Aquacultural Engineering 4 (1985) 155-174

Ammonia Production and Oxidation During the Culture of Marine Prawns and Lobsters in Laboratory Recirculation Systems
J.F. W i c k i n s
Ministry of Agriculture, Fisheries and Food, Directorate of Fisheries Research, Fisheries Experiment Station, Conwy, Gwynedd LL32 8UB, UK

ABSTRACT Data obtained during the culture o f tropical prawns Penaeus sp. and European lobsters Homarus gammarus (L.) in 20 laboratory recirculatT"on systems over a number o f years were used to obtain gross estimates o f (a) the rates and diurnal patterns o f ammonia excretion, (b) food utilization and wastage, and (c) nitrification rates in biological filters. Ammonia excretion in Penaeus monodon decreased with increase in animal size from 0.93 to 0.30 mg N g -~ day -1 at 1.6 and 2 7 g live weight, respectively. Large lobsters (300 g) excreted 0.3 mg N g -x day -1. Peak excretion times were approximately 3, 9 and 15 h after a single morning feed for prawns and after 6 and 12 h for lobsters. Uneaten food solids amounted to 11% (lobsters) and 32% (prawns) o f the daily ration and, for 4 g animals, only 69% (lobsters) and 45% (prawns) o f the food nitrogen in the daily ration was converted into crustacean flesh. Nitrification rates ranged from O. 03 to 0.43 g N oxidized per m2 o f filter media surface per day. Some evidence for periodicity in nitrifying activity was also found and is discussed in relation to the published literature.

INTRODUCTION E c o n o m y of water during the commercial culture o f aquatic organisms is frequently achieved through re-aeration or re-oxygenation. Where close control of other specific environmental factors, particularly temperature, is required, the water may be recycled between the culture tanks and specialized water treatment units. Percolating bio155 O Crown Copyright, 1985.

156

J. F. Wickins

logical filters are among the treatment units widely used in recirculation systems but factors governing their design and performance in heavily stocked seawater systems are not well understood (Tiews, 1981). Besides oxygen depletion, major changes occur in ammonia, nitrite and nitrate levels of recycled seawater. Under extreme conditions, production of waste metabolites by microorganisms in the system can equal or even exceed that from the cultured organisms (Moffett and Fisher, 1978; Wickins, 1982). The successful design and operation of suitable water treatment units depends upon knowledge of the cultured animals' feeding behaviour, excretion patterns and tolerance to recycled water. Since oxidation of wastes is accomplished by microbial slimes in biological filters, the nutritional and environmental requirements of those slimes have also to be met. The metabolic activity of all the organisms within the culture system will vary according to the feeding practices adopted, and unless these are in sympathy with the physiological and behavioural requirements of the cultured species, fouling and unnecessary loads will be placed on the treatment units. Even so, considerable advances have been made during the last decade in water treatment methods for the intensive culture of fresh and brackish water fish (Tiews, 1981). The rate and temporal pattern of excretion in fanned crustacean species are generally less well documented than those for fish. In addition, feeding is often less efficient in captive decapods due to their variable requirement for food within the moult cycle and their habit of external mastication (Wickins, 1976a, 1982). In this paper, data obtained over a number of years during the culture of tropical prawns Penaeus spp. and European lobsters Homarus gammarus (L.)in laboratory-scale recirculation systems were used to estimate (a) the rates and diurnal patterns of ammonia production, (b) food utilization and wastage, and (c) gross nitrification rates in biological filters. Some evidence for periodicity in nitrifying activity was also found and is discussed in relation to published literature.

METHODS Excretion measurements Measurements of ammonia excretion were made on 16 intermoult prawns (Penaeus monoclon) of 1.5-27 g live weight held as populations

Ammonia production and oxidation during prawn and lobster culture

157

or as individuals in tanks (28 18 16 cm water depth) at densities o f 3 9 6 - 5 9 5 g m -2, i.e. 1-10 prawns per tank. Filtered seawater was delivered to the tanks at 5 litres h -1 giving 75% replacement in 2 h. Effluent from each tank was collected in a vessel from which samples were taken for analysis every hour for 24 h periods. Each vessel was emptied and cleaned after the samples had been taken. The prawns were either starved for 2 days and then fed with mussel or starved for 3 days prior to making each determination. The results gave information on weight-specific excretion rates and on the pattern o f excretion after a single feeding period. The pattern of excretion in two lobsters (300 g live weight) fed with shrimp was determined by differences between water samples taken from the inlet and outlet of each individual lobster rearing compartment (Beard et al., in prep.) and also by measuring the accumulation of ammonia in static water tests.

