut, sea otter, porpoise, and
albatross 54umond 1()B0AB7. The sea kayak or some similar craft may thus have been in use in
9outh Alaska at a very early date indeed, for F as mentioned above F open skin boats have not
been used in historical times as actual hunting craft in <est or 9outh Alaska.
= T'( 6a$k(n7i( ,ro#p
The #ittigauit site in the %ackenie delta has produced kayak parts which date to the 1.th or
16th century A.4.. The find collection includes two wooden end ,horns/, one 5straight7 deck
beam, and a portion of a very thin 5@) x 6 mm7 board which has been identified as a possible
bottom stringer fragment. The shape and measurements of the fragments are reminiscent of the
historical %ackenie type with its curved horns and thin, wide bottom stringers 5%c8hee
1(D@0.@, )@7.
#ittigauit is a coastal site, and the ma-or part of its faunal assemblage 5D(Y=)DY depending on
the stratum7 is beluga; the next most common species is the caribou 5DY=1DY7, while seal
accounts for less than AY in all strata 5op. cit.0A@=A.7. The subsistence system of the local
historical #ittegaryumiut Eskimo F which was based on massive beluga drives F seems thus to
derive at least from the 1.th century, and the %ackenie type kayak may have been used already
at this time specifically for beluga hunting 5op. cit.0(C7. The meager amount of seal in the faunal
material is also concurrent with the historical situation0 due to the productive beluga drive
system, seal hunting does not seem to have played any notable part in the economy; in addition,
we may assume that at least a part of the recovered seals were originally caught on the winter
ice.
The permanent log house villages and high population density of the %ackenie Eskimo F both
uni!ue among "anadian *nuit F have been considered to be linked specifically to the efficient
mass hunting of beluga. Jery little is known about the history of the %ackenie 4elta people
before the beginning of the >beluga period> in the 1.th century; even the connection with the
Thule culture is unclear. Though the finds from the beluga period do contain several artefact
types of Thule and the preceding $irnirk culture 5especially harpoon heads7, %c8hee sees the
finds as being characteried by the flavour of an independent local culture with possible
influences from <est Alaska 5%c8hee, op. cit.0(C=(A; 1(D601(B=1(17.
Kn the basis of his typological analysis, %c8hee considers it possible that the ancestors of the
%ackenie Eskimo represented a pre=Thule inland population which was only superfluously
affected by the Thule expansion but had older, more permanent connections overland with the
<est Alaskan cultures, especially ;orton 5loc. cit.7.
Thus, though the %ackenie kayak or its typological predecessor may have been used for beluga
hunting only since the middle Thule period, it is !uite possible that several of the features which
distinguish the %ackenie kayak from the Thule=derived Arctic types 5e.g., the elevated deck
stringer and the uni!ue bottom construction7 may actually be derived from an older type which
was used by the original inland population and which may also have absorbed some ;orton
features. %c8hee has actually also proposed that the %ackenie kayak may have diffused to its
present area as late as the 1ABB2s directly from <est Alaska 5Arima 1(D.0(17, but the structural
differences between the %ackenie type and the 1.th century find from ;ukleet mentioned
above are so great that this explanation does not seem very probable.
A@
T'( T'r((&!rin,(r(+ Typ(
= T'( Ea! Cana+ian ,ro#p
The origin of the East "anadian and 8reenlandic three=stringered types is also somewhat unclear
as, according to present data, the angular=hulled three=stringer construction does not seem to
have appeared in Alaska either in historic or prehistoric times. The presence of the pattern in the
#oryak and also possibly the Heindeer "hukchi kayaks could indicate that it was formerly also
known on the eastern side of $ering 9trait; so far, however, no concrete evidence of this has been
forthcoming. As noted above, both ,forefathers/ of the Thule culture F 3unuk and $irnirk F seem
to have used round=hulled multiple=stringer pattern kayaks, as did also Thule itself during the
classical phase both in Alaska and in "anada. *t is thus !uite improbable that the triple=stringer
pattern could have been imported to eastern "anada by the Thule people. *t is also e!ually
difficult to imagine that the triple=stringer pattern could have developed from the structurally
superior multiple=stringer pattern used by classical Eastern Thule0 if classical Thule was still
using the multiple=stringer pattern in its caribou kayaks, why should only its sea kayak type have
degenerated to a more primitive three=stringer typeR The possibility that post=expansion eastern
Thule adopted the three=stringer pattern sea kayak as a complete implement from some other
culture seems a more reasonable scenario; however, this raises the !uestion, who could the
original users of the three=stringer type have been and what kind of process could have led to the
adoptionR
The nearest neighbours of the classical phase Thule people in East "anada were the *ndians and
the 4orset people. Adopting a new sea kayak type from the *ndians seems unlikely, for the only
truly maritime *ndian culture known from the area F the %aritime Archaic F had disappeared
some C.BB=ABBB years before the Thule expansion 5cf. Tuck 1(D.0(@; +ithugh 1(DC016A7 and
the *ndians of northeastern "anada during the Thule period were mostly inland caribou hunters.
The exception was the so=called 3oint Hevenge complex, dating from c. .BB A.4. on the central
:abrador coast, which had a coastal=inland subsistence cycle and possibly practiced maritime
hunting; however, the Thule culture seems to have spread to the 3oint Hevenge area only in the
16th century or c. ABB years after the end of the classical phase 5cf. +ithugh 1(DC01CD=), 1AC=A,
1A., 1..=D, 1.(, 16C, 16@, 16D, 1(.7. At this time, the three=stringer pattern was already in use in
the eastern Thule sphere all the way to 8reenland. Also, there is neither ethnographical nor
archaeological evidence that the "anadian *ndians ever used kayaks; during historical times all
East "anadian *ndian tribes used open bark canoes.
A much more probable source is the 4orset culture, which like Thule was an Eskimoid maritime
culture. 9ome time in the third millennium $.". a branch of the <est Alaskan A9Tt complex
1
known as *ndependence * migrated to the Arctic Archipelago and northern 8reenland; its
successors *ndependence **, 9ar!a! 5in 8reenland7, and 3re=4orset spread south, and from them
sprung the actual 4orset culture during the early part of the last pre="hristian millennium. At its
largest, the area of the 4orset culture reached from $anks *sland to central East 8reenland and
south to ;ewfoundland; by the time of the Thule expansion, however, this realm had shrunk and
1
+rom Arctic 9mall Tool tradition. An Alaskan representative of this complex was the
aforementioned 4enbigh, also known by its full name, the "ape 4enbigh +lint "omplex.
A.
the 4orset settlements had apparently totally disappeared from ;ewfoundland and 8reenland
5%ap J**7.
4orset was originally thought to be primarily an inland culture which practiced sea mammal
hunting only during the winter from the ice and possibly did not use boats at all; some authorities
even suggested that the 4orset people were *ndians 5cf. %athiassen 1(A601C(=1AB, passim7. :ater
research has shown that the 4orset people were in fact mainly dependent on maritime hunting
and that both seals and walrus were hunted also in the summer, presumably from boats. Almost
all known 4orset summer dwelling sites are situated on the coast and their analyed faunal
refuse consists mainly of the bones of various seals and walrus. <inter sites are extremely rare;
the 4orset people are thought to have wintered like many historical "entral Eskimo groups, in
snow houses on the sea ice close to polynias or permanent leads fre!uented by sea mammals.
