Reproductive biology of the tropical fish Trichogaster pectoralis (Regan

J. Fish Biol. (1982)21, 157-170

Reproductive biology of the tropical fish Trichogaster pectoralis (Regan)

A. J. HAILS* AND Z. ABDULLAH Department ojBiology, Agricultural University ojkfalaysia. Serdang, Selangor, Malaysia

(Accepted 27 October 1981)

The breeding cycle of an air-breathing Anabantid, Trichoguster pectoralis was studied in Malaysia. The species was found to breed all the year but with pronounced peaks ofactivity associated with the two wetter periods in the year. The advantages of this timing and the cues which stimulate gonadal activity are discussed. From the study ofoocyte distribution in selected ovaries it was concluded that T. pectoralis is a total spawner although a batch of ripe oocytes may be released over an extended period of time. Fecundity was found to be related to length of fish by the formula (log,, Egg no.)= -4.6854+4.3080 log,, standard length (s.L.) (mm).

I. INTRODUCTION

In tropical regions temperature and daylength, two major reproductive cues in temperate regions, show little seasonal variation. Of more relevance is the pattern of rainfall: therefore, it is useful to subdivide tropical habitats into equatorial and floodplain regions. The former, characterized by high rainfall all year often show two rainfall peaks associated with the March and September equinoxes. With increasing latitude north and south of the Equator, this grades into areas of diminished rainfall where annual flooding takes place, followed by periods of drought. Thus, the tropical habitat is far fr6m a climatically homogenous zone and this may partially explain the variation in the literature with regard to the timing of reproduction in the tropics. The majority of fish species spawn at periods of high water in the floodplain areas, this being a time of high productivity of the water (Lowe-McConnell, 1975). In common with the floodplain areas, in equatorial regions breeding takes place in many species during the peak rainfall periods but, since rainfall occurs all the year and the environmental changes are thus less dramatic, many fish species breed all year round with peak reproductive activity during the wetter periods.

Malaysia, where the present study was carried out, is an equatorial region. From a number of studies covering a range of animal groups, breeding periods range from complete aseasonality (e.g. some Amphibians: Berry & Varughese, 1968; and Tyto afba (Class Aves): Lenton, 1980) to slightly seasonal with peak breeding in the two wetter periods (e.g. Molluscs: Berry, 1975; Betta pugnax (Class Pisces): Ang, 1971; and Osteochilus hasselti (Class Pisces): Yap, 1978) to highly seasonal with one breeding peak during the first wet period of the year (e.g. many bird species: Medway & Wells, 1976; Langham, 1980).

The aim of this study was to investigate the breeding cycle of Trichoguster *Present address: Zoology Dept., University of Malaya, Lembah Pantai, Kuala Lumpur 22-1 I ,

Malaysia.

0022-1 112/82/080157+ 14$03.00/0 I 5 7

0 1982 The Fisheries Society ofthe British Isles

158 A . J . H A I L S AND Z. A B D U L L A H

pectoralis (Regan), an Anabantid species, introduced to Malaysia from Thailand in the 1920s and now abundant in paddy-fields and ponds and of some economic importance as a food source (Tan et al., 1973). In addition, two related factors were studied to extend knowledge of the breeding biology of this species. The ovarian cycle was examined to determine if the distribution of oocyte sizes conformed to the spawning types, Vide, Hickling & Rutenberg (1 956) and Qasim & Qayyum (1961). Fecundity was also examined: since this species exhibits parental care (the male builds a bubble nest in which the eggs are laid and guarded by the male), it might be expected that fecundity would be low (Welcomme, 1967; Lowe-McConnell, 1975). This has been shown mainly with Tilapia species, although some conflicting evidence came from Wootton (1973) working on sticklebacks, Gasterosteus aculeatus.

11. STUDY SITE

The sampling area was in the campus of the Agricultural University of Malaysia, 18 miles south of Kuala Lumpur and consisted of an old mining pool and a series of artificial ponds linked to the pool by a stream. Water is present in all three areas all the year, although there are brief periods when the water level in the connecting stream is < 15 cm deep.