Culture systems
Observations were made on 20 recirculation systems ( 8 7 5 - 1 0 627 litres capacity) operated for periods of 2 8 - 3 5 2 days (Fig. 1). The system contained stocks o f adult or rapidly growing juvenile marine prawns or lobsters. Data from systems in which experimental treatments or malfunction either caused mortality or severely reduced growth rates were not used. Sali.nity and temperature, respectively, were 20-34~/oo

~ Piutmo o ~ m v ~ ~r, ls~udi~

'

" "+ +
Tank drain

+i
.

!)

Air

Fig. 1.

The general arrangement of tanks and water treatment units in laboratory recirculation systems.

.~ Oo

TABLE 1

The V o l u m e s (litres) o f Culture Tanks and Water T r e a t m e n t U n i t s in t h e L a b o r a t o r y R e c i r c u l a t i o n S y s t e m s ( T h e T w o

C o l u m n s Referring t o S y s t e m s C o n t a i n i n g L o b s t e r s are M a r k e d (L)) 1 1 8 565 6960 1 605 19 21 85 52 22 44 41 59 1512 4 200 1 515 573 663 590 637 680 65 386 41 1 (L) 1 7 245 5733 1 1 3 953 753 2 4 2 907 1392 1 1 2 603 2030 4 4 1508 845 1 1 1 442 852 3 3 1 075 438 1 1 1 052 372 3 (L) 3 946 560 1 1 875 400 475 54

1 3 10 627 10425

202

2b

202 L

213 P

760 P

1 000 L

123 L

213 L

213 L

189 P

190 P

80 L

136 P

75 L

175 W W . . . .

170

95

26

170

160

160

156 DEP

182 DEP

216 -

160 W

150 -

Number of systems Number of trials Total system volume Total volume of water in culture tanks Totalvolume of water treatment unitsa Treatment unit volume as a percentage of the system volume Percolating filter volume (litres) Filter media (L = limestone, 45% voids; P = plastic, 93% voids) Hydraulic load to filter (m 3 m-3 day-I) Mechanical filtration, DEP = diatomaceous earth pressure filter; W = synthetic wadding Dispersed air flotation Crustacean biomass/system (kg) ~/ 1.8 ~/ 30 . . 2.0-3-5 . 0-4 . 0.2 . 2.0 . 1.2 . 1.9 2.1

0.7-2.6

~/ 0.04-0.2

0.04-0.8

a Includes volume of percolating filter. b Animals reared in water reservoir not in separate tanks.

Ammonia production and oxidation during prawn and lobster culture

159

and 28 -+ 2C for prawns and 28-32~/0o and 20 + 2C for lobsters. Partial water changes were made once or twice each week. Further details of the systems and their management are given in Table 1, and by Beard and Forster (1973), Beard et al. (1977), Richards and Wickins (1979), Sedgwick (1979), Beard and Wickins (1980) and Wickins and Beard (1978). Feeding and calculations o f 'load' Food was given once each day in the morning and was adjusted so that some remained uneaten the next day. Generally a 50:50 mixture of fresh mussel Mytilus edulis L. and frozen shrimp Crangon crangon (L.) was given but where pellets were fed to the prawns these were obtained from the Taiyo Gyo Gyo Co. Ltd of Japan (Sedgwick, 1979). Determination of gross food utilization was made from the difference between the blotted weights of fresh food supplied and of the uneaten food removed by hand-net the next day (Beard and Forster, 1973). To facilitate comparisons between diets and between species, the food and animal biomass data were converted to values representing their nitrogen content. To this end the approximations of water and nitrogen content of the mussels and shrimps (annual means), pellets and juvenile lobsters and prawns shown in Table 2 were used.

TABLE 2

Item

Water (%)

g N kg -1 dry weight ( N-Prtein~6.25 /

93.9 100.6 80-0 107-0 125.5

Reference

Mytilus edulis Crangon crangon Pelleted feed Juvenile lobsters

Prawns (Penaeus spp.) (mean value)

83-0 72-0 5-0 67-5 73-0

Dare and Edwards (1975) This paper Sedgwick et al. (1979) Beard, Richards and Wickins (in prep.) Guary et al. (1974), Gerhardt (1978), this paper

160

3". F. Wickins

For example, if a population of 100 g biomass of prawns (mean weight 5 g) were fed daily with 37.5 g mussel, plus 37.5 g shrimp, the gross amount of nitrogen given would be calculated thus: Mussel: 37.5 X
(100 -- 83)