*mplements made from more fragile organic materials such as wood and sealskin are fairly rare
in 4orset finds. ;evertheless, the last few years have produced evidence that the 4orset culture
did indeed know the kayak before the Thule invasion. 'ouse ZD1 from the 4orset site of
;unguvik on ;orthern $affin *sland produced a series of baleen ribs 5+ig. 16b7, apparently from
a toy kayak, which resemble to a remarkable degree the ribs of the historical East "anadian
types. The bottom and sides of the ribs are flat and the bends are sharp; in addition, the
proportions indicate that the vessel had the narrow, deep prow and broad, shallow stern typical of
the historical East "anadian types. 'ouse ZD1 has been dated to the 1Ath century A.4., i.e., to a
time when the Thule culture was still spreading to $affin *sland. 'owever, house ZD6 on the
same site produced a full=sie, albeit broken, angular rib
1
5+ig. 16c7, and the uncalibrated "=1@
dates for this house go back to the @th to 6th century A.4., clearly earlier than the Thule
expansion 5%ary=Housseli[re 1(D(0CC=C67.
The kayak does not seem to have been new to the 4orset culture, for kayak parts dating to the
Early 4orset phase have been found on southern $affin *sland 5cf. %axwell 1().01AD7. *t thus
seems !uite probable that the three=stringered kayak is no newcomer to its present area; on the
contrary, it would rather seem to represent the original local type which has been partly replaced
by a newer import F the Arctic group F but has nevertheless continued its existence as a loaned
feature in the Thule culture and its descendants though its original 4orset users have disappeared
long ago 5cf. also Arima 1()D0DC=D@7.
The reason why the Thule people should have adopted a new kayak model is unclear; however, it
seems that the motive may have had to do with hunting techni!ue. <hen umiak whaling became
common in the $ering 9trait area during the 3unuk phase, the kayak completely disappeared
from the 4iomedes and 9t. :awrence *sland which did not have indigenous populations of
caribou. <e may presume that also Alaskan proto=Thule, upon developing a whaling=based
economy with the help of 3unuk=derived technology, probably also gave up kayak sealing
although retaining the craft for caribou hunting. <ith the onset of the :ittle *ce Age the eastern
Thule population had to follow the central groups and give up whaling, shifting instead very
rapidly to kayak sealing. The adoption of the 4orset type would most likely have happened in
this situation, for several things F e.g., the typological development of Eastern Thule harpoon
heads F indicate that the Thule people may have adopted their kayak sealing techni!ue and
technology directly from the 4orset people 5#ankaanp\\ 1()@01C=1); 1().7.
1
Eugene 1. Arima 5pers. comm.) has expressed doubts about the identification of this find, and after
seeing the ,rib/ at the ;ational %useum of "anada, the present author must concur. The ob-ect
appears too thin and wide to be a kayak rib; it may be part of a box, etc.
A6
= T'( Gr((n%an+ ,ro#p
The development of the 8reenlandic types seems to be connected with the *nugsuk culture 5c.
1ABB=16BB A.4.7 which developed in 8reenland from modified Thule and was specifically
characteried by a highly developed kayak hunting technology. Iudging from finds of kayak
parts and models, early 8reenland kayaks seem to have resembled the East "anadian types with
their three=stringered flat bottoms and simple end structures 5cf. 'oltved 1(@@ 3l. 1D0(7. The
typical modern 8reenland pattern with separate end formers and round ribs only seems to have
come into use during the European contact phase which began in the 16BB2s, for both drawings
in early literary sources 5e.g., bishop Hesen2s chart from 16B., Egede 1D@1, "rant 1DDB7 and
many of the <est 8reenland kayaks brought to Europe by whalers beginning in the 16BB2s 5cf.
;ooter 1(D1; 3etersen 1()6a0.(7 still represent the old pattern of construction.
*n 1(C1 a 4anish expedition found the remains of a kayak of <est 8reenland type in <ashington
:and in ;orth 8reenland 5cf. #och 1(CA0DC7. The find did not contain any imported European
implements typical of the contact phase, but it did produce sealing implements typical of the
*nugsuk culture 5cf. %athiassen 1(C)a01(C=C1C7. *nugsuk, however, was still unknown at this
time, and it was not until the following year that Therkel %athiassen excavated the *nugsuk type
site 5cf. %athiassen 1(AB7. 9ince it was known that the ;orth 8reenland Eskimos had not used
the kayak in the early 1)BB2s and that the historical kayak type of the area was an East "anadian
type imported in the 1)6B2s from $affin *sland, %athiassen in his analysis of the find assumed it
to represent a kayak left by some <est 8reenland Eskimo who had strayed north 5%athiassen
1(C)a0CB(=C117.
According to present archaeological data, however, the kayak was known in ;orth 8reenland
already during the classical Thule phase 5cf. 'oltved 1(@@0CAB, C)@, 3l.@C01., 3l.@(01.7, and it
must also naturally have spread via the Thule district to southern 8reenland, for all migrations
from "anada to 8reenland took place over 9mith 9ound. *mplements such as harpoon >wings>
and line tray parts representing developed *nugsuk kayak sealing technology have been found in
several sites in ;orth 8reenland 5e.g., 'oltved 1(@@0CAB=CAA, 3l.1D01C, 1A; 3l.1)01@=C@; <issler
1(1) +ig. AB7, indicating that the kayak sealing system typical of *nugsuk had also spread to the
Thule 4istrict, although it was later forgotten.
*t is thus !uite possible that the kayak in !uestion was not a <est 8reenland boat brought to the
;orth by chance but a type which was actually in general use in the area during *nugsuk times,
until climatical deterioration caused by the :ittle *ce Age made open sea hunting unproductive
and forced the population to concentrate on land and ice hunting. The special interest of the find,
however, lies in the fact that among the wooden parts was also the whale rib masik which had
lashing holes for three deck stringers; thus, the multiple deck stringer pattern seems to have been
in use at least in <est 8reenland already in the pre=contact *nugsuk period.
*n East 8reenland, the 4orset type may have remained in use longer than on the <est "oast. The
first representatives of early modified Thule arrived in East 8reenland in the 1ABB2s from Elles=
mere *sland via the northern route, without traversing <est 8reenland. <hile the *nugsuk culture
was developing in <est 8reenland, cultural development in East 8reenland followed its own
course until expanding *nugsuk rounded "ape +arewell and spread to East 8reenland around the
turn of the 16th century 5$andi ? %eldgaard 1(.C0C)=A17. A kayak model of undetermined age
from the 9yttenkilometern]sset site in ;ortheast 8reenland clearly resembles the East "anadian
types as it has a strong, high prow and a low, slowly curving stern and also has the cockpit
AD
positioned clearly behind the centre point 5+ig. 1Dc7; indeed, the publisher of the find, Thomas
Thomsen, compares the model to the sea kayak of the Aivilik Eskimo of western 'udson $ay
5Thomsen 1(1D0@1(=@CB7. The same shape can also be found in a model of the old East
8reenland kayak made for Thalbiter in Ammassalik, although the bow is not angular and the
cockpit is almost in the middle 5Thalbiter 1(1C +ig. (B7. The Ammassalik oral tradition
mentioned by Thalbiter about kayaks formerly having three=part ribs probably also relates to
this old type.
A)
SYNTHESIS
T'( Pr('i!ory o) !'( Typ( Gro#p
$ased on the analysis of structural traits, historical kayaks may be divided into three main type
groups0
1. +lat decked kayaks with three hull stringers 5the East "anadian and 8reenlandic types and
the #oryak type7.
C. +lat decked types with multiple hull stringers 5the Arctic types7.
A. Hidged decked types with multiple hull stringers 5the $ering 9ea and 9outh Alaskan types7.
The %ackenie and Heindeer "hukchi types are not directly assignable to any group. The
%ackenie kayak seems to be closest to group C. but has unorthodox end and deck structures.
The Heindeer "hukchi kayak, on the other hand, is too poorly described to be assigned with
confidence to any group; the model shown in Iochelson2s illustration would seem to be closest to
group 1. but the actual pattern of the hull structure remains in doubt.