111. MATERIALS AND METHODS

Adult fish were caught using a cast net on the 15th of each month from June 1979-May 1980. Fishing was continued until 10 males and 10 females had been caught although in some months this proved impractical. Fish were immediately transported to the laboratory, killed, weighed (g), total length (T.L.) and standard length (s.L.) measured (mm) and the gonads dissected and weighed.

Ovaries of a range of sizes and developmental stages were prepared for histological study by fixing in Bouin’s fluid for 48 h. Material was prepared for sectioning using standard procedures, stained with Harris’ haematoxylin and eosin and sectioned at 6-8 pm. For the oocyte diameter determinations serial transverse sections were taken at intervals to provide a maximum of 10 sections for each ovary. All nucleated oocytes in each slide were measured using a projection microscope. When non-circular oocytes were encountered, the largest and smallest diameters were measured and a mean taken.

Potential and residual fecundity were estimated by counting ripe oocytes in gonads from ripe and spent fish. Both types of ovaries were collected fresh and opened under water, the ripe oocytes were teased from the ovarian wall. In a few cases formalin preserved ovaries were utilized.

IV. RESULTS BREEDING CYCLE

The gonadal index (either gonad weightkotal weight or gonad weighthomatic weight) is used to monitor breeding activity in fish. Many workers, e.g. Le Cren ( 1 95 1 ), have emphasized the dangers of utilizing this procedure when the index varies with the length of fish (Delahunty & De Vlaming, 1980; Mann, 1974). In this study, this factor was investigated by examining standard length/weight relationships of adult females using both total weight and total weight less gonad weight. The equations, obtained by the least squares method were:

log,, (total weight -gonad weight)= -4.5098 + 2.982 1 log,, S.L. (mm) (r=0-88; n = 138)

REPRODUCTION I N T. PECTORALIS 159

log,, total weight= -4-5650+3.0222 log,, S.L. (mm) (r=0-85; n= 138)

Both equations were significant at the 0.1% level. An analysis of covariance between the two lines indicated that the two slopes were not significantly different while differences between the elevations were significant at the 1% level. Thus, it would appear that at all lengths of mature fish, gonad weight is a constant proportion of body weight and justifies the use of changes in gonadal index (GI) as an indicator of the breeding cycle. The monthly means for GI for males and females are shown in Fig. 1. There are two peaks of gonadal activity which coincide for males and females, one in August, September, October, November and another in March, April and May.

There are at least two problems associated with utilizing mean GIs in fish which are not highly seasonal breeders. The first is illustrated by the high standard errors of the mean GIs in Fig. 1: this indicates that while there are

0.08 I I I I I I I I I I I

J J A S O N D J F M A M

Mcmths FIG. I . Mean+ S.E. GI for (a) females and (b) males.

160 A. J . HAILS AND 2. ABDULLAH

TABLE I. Macroscopic and microscopic characteristics of ovaries at different developmental stages

Developmental stage Macroscopic characteristics Microscopic characteristics

Resting Very small, dark red/purple tough ovarian wall. GIZO- 1.9

just visible. Early development Pale yellow, small, oocytes

GI"2-0-4.9

Enlarged Pale yellow, ovary fills 1 - 2 abdominal cavity.

Ovary very large and turgid, fills abdominal cavity.

Ovary less turgid, patches of red on ovary wall give speckled appearance.

T & GI 5.0-7.9 Ripe

GI = 8.0 + Partly spent

GI"5.0 - 9.0 Spent Ovary very shrunken and

dark, only a few large oocytes visible

Stage* I oocytes only

All stages (I-IV) present, stage IVs predominate, mostly in early stages of yolk granule deposition

Many large stage IVs present, up to 787 pm in diameter

in diameter Stage IVs up to 787 pm

Stage IVs still present, large spaces present in septa

Only a few scattered IVs present, many IVs undergoing atresia

*Stage 1 Densely staining cytoplasm < 150 pm diameter. Stage 11 Yolk vesicles around periphery of cytoplasm "1 50-300 pm. Stage 111 Yolk vesicles throughout cytoplasm"200- 330 pm. Stage IV Yolk granules present in cytoplasm> 330 pm.