100

6.37 = 6-37 g dry wt, which contains - 93-9 = 0.598 g N, plus 1000 Shrimp: 37.5
(100 -- 72)

I00
10.50
1000

= 10-50 g dry wt, containing

X 100.6 = 1.056 g N, giving

Total: 1.056 + 0.598 = 1.65 g N (fed per 100 g prawns held per day) The gross nitrogen load on nitrifying filters (the estimated amount of total ammonia nitrogen to be oxidized) was derived (see below) from measurements of food consumption and biomass increase during short (2-4 week) intervals for six populations of 10-488 prawns, and intermittently over a 2-year period for nine individual lobsters. To simplify the calculation, growth was assumed to be linear during such short intervals and results were expressed as average nitrogen values for groups of similar sized animals. The nitrogen load, or nitrogen not incorporated into crustacean flesh, included that excreted by the cultured animals but not that contained in those fragments of cast exuvia removed and weighed with the uneaten food. Many exuvia were wholly or partly eaten and their nitrogen content thus recycled. An example of the calculation of gross increase in nitrogen content for a population of prawns (mean weight 5 g) whose biomass increased by 227 g (from 258 to 485 g) in 28 days was: 227 X (100
--

73)

(100)

61-29 125.5 = 7-69 g N = 61.29 g dry weight, which contained - 1000

representing 7.69 28 = 0.27 g N day -l

A m m o n i a production and oxidation during prawn and lobster culture

161

The total weight of nitrogen fed (less that removed as uneaten food) was 4.99 g day -~ and the difference 4.99 -- 0-27 = 4.72 g N day -~ was the gross nitrogen load on the nitrifying filter.

Frequency of sampling and analytical methods

In each system, measurements of temperature, salinity and pH were generally made two or three times per week and ammonia and nitrite once or twice each week before feeding and routine water changes. The analytical methods used were those described by Wickins and Helm (1981). Additionally, measurements of ammonia nitrogen, nitrite nitrogen and oxygen were made at hourly intervals for up to 10 24-h periods in a 7245 litre lobster culture system over an 8-month period. Samples were taken from the water inlets to individual lobster comparments and, since the water flows to compartments and biological filter were in parallel (Richards and Wickins, 1979), the samples were also representative of the water supplied to the nitrifying bacteria in the filter. In this paper the term 'ammonia nitrogen' refers to the sum of ionized (NH~-N) and unionized (NHa-N) ammonia nitrogen and is designated, for convenience, total NH4-N.

RESULTS

Excretion of total ammonia nitrogen

Specific ammonia nitrogen excretion rates for the prawn Penaeus m o n o d o n decreased with increase in animal size from 0.93 to 0.30 mg N g-~ day -~ at 1.6 and 27 g live weight respectively (Fig. 2). Large lobsters (300 g) excreted on average 0-3 mg N g-i day-~ but no data were obtained for post-larval and juvenile lobsters. Similar patterns of excretion were observed both for prawns and for lobsters (Fig. 3). Ammonia excretion in prawns fed between 08.00 h and 10.00 h was greatest between 11.00 h and 13.00 h and, in small prawns (1-16 g), peaked again at 18.00 h. For all prawns a further peak occurred at 24.00 h. Excretion was lowest from 01.00 h until feeding began again after dawn. Peak excretion times for lobsters fed at 09.0010.00 h were 16.00 h and 22.00 h, although the first peak was not as marked as that for prawns.

162
2.0

J. F. Wickins

IO

I I

1.01

0.5

E E

0.25
I-

0.10

10

15

20

25

30

Live w e i g h t (g)

Fig. 2.

The live weight specific ammonia excretion rates ofPenaeus monodon (e). The regression line is: in N = in 1-002 -- 0.045 W (r = 0.93, degrees of freedom = 12, P'~ 0.001)

where N = mg total NH4-N g-1 day-1 and N = live weight (g). For comparison, data for Penaeus indicus (o) (Gerhardt, 1978) and Macrobrachium rosenbergii (x) (Wickins, 1976a) are included. Prawns starved for 3 days exhibited similar but lower peaks to prawns fed once after 2 days starvation. F o r up to 7 h after the meal the fed prawns excreted 50% more a m m o n i a nitrogen than the starved prawns b u t this difference fell to a r o u n d 30% after 24 h.

Food utilization and wastage

F o o d wasted The weight o f f o o d remaining u n e a t e n 24 h after feeding was determ i n e d for three p o p u l a t i o n s o f prawns and eight groups of individually

Ammonia production and oxidation during prawn and lobster culture

5o
g"

163

.- Jr" 4
', l

,,o,,
l

30(

2o
'~ 10

,t
.~.