The relative structural interdependence of features, the complexity of the various designs, and
the geographical distribution of types all would seem to indicate that group 1. represents the
oldest and group A. the youngest structural pattern; group A. could also be construed as a
developed variant of group C. The historical distribution of patterns could be explained through
three waves of features originating in Alaska, but groups 1. and C. might also be seen as
developing independently, one in the east, the other in the west.
The results of the structural feature analysis conform with the archaeological data, although
many details remain unclear due to the paucity of finds.
Kn the basis of archaeological finds, group 1. may be connected with the 4orset culture and is
thus representative of the oldest and most original kayak type in East "anada. "urrent
archaeological data does not tell us whether the three=stringered type was developed in "anada
by 4orset or its predecessors or whether it was imported from Alaska by *ndependence *, for
there are no kayak finds older than the 4orset phase from either area. The possibility that the
three=stringered type might have spread east after the *ndependence * phase is almost nonexistent
since 4orset and its predecessors seem to have had negligible contacts with Alaska before the
Thule expansion. 'owever, there are so many structural similarities between the three=stringered
and the multiple=stringered patterns that totally disconnected origins seem unlikely.
8roup C. is connected with the Thule culture which spread into "anada and 8reenland beginning
in the 11th century and is thus younger at least in the east. *n Alaska the round=hulled, flat=
decked lines of this group go back at least to the Kld $ering 9ea culture, possibly even to
;orton, but earlier than this there is no clear archaeological proof of kayak use from <est and
;orth Alaska although both 4enbigh and "horis seem to have actively practiced maritime
hunting.
*f the 4orset kayak was not an independent invention but was descended from a type brought by
*ndependence * from Alaska, the three=stringer construction may be assumed to represent a
A(
pattern used by the Alaskan A9Tt complex and thus at least by 4enbigh. 'eath considered the
presence of the three=stringer pattern in the #oryak kayak to possibly indicate that the pattern
had spread from the 9ea of Kkhotsk along the Arctic coast to 8reenland 5'eath 1(D)0CC7.
'owever, according to 9oviet archaeologists the maritime adaptation of the #oryak did not take
place until c. the beginning of the "hristian era 5Jasilievski ? 8olub-ev 1(D601D@7. 9ince the
Eskimo settlements of 9iberia at this time belonged to the sphere of the Kld $ering 9ea culture
5which used round=bottomed kayaks7, the #oryak could not have adopted the three=stringer
pattern from the Eskimo in the course of their maritime adaptation although most of their other
maritime hunting paraphernalia does seem to be of Eskimo origin 5cf. 8urvich 1(D(0A@7.
The only viable explanation for the preservation of the pattern in the #oryak kayak would seem
to be a much earlier spread F perhaps even from 4enbigh F of the three=stringered kayak to the
inland tribes of the "hukchi 3eninsula, and its use in this area possibly as a reindeer hunting boat
up to the time of the maritime adaptation. The reindeer kayak of the Heindeer "hukchi might
thus actually be closer to the original A9Tt kayak than the #oryak pattern, which makes the
paucity of data on this type all the more lamentable.
The origin of group A. must still remain open since archaeological kayak finds from <est and
9outh Alaska do not date back further than the 1.th century A.4., and the same basically applies
also to group C. Although there are grounds for assuming that the three=stringer construction may
represent the oldest pattern also in ;orth and <est Alaska, both the ridged deck and the multiple=
stringer hull may have actually originated in the Alaska 3eninsula F #odiak *sland area or the
Aleutian *slands where the maritime economic pattern has much older traditions. +or the same
reason, it is also impossibly to determine which group has chronological priority.
T'( Ori,in o) !'( Typ( Gro#p
%any inland tribes of Alaska and ;orthwest "anada have used so=called kayak=form canoes in
which the bark or skin sheathing is supported by a compound lens=shaped bottom frame instead
of the separate slats usually found on *ndian canoes. The transverse profile of these kayak=form
canoes is similar to that of the East "anadian kayaks, for the bottom and sides are flat and are
-oined at a sharp angle; like the :abrador and 8reenland kayaks, these canoes also have a distinct
corner at the foot of the bow and stern 5cf. Adney ? "hapelle 1(6@01.)=16)7.
The #utenai and 9huswap *ndians of $ritish "olumbia and <ashington 9tate and the 8olds of
the Amur Hiver in 9iberia used another pattern, the >pointed> canoe, which had a pro-ection
resembling the ram of an old warship at either end. The internal frame of these canoes was
constructed of very thin wooden stringers and ribs which were tied together with lashings. The
form of these craft differs notably from both the round=ended *ndian canoes and the kayaks, but
the structure of the bottom frame is nevertheless !uite close to that of the multiple=stringered
kayaks 5cf. %ason ? 'ill 1(B1; Adney ? "hapelle 1(6@016)=1DA7. 'ornell presumed that the
kayak developed from the >pointed> canoe through an extension of the sheathing covering the
fore and aft pro-ections toward the centre to form an actual deck 5'ornell 1(@601D(=1)B7.
&nfortunately, this evolutionary theory does not explain the origin of the kayak gunwales since
the thin gunwale strips of the canoe must F following 'ornell2s scheme F have gradually
developed into the cockpit coaming and the hull itself would have become a rounded cigar
supported by hooplike ribs.
@B
Although the three=stringer pattern seems more >archaic>, this does not necessarily mean that it
is the direct ancestor of the multiple=stringer pattern. The oldest 9outh Alaskan Eskaleut cultures
and the <est=;orth Alaskan A9Tt do not seem to have been in close contact before the ;orton
period 5cf. 4umond 1(DD0)67, and it is thus !uite possible that the two patterns emerged through
parallel development already before the eastern spread of the A9Tt.
The multiple=stringered kayak may have been developed by the 3aleo=Aleut culture on the basis
of a 9iberian multiple=stringered open canoe as early as c. DBBB=6BBB $.". and could have
spread to the mainland c. @BBB $.". with the Kcean $ay culture. Through chance contacts
between the Kcean $ay people and A9Tt groups fre!uenting the Alaska 3eninsula, the idea 5if
not the actual design7 of the kayak may have been transferred to the A9Tt sphere; archaeological
evidence of A9Tt on the Alaska 3eninsula dates back at least to c. 1(BB $.". 5cf. 4umond
1))101CB=1A17. 'owever, as there were no actual trade contacts between the two groups, the
A9Tt people may not have adopted the 9outh Alaskan multiple=stringered pattern as such, but
may instead have developed their own three=stringered type by adding a full deck to a >kayak=
form> river canoe type which they were already using.
The spread of the multiple=stringer pattern to <est and ;orth Alaska would most likely have
been set into motion by the rise of maritime hunting and the closer contacts between <est and
9outh Alaska which characterie the ;orton stage 5cf. 4umond 1(D.016(=1DB; #. <orkman
1(DD01@; <. <orkman 1()B0)); 1()C, passim7. These trends would have facilitated the diffusion
of the 9outh Alaskan kayak west of the Alaska 3eninsula, and we may assume that the multiple=
stringer pattern had reached 9eward 3eninsula by the beginning of the Kld $ering 9ea phase and
;orth Alaska by the beginning of $irnirk. 'owever, though now extinct in Alaska, the three=
stringer pattern survived in relative isolation in "anada and possibly 9iberia, where it had spread
already during the A9Tt phase.
The ridged deck may also have originated in the Aleutians or 9outh Alaska fairly early on,
though F in accordance with the scenario presented above F hardly before the ;orton phase. *n
any event, it seems to have spread to <est Alaska fairly late, probably only toward the end of the
first millennium A.4. or slightly later, since the historical northern limit of this pattern 59eward
3eninsula7 coincides with the southern limit of umiak whaling.