I T

FIG. 2. Meanks.D. GI at different developmental stages. ED, Early development; EN, enlarged; R,

discernible breeding peaks, there is considerable variation in gonadal activity. It is therefore questionable if the mean GI reflects the gonadal activity of the majority of the population. The second problem relating to the use of GIs is discussed by Delahunty & De Warning (1980) who note that with fish which do not shed all the ripe eggs at one time, GIs may not distinguish between partially spent fish and those with developing ovaries.

ripe; PS, partly spent; REGIS, regressinghpent; RES, resting.

REPRODUCTION 1N T. PECTORALIS 161

Both of these problems can be investigated by identification and then classi- fication of the ovarian stages. In Table I, a broad classification is given correlating macroscopic appearance of the ovary with its histological features. All ovaries were classified on this basis. In Fig. 2, mean GI k S.D. at different developmental stages of the ovary is shown and this illustrates the problem of distinguishing between different developmental stages on the basis of GI. It is therefore important to determine if this problem affects the conclusion drawn from the cycle of GIs. This is investigated in Fig. 3(a) where it is clear that two periods with a high percentage of ripe females occur, coinciding with the two peaks of high mean GI in Fig. 1. Similarly, the period of high gonadal quiescence in Fig. 3(b), when a large proportion of the population is either in spent or regressing condition, coincides with the period of very low mean GIs in January-February in Fig. I . It is worthy of note that December, although a period with a high mean GI for females, is not a period of high gonadal activity but of regressing ovaries, presumably moving into a resting state in January-February. This is in agree- ment with the low mean GI for males in December.

.- a , #

0 I

(b) 1 00

:fJ 2 0 I J J A S O N D J F M A M

0

FIG. 3. Percentage of (a) ripe and (b) restinglspenthegressing females over the year.

OVARIAN CYCLE The distribution of oocyte diameter over a range of ovary sizes is summarized

in Fig. 4. At all ovarian stages described as developing, a group of ripening oocytes is discernible. This group grows in size and relative proportions (presumably as more smaller ones move into the ripening group) as GI increases. Thus in (c) which is designated as early development, the median size class of the developing group is 23-24 units (5 18-562 pm). As further ripening occurs, GI increases as in (b) an enlarged ovary, the median size class now being 25-26 units (563-607 pm). Finally, in a fish described as ripe, (a) the median size class is 27-28 units (608-652 pm). The ovary (d) was macroscopically classified as partially spent. Here the proportion of large oocytes (25 units and greater) is reduced when compared with a ripe ovary. This is shown in Table I1 where the proportion of oocytes of 25 units or greater for ripe, developing and spent ovaries are given.

I62 A. J. H A I L S A N D Z. A B D U L L A H

20

10

I

40

30

E p 20 .& ae

10

40

30

20

10

( b ) Enlorged GI = 4.40

h

n ( c ) Early development GI = 2-59

( d 1 Porliolly spent GI = 5-63

( e ) Spent/regressiq GI = 3-34 I

I h A.-LbIL\ 2 4 6 8 10 12 14 16 18 20 22 24 2628 30 32 34

Micrometer wits

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 Micrometer units

FIG. 4. Distribution of oocyte diameters in a range of ovary stages. (a) Ripe, GI=8.08; (b) partially spent, GI = 5.63; (c) spenthegressing, GI= 3.34; (d) partially spent, GI = 5.63; (e) spent/ regressing, GI = 3.34.

TABLE 11. Percentage of large, ripe oocytes present in ovaries at difierent developmental stages

Reproductive O/o Stage IV oocytes stage > 563 pm in diameter GI

10.10 Ripe 8.08 Ripe 7.15 Developing 6.90 Partly spent 5-63 Partly spent 4.40 Developing

83.7 80.9 82.1 61-8 5 2 4 72.3

The ovary (e) was classified as spenth-egressing. The majority of stage 11,111 and IV oocytes when examined under the microscope were clearly in the early stages of atresia and there was additionally a number of oocytes in advanced stages of atresia. This latter group is not included in the graph of oocyte distribution.

TA

BL

E

111.