\\

.~.

'~-~.,.~_.o..~..~.~
(b)

0 800 . 1000 '

I 1200

1400

' 1800

, 1800

2/ 0
Time

2 2=00

2 4100

' 0200

' 0400

06 ~0

'0 080

' 1000

Fig. 3. The temporal pattern of ammonia excretion rate. In Penaeus monodon, (a) 1-10 g live weight, n = 9; (b) 16-17 g, n = 3; (c) 27 g, n = 4; and in Homarus gammarus, (d) 300 g, n = 4. held lobsters over periods of 16 and 98 weeks, respectively. The wet weight of food removed from prawn tanks ranged from 0 to 26% (mean 11%, n = 277) o f the daily ration and from lobster compartments from 4 to 66% (mean 32%, n = 95). The total 'wastage' of nitrogen was also estimated by determining the percentage of food nitrogen (the daily ration) not converted to flesh nitrogen, in order to account for that lost by leaching and fragmentation of food, as well as that left uneaten. The daily ration values were averaged over 2 - 4 week periods to coincide with measurements of animal biomass. The results, which are specific to the feeding methods used (see Discussion), showed that losses were again greater for populations o f prawns than for individually held lobsters. When plotted against animal size, however, the percentage of food nitrogen not converted to flesh nitrogen increased slightly with size according to the regression equations: Prawn (0.3-12 g live weight): % = 48.35 + 1.75 I4' (r = 0-99, degrees o f freedom = 24, P "~ 0-001) Lobster (0-2-200 g live weight): % = 30.86 + 0.18 I4I (r = 0-61, degrees of freedom = 84, P "~ 0.001)

(1)

(2)

where W = mean live weight (g). For 4 g prawns and lobsters, respectively, wastage was approximately 55 and 31% of the daily ration.

164

J. F. Wickins

Gross food consumption The difference between the dally food ration and the amount recovered the following day represents the food consumed not only by the cultured species but also by heterotrophic microorganisms (microbial biomass) in the culture system. It includes that lost through dissolution. The gross consumption per 100 g of animals held, measured in 11 culture trials with prawns and lobsters, decreased with increasing size according to the regression equations: Prawns: In y = In 0.963 -- 0-306 In x (r = 0.81, degrees of freedom = 22, P ,~ 0-001) Lobsters: l n y = In 0.203 -- 0-317 l n x (r = 0.60, degrees of freedom = 95, P '~ 0.001) where y = g N consumed per 100 g of animals held per day and x = mean live weight (g). The slopes were not significantly different (P > 0.05). Gross nitrogen load The difference between the food consumed within a recirculation system and that incorporated in the tissues of the cultured species represents a load on the water treatment plant. The calculated amounts of nitrogen to be oxidized (see Methods) for different sizes of animal and different husbandry practices are shown in Fig. 4. Taking two widely separated examples from the prawn data, feeding a population of 25 prawns of mean weight 4 g (i.e. 100 g biomass) will produce at best (pelleted feed, low stocking density) 0.35 g N day -1, 95% confidence limits 0.25-0.48 g, and at worst (fresh foods, high stock density) 0.87 g N day -x, 95% confidence limits 0-60-1.29 g, to be oxidized by the filter. For lobsters of similar size the figure was 0-05 g N per 100g held per day -1, 95% confidence limits 0.03-0-09 g. A general estimate of the proportion of this nitrogen that is excreted as ammonia nitrogen in prawns may be obtained by comparing the daily amount of nitrogen to be oxidized (Fig. 4) with that excreted by (4) (3)

Ammonia production and oxidation during prawn and lobster culture

2.0

165

1.0

t~
~ 0.2

x x (b) o

(a)

(c)

0.1 o
o

o
OO O O .L _,

005

g,
o

.
O OO~ dO O

(tl) 0.02 , , I I I lO . . . . . . . . . . 1

100

Live weight ( g )

FiB. 4

Examples of the differences in the gross nitrogen load on biological filters due to diet, feeding frequency, animal size and species.

(a) Prawns, fresh foods, one feed per day, population> 50 m -2, regression: In y = In 1.21 -- 0-23 In x (r = 0-89, degrees of freedom = 24, P ~ 0-001). (b) Prawns, pellets, one feed per day, population > 50 m -2, regression: l n y = In 0-94 -- 0.29 In x (r = 0.81, degrees of freedom = 17, P ~ 0.001). (c) Prawns, pellets, two feeds per day, population < 20 m -2, regression: l n y = In 0-58 -- 0-37 In x (r = 0-99, degrees of freedom = 3, P ~ 0.001). (d) Lobsters, fresh foods, one feed per day, individually held animals, regression: In y = In 0.06 -- 0.14 In x (r = 0-66, degrees of freedom = 85, P ~<0.001).