&miak whaling became the dominant mode of subsistence in the $ering 9trait area during the
time of the 3unuk and Thule cultures, starting c. )BB A.4.. Knce in full swing, it rendered kayak
sealing almost superfluous; the kayak probably remained in use mainly as a caribou hunting boat
and seaworthiness was conse!uently no longer a prere!uisite. *f the ridged deck had spread to
<est Alaska before the 3unuk stage, it probably would also have spread all the way to ;orth
Alaska since kayak sealing seems to have played an important part in the subsistence economy
of the $irnirk culture inhabiting the northern coasts. 'owever, due to the success of whaling, the
3unuk and Thule cultures no longer had use for a new sea kayak type. Thus, the spread of the
ridged=decked type stopped at the southern limit of whaling on 9eward 3eninsula and the ;orth
Alaskan caribou kayak retained its flat deck.
@1
FIGURE CREDITS
+ig. 1.
a. The author, after 3etersen 51()6b0 A)=A(7
b. The author, after Arima 51()D0 +ig. CD7
c. Eimmerly ? 9kean 5Arima 1()D0 +ig. )7
+ig. C.
a. Arima 1(D.0 +ig. C1 5text removed7
b. Eimmerly 1(D(0 xvii 5text removed7
c. The author, after :angsdorff 51)1C7
+ig. A.
a. The author, after Iensen 51(D.0 Tavle *J, 9nit $7
b. The author, after Arima 51()D0 +ig. CD7
c. The author, after "hapelle 51(6@0 +ig. 1D67
d. The author, after Eimmerly ? $arker 5Arima 1()D0 +ig. D7
+ig. @.
a. Arima 1(D.0 +ig. CC 5text removed7
b. The author, after Taylor 51(D@0 +ig. CC7
c. The author, after Eimmerly 51(D(0 +ig. DA7
d. The author, after "hapelle 51(6@0 +ig. 1D(7
+ig. ..
a. The author, after "hapelle 51(6@0 +ig. CB)7
b. The author, after Iensen 51(D.0 Tavle **7
c. The author after "hapelle 51(6@0 +ig. CB17
d. 'oltved 1(6D0 +ig. ...
+ig. 6.
a. The author, after Arima 51()D0 +ig. AB7
b. The author, after "hapelle 51(6@0 +ig. 1(D7
c. The author, after Arima 51()D0 +ig. A.7
d. The author, after Arima 51()D0 +ig. @B7
+ig. D.
a. The author, after Arima 51()D0 +ig. @67
b. The author, after "hapelle 51(6@0 +ig. 1((7
+ig. ).
The author, after Eimmerly ? $arker 5Arima 1()D0 +ig. D7
+ig. (.
a. The author, after "hapelle 51(6@0 +ig. 1(A7
b. The author, after Arima 51()D0 +ig. 1C7
c. The author, after "hapelle 51(6@0 +ig. 1(C7
d. The author, after "hapelle 51(6@0 +ig. 1)67
@C
+ig. 1B.
a. The author, after "hapelle 51(6@0 +ig. 1)17
b. The author, after "hapelle 51(6@0 +ig. 1)C7
c. The author, after "hapelle 51(6@0 +ig. 1)B7
d. The author, after "hapelle 51(6@0 +ig. 1DD7
+ig. 11
a. The author, after "hapelle 51(6@0 +ig. 1D)7
b. The author, after "hapelle 51(6@0 +ig. 1D(7
+ig. 1C.
$ogoras 1(B@=B(0 +ig. @Dc.
+ig. 1A.
Iochelson 1(B.=B)0 +ig. )A.
+ig. 1@.
$ogoras 1(B@=B(0 +ig. @Db.
+ig. 1..
a. +ord 1(.(0 +ig. D(a=b.
b. 8iddings 1(6@0 3l. C).
+ig. 16.
a. %axwell 1()A0 +igs C=A.
b. %ary=Housseli[re 1(D(0 +ig. @.
c. %ary=Housseli[re 1(D(0 +ig. ..
+ig. 1D.
a. +ord 1(.(0 +ig. D)a=c.
b. "ollins 1(AD0 3l. )A06.
c. Thomsen 1(1D0 3l. SSJ0..
All maps by the author.
@A
@@
@.
@6
@D
@)
@(
.B
.1
.C
.A
.@
..
.6
.D
.)
.(
6B
61
KAYAK 6EASURE6ENTS AND LENGTH89IDTH RATIO TABLES
%easurements used in scatterplots a and b
area 5a^average7 l x w cm lGw source
Gr((n%an+ !yp(
Ea! Gr((n%an+
5a7 .). x @) 1C,C Iensen 1(D.0CB
..B x @. 1C,C Thalbiter 1(1C0A)C
... x @6 1C,1 Thalbiter 1(1C0A)C
.AD x @D 11,@ Thalbiter 1(1C0A)C
So#!' Gr((n%an+
.(. x .1 11,D Adney ? "hapelle 1(6@0+CB)
+redriksdal .A1 x .A 1B,B Iensen 1(D.0T *J
;anortalik 5a7 ... x @( 11,A 3etersen 1()6a0@@
IulianehMb 5a7 ..) x .1 1B,( Iensen 1(D.0CB
IulianehMb .A. x .A 1B,1 3etersen 1()6a0+@@
9(! Gr((n%an+
+redrikshMb 5a7 .C) x @( 1B,) Iensen 1(D.0CB
+redrikshMb .A@ x .B 1B,D Iensen 1(D.0C6
8odthMb 5a7 .CC x .B 1B,@ Iensen 1(D.0CB
8odthMb 5a7 .C6 x @( 1B,D 3etersen 1()6a0@A
8odthMb 5a7 .AC x .B 1B,6 3etersen 1()6a0@A
8odthMb 5a7 .1A x .B 1B,A 3etersen 1()6a0@A
8odthMb 5a7 .A@ x .B 1B,D 3etersen 1()6a0@A
8odthMb 5a7 .1B x @) 1B,6 3etersen 1()6a0@A
9ukkertoppen 5a7 .11 x .1 1B,B 3etersen 1()6a0@A
9ukkertoppen 5a7 .1. x @( 1B,. Iensen 1(D.0CB
9ukkertoppen 5a7 .1. x .1 1B,1 3etersen 1()6a0@@
#angaamiut 5a7 .AA x @D 11,A 3etersen 1()6a0@A
#angaamiut .AB x @( 1B,) Iensen 1(D.0CC
'olsteinsborg 5a7 .C1 x .C 1B,B Iensen 1(D.0CB
'olsteinsborg 5a7 .C. x .A (,( 3etersen 1()6a0@A
'olsteinsborg 5a7 .AB x .1 1B,@ 3etersen 1()6a0@A
'olsteinsborg 5a7 .1A x .1 1B,1 3etersen 1()6a0@A
Egedesminde @6) x .A ),) Iensen 1(D.0C.
Egedesminde .1A x 6B ),6 $irket=9mith 1(C@0C6)
Egedesminde .AD x 66 ),1 $irket=9mith 1(C@0C6)
Egedesminde .A) x 6. ),A $irket=9mith 1(C@0C6)
Egedesminde .AD x 61 ),) $irket=9mith 1(C@0C6)
Iakobshavn @(D x .B (,( 3etersen 1()6a0@@
&pernivik 5a7 .C@ x .C 1B,1 3etersen 1()6a0@@
.C1 x @( 1B,6 Adney ? "hapelle 1(6@0+CBD
.BA x .1 (,( Adney ? "hapelle 1(6@0+CB6
@(D x .A (,@ author2s data 5+;%7
Thule .(. x .@ 11,B 'oltved 1(6D0D)
6C
Ea! Cana+ian !yp(
Nor!' Gr((n%an+
R .A@ x .6 (,. Adney ? "hapelle 1(6@0+CB.