Func

tiona

l equ

atio

ns re

latin

g eg

g nu

mbe

r to

leng

th a

nd w

eigh

t of f

ish

Dep

ende

nt

Inde

pend

ent

varia

ble

varia

ble

Line

ar re

gres

sion

@

=a + b

x)

Line

ar re

gres

sion

(l

ogy=

a+b

logx

)

No.

eggs

S.

L. (

mm

) y=

-97

1 19

*72+

956.

92 S.

L.

No.

egg

s To

tal w

t. (8

) .v=

165

9.56

+369

.20

Tot

. wt

r = 0.

94

P <

0.00

1

r=0.

89

P<O.

OOl

r=0.

86

P<O

.O01

r=0.

32

P>

0.05

r=0.

88

P<O

.OO

l

r= -

0.13

P>

0.05

No.

egg

s So

mat

ic w

t. (8

) y=

104

3.90

+413

.52

Som

. wt

No.

resi

dual

egg

s S.

L. (

mm

) y=

- 1

9257

+ 15

3.30

S.L

.

No.

egg

s mm

-I

S.L.

(mm)

No.

egg

s g-I

Tota

l wt.

(g)

y=22

1-47

+ 1-

99 T

ot. w

t

y=

- 52

1.8

1 + 5-

60 S.

L.

logy

= -4

.685

4+4’

3080

l0g

S.L

. r=

0*93

P<O

.OO

l l0

gy=2

.503

9+ 1

.075

6 lo

g To

t. w

t r=

0*91

P<

O.O

01

r=0.

89

P<O

.OO

l lo

gy= - 1

0.72

51 t6

.557

4 l0

gS.L

.

logy

= -4

-872

5+3.

3957

l0

gs.L

.

logy

= 1-

7336

+0.4

468

log

Tot.

wt

logy

=2.4

398+

1.1

008

logS

om. w

t

r=0.

32

P>0.

05

r=0-

88

P<O

-OO

l

r= -

0.05

P>

0.05

164 A. J . HAILS AND Z. ABDULLAH

FECUNDITY An attempt was made to estimate residual fecundity which can be substantial

(Macer, 1974; Ruelle, 1977) and may lead to an over-estimate of fecundity. Bagenal (1978) suggests that fecundity is best expressed as a function of length rather than weight of fish since the total weight is clearly affected by gonad weight: in situations where ripe gonad weight is not a constant proportion of total weight or where fecundity has been estimated in gonads not fully developed, then it would be misleading to express fecundity as a function of fish weight. However since neither of these situations exist in this case, fecundity has been expressed as a function of length and weight of fish. The functional regression (following the method of Ricker, 1973) has been used to relate length and weight of fish to egg number (Table 111). Fecundity is closely related to both length and weight of fish, egg numbers increasing with length and weight. The regression line for residual fecundity was not statistically significant (P> 0.05) indicating no relationship between the number of residual eggs and fish size. In Table 111, fecundity has been expressed per unit length and unit weight of fish (relative fecundity as defined by Bagenal, 1978). Egg number mm-l increased with length of fish while egg number g-1 was not related to weight of fish (the mean egg number g-1 for all fish was 39 1 k 12.7 s.E.). Although potential fecundity (i.e. total egg number) was used in the above analysis, similar results were obtained using effective fecundity (potential-residual fecundity).

V. DISCUSSION BREEDING CYCLE

The data relating to changes in ovarian GI show that the mean monthly GI reflects in broad terms the gonadal activity of the females in the population. It is possible, however, to mask information in a species where gonadal activity is not completely synchronized and this is shown in the December figures where although mean GI is quite high, the majority of fish are in a spent or resting phase.

The species studied conforms to the broad pattern of breeding in tropical species, with peak breeding occurring at the times of peak rainfall (Figs 1, 5 ) although females in breeding condition can be found at all times. Thus, even in the relatively stable conditions of this environment when compared with the floodplain areas, breeding is still linked with annual rainfall. This result is similar to that obtained in the Kuala Lumpur area by Ang (1 97 1) with Betta pugnax and matches the peak breeding months noted by Yap (1 978) working on Osteochilus hasselti. It contrasts with Tan (1969) with T. trichopterus in Singapore who reported aseasonal breeding. However, the rainfall pattern in Singapore although higher in October, November and December, shows less