P. m o n o d o n

(Fig. 2). Assuming t h a t a m m o n i a nitrogen is 75% o f the total nitrogen excreted by prawns (see Discussion), then the total nitrogen excretion f r o m , for example, 25 4 g prawns will be 0-084 + 0-75 = 0.112 g N 100 g-~ day -1, or a b o u t 13-32% o f the total, depending on diet and stocking density. Table 3 indicates t h a t excreted nitrogen became a more i m p o r t a n t c o m p o n e n t o f gross nitrogen load when the prawns were fed more efficiently w i t h pellets ( b u t see Discussion). F o r large 300 g lobsters the p r o p o r t i o n o f total nitrogen excreted was e s t i m a t e d to be a b o u t 13% o f the gross nitrogen load. No comparable d a t a were available for post-larval and juvenile lobsters.

166

J. F. l~ickins

TABLE 3 Excreted Nitrogen as a Percentage of Total Nitrogen Load in Marine Recirculation Systems

Density (m-2)

Food

Feeds per day 0.8

Mean live weight ( g) 4.0 10.0 20.0

9.1 -

300
13.3

Prawns Prawns Prawns Lobsters

>50 >50 <20 Individually confined

Fresh Pelleted Pelleted Fresh

1 1 2 1

10.3 1 2 . 8 12.0 12-8 1 7 . 7 17-7 20-1 32.0 . . . .

Nitrification

Diurnal changes in nitrogen content o f recycled seawater Diurnal changes in ammonia production from excretion (Fig. 3) and microbial degradation o f food particles in turn affect nitrifying filter performance and, in particular, the concentration of nitrite in the water. Hourly measurements of ammonia and nitrite nitrogen and oxygen (Fig. 5) give an idea of the extent of the fluctuations that occurred. A m m o n i a concentration increased slowly to about 0.3 mg total NH4-N litre -1 at 16.00 h concurrent with the peak in ammonia excretion rate and declined to around 0.1 mg litre -1 before dawn. No second peak was observed. Nitrite concentration built up slowly from 0-02 mg NO2-N litre -1 (08.00 h) to 0.06 mg litre -~ (18.00-07.00 h) and then declined. After an initial decrease, oxygen took 16-18 h to return to pre-feed levels. Animal biomass and the frequency with which system water was partially replaced had little effect on the means and ranges of nitrogen concentrations which were probably more readily influenced by filter size and hydraulic load. Overall filter performance (nitrification) Although the m a x i m u m nitrifying capacity o f the biological filters could not be measured, it was possible to obtain an estimate o f the gross nitrification rates which occurred by consideration of the total daily nitrogen load on the filters (Fig. 4) and the area available for

Ammonia production and oxidation during prawn and lobster culture

167

o~ E
t'4

6.5

o 6.0 o~ E Z
-r

0.25

0.0 0.10

-&
E v z 0.05

g
Z 0.0 I 08 O0 I 12 O0 I 16 O0 | 20 O0
Time

24

I O0

I 04 O0 08

| O0

I 12 O0

16

I O0

(h)

Fig. 5.

The diurnal fluctuations of oxygen, ammonia and nitrite nitrogen levels in a laboratory 7245 litre lobster culture system.

bacterial attachment, that is, the specific surface area of the filter medium. To ensure that each filter was working within its nitrifying capacity, only data collected while ammonia and nitrite nitrogen concentrations remained below 0-8 mg litre -~ for several months were used. The routine partial water changes only reduced the nitrogen load slightly and only on one or two days each week. For example, if the recirculating water contained 0.2 mgNH4-N litre -~ and the replacement water 0-03 mg litre -~, then a weekly renewal of 500 litres of water from a 1000 litre capacity system would remove 500 X ( 0 . 2 - 0 . 0 3 ) = 0.08 g NH4-N week -~. This loss is insignificant compared with the nitrogen loading, as illustrated by the daily production o f 'load' nitrogen from a typical system containing 2 kg of 10 g prawns fed moderately carefully (Fig. 4, line (b)), viz. 0-48 X 20 = 9.6 g N or 67-2 g N week -~. In Fig. 6 the estimated gross nitrification rate (g N oxidized m -2 day -~) is plotted against the growth rate o f the cultured animal expressed as the daily increase in nitrogen content of the total population held. Overall, nitrification rates increased from 0-03 to 0-43 g N m -2 day -~ and