6BC x .) 1B,@ Adney ? "hapelle 1(6@0+CBB
@(1 x .6 ),) Adney ? "hapelle 1(6@0+1((
.B@ x .6 (,B 3etersen 1()6a0.@
.6B x .6 1B,B $oas 1(B11BD01A
.). x .6 1B,@ 'oltved 1(6D0D)
.AB x .@ (,) 9teensby 1(1B0A.(
9(!(rn H#+on Bay & Ba))in I%an+
*glulik .6( x .( (,6 %athiassen 1(C)a0(@
*glulikR 6.) x D1 (,A Arima 1()D0+AB
.@C x .( (,C Arima 1()D016A
3ond *nlet .). x 6C (,@ %athiassen 1(C)a0(C
"ape 4orset 6)6 x 6@ 1B,D Adney ? "hapelle 1(6@0+1(6
"umberland 9ound 6BB x .D 1B,. $oas 1(B1=BD0(,1A
9avage *slands 6@B x 6C 1B,A $oas 1(B1=BD01A
R 61. x 6D (,C Arima 1()D0+AC
La"ra+or
666 x .( 11,A Adney ? "hapelle 1(6@0+1()
DAA x .( 1C,@ Adney ? "hapelle 1(6@0+1(D
*vuyivik DAD x 6@ 11,. Arima 1(6@0C@B
*vuyivik DBC x 6( 1B,C Arima 1(6@0C@B
&ngava $ayR DDD x 6D 11,6 Arima 1()D0+A6
"ape 9mith D1B x D1 1B,B Arima 1()D0+A(
Ea!(rn H#+on Bay
@)) x D. 6,. Arima 1()D0+@1
.@( x D1 D,D Arima 1()D0+@B
B(%$'(r I%an+
.6@ x DC D,) Arima 1()D0+@C
6(A x D. (,C Arima 1()D0+@6
@() x D( 6,A +reeman 1(6@0DB
Ar$!i$ !yp(
Cari"o# Ekimo 5length without >horns>7
.DB x @6 1C,@ Adney ? "hapelle 1(6@0+1(A
.AB x @6 11,. Arima 1()D0+C1
6CB x @D 1A,C Arima 1()D0+CB
66D x @6 1@,. Arima 1(D.0+CB
6.B x @D 1A,) $oas 1(B1=BD0D6
@C( x A( 11,B $irket=9mith 1(C(0*,1)D
6A
.11 x @@ 11,6 $irket=9mith 1(C(0*,1)D
N(!i%ik
..C x @6 1C,B Adney ? "hapelle 1(6@0+1(@
616 x @6 1A,@ Arima 1()D0+1C
6A( x @6 1A,( Taylor 1(D@0111
616 x @. 1A,D Taylor 1(D@0111
.)( x @D 1C,. Taylor 1(D@0111
Copp(r Ekimo
.D6 x @D 1C,A Adney ? "hapelle 1(6@0+1(C
D11 x @B 1D,) Arima 1()D0+11
6.) x A( 16,D Ienness 1(@601@B
Nor!' A%aka
"ape $arrow .C6 x @D 11,C Adney ? "hapelle 1(6@0+1)D
#otebue 9ound .CD x @D 11,C Adney ? "hapelle 1(6@0+1)6
"ape #rusenstern .C6 x @6 11,@ ;elson 1)((0CC1
@.) x @6 1B,B Ienness 1(@601@B
@B@ x @6 ),) Ienness 1(@601@B
6a$k(n7i( !yp(
@)) x @( 1B,B Adney ? "hapelle 1(6@0+1))
@@A x @( (,B Arima 1()D0+)
.B1 x @) 1B,@ Arima 1()D0+D
A(B x @) ),1 Arima 1()D0+.
@D) x .B (,6 Ienness 1(@601@B
@)A x @6 1B,1 Ienness 1(@601@B
B(rin, S(a !yp(
S(:ar+
"ape Espenberg @AD x 61 D,C ;elson 1)((0CC1
#ing *sland @6) x 6. D,C Adney ? "hapelle 1(6@0+1)1
#ing *sland @6. x DC 6,. ;elson 1)((0CCB
Nor!on
.1) x 6B ),6 Adney ? "hapelle 1(6@0+1)C
.1B x DB D,A ;elson 1)((0CCB
N#ni*ak
'ooper $ay @61 x D) .,( Eimmerly 1(D(0+DA
;univak @.6 x D6 6,B Adney ? "hapelle 1(6@0+1)B
;univak @6B x D@ 6,C ;elson 1)((0C1(
;univak @6B x D@ 6,C ;elson 1)((0C1(
Bri!o% Bay ;<=
@6B x D@ 6,C Adney ? "hapelle 1(6@0+1DD
6@
So#!' A%akan !yp(
A%(#!
.@B x .1 1B,6 Adney ? "hapelle 1(6@0+1D)
+ox *slands .@C x @6 11,) Author2s data 5EtholTn coll.7
.6A x @C 1A,@ :-apunova 1(6@0CC(
.@B x .A 1B,C :-apunova 1(6@0CAB
.6C x @D 1C,B :-apunova 1(6@0CA1
6.
Kayak hull measurements by use environment
35
40
45
50
55
60
65
70
75
80
350 400 450 500 550 600 650 700 750 800
Length cm
W
i
d
t
h

c
m
Open sea
Sparse floe
Dense floe
Estuary
River
Ratio 1:8
Ratio 1:12
Wit! 48
Kayak hull measurements by hunting method
35
40
45
50
55
60
65
70
75
80
350 400 450 500 550 600 650 700 750 800
Length cm
W
i
d
t
h

c
m
Stealt!
"!ase
Wit! 48 #$
66
App(n+i>
Comm(n! on !'( !a"%( an+ $a!!(rp%o!
As may be seen from the table of measurements and lengthGwidth ratios and the scatterplots, the
various kayak types form three distinct groups in which certain measurements correlate with
certain environmental factors0
1. Types used in pinniped hunting 5stealth techni!ue7 among ice floes 5eastern 'udson $ay,
$elcher islands
1
, and all $ering 9ea types7 are usually @.@ to ..6 metres long and c. 6) to )B
centimetres wide; the range of deviation in length 5_1CY from the median7 is moderate and that
in width 5_)Y from the median7 is small. The lengthGwidth ratio varies between c. 601 and c. )01
and the range of deviation is moderate 5_1@Y7.
C. Types used in open water 59outh 8reenland, :abrador, Aleutians
C
7 and relatively ice free
areas 5the rest of 8reenland, western 'udson $ay ? $affin *sland7 are c. @.D to D.) metres long
and @. to D1 centimetres wide; the deviation both in length 5_C.Y7 and width 5_CC..Y7 is fairly
high. Though the lengthGwidth ratio varies in the main between )..01 and 1C01, the deviation
5_1DY7 is moderate compared to that of the actual measurements.
A. Types used mainly for caribou hunting 5chase techni!ue7 in inland and estuarine waters 5the
Arctic types7 are c. @ to D metres long without the ,horns/ and c. A( to @D centimetres 5mostly @6
to @D centimetres7 wide; the deviation in length is very high 5_CDY7 while the deviation in width
is very low 5_(Y, in (BY of the examples only c. _CY7. The lengthGwidth ratio varies between
(01 and 1D01, the deviation being very high 5_A1Y7.
The structuring force of the environment seems to affect the absolute measurements and their
ratios differently in different groups0
8roup 1 is characteried by the effect of two factors F the ice=filled sea and the stealth hunting
method F on the absolute length and width measurements respectively and separately; the
smallish deviation in the lengthGwidth ratio is an outcome of the limited deviation in the absolute
measurements themselves, not of the optimality of any specific ratio, as may be gathered from
the fact that the cluster in the scatter diagram is non=directional 5or, if anything, vertical rather
than diagonal7. The shape of the cluster indicates that the demands structuring the measurements
F agility and stability F do not conflict with one another; both may be optimised without
-eopardiing the other and compromises have been unnecessary.
