Proposed Culture of Scylla serrata in Mauritius

Introduction

The island of Mauritius is small but densely populated.

There are few natural resources and sugar, at present, is the main

source of income. In order to broaden the basis of the economy,

the government 15 encouraging the sugar indus t ry to diversi f y , and

one of the fields being developed is aquacult ure . Tour ism is r a p­

idly increasing o n the island and several hotels a r e bein g r e b uil t

and new ones opened. Tourist demand, in ttirn. has stimulated

interest in aquaculture and three organisms are being ~ onsid e red .

namely prawns (Macrobrachium rosenbergi), oys t ers (Crassostrea

cucullata) and the crab Scylla serrata.

Twenty two of the sugar companies have now combIn e d to

set up a hatchery which will produce post larval !v1acrobrac Lium to be

reared in ponds on indivldual sugar estates. One of th e larg e st

companies, Flacq United Estates Ltd., (FUEL) has start ed its own

hatchery with a planned annual capacity of 3 x 106 pos t l arval

Macrobrachium. } UEL has two estates, hath of which will eventually

h ave ponds. Thus far they have constructed 16 ponds totalling 5 ha .

e n o ne estate • It is planned to increase this to 30 ponds.

. ~L~c r obrachiurr are already being produced and supplied to the com­

p any 's own hotels.

The island has few estuaries or areas suitable for natural

populat ions of Scylla. There is however, a large d e mand for crabs

conse quent upon t h e increased tourist traffic and a dec line in th e

supply o f l obsters from the local reefs - alle gedly through over-

fishing. The po s Ition is thus analogous to that of Hawaii: in bot h

isl a nds there is a large demand by tourists wh o are prepared to pay

high prices, but there is a limited natural supply. These conditi ons

favour aquaculture since a high price can be obtained for the produc t

a nd there is no competition from natural supplies.

Areas suitable for culture of Scylla

Most of the island has a fringing coral reef, about 1 km

offshore. This reef encloses a calm lagoon which is mainly 3 to 10

m in depth. The coastline is chiefly rocky, but there are many

beaches. Along the rocky coasts about 15 inlets have had their

entrances to the lagoon walled off. A narrow entrance permits some

- 2 -

tidal exchange. The walled-off inlets are referred to as bara­

chois, and altogether they have an area of about 80 hectares.

The inlets have rocky sides, usually with a dense fringe of man­

groves (Rhizophora), and a sandy mud bottom. The barachois were

originally constructed as fish ponds but fell into disuse and

many have been allowed to deteriorate. In some cases, the ponds

are used for the rearing of oysters (Crassostrea cucullata)

which are collected as juveniles from the wild. Introduction of

other oysters (2. virginica and gigas) has not been successful

and further imports have now been banned because of the danger of

introducing oyster diseases or parasites. It i s pr opo se d t o u t i ­

lise the barachois as areas for rearing of fish (Siga~~), oysJ;;ers

and Scylla (not simul~aneously, since Scylla eats o ysters and c an

be a serious pest).

Most barachois have no streams or rivers enterin~ into

U.em but the porous nature of the soil anli old la .... va of which the

lsland is composed results in numerous freshwater s p rings along

z. lJe shore. Thus most barachois have fresh water entering into

them. We measured salinities in several of the barachois. One i n

which tne owner was attemptinr to obtain a self-replicA tin c pop uln­

~ io~ of Scylla, by holding ovigerous females had a salinity of 22~~

Thi s salinity, coupled with high temperatures, would, however,

resul t in mortality of first stage zoeae as I have been shown

experimentally. It was not surprising, therefore, to h e ar that no

success had been achieved.

In two barachois owned by FUEL and regarded as possible

Scylla rearin~ sites, there was considerable s nlinity layering

pr ~ s ent. One had a surface salinity of 8 0/00 with 28 0/00 on the

bottom at 2 m. The second had 2 0/00 on the surfac~ and 25 0/00 on

the bottom at 1 m. These salinities should not pre s ent a direct

prohlem to Scylia, since it has been found experimentally (Heesman\

unpublished) that Scylla can survive in salinities below 5 %0.

~e had no facilities for measuring oxygen concentration and thus do

not know whether the strong salinity layering was associated with

low oxygen conditions or whether tidal exchange was adequate to

prevent this from occurring. If Scylla are held in large numbers

in the ponds deoxygenation problems may arise since the large input

of organic food could result in an increased oxygen demand by the

bottom waters.

- 3 -

Possible culture methods for Scylla

If the barachois prove to be suitable for rearing of

Scylla a source of juveniles has to be found or ~reated. At

present, there is a rather low level rearing project being

carried out in one barachois by a private ind iv idual. Juvenile

Scylla are collected from two areas in the DoE. of the island,

where they are found beneath rocks and betwe en mangrove roots in

the intertidal zone. They are put into a walled-Qff enclosure

in the barachois and fed on fish. Because of the rocky sides I'Ul d

the presence of mangroves, the only feasible way of remov i ng

crabs from the pond is by trapping. Crabs of 14 t o 17 em c a r a ­

pace width are removed from the enclosures tied up and marketed.

Discussion with the owner suggested that it is not an economi c

proposition. He thought he was recovering 10 to 15% of the crabs

and was losing the re~t through escapes or poaching . The possib­

ility of mortality or of large crabs eating small ones had not

been considered. He paid 1 rupee (about 15 c) per juvenile and

sold the adults for 10 rupees. Thus a recovery of 10% of crabs

would not be adequate to cover the co~ts of the fish whi c h h ad

be en supplied as food.

The area in Mauritius which is available for catching

o f juvenile Scylla is limited and it is unlikely that it could

supply sufiicient material for a large commercial operRtion. The

c osts of juvenile crabs are also high unless reco very rate co uld

be substantially improved. The alternative is to r e ar Scylla in

a hatchery and release postlarvae into the barachois.