168
0.5

J. F. Wickins

A 0.2 E z
OI

0.1

o"
Z

0.05

0.02 I 0.1
Growth rate

0.01

I 1.0 (gNd - 1 )

Fig. 6. The relation between estimated gross nitrification rate in plastic media filters and growth rate of cultured prawns. Minimum input concentration of ammonia nitrogen 0.1-0.3 mg litre-l; hydraulic load 153-182 m 3 m - 3 day -~.

showed no tendency to decrease with increasing animal growth rate, as might be expected if the nitrifying capacity o f the filter were exceeded. The nitrate resulting from nitrification was maintained between 3 and 55 mg NO3-N litre -~, mean levels, by dilution, which resulted in only a slow increase in nitrate levels over the culture period. Despite the regular, partial water changes and careful control of temperature, salinity and pH, cyclic fluctuations in biological filter performance were observed in 25 trials made under different load conditions in 20 separate systems. The magnitude of the fluctuations in ammonia and nitrite levels was irregular, between 0.1 and 0.6 mg N litre -1, and occurred in systems with mature, well-established nitrifying filters. The mean interval between the fluctuations was 3-5 weeks --standard error = 0.1, n = 114 for ammonia and 3.9 weeks-+ standard error = 0-2, n = 88, for nitrite nitrogen. The frequency distributions of the observed intervals for ammonia and nitrite are shown in Fig. 7.

Ammonia production and oxidation during prawn and lobster culture

169

40-

35-

(a)

30-

25-

20-

15-

10-

ol
25

(b) 20

15

10

10

Interval (weeks)

Fig. 7.

The frequency distribution of fluctuations in pre-feed (a) ammonia and (b) nitrite levels in 20 laboratory recirculation systems.

170

J. F. lCickins

DISCUSSION

Food wastage and excretion

Most of the culture systems discussed in this paper contained stocks of crustaceans that were being grown either for breeding or to determine maximum growth rates in captivity (Arnstein and Beard, 1975; Beard et al., 1977; Wickins and Beard, 1978; Richards and Wickins, 1979; Beard and Wickins, 1980). A mixture of fresh mussel and frozen shrimp was the best diet available for this purpose, but the daily preparation of the mussel meats and the manual feeding methods used were so time-consuming that only one feed per day was practicable despite the wastage this caused; 11 and 32% of the daily ration for prawns and lobsters, respectively, were removed uneaten. The lower figure for prawns is misleading because considerable fragmentation and dispersal of food occurred in the prawn cultures due to communal feeding, feeding behaviour and, doubtless, the high temperatures involved. Overall, more than 50% of the nitrogen in the daily ration was not converted into prawn flesh. A similar value (58%) was reported for trout grown from 48 to 134 g (mean weight) in recycled water (Van Toever and MacKay, 1981). Greater wastage by juvenile prawns was found by Gerhardt (1978) who fed 0.5-4.0 g Penaeus indicus with a pelleted diet at 2-5% of the live weight per day for 2 months. Uneaten food fragmented and was oxidized within the sub-gravel filter. Approximately 80% of the food nitrogen given in his experiments was not converted into prawn flesh and represented the nitrogen load on the filter. Although the present experiments were not designed to test feeding frequency and water exchange rates, Fig. 4 illustrates how fresh foods, infrequent feeding and high stocking densities can place a high gross nitrogen load on water treatment units in the recirculation systems that contained prawns (see also Wickins, 1976a, 1982). More frequent feeding with pelleted diets is known to improve growth in prawns (Sedgwick, 1979) and at low population densities seems likely to reduce the risk of fouling. The conclusion that excreted nitrogen becomes a more important component of gross nitrogen load as prawns are fed more efficiently with pellets (Table 3) must be viewed with some caution, since the estimates were based on the assumptions that (a) ammonia nitrogen is 75% of excreted nitrogen (Snow and Williams, 1971; Gerhardt, 1978)

Ammonia production and oxidation during prawn and lobster culture

17 t

and (b) similar excretion rates occurred in prawns reared on different diets and stocking densities. This assumption is debatable since, although excretion rate in juvenile Macrobrachium rosenbergii appeared not to be influenced by diets of plant, animal or pelleted material (Nelson et al., 1977), excretion in Crangon franciscorum was shown to be influenced by diet (Nelson et al., 1979).