8roup C, on the other hand, is specifically characteried by the lengthGwidth ratio, which has a
deviation only slightly larger than in group 1 though the deviation in the actual measurements is
double to almost triple. The importance of the ratio is evident from scatterplot 1, where the
cluster formed by the open sea and relatively ice free groups is clearly directional in spite of the
high degree of measurement deviation. Kne receives the impression that the demands structuring
the absolute measurements = sturdiness and stability on the one hand, good tracking and speed on
the other = are in mutual conflict and have had to be compromised through an optimum ratio
which has been maintained in spite of changes in one or other of the actual measurements.
1
The 6(A cm long kayak in Arima 1()D +ig. @6 is atypical and has been omitted from the
scatterplot and calculations.
C
&nfortunately, no measurements were available for full=sie 3acific Eskimo baidarkas.
6D
8roup A represents a cluster where one absolute factor F width F is critical while the other F
length F tolerates a high degree of variation. This phenomenon is even more apparent in
scatterplot C, where types used in chase type hunting F i.e., caribou drives 5Arctic7, beluga drives
5%ackenie7, and sea otter hunts 5Aleut7 F are distinguished from types used in stealth type
pinniped hunting. %inimiing width is clearly the most important goal; both the maximum and
the minimum measurements depend on the width of the user2s own pelvis. :ength, and with it
the lengthGwidth ratio, varies !uite strongly, presumably in reaction both to specific differences
in the use context and to aesthetic views 5e.g., the ,horns/7
1
; on the other hand, the lengthG width
ratio is usually over 1101, i.e., even short kayaks are fairly narrow in shape. The horiontal
cluster gives the impression that the measurements of these kayaks are structured mainly by one
specific demand F speed F and the main limiting factor is the user2s own girth.
1
*t may be noted that of the kayaks identified as ,chase/ types in scatterplot C, the %ackenie
kayaks, the ;orth Alaskan kayaks 5also used in ice lead sealing7, and the Aleut kayaks, which were
all used at least to some extent on the sea, fall at the >short> end of the cluster within the )..=1C
ratio margin, while dedicated inland types such as the ;etsilik and "opper Eskimo kayaks are
generally longer.
6)
REFERENCES
Unp#"%i'(+0
'EAT', IK'; 4. 51(DB7 A "omparative 9tudy of #ayaks. ;ational %useum of "anada,
Kttawa.
#A;#AA;3``, IAH%K 51()@7 #ulttuureiden v\linen vuorovaikutus arktisessa #anadassa
1BBB F 1.BB A.4. N"ultural interaction in Arctic "anada 1BBB F 1.BB A.4.O 9eminar
paper. 'elsinki &niversity 4ept. of Archaeology.
P#"%i'(+0
A4;E1, E4<*; TA33A; F "'A3E::E, 'K<AH4 * . 51(6@7 The $ark "anoes and 9kin
$oats of ;orth America. 9mithsonian *nstitution, &nited 9tates ;ational %useum.
$ulletin CAB, %useum of 'istory and Technology. <ashington.
A*8;EH, IEA; 9. 51(D67 Early 'olocene Evidence for the Aleut %aritime Adaptation. F
Arctic Anthropology 1A 5C7. %adison.
AH*%A, E&8E;E 1. 51(6@7 ;otes on the kayak and its e!uipment at *vuyivik, 3.W. F ;ational
%useum of "anada $ulletin 1(@, "ontributions to Anthropology 1(61=6C 3art **.
Kttawa.
aaaaaaaa 51(D.7 A "ontextual 9tudy of the "aribou Eskimo #ayak. "anadian Ethnology
9ervice 3aper ;o. C.. ;ational %useum of %an %ercury 9eries. Kttawa.
aaaaaaaa 51()D7 *nuit #ayaks in "anada0 a Heview of 'istorical Hecords and "onstruction.
"anadian Ethnology 9ervice 3aper ;o. 11B. ;ational %useum of "anada %ercury
9eries. Kttawa.
AH&T1&;KJ, 9.A. F :EJ*;, %.8. F 9EH8EIEJ, 4.A. 51(6@7 Ancient $urials on the
"hukchi 3eninsula. F %*"'AE:, 'E;H1 ;. 5ed.7 The archaeology and
geomorphology of ;orthern Asia. Arctic *nstitute of ;orth America, Anthropology of
the ;orth, Translations from Hussian 9ources .. Toronto.
$A;4*, 'A;9=8EKH8 ? %E:48AAH4, IbH8E; 51(.C7 Archaeological investigations on
"lavering b, ;ortheast 8reenland. <ith an Anthropological Appendix by H. Hay and I.
$alslev IQrgensen. %eddelelser om 8rQnland 1C( 5@7. #Qbenhavn.
$*;+KH4, :E<*9 H. 51(D)7 ;unamiut Ethnoarchaeology. 9tudies in Archeology. ;ew 1ork.
$*H#ET=9%*T', #AI 51(C@7 Ethnography of the Egedesminde 4istrict, with aspects of the
general culture of <est 8reenland. %eddelelser om 8rQnland $d. :SJ*. #Qbenhavn.
aaaaaaaa 51(C(7 The "aribou Eskimos, %aterial and 9ocial :ife and their "ultural
3osition. *. 4escriptive 3art; **. Analytical 3art. Heport of the +ifth Thule Expedition
1(C1=C@, Jol. J. "openhagen.
6(
aaaaaaaa 51(.A7 The "hugach Eskimo. ;ationalmuseets 9krifter, Etnografisk H]kke J*.
#Qbenhavn.
$:A"#, HK$EHT +. 51(D@7 8eology and the Ancient Aleuts, Amchitka and &mnak *slands,
Aleutians. F Arctic Anthropology 6 5C7. %adison.
$KA9, +HA;E 51)))7 The "entral Eskimo. F 9ixth Annual Heport of the $ureau of
Ethnology 1))@=2).. <ashington.
aaaaaaaa 51(BD7 The Eskimo of $affin :and and 'udson $ay *=**. $ulletin of the
American %useum of ;atural 'istory Jol. SJ. ;ew 1ork.
$K8KHA9, <A:4E%AH 51(B@=(7 The "hukchee. The Iesup ;orth 3acific Expedition.
%emoir of the American %useum of ;atural 'istory Jol. J**. ;ew 1ork.
":AH#, 4K;A:4 <. 51(D@7 "ontributions to the :ater 3rehistory of #odiak *sland, Alaska.
Archaeological 9urvey of "anada, 3aper ;o. CB. ;ational %useum of %an %ercury
9eries. Kttawa.
"K::*;9, 'E;H1 $. 51(AD7 Archaeology of 9t. :awrence *sland, Alaska. 9mithsonian
%iscellaneous "ollections Jol. (6. <ashington.
"HA;TE, 4AJ*4 51D).7 'istoria om 8rQnland. $iblioteket for %enniskesl]gtens 'istorie,
vol. *J. "openhagen.
"&HT*9, E4<AH4 9. 51(AB7 The ;orth American *ndians vol. CB. 51. reprint 1(DB7. ;ew 1ork
? :ondon.
4&%K;4, 4K; E. 51(DC7 The Alaska 3eninsula in Alaskan 3rehistory. F JK8ET, +.<. ?
9TE3'E;9K;, H.:. 5eds.7 +or The "hief0 Essays in 'onor of :uther 9. "ressman.
&niversity of Kregon Anthropological 3apers ;o. @. Eugene.
aaaaaaaa 51(D.7 "oastal adaptation and "ultural "hange in Alaskan Eskimo 3rehistory. F
+*TE'&8', <*::*A% 5ed.7 3rehistoric %aritime adaptations of the "ircumpolar
Eone. The 'ague F 3aris.
aaaaaaaa 51(DD7 The Eskimos and Aleuts. :ondon.
aaaaaaaa 51()B7 A "hronology of ;ative Alaskan 9ubsistence 9ystems. F #KTA;*,
1K9'*;K$& ? <KH#%A;, <*::*A% $. 5eds.7 Alaska ;ative "ulture and
'istory. 9enri Ethnological 9tudies no. @. Ksaka.
aaaaaaaa 51()17 Archaeology on the Alaska 3eninsula0 The ;aknek Hegion, 1(6B=1(D..