Existing and proposed facilities for the rearing of

Macrobrachium should be suitable for a Scylla larval rearing pro­

gramme, since tanks, compressed air, running sea water, staff and

general larval rearing technology is already availabl e . Experi­

mental work is obviously necessary in order to develop a suitable

method for Scylla, e.g. optional salinities, temper a tures, stocking

densities and food. Further experimental work is required to

establish whether it may be possible to release megalopas rather

than the postlarvae. The megalopa is the stage which recolonises

estuaries and thus, if released into a suitable area, they would

probably metamorphose and settle down. Since this stage is canni­

balistic, it would be extremely advantageous to be able to release

it into the larger volume of the barachois, thus achieving greater

-JO---

~------60----~··

, 23

~ +--30--

•

1

I

~.~------------------tSO---------------------·

ind

Top

Side _

Fie .1. ~rap llsed for e~tehing ~eylla serrata in Mauritius. It is

ma de of 5 em galvanised me s h with no fra me. The resultins trap

i s very liGht but suited to the sheltered conditions in which it

is used. Cr~ bs are r e leased by unlacing on corner and shaking

.hem out. Measurements in cm.

- 4 -

spatial separation between individuals.

There are two fUTther problems to be solved apart

from larval rearing. Firstly a suitable food source has to

be found. Protein is scarce on the island and imported protein

such as fish meal is expensive. Possible sources are offal from

a chicken and a pig factory. This could be combined with fish

caught in gill nets in barachois used for oysters. It may be

feasible to combine the food into a large dried pellet form .

Secondly, there is the major question a s to whetb.er

the barachois will prove to be suitable for rearing high de nsity

populations of Scylla. The low return from the one eMisting

pond is not an indication of what could be achieved under good

management. Poaching can be reduced or virtually c hecked by

proper security methods. FUEL have constructed a high s(;curity

fence around their barachois and the area inside is patrolled by

guards with dogs.

Continuous stocking may also result in low returns ,

since I have found that one of the main food sources of ~112

is small crabs. Although the rocky shores and the mangr ove roots

offer a large area for concealment for small crabs, they would b ~

vulnerable to predation when foraging for food. Clearly, the

bio logy of Scylla in the barachois must receive a great deal of

attention in order to establish whether the project will be

economic.

Acknowledgements

I wish to express my thanks to Flacq United Estates

Ltd., for inviting me to Mauritius and especially to Mr. P. de

B. Baissac of FUEL for his assistance during my visit.

Natural Food, Foregut Clearance Rate and Activity of the Crab

SCYLLA SERRA'l'A

B.J. Hill

Zoology Department, Rhodes University, Grahamstown, South Africa.

Research carrie d out in Australia 1974 - 1975 with the aid of grants

from e.S.I.R. and Fishc' ries Development ~orporation, at t.l1e

laboratory of the ~ue ensland State Fisheries, Deception B~y.

2

Abstract

The natural diet and rate of foregut clearance and the diurnal

acti vi ty of the crab Scylla. seA.r~ was determined. The gut volume

is related to size of crab as gut volume (ml) = 0.07eO.033 x, where

x = carapace width in mm. Fifty per cent of crabs collected in

Aus,tralia. and South Africa contained mollusc remains and about 21%

~ contained crustacean remains - chiefly grapsid crabs. Fish remains

were rarely found and it is concluded that.§.. serrata does not

normally catch mobile forms such as fish and penaeid pra.'1Ils o Gut

clearance of organic tissue was rapid and almost complete a£ter 12 h.

Fish bone was retained for a mean time of 2 to 3 days, and shell for

5 to 6 days. Time-lapse photography was used to record activity, it

was found that visible light flashes reduced activ:l.ty and so infra-

red lighting l'1as used o It was found that .§. p,.e.rrata are buried

during the day, they emerge at sunset and spend the night feeding,

walking around and remaining stationaryo Feeding occurs intermittently

throughout the night even tlhen unlimited food is available. If no

food was present the amount of time spent emerged l'laS halved.

3

Introduction

Scylla serrats is a large portunid crab found throughout much of

the Indo-Pacific region. It is generally estuarine, occurri~~ both

intertidally and subtidally. Although it is a large prominent

animal, little is kno\'m of its behaviour and nothing has been puld.i ,';')Z1-ac1

on its feeding habits. It is commonly caught in traps baited with fi~~

but it is not known whether fish forms part of i.ts natural diet.. There

is an increasing interest in the aquaculture of §. .. !!.errata (Brick, 1974)

and in this field it "lould be of value to knOll uhich types of food axe

eaten by the crab under na.tural condi tiona. A study of the foregut

contents of .§.. serrate. "18S therefore ini tinted using crabs caught ix,.

South Africa and Australia. 1m attempt nas made to extend the

information gained from gut analysis to give an indication of uhen

crabs had eaten by measuring the volume of gut contents and the rate

at \Thich organic tissue is cleared from the foregut" In adcLiti. on t he

clearance rate of shell and bone \'Tere determined in order -1:;0 establi sh

i'lhether their occurrence accurately reflects the frequency 1lith uhich

they are eateno As a means of supplementing the information on

feeding, laboratory studies "Tere made of activity using time~lapse

photography.

Materials and Methods

Foregut Volume

It was nocessary to knOlT the theoretical maximum volume of the foregut

of SCjlll<t1 flerrata at various sizes in order to establish the relativo

fullness of the gut of crabs ~lhich had been feeding. The foregut 'l'mll

is lined l'ri.th chitin and is comparatively inelastic but soft and, uhen

4

the gut is empty, the anterior wall collapses into the gut. The

maximum volume of the foregut was measured on 26 freshly killed

crabs (47 to 178 mm carapace width). The foregut was exposed by

dissection, and isolated from the midgut by cutting it off

posterior to the filter-chamber. All muscle attachment s were cut,

the oesophagus was ligatured and cut after which the for;)gu'~ cou.ld

be lifted out of the crab. It was then filled with water injected

through the cut end of the filter chamber and the t otal volume of

gut wall + content measured by displacement in a measuring cylinder.

The foregut rTas then punctured and completely emptied and the volume

of the gut wall measured by displacement. The difference gave

foregut volume.