Nitrification
Diurnal variations in the chemistry of recycled water (Rosenthal, et al., 1981) dictate the time when samples must be taken to obtain the specific information required. The measurements presented here were obtained from samples collected prior to feeding and water changes and represent, for example, minimum levels of nitrogen to which the microorganisms in the biological filter are exposed. They do not represent the peak levels to which animals might be exposed in the culture tanks which, although higher, rarely exceeded 0.7 mg total NH4-N litre -1, 1-0 mg NO2-N litre -1, and 75 mg NO3-N litre -1 (see also Wickins, 1976b, 1981). Knowledge of the varying diurnal demands made on water treatment units allows appropriate periodic adjustments to be made, for example, to aeration rates and hydraulic load to mechanical and biological filters. To check for the occurrence of harmful levels of metabolites, the time of sampling for each substance must be determined in relation to feeding times. The estimates of the overall nitrifying performance of percolating filters obtained in this study are likely to substantially underestimate true nitrification rates which could be as much as two to three times higher. This is because two assumptions are inherent in the estimates of gross nitrification rate: firstly, that the microbial biomass within the system contains only a small amount of the nitrogen in relation to the gross nitrogen load, and secondly, that the nitrifying organisms are evenly distributed throughout the filter medium. The assumption that the microbial biomass contains relatively little of the nitrogen would seem reasonable for the filters which contained plastic medium which had a high proportion of voids (93%). In contrast, graded gravel medium (nominal maximum size 2-5 cm) which has a similar specific surface area to the plastic medium (~ 160-200 m 2 m -3) contains 40-50% voids and is thus more likely to support greater quantities of microbial biomass (Wickins and Helm, 1981). However,

172

J. F. Wickins

the microbial film attached to the surfaces within percolating filters undergoes periodic, partial sloughing, some of the fragments being washed from the filter as suspended film-humus solids. Short-term (nonseasonal) cyclic sloughing and variations in filter efficiency (Bruce and Merkens, 1973) are well-known occurrences in wastewater treatment plants, and periodicity ranges from about 14 to 30 days (Heukelekian and Crosby, 1956; Sanders, 1966; L. F. Reynolds, 1974, pers. comm. in Howell and Atkinson, 1976). In the laboratory marine filters, a similar periodicity (24-27 days) in pre-feed ammonia and nitrite levels was observed for which no explanation more suitable than sloughing could be offered. The occurrence of such sloughing would lead to the loss of an unknown amount of nitrogen (microbial biomass) over long culture periods. The second assumption about even distribution of nitrifying organisms is in fact unlikely to be true for deep filters because heterotrophic microorganisms compete successfully with nitrifiers for space and would tend to displace nitrifiers to the lower regions of the filter. The proportion of the filter they would occupy is therefore unknown but presumably would depend upon the amount of organic oxidation occurring in the filter (S~irner, 1981). Any organic oxidation would thus lead in turn to overestimates of filter nitrifying capacity. For these reasons, and because removal of organic material by separate mechanical filtration and foaming can affect filter performance, no attempt was made to calculate carrying capacity. Experience has "shown, however, that percolating filter sizes based on nitrification rates of 0-3-0-5 g N m -2 day -1 have been satisfactory for laboratory culture systems (Wickins and Beard, 1978; Richards and Wickins, 1979). Indeed these systems provided sufficiently good environments to allow: (a) four species of penaeid prawns to be reared to maturity within 8-18 months and induced to spawn (Arnstein and Beard, 1975; Beard e t al., 1977; Beard and Wickins, 1980); (b) batches of lobsters to be reared at 3-month intervals, regardless of season, from egg to market size (c. 400 g) in less than half the time normally taken by natural populations (Richards and Wickins, 1979).

ACKNOWLEDGEMENTS The assistance of a student, Mr Romeo Milan, who measured prawn excretion is duly acknowledged.