&niversity of Kregon Anthropological 3apers ;o. C1. Eugene.
aaaaaaaa 51()C7 Trends and Traditions in Alaskan 3rehistory0 The 3lace of the ;orton
"ulture. F Arctic Anthropology 1( 5C7. %adison.
419K;, 8EKH8E 51()67 $aidarka. Edmonds.
DB
E8E4E, 'A;9 51D@17 4et 8amle 8rQnlands ;ye 3erlustration eller ;aturel='istorie.
"openhagen.
+*TE'&8', <*::*A% 51(DC7 Environmental Archaeology in 'amilton *nlet, :abrador.
9mithsonian "ontributions to Anthropology ;o. 16. <ashington.
+KH4, IA%E9 A. 51(.(7 Eskimo 3rehistory in the Jicinity of 3t. $arrow, Alaska.
Anthropological 3apers of the American %useum of ;atural 'istory @D 517. ;ew 1ork.
+HEE%A;, %*:TK; %.H. 51(6@7 Kbservations on the #ayak "omplex, $elcher *slands,
;.<.T. F ;ational %useum of "anada $ulletin 1(@, "ontributions to Anthropology
1(61=6C 3art **. Kttawa.
8A4, +*;; 51()@7 8rQnland. 3olitikens 4anmarkshistorie. crhus.
8E*9T, KTTK <*::*A% ? HA*;E1, +HKE:*"' 8. 51(A67 Archaeological Excavations at
#ukulik, 9t. :awrence *sland, Alaska. %iscellaneous 3ublications of the &niversity of
Alaska Jol. **. <ashington.
8*44*;89, IA%E9 :K&*9 51(6@7 The Archaeology of "ape 4enbigh. $rown &niversity
$icentennial 3ublications, 9tudies in the +ields of 8eneral 9cholarship. 3rovidence.
aaaaaaaa 51(6)7 Ancient %en of the Arctic. :ondon.
8&E%3:E, 4.:. 51(6D7 The 3acalik #ayak of the $elcher *slands. F ;ational %useum of
"anada $ulletin CB@, "ontributions to Ethnology J. Kttawa.
8&HJ*"', *.9. 51(D(7 An ethnographic study of cultural parallels among the aboriginal
populations of northern Asia and northern ;orth America. F Arctic Anthropology 16 517.
%adison.
'A<#E9, E.<. 51(167 The :abrador Eskimo. "anada 4epartment of %ines 8eological 9urvey
%emoir (1.; ;o. 1@. Anthropological 9eries. Kttawa.
'EAT', IK'; 4. 51(D)7 9ome comparative notes on kayak construction. F E*%%EH:1,
4AJ*4 <. 5ed.7 "ontextual 9tudies of %aterial "ulture. "anadian Ethnology 9ervice
3aper ;o. @A. ;ational %useum of %an %ercury 9eries. Kttawa.
'K:%, 8&9TAJ 51(117 Ethnological 9ketch of the Angmagsalik Eskimo. F T'A:$*TEEH,
<*::*A% 5ed.751(1@7 The Ammassalik Eskimo. %eddelelser om 8rQnland $d.
SSS*S. #Qbenhavn.
'K:TJE4, EH*# 51(@@7 Archaeological *nvestigations in the Thule 4istrict. *. 4escriptive
3art. %eddelelser om 8rQnland 1@1 517. #Qbenhavn.
aaaaaaaa 51(6D7 "ontributions to 3olar Eskimo Ethnography. %eddelelser om 8rQnland
1)C 5C7. #Qbenhavn.
'KH;E::, IA%E9 51(@67 <ater transport, origins ? early evolution. "ambridge.
D1
*;89TA4, 'E:8E 51(.@7 ;unamiut. $landt Alaskas *nlands=Eskimoer. #Qbenhavn.
IA"K$*, A. 5bearb.7 51(AD7 "arl 'einrich %ercks ethnographische $eobachtungen dber die
JVlker des $eringsmeeres 1D)( F (1. F $aessler Archiv vol. CB. $erlin.
IE;9E;, 3. 9"AJE;*&9 51(D.7 4en grQnlandske ka-ak og dens redskaber. #Qbenhavn.
IE;;E99, 4*A%K;4 51(@67 %aterial "ulture of the "opper Eskimo. Heport of the "anadian
Arctic Expedition 1(1A F 1), vol 16. Kttawa.
IK"'E:9K;, <A:4E%AH 51(B.=B)7 The #oryak. The Iesup ;orth 3acific Expedition.
%emoir of the American %useum of ;atural 'istory vol. J*. ;ew 1ork.
#A;#AA;3``, IAH%K 51().7 'arpoon heads, kayaks, and the 4orset = Thule succession.
F 9uomen Antropologi 1G1().. 'elsinki.
#K"', :A&8E 51(CA7 Hesultaterne af Iubil]umsekspeditionen ;ord om 8rQnland i 1(C1. F
;aturens Jerden, +eb. 1(CA. #Qbenhavn.
de :A8&;A, +HE4EH*"A 51(.67 "hugach 3rehistory, The Archaeology of 3rince <illiam
9ound, Alaska. &niversity of <ashington 3ublications in Anthropology Jol. 1A. 9eattle.
von :A;894KH++, 8EKH8 'E*;H*"' 51)1C7 $emerkungen auf einer Heise um die <elt in
den Iahren 1)BA bis 1)BD. +rankfurt am %ayne.
:A;T*9, %AH8AHET 51(@67 The 9ocial "ulture of the ;univak Eskimo. Transactions of the
American 3hilosophical 9ociety, n.s. vol. A. pt. A. 3hiladelphia.
aaaaaaaa 51(DC7 +actionalism and :eadership0 A "ase 9tudy of ;univak *sland. F Arctic
Anthropology ( 517. %adison.
:A&8':*;, <*::*A% 9. 51(6C7 $ering 9trait to 3uget 9ound0 dichotomy and affinity
between Eskimo=Aleuts and American *ndians. F "A%3$E::, IK'; %. 5ed.7 3re=
historic "ultural Helations between the Arctic and Temperate Eones of ;orth America.
Arctic *nstitute of ;orth America, Technical 3aper no. 11. %ontreal.
aaaaaaaa 51()B7 Aleuts, 9urvivors of the $ering :and $ridge. ;ew 1ork.
:IA3&;KJA, H.8. 51(6@7 Aleutskie baidarki. F 9bornik %ue-a Antropologii i Etnografii
SS**. %oskva = :eningrad.
%AH1=HK&99E:*EHE, 8&1, K.%.*. 51(D(7 A +ew 3roblems Elucidated... and ;ew Wuestions
Haised by Hecent 4orset +inds in the ;orthern $affin *sland Hegion. F Arctic, vol. AC
no. 1. "algary.
%A9K;, KT*9 T. ? '*::, %EH*4E; 9. 51(B17 3ointed $ark "anoes of the #utenai and
Amur. F 9mithsonian Annual Heport for 1)((. <ashington.
%AT'*A99E;, T'EH#E: 51(C)a7 %aterial "ulture of the *glulik Eskimos. Heport of the +ifth
Thule Expedition 1(C1=C@, Jol. J* ;o. 1. "openhagen.