Rate of Foregut Clearance

The rate of clearance of organic tissue, bivalve shell and fish bone

was determined in Scylla serrata of 90 to 150 mID carapace width~

The crabs were obtained from traps in Moreton Bay (153cE, 270 S)9

Queensland, Australiao They were starved in laboratory tanks for at

least a week before experimentation. Experimental crabs 1-Iere kept

individually in 1.1 m diameter tanks containing 0.5 m depth of

filtered sea water which was aerated and continuously replaced at

300 to 400 I per h. A cement tile shelter was put into each tE'nk and

crabs spent most of the d~light hours in these shelters. The tanks

were kept in a temperature controlled laboratory but variations in

the temperature of inflOlling "'Tater did cause some temperature

fluctuation. During organic food clearance experiments temperatures

ranged between 1SoC and 220 C and during shsll and bone clearance

between 22°C and 24°C. Salinities varied between 33 and 35~oo. The

organic food used was abdomens of the prawn P@aeWJ. I!lebejus. 'rhe

5

cuticle was peeled off the abdomen, surface moisture dried using a

paper towel, the abdomen was weighed and 8 to 10 g put into each

tank at dusk. The crabs were observed every 10 mins until they

began feeding. Ten mins after commencemen.t of feeding all uner:;;ten

food was removed from the tank and dried to constant weight at oooe.,

Addi tional. samples of prawn abdomen were \,leighed fr ;9sh and after

drying to provide conversion factors between wet and dry weights.

Crabs, in batches of 10, were killed ~ immersion in ice-water at

various time intervals after feeding. The food remaining in the

foregut was removed and dried to constant weight. The clearance

rate of shell from the gut was determined using crabs which had fed

on the bivalve Plebidonax del.tQide§. Ten live bivalves, in the

size range 18 to 22 mm shell length, uere put into each tank ovornight

together with a single crab. Crabs which ate all 10 bivalves uere

sacrificed at various time intervals in batches of 10 and the 8hell

still present in the foregut t'Tas dried and weighed. CleaJ."ance r at0

of fish bone nas determined using the same technique but in crabs

which had fed on heads of 10 to 14 cm specimens of the fishes C0:tr....§ill,

ovatus or 1-1:U£il sp. Heads alone were given since they contain a

large proportion of bone. Only crabs which ate 10 to 15 g of fish head

were used.

AnalYSis of Gut Contents

The contents of the foreguts of 67 Scylla serrata (size range 25 to

176 rom carapace width) caught in Moreton Bay, Australia were removed

and analysed. A. further 40.2.0 serrata (si3e range 12 to 170 mm

carapace width) collected in the KOl-n.e and Kleinemond estuaries

(270 E, "OS) in South Africa 1-lere also analysed. The following

6

methods were used to collect crabs. In Australia a 6 m otter trawl

operated in water 1 to 3 m deep caught 17 crabs at night and 2

during the day. A further 16 were taken in seine or tangle nets in

shallows at night and 32 lTere collected during the day in the

intertidal zone whore the crabs l'lere either buried in mud OY' beneath

rocks. In South Africa. 14 crabs, including 5 ovigerous females were

collected by means of a scoopnet from shallows at night, 10 juveniles

(12 to 55 mm) from scoopnetting in s. weed bed in the day, and 10 from

intertidal burrows in the day. A further 6 "Tere cc\Ught in a specially

modified trap in which crabs could not reach the bait. Crabs

collected in conventional traps cannot be used for gut analysis since

they may have eaten food in the traps. Crabs ~7ere either frozen or

preserved in formalin immediately after collection., Crabs tr(.l.~11ed

in Australia were hO~7ever kept alive until return to the laboratory,

this involved a delay of un- to 3 h betl'TOen capture and freezing., Gut

analysis of South Mrican crabs involved an estimation of the rela.ti";re

fullness of the foregut together 'rl th identification of the material

present. In Australia, if the gut appeared to be more than 50% fUll p

the oesophagus was ligatured and the volume of the contents measured

by displacement in a measuring cylinder. The gut contents ",ere then

identified after "mich they were dried to constant weight at 800 Ce

This was follol'Ted by decalcification in 10% HCI, tlashing with distilled

water, centrifugation, decanting of supernatant fluid and re-drying to

constant \might. The results were used to calculate the weight of

shell and the ,{,TOight of the shell-free gut contents. Dry weights were

used in all analyses since Steigenberger and Larkin (1974) found that

initial hydrolYSis of food in stomachs led to inaccuracies, in wet

weights.

7

Catching of Live Penaeid Pra~mB

The rate at uhich Scylla serrata feed: on live penaeid pra~ms was

investigated in an experiment in which 8 crabs (carapace l'Tidth 90

to 125 mm) nere held individually in tanks as described above for

clearance experimentso Each tank also contained 10 live E~~~ .

.BIn bejWl (10 to 15 rom carapace length). Crabs nere f ed daily on

E,. pIe he/in!'} abdomens for 10 dayso The prauns were then re!ilwed and

the crabs starved for 5 days. Ten live prmms 11ere then put i nto

each te.nk for a further 10 days during l1hich time the crabs '(Jere not

fed. The number of prawns in each tank uas counted every mor ning nnd

made up to 10. Two control tanka each contained 10 prauns but no

crabs.

Activity Recording

Activity and feeding of ScYlla Aerrata \'Tare recorded using time-lapso

photography. Crabs ''\1ore put individually into a rec ta..'1gular fibre~

glass tank p 0.9 x 1.8 m, containing 5 em depth of mmd and 10 em of

'Hater. The salinity ues 33 to 35<roo end tempGratul'c 20 to 22°0 0

A Shackman Nark III 35 mm Autocamera with an 18 mm uide-angle l ons

uas mountod on e frame over the tank. The camera nns opero.ted by a

Sheckman automatic timer Type TU1, set to trigger the camera and. hro

electronic flashes at 15 min intervals. A clock in the field of vieu

identified tho time of each photographo Initially records were made

using visible light flashes and llford. FP 4 film. In vieu of possible

effects of visible light flashea on behaviour of .2. serrata it tre.s

decided to duplicate records using infra-red photography. The

electronic flashes \Jere covered uith gelatine filters (Kodak No 87)

which do not transmit uavelengths shorter than 740 nm. The film used

8

was Kodak liS Infra-red No 2481" lui tial 24 h experiments indicated

that crabs spent the daylight hours buried in the sand from uhich

they rarely emerged, thus experimental recordings iJero made from

1700 h to 0700 h. The tank was first housed in a light-proof

laboratory and subjected to a 12 h light - 12 h dark cycle of

artificial light. This ~las discontinued as it mm found that Hhen

the lights nere sl1itched on at 0600 h, many crabs emerged f r om the

sand if buried and. moved about the tallk before rGburyi~.. Tho tanlc

was then moved into a laboratory with a natural 1i~ht cY'clo o Thin

unforttmntely resulted in loss of control of deW longth.. Tho

illumination at the surface of the lrater \'Tas about 140 lux 0.1; middClY,

this doclined to 25 lux 1 h before StIDset and nas about 35 lux 1 h

after smwisoo Complete darkness occurrod from 30 mi n aftnr s unoct

to 30 flin bofore sunrise" Crabs iJere put into the tank 10 h be{'o:<'c

record:in.,,~ cOl!'.1I\onccd o Two condi tiona nero recorded, firstly onc in

uhich thoro uas no food in the tank fu"ld secondly ono in uhich '100

live bivnlV0s (Pl,,ohido~ deltoides) 'Hero put into a shallow d i sh i .n

the c~ntro of tho tank and left overnight" Each experiment 1183

repented nsing at least 10 different crabs in both visible end infra­

r8(1. light " Tho n1.llll.her of bivalves enten overnight uas noted.

Activity film uas analysed by exe.mining each photographic frame and

recordine uhethor tho crab 'UE1.S buried or ernorgecl p uhether it hen

mO'\Ted sinco the previous frame and uhethol' or not it uas fecdioGo

Vi Dible light recordings VIere made t1ith crabs of 85 to 143 mm cnrapace

uidth (moon 122 rom) and infra-red with crabs of 105 to 141 rrun carapnce

~1idth (moan 129 rom) 0 P.J.l crabs had fed one to 6 days prior to being

introducod. into the tank.

9

Results

Poregut Volume

Foregut volume of SCyJl~ serrata was found to be exponentially

related to ca.rapace width. Carapa.ce width uas used as a measure

since it Has felt to be more meaningful than ",eight in t his CC,3e.

In.§.. se!,rata larger than 120 mm carapace width, there is

differential grol'1th of the Clm-IS betueen the sexes . I~easurements

indicated that in large male crabs (170 rom carapace ~ddth), the

claws make up about 45% of the total l'leight, uhereus they contribute

about 221b of the \-Teight of an equivalent l'ddth female. The vreight

of body minus clat-ro does not vary greatly bahreen male and female

although tho latter are slightly heavier. Poregtlt volume is directly

related to ueight of body (minus claus) but since the lcltter paX'cilioter

is not alunys a convenient one to measure, carapace 'Width UM s elect ed ..

0 .. 033 x Foregut volumG is reInted to carapace l'ddth as y ::::: 0.07e ~ uhere

y = gut volume in m1 and x ::: carapace l'Tidth in rom, (r = 0 0 940, n ::: 26)"

According to this relationship, crabs of 80, 120 and 160 rom carapace

width uould have foregut volumes of about 1, 3.6 and 13.7 ml

respectively ..

Foregut Clearance

The mean fr0Dh '[wight of pratm abdomen eaten in organiC food

cloara~co experiments was 7.0 g (standard deviation 2.52, n = 50),

crabs eating less than 2 0 0 g "lere not used. The l'might of food in

the foroGlrt of Bcyll~ serrata killed immediately after feeding

accounted for only 7y}~ of the difference in \-might between food put

into tho tanks and that taken out. Dagg (1974) found a similar

phenomonen uben dealing nith feeding in predatory ampbipods and

10

attributed it to loss of tissue fluids and small particles of food

\·rlrich llere not recovered from the tank. The proportion of food

apparently eaten lihich could be accounted for by foregut content

declined rapidly with time and after 2 h, 22% could be found i n tho

foregut (Pig. 1). After 12 h, although all crabs still rC '~ained

some organic tissue in the foregut, this had declined further to 4%&

Thus the rate of clearance of organic tissue from the foregut of

.2,.. senatn is rapid.. The presence of organic material i.n the gut of

crabEl caught in the wild is an indication that the crabs hnve recently

beon feedin:!~ Crabs in uhich there is a large quantity of organi c

tiosuo profJtt.T1l.ubly mus t have ea.ten l'ti. thin a feu hours of captur~"

Exporit10nts on tho rate of clearance of shell and fish bone

nere analysed in tuo uuyso Firstly the number of filUm.ala in uhi ch

shell Or bOlli) uas present after various time intervnlsl)was recorded ..

Secondly~ the mean dry l'1oight of bong or shell retained at these times

was det81~ncd.. Fifty per cent of Scylla ~£rata retained fi3h bon8

for 2 to :5 days and sholl for 5 to 6 dayo (Fig.. 2) 0 Tho twight of

fish-bono l~cma.ining in the foregut decrcD.3ed after 2 dayso This UEW

not the CD.~e for shell p o£ter 8 days the moan T-might of shell retaipcd

vas not louc.r than that a:fter shorter periods .. The amount of sholl

remaininG io not correlated l'ti. th time (r == 0 0 006 p Fig. 2) 0 The

crabs uocd in these experiments lfere in the internoul t stage and it

is u.n..Ukely that they uould be taking up large amounts of c alcimn at

at this ste.ge o 1080 of shell from the foregut is probably thr0t15h

regurgitnt10n of total contento

Tho retention tine of fish bone U8.S sufficiently long for it

to be detected in th~ foreguts of crnbs if they did eat fish.. Since

100

80

UJ C> ;5 60 z UJ u a:: 40 Ie. t\

20 {----i--+

.0 0 2 ·4 6 8 10 12

HOURS Fig.1. Scylla serrata: Amount of food remaining in the foregut

expressed as a percentage of the weight of fOOd apparently eaten.

Points are means, vertical bars indicate two standard errors;

n = 10 at each time interval. 1 00 ~

80 w (!)

~ 60 z UJ x u 1.0 a:: UJ CL BONE

20

0

500 x

01 400 E

t- 300 ::r: • (!)

UJ 200 • 0-00 6 3:

• ·SHELL 100

a a 2 3 I. 5 6 7 S 9

DAYS

Fig.~. Sc~lla serrata: Rate of clenrance of bivalve shell f~Bgments

(aots) :~ n(j f i sh bones (crosses) from the 1'oregut. Upper graph shows

the percentage 01 cra bs still retoining shell or bone ~t each time

interva l (n = 10). Lower graph shows mean weight of shell or bone

present in foregut.

of weight of shell or

on m~8ns. Corre18tlon

Lines fit+ed by regression analysis, in the case

bone, ana lysis based upon original data not

coef ~ icients are shown next to each line.

11

the retention time was not as long as shell, analysis based upon

hard parts in the foregut would be biased in favour of shelled

animals.

Analysis of Foregut Contents

Sixty-five per cent of §cylW- serrat~ caueht in South Mrica and

64% of thoso caught in Australia had identifiable material in t he

foregut So In both South African and Australian crabs, 50% bad.

mollusc remains, but wbereas in South African crabs gastropods were

most common (47% "Jith gastropods, 15% lrlth bivalves) 9 in Australian

crabs, bivalves, chiefly mytilids predominated (30% ~dth bivalves, 201" t:.) tri.th gastropods). In South African crabs the most common

gastropod nan pass!!; .1n.>nus~)iQnn, a small onnil commonly found crawling

on the fn.trface of the substratoo Bivalvo genera eaten in South

genera Hora not idontified in Australiao Crustacea lolere f01,\i1d in

2205% of South African ,6,,, perTrtta, flnd 2a:~ of Austl'illia. They were

Chiefly the remains of sEiall grapsid crabs and hermit crabs. In

one case the remains of 7 hermit crabs were found in a 126 mm

.§.o £le:t:I.ntA.o Fish rcmainn \Tere found in :3 of the Australian crabs

examined.. Fish rem.!1ins nere 81so fotmd in tuo of the ovigerous

females collected in the Kleinemond estuary. Four of these females

contained gastropod remain~o No other £~ serrata collected from

South Afr:i.ca had fish remains in the foregut ..

SlTIBll amounts of plant remains nere found in 4 adult Sq,ylla

sen-ata from Australia.. It is not knol'm uhether tbis pla.'tlt material

forma part of the normal eliot of adults or whether it uas merely

swallowed along with other food.. The foregut of one juvenile from

South Af1~ca was nearly full of fresh plant material and in this case

12

it appeared that the crab had been feeding on aquatic macropbytes.

Three crabs l·Jhich had recently maul ted and still had very

soft exookeletons were found to have the foreguts full of a 'variety

of debris consisting of hits of old eroded shell, BEl.all pobbles,

sand and plant detritus. This material had apparently been mmHmred

subsequent to moult and lIas not regarded as food.

In those Scylln .§erratl\ in uhich tho foregllts were less than

50% full, inorganic material made up nearly 100;; of the content. In

9 crabs the foreg-at uao more than 7r;jb full, and in these the inorganic

component nas 12 to 43% of the contento Thus~ along with inorganic

tissue, crabs tend to swallow a great deal of apparently indigestible

material. This m~terial consists of shell fragments? gastropod

operculnp pie cos of crustacean exonkoleton Q1J.(l fish bones " al l of

uhich ero normally firmly attnched to orgru'lic tisouo o Tho amount 0:(

organic content in the foregut had limited value as an indicntor of

the time of day or night uhen the crnbs mainly foofl" Thin Has dUf>

to the small proportion of crabs uhich had full or nearly full

foreguts = 6 from South Africa and 9 from Australia c Five of the

crabs collocted from the intertidal zone during tho afternoon ,just

after retreat of the tide had full or nearly full foreguts suggestiXl~

that intertidal £0 !lQ.:;:T'E\ta may feed during the day 't'Then submerged"

Foux crabs collected from trmrls a.t night also had full stomachs

shol1ing that subtidal crnbs do feed at n.i.ghto Unfortunately

daylight trauls yielded too felT crabs for any statement to be made

about doytimo feedin..g by subticlal crabs e

13

Catching of Live Penaeid Prawns

Live pra-una kept in tanks in "mich !?cylla serrata were being fed,

disappeared at a mean ra.te of 0.46 pravma/tank/day (standard

deviation 0.28). They disappeared at a mean rate of 0.54 prawns/

tank/day (standard deviation 0 .. 21) from tanka containing unfed

era bs. No dead prawns l'1ere found in ta.nks wi th ~rabs and thus it

'\'laS assumed that missing prmms had been eaten by the crah80 Dead

prmIDs uere found in the control tanlro uhere the mean rate at uhich

they died 1m3 0,,45 prmms/tank/day. This rate is not significantly

difforent from th'J rate at lThich prmms disappeared from tanks

containing either fed or unfed crabs o It is concluded that li$ Ae~rnta

did not catch live penaeids even uhen not fed for up to 15 dayn bu t

that tho crabs did eat prm'lns uhich beconm incapacitnted OJ.' died ~

Activity Records

The renul to of activity recording under infra-red light are sho"m

diagramnatically in Fig. 3. Scylla sorrata emerged from the semd

in the honr after sunset and buried again shortly before or l1ithin

30 min after sunrise. They are thus clearly nocturnal animals. No

experiments lTere undertaken to establish uhether emergence uas merely

triggered by 10\'1 light levels or uhether an enogenoua rhythm exists

as claimed for the crab GoneRlax rhomboidElE!, by Atkinson (1974)0

The mean time spent emerged by Sc,vlla serrata was affected by

the presence or absence of visible light and/or food (Table 1)0

Visible light flashes were associated vnth. a 25 to 30% drop in the

number of hours emerged. Unfed crabs under visible light emerged

after sunsot but subsequently reburied much earlier than those under

infra-red" At midnight less than 40% 'l"Tere still emerged under visible

light and this dropped to leaa than 20)~ after 0230 h. Under infra-red

UNFED CRABS FED CRABS

II' III "" II J J f ""I"I! Ii I" I "I! I" ! 40

"" II " ,,'I ""'1'" W ",,! I!! I' I llUl 'II!! 11

II II! I'I! IJ " "I I" iii 'I ! I I' 42

111111 !J L III 1111 U. f Ii It pi I""" " ! f ! I" " 11 " "" " 'I "II 52

1111 II I 1 II II ~ I' II ! II' I " II! Ii " IIII!II II' " ! I' I' P'" '11'1 53

1111111111111 I I II 1"'If( '1'111"1' U 11

U. II " II' I II iii III II! 'II '111111 II 'I' II " III" I" " II' I 13

.1 I'll 1 I! 'I II 15

111111111 w.. W. LLJ II "I "III! II 'I ! I' II II 1,,111 '1 111 I III 51

1111111111111 111111111 W. 111"1" '111j1 I'll 1m ! I' 65

IP'h'jI"lfllIllI! ! 11"1111111 11111 I' 65

I ,t

i i , i i , ! i , , 1 • i i • i i , • , , J , I

17 19 21 23 01 03 05 07 17 19 21 23 01 03 05 07

TIME (h J TIME (h)

Fig. 3. Scylla serrata activity as re'corded by time-lapse

photography at 15 min interv,als. Each line SUnlmc'lrises activity

of a single crab. Unfed crabs are grouped on left, fed crabs on

right. Thickened horizont31 lines indicate when crab was emerged

from sand; vertical bar above line ShOWS that crab had moved since

previous photogr8ph; vertical bar below line shoVls crab \Il 8 S

fn nding when pho to gr Aph V9as taken. Numbers to right of record of

f~d crabs a r e the number of bivalv~s eaten by eaqh crab. Mean time

of sunset and sunrise are arrowed.

14

100% of the crabs ""rare emerged from 2330 h until 0500 h. When food

was present the time spent emerged approximately doubled, both under

visible and under infra-red light (Table I). At: test shoued thnt

the increase in emergence time is significant at the p«Oo002 level

for both light condi tiona 0 Under viai ble light craba ai:e a moan of

42.8 bivalves (standard deviation 15.1) 0 Ulnder infra~reci t:10Y ate a

mean of 38 0 0 (stnndard deviation 21.7). The difference i (ol not

statistically significanto Although the nUllibel' of bivalves oaten UfiIJ

similar g visible light s:lenificantly suppressed eL10rgence tin~o uhen

food \'18S prcsc.n'G in tho tank (p< 0000?) 0

Hhon no fooe1 uas avo.ilcblo in th,) tank, §(;ylla. .Q~rrD.tA. under

infra-red did not remain continuously emoX'G'ed all niGht but occiloionally

buried f or a uhilo bofore rc-cILorgi.nc;' (ri:1o 3)" '1'hi8 bohnvi Our onl y

occlll""'Tecl on one occasion ""Thon food lTOD prcsent in tho tank compared

t 16 ' 1 · f d t In 7~~ f S '-o occas:.I.OllS ,nan no 00 lIas presen 0 .h~ 0 cases,,L.o .£1<:'...!T8.-{,t:\

located tho food ui thin 15 min of emel'genco o Subsequently the c};f!.b:J

left the food and retu:n:\O(l to feed at intervals throuehout the night "

T\'JO types of behaviour Here evident uhon crab:::! uor.:: eUlorgocl"

Firstly UQl1::ing oJ.'ound the tank, ro1d secondly remaining atatioDa:t'Y

in ono pleco, frequently for several hOUl:D (Figo 3) 0 Uhilo

statioD3ry the crabs occasionally ptillhed their logs tmder thn sand

but they did not cover them8elves" In both lic:ht and food conditior~n f

the percentage of photographic frames uhich shol'Tcd that tho crabs bad

moved ''/'as ebot\t 6476 (Table I) 0 This suggests that the mean

proportion of time emerged lThich is apen'~ ual!d.ng a.bout is relatively

constanto

Table I. Sgvlla serrata emergence and locomotory activity as

recorded by time-lapse photography under conditions of either

visible or infia-red lie;ht, with or l·Ti. thout food (100 Plei2i§.9.r:£~::.

deltoiden) in the tank.

'foot!.

Light regime Length of Mean time Percentage Percentage of

night (h) emerged (h) of night of frames

emerged sho'l<1ing moves

Infra-red f.lble",t 11.7 6.4 55 65

Infra-rod Prt)u.t 1202 12.6 103 63

Vioible \1t.Jl,o<t 10 0 8 4.6 43 66

Vidblo p(~.it. ... t 11.3 9.6 85 63

15

Discussion

The result3 of analysis of the contents of the foreguts of Scylla

serrata sholT that it is a predator of sessile or sImI moving benthic

macro-invertebrates, chiefly molluscs. A similar role "las reported

for 3 species of large crabs in British uaters by r'lu"-ltz ~ 21 (1965L

Differences in the contribution of molluDc types to the diet of

.§.. ~0~rt\t.a betueen South Africa and Australia probably r esul t from

differencoD in tho availability of prey. Bivalves are not common in

South African estuaries t"Thereas surface-duelling gastropods such as

NasRa krf}.l~~i.~"l1.l are abundant. The opposite appears to be true of the

Australian fli. tuation uhere small mytilids commonly occur on solid

substrateflo Ennis (1973) found that the percentage occurrence of

various prey flpccies in lobster stomachs reflects the relative

abund£mc8 of proy specios in the habitat. Fish bonoo are seldom fow."'ld

in .2, .. i!.~'lt,a foreguts, nnd it appears that the crabs rarely feed on

fioh even allotling for the faster rate of clearance of fish bone as

compared to shollo Obse!'Vations in tanks indicated that ,.,hon crabs

'tTere given l1hole dead fish to feed on, the heads nere frequently not

eaten. The rest of the fish t"1ruJ eaten hOt"1ever and scales, vertebrao

and ribs nore s17al1oued. Thus if fish nore bsing oaten their roma:J.nD

could be expected to occur in the forcguto Crabs mai.ntained in tanks

together Hith live fish did not attempt to catch them unless the fish

ucre injtlt'odo Neve:rthelcss, So £IEn~rata readily fed on cut-up fioh9

evon emerGing from sand or sholters during the day in order to do so.

They did not o~orGe dt~ng tho day to feed on live bivalves o Fish is

successfully used as a bait in traps and Fielder (1 %5) has pointed

out that the efficiency of a fishery utilising baited traps relies

upon offering biat uhich is moro attractive than natural food o 'l'hus

., \

16

gut contents probably reflect availability rather than preference.

AI though fish are common in estuarine areas, their non-availability

to 2. serr~ta may be related to the difficulty of catching highly

mobile forms in the muddy waters typical of estuarine and mangrove

areas. Experimentally it 'tras shOlm that §.. p..,errata did not catch

live penaeids. Penaeids are capable of rapid escape movements and

like fish are apparently too mobile for §._ serrata. Thus~. serrata

appears to be restricted in diet to slow~moving benthic animals

although they \Till eat mobile animals uhich become incapacitated or

dio.

Plant material "ras uncommon in Scylla ~tl! foreguts • . The

single crab in uhich the foregut lTas full of plant ma.terial uaa a.

15 rom crnbo Although small crabs (loss than 30 mm) also contained

mollusc und crustacoan remains, it is possible that plant material

may be part of the natural diet of juvenile stages o Veerannun (1974)

maintained small~. serrata (less than 10 g) in tanks for exporimoutal

purposes and stated that they uere fed exclusively on a ple.nt. dioto

The claus of .§£Y.~ serrata are large and pouorful and 11011

suited to crushing hard-shelled prey.. Ebling et a~ (1964) found that

ga.stropods crushed by the chelae of 80 mm l'ortunu..9 11~!: required

forces of. up to 40 kg to break tho shellso All the bivalves oaten in

exporirr'3nts described in the present paper uere thoroughly crushed

by the cholao before being enten. Subsequent 6ue.llol'Ting is rapid

and apP8rontly not discriminatory since indigestible material such

as shell end bone enter tho foregut. .2.~ serru.!.2, has a large

functional eastric mill in the roof of the foregut. As pointed out

by Vonk (1960), ID.o.lacostracarul ui th n uell developed gastric mill do

not chou their food uith the mouthparts, food cheuing is El. function

of the gastric mill.. Th8 food in the foregut of Q. nerrata killed

17

immediately after feeding on prawn abdomens, was found to be cut

into pieces 3 to 4 mm square and 1 to 2 mm thick. After 2 h the

food had been ground up into a paste. In the foregut food is also

mixed ui th enzymes from the midgut (Vonk, 1960). The reoul ting

mixture of solubilised and suspended fine material is passed back

to the midgut and digestive gland. Thus the clearance rates

reported in the present paper do not reflect digestion rates but

rather the rate of removal of food from the foregut. This rate

is rapid for organic caterial and should enable tho crabs to fill

the foregut more than once a night. Activity records shollod that

tho crabn did feed several times a night although tho extent to

which tho gut uao filled on each occasion is not }mmm.

Scylla oerrl1.tu opends most of the night ei the!' walking over the

bottom or stationnry but not buried. Locomotory behaviour lIould lK\

of obvious importanco in loca.ting elou moving or sessile prcyo ,'/hile

stution?ry but not blC."iod th9y are still in a position to catch proy

such aD Cl;'ubs orhorrnit crabs uhich may uander l'1itlrin catching

distanco o Tho stntionary periods may th0rcfore represent part of the

hunting bchEl.viour.. There is no evidonco to suggest that.2,. sarrain

catches prey from a buried position as cJnimed for the portunid crab

Qy8.1iBof) D~~11l1:p8n8,!fl. by Caine (1974). Some of tho bivalves fotmd

i n n.o ~_r.I!:-,t'1 f:;;OH Sonth Africa are normally found burrowed in mud or

sarsl inclicatine that .£l.~ 8e~!':.nta is capable of locating buried bivalves

and (liec;jr>;~ them upo This b8h~viour 'Has not evident in Australian

§;. ser:r.~t<\ lThoTO only one cC'o.b contained F.l. burrouine bivalve althouc;h

the latter PTO nbundant in areas whet'e §..O ~rJ:~t~ \'l08 caught.

18

vThen food was present in the activity tank, Scylla ~errata

spent t,·rl.ce as much time emerged as compared to \-Then there .. ms no

food present. An increase in activity when food is present lIas also

reported for the crab Jiemig;:aJlflt!§. oregonc1'1sia by Symons (1964) 0

Such an increase is understandable uhcn it leads to prey captu:.t:'c ,

but this uao not the case in the present study uhero i.ncreaaed

activi ty uas not a requirem8nt for obtaining food o In nddi tion the

crabs did not merely foed on the available bivalves and then buryo

On the contrary, after feeding the crabs spent hours ualking around

the tllllko It is not clear why this shoulcl occur but it could boO

important in non-feeding aspects of behaviour such aD maintenance of

territoryo

It is not lmatm at present hOll much food Sc:y~ s~errnta

consumes under natural conditions. The numbers of bivalves eaten

during tho acti~~ty OKporim3nts cannot be used as an absolute

indication since as \"Tolna and Doan (1CJ72) .found, tho number of prey

ea.ten by crabs ia rela.ted to the number a.t risk" Nevertheless it

appears that they can eat 1m-go numbers of prey in a nighto

,2,0 serrD.ta io a large crab and, as Stephenson (1962) pointed out,

the only large crab uhich has successfully invaded estuaries in the

Indo-Pacii'ic regiono Its large size and abundance may make it an

important predator of estuarine molluscs a~d crustacoao }funtz~. 0..1

(1965) found that predatory activities of largo crabs in Brit.ish

waters Here an important factor in determining the distribution of

pray specios., Conversely, the apparently restricted diet of ('1017-

moving b::mthic macro-invortebrates may be a factor in limiting tho

distribution of ,20 serrnt~ uithin estuarieso

19

Acknowledgements

Financial support for this research was provided by Rhodes

University, the South African C.S.I.R. and the South African

Fisheries Development Corporation. In Australia the following

organisations placed facilities at ~ disposal~ Queensland state

Fisheries Laboratory at Deception Bay, the C.S.I.R.O., and the

.. University of Queensland. The assistance of staff of these

institutions is gratefully acknowledged. Mr. A. Tucker kindly

provided all crabs used in experimental l'lork. Miss p. Bishop

carried out most of the gut analysis of crabs collected in South

Africa.

20

Literature Cited

Atldnson, R.J .A.:: The activity rhythm of Gone.:e~ rhomboides.

Mar. Behav. Physiol., £,325-335 (1974).

Brick, R.lf .. : Effects of l,ater quality, antibiotics, phytoplankton

and food on the survival and development of larvae of §m 1a

serrata. Aquaculture, 2, 231-244 (1974).

Caine, E.A.: Feeding of Q:valipes ,emadulpens;ts and morphological

a.daptations to a burrowing existence.

Lab. Hoods Hole, ill, 550-559 (1974).

Biol G Bull. mar. bioI.

Dagg, HuT,,: Loss of prey body contents during feeding by an aquatic

predator.. Ecology, 22, 903-906 (1974).

Eblingg F"Jo, J.Ao Kitching, L .. Muntz and. C~lt> Taylor: The ecology of

Lough In0 XIII. ExperiIT.3ntal obnervations of the dcstruction

of llTIi1u.s edulis and MucelJ,.§; lapi1}us by crab.. J 0 Animo Ecol., f

21, 73-82 (1964).

F~sis, G.P.: Food, feeding and conQition of lobsters, Homa~

£30!icanus throughout the seasonal cycle in Bonavistn Bay,

NCilf oundland 0 J. Fish. Res. Bd. Cano, 2Q, 1905-1909 (1973).

Fielder, DoR.: The spiny lobster Jasus ~.ndei in South Australia

III. Food, feeding, and locomotory activity. Aust. J. mar.

Frcshuat c Res .. , .1§.p 351-367 (1965).

Muntz, L., FoJ" Ebling and J.A.. Kitching: The ecology of Lough Ine XIV.

Predatory activity of large crabs. J. Anim. Ecol., 21, 315-329

(1965).

21

S-teigenberger, L.W. and P.A .. Larkin:: Feeding activity and rates of

digestion of northern squawfish (£,J;.x..chocheilus). J. Fiah.

Res. Ed. Can., 21, 411-420 (1974).

Stephenson, W.: Evolution and ecology of p 'Jrtunid crabs vii th special

reference to the Australian species. In: The Evolution of

Living Organisms. pp 311-327. Ed. by G. Ii. Leeper" r·1elbourne

Uni versi ty Preas, 1>1el bourne (1962) 0$

Symons, P .E.K.: Behavioural responses of the crab Hemi~.Rsus

oregonensis to temperatu:re, diurnal light variation and food

stimuli. Ecology,~, 580-591 (1964).

Veerannan, K.M.: Respiratory metabolism of crabs from marine and

estuarino habitats: an interspecific camparisono Mar. BioI.,

Vonk, HoJ.: Digestion and metabolism. In: The Physiology of

Crustacea i. pp 291-316. Ed. by T.H. Ylaterman. Academic

Press, London & New York (1960).

Walne, P.R •. and G.J. Daan:. Experiments on predation by the ahore crab

Carcin~ maenas on I{ytil~ and &i!ceng.ri~ J. Cons. perm. into

Explor. Mer., ~, 190-199 (1972).

Dr. B.J. Hill Zoology Department Rhodes University Grahamstown 6140 South Africa..

Exploitation of §~Llla serrata

in South Africa

There are three obvious ways in which §c~~a serrata can be

exploited in South Africa.

1. Fishing of Natural population~

The Scylla populations in Eastern Cape estuaries vary consider­

ably from year to year. They therefore do not lend themselves to

establishment of a fishery in the region, although in good years small

loc~l fishing operations could be undertaken 8S an extension of other

fishing acti vi ties. The potential of Natal estuaries is not knoY:n and

requires investigation.

2. Culture of crabs from egg to adult

A great deal more research is required before this could be

entered into. No work has been carried out into the merits of pond

versus tank culture. It appears from recent work in Fiji that in

pond culture it may be feasible to combine .2.cylla rearing with panaeids

and fish. There is no university or marine institution in South Africa

aj!art from the l! .... D.C., ~"hich is equipped to handle this ty?C of research.

It would appear that if the work is to be carried out, the F.D.C. would

be the most suitable organisation to do so. Two main areas of future

research are larval rearing and feeding.

3. stocking of Estuaries ',11th E?ubseg,uent Fishing

The population of SC111a in Eastern Cape estuaries could possibly

be stabilised to a large extent by annual stocking with }ostlarvae.

Advantages: 1. Growth rates of ~yllp- in Eastern Cape estuaries are

as good as those reported anywhere else in the 'ilOr ld, despite their

gener rJ.lly lower temper8.ture regime. Crabs could be harvested at an age

of 12 to 18 months.

2. Crabs would be utilising natural food sources (molluscs and

crustacea) and would not be competing with man for expensive ~rotein.

Further research into the biology of S~lla in the Eastern Cape is

necessary in order to establish whether variations in Fopulation size are

due to recruitrllents or to the ability of the estuaries to sustain large

popUlations of crabs.

Disadvantages:

could be high.

1. A hatching facility would be required and the cost

2. There might be considerable r esista nc e from local sport

fishermen who intensely dislike ~~~t sinc e it takes baited lines and

interferes with their fiShing.