Ammonia production and oxidation during prawn and lobster culture REFERENCES

173

Arnstein, D. R. & Beard, T. W. (1975). Induced maturation of the prawn Penaeus orientalis Kishinouye in the laboratory by means of eyestalk removal. Aquaculture, 5, 411-12. Beard, T. W. & Forster, J. R. M. (1973). A growth experiment with Penaeus monodon Fab. in a closed system. ICES C.M., K: 39, 6 pp. (mimeo). Beard, T. W. & Wickins, J. F. (1980). The breeding ofPenaeus monodon Fabricius in laboratory recirculation systems. Aquaculture, 20, 79-89. Beard, T. W., Wickins, J. F. & Arnstein, D. R. (1977). The breeding and growth of Penaeus merguiensis de Man in laboratory recirculation systems. Aquaculture, 10, 275-89. Bruce, A. W. & Merkens, J. C. (1973). Further studies of partial treatment of sewage by high-rate biological filtration. War. Pollut. Control, 72,499-527. Dare, P. J. & Edwards, D. B. (1975). Seasonal changes in flesh weight and biochemical composition of mussels (Mytilus edulis L.) in the Conwy estuary, North Wales. J. exp. mar. Biol. Ecol., 18, 89-97. Gerhardt, H. V. (1978). The culture of Penaeus indicus Milne Edwards in experimental closed systems with special reference to water quality. MSc Thesis, Rhodes University, Grahamstown, South Africa, 86 pp. Guary, J. C., Kayama, M. & Murakami, Y. (1974). Lipid class distribution and fatty acid composition of prawn, Penaeus japonicus Bate. Bull. Jap. Soc. Scient. Fish., 40 (10), 1027-32. Heukelekian, H. & Crosby, E. S. (1956). Slime formation in polluted waters. II. Factors affecting slime growth. Sewage Ind. Wastes, 28, 78-92. Howell, J. A. & Atkinson, B. (1976). Sloughing of microbial film in trickling filters. Wat. Res., 10, 307-15. Moffett, W. L. & Fisher, W. S. (1978). Ammonia production rates ofArtemia salina under various culture conditions. J. Fish. Res. Bd Can., 35, 1643-8. Nelson, S. G., Knight, A. W. & Li, H. W. (1977). The metabolic cost of food utilization and ammonia production by juvenileMacrobrachium rosenbergii (Crustacea: Palaemonidae). Comp. Biochem. Physiol., 57A, 67-72. Nelson, S. G., Simmons, M. A. & Knight, A. W. (1979). Ammonia excretion by the benthic estuarine shrimp Crangon franciscorum (Crustacea: Crangonidae) in relation to diet. Mar. Biol., 54, 25-31. Richards, P. R. & Wickins, J. F. (1979). Ministry of Agriculture, Fisheries and Food, Lobster Culture Research. Lab. Lear., Direct. Fish. Res., Lowestoft, (47), 33 pp. Rosenthal, H., Andjus, R. & KrOner, G. (1981). Daily variations of water quality parameters under intensive culture conditions in a recycling system. In: Aquaculture in Heated Effluents and Recirculation Systems. Vol. 1, ed. K. Tiews, Heenemann Verlagsgesellschaft, Berlin, pp. 113-20. Sanders, W. M. Ill. (1966). Oxygen utilization by slime organisms in continuous culture. Air and Wat. Pollut. Int. J., 10, 253-76.

174

J. F. Wickins

S~'rner, E. (1981). Removal of dissolved and particulate organic matter in high-rate trickling filters, l~at. Res., 15, 671-8. Sedgwick, R. W. (1979). Effect of ration size and feeding frequency on the growth and food conversion of juvenile Penaeus merguiensis de Man. Aquaculture, 16, 279-98. Snow, N. B. & Williams, P. J. le B. (1971). A simple method to determine the O:N ratio of small marine animals. J. mar. biol. Ass. UK, 51,105-9. Tiews, K. (Ed.) (1981).Aquaculture in Heated Effluents and Recirculation Systems, Vols 1 and 2, Heenemann Verlagsgesellshaft, Berlin. Van Toever, W. & MacKay, K. T. (1981). A modular recirculation hatchery and rearing system for salmonids utilising ecological design principles. In: Aquaculture in Heated Effluents and Recirculation Systems, Vol. 2, ed. K. Tiews, Heenemann Verlagsgesellshaft, Berlin, pp. 403-13. Wickins, J. F. (1976a). Prawn biology and culture. In: Oceanography and Marine Biology: An Annual Review, Vol. 14, ed. H. Barnes, Aberdeen University Press, pp. 435-507. Wickins, J. F. (1976b). The tolerance of warm water prawns to recirculated water. Aquaculture, 9, 19-37. Wickins, J. F. (1981). Water quality requirements for intensive aquaculture: a review. In: Aquaculture in Heated Effluents and Recirculation Systems, Vol. 1, ed. K. Tiews, Heenemann Verlagsgesellschaft, Berlin, pp. 17-37. Wickins, J. F. (1982). Opportunities for the farming of Crustacea in western temperate regions. In: Recent Advances in Aquaculture, eds J. F. Muir and R. J. Roberts, Croom Helm, London, pp. 87-177. Wickins, J. F. & Beard, T. W. (1978). Ministry of Agriculture, Fisheries and Food, Prawn Culture Research. Lab. Leafl. Direct. Fish. Res., Lowestoft, (42), 41 pp. Wickins, J. F. & Helm, M. M. (1981). Sea water treatment. In: Aquarium Systems, ed. A. D. Hawkins, Academic Press, London, pp. 63-128.