DC
aaaaaaaa 51(C)b7 Eskimo Helics from <ashington :and and 'all :and. F %eddelelser om
8rQnland $d. :SS* 51(C(7. #Qbenhavn.
aaaaaaaa 51(AB7 *nugsuk, a %ediaeval Eskimo 9ettlement in &pernivik 4istrict, <est
8reenland. F %eddelelser om 8rQnland $d. :SSJ**. #Qbenhavn.
aaaaaaaa 51(A67 The Eskimo Archaeology of Iulianehaab 4istrict, with a brief summary of
the prehistory of the 8reenlanders. %eddelelser om 8rQnland 11) 517. #Qbenhavn.
%AS<E::, %KHEA& 9. 51()A7 A "ontemporary Ethnography from the Thule 3eriod. F
Arctic Anthropology Jol. CB 517. %adison.
aaaaaaaa 51().7 3rehistory of the Eastern Arctic. ;ew <orld Archaeological Hecord.
Krlando.
%c8'EE, HK$EHT 51(D@7 $eluga 'unters. An archaeological reconstruction of the history
and culture of the %ackenie 4elta #ittegaryumiut. ;ewfoundland 9ocial and
Economic 9tudies ;o. 1A. *nstitute of 9ocial and Economic Hesearch, %emorial
&niversity of ;ewfoundland. Toronto.
aaaaaaaa 51(D67 <est Alaskan influences in %ackenie Eskimo culture. F 'A::, E4<*;
9. 5ed.7 "ontributions to Anthropology0 The *nterior 3eoples of ;orthern Alaska.
Archaeological 9urvey of "anada 3aper ;o. @(. ;ational %useum of %an %ercury
9eries. Kttawa.
%&H4K"', IK'; 51)(C7 Ethnological Hesults of the 3oint $arrow Expedition. F ;inth
Annual Heport of the $ureau of Ethnology 1))D=2)). <ashington.
;E:9K;, E.<. 51)((7 The Eskimo About $ering 9trait. F Eighteenth Annual Heport of the
$ureau of American Ethnology 1)(6=2(D. <ashington.
;KKTEH, 8EHT 51(D17 Kld kayaks in the ;etherlands. %ededelingen van het Hi-ksmuseum
voor Jolkenkunde ;o. 1D. :eiden.
K98KK4, "KH;E:*&9 51(AD7 Ethnography of the Tanaina. 1ale &niversity 3ublications in
Anthropology ;o. 16. ;ew 'aven.
3ETEH9E;, '.". 51()6a7 9kinboats of 8reenland. 9hips and $oats of the ;orth, vol. 1. The
;ational %useum of 4enmark, The %useum of 8reenland ? The Jiking 9hip %useum
in Hoskilde. Hoskilde.
aaaaaaaa 51()6b7 !aanniornermut ilitsersuut = *nstruktion * ka-akbygning = *nstruction in
#ayak $uilding. Hoskilde.
3ET*TKT, E.+.9. 51))D7 :es grands Es!uimaux. 3aris.
3KH9*:4, %KHTE; 3. 51(1C7 eber einige 8er\te der Eskimo. F Eeitschrift fdr Ethnologie
S:*J. $erlin.
DA
aaaaaaaa 51(1.7 9tudies on the %aterial "ulture of the Eskimo in <est 8reenland.
Arbe-der fra den 4anske Arktiske 9tation paa 4isko. %eddelelser om 8rQnland $d. :*.
#Qbenhavn.
HA*;E1, +HKE:*"' 8. 51(@17 Eskimo 3rehistory0 The Kkvik 9ite on the 3unuk *slands.
Anthropological 3apers of the American %useum of ;atural 'istory AD 5@7. ;ew 1ork.
HA9%&99E;, #;&4 51(A17 The ;etsilik Eskimos. 9ocial :ife and 9piritual "ulture. Heport
of the +ifth Thule Expedition 1(C1=C@ Jol. J*** ;o. 1=C. "openhagen.
HA1, 4KHKT'1 IEA; 51()B7 Artists of the Tundra and the 9ea. 9eattle and :ondon.
HK$EHT=:A%$:*;, IKf::E 51()B7 :e #ayak alToute F vu par son constructeur et utilisateur
F et la chasse g la loutre de mer. F Kb-ets et %ondes, tome CB 517.
9"'&:TE=:KHE;TEE;, ".<. 51(B@7 Eskimoernes *nvandring i 8rQnland. F %eddelelser om
8rQnland $d. SSJ*. #Qbenhavn.
9TEE;9$1, '.3. 51(1B7 "ontributions to the Ethnology and Anthropogeography of the 3olar
Eskimos. %eddelelser om 8rQnland $d. SSS*J. #Qbenhavn.
TA1:KH, I. 8AHT' 51(D@7 ;etsilik Eskimo %aterial "ulture. Kslo = $ergen = TromsQ.
aaaaaaaa 51()B7 "anoe "onstruction in a "ree "ultural Tradition. "anadian Ethnology
9ervice 3aper ;o. 6@. ;ational %useum of %an %ercury 9eries. Kttawa.
T'A:$*TEEH, <*::*A% 51(1C7 Ethnographical "ollections from East 8reenland
5Angmagsalik and ;ualik7 made by 8. 'olm, 8. Amdrup and I. 3etersen and described
by <illiam Thalbiter. F T'A:$*TEEH, <*::*A% 5ed.7 51(1@7 The Ammassalik
Eskimo. %eddelelser om 8rQnland $d. SSS*S. #Qbenhavn.
T'K%9E;, T'K%A9 51(1D7 *mplements and Artefacts of the ;orth East 8reenlanders. +inds
from 8raves and 9ettlements. %eddelelser om 8rQnland $d. S:*J ;r. .. #Qbenhavn.
T&"#, IA%E9 A. 51(D.7 3rehistory of 9aglek $ay, :abrador0 Archaic and 3alaeo=Eskimo
Kccupations. Archaeological 9urvey of "anada 3aper ;o. AC. ;ational %useum of %an
%ercury 9eries. Kttawa.
T&H;EH, :&"*E; %. 51)(@7 Ethnology of the &ngava 4istrict, 'udson $ay Territory. F
Eleventh Annual Heport of the $ureau of Ethnology 1))(=2(B. <ashington.
JA9*:IEJ9#*, H.9. ? 8K:&$IEJ, J.A. 51(D67 4revnie 3oseleniia na 9ahalina. ;ovosibirsk.
<'*TTA#EH, ".E. 51(AD7 Arctic Eskimo, a record of fifty years2 experience and observation
among the Eskimo. :ondon.
<*99:EH, ":AH# 51(1)7 Archaeology of the 3olar Eskimo. Anthropological 3apers of the
American %useum of ;atural 'istory, Jol. SS** 3t. ***. ;ew 1ork.
D@
<KH#%A;, #AHE; <. 51(DD7 "hugachik *sland0 A #achemak Tradition 9ite in &pper
#achemak $ay, Alaska. F Anthropological 3apers of the &niversity of Alaska 1)5C7
<KH#%A;, <*::*A% $. 51()B7 "ontinuity and "hange in the 3rehistoric Hecord from
9outhern Alaska. F #KTA;*, 1K9'*;K$& ? <KH#%A;, <*::*A% $. 5eds.7
Alaska ;ative "ulture and 'istory. 9enri Ethnological 9tudies no. @. Ksaka.
aaaaaaaa 51()C7 ;orton and the <estern #enai 3eninsula. F Arctic Anthropology 1( 5C7.
%adison.
E*%%EH:1, 4AJ*4 <. 51(D)7 #ayaks of 'ooper $ay, Alaska. F E*%%EH:1, 4AJ*4 <.
5ed.7 "ontextual 9tudies of %aterial "ulture. "anadian Ethnology 9ervice 3aper ;o.
@A. ;ational %useum of %an %ercury 9eries. Kttawa.
aaaaaaaa 51(D(7 'ooper $ay #ayak "onstruction. "anadian Ethnology 9ervice 3aper ;o.
.A. ;ational %useum of %an %ercury 9eries. Kttawa.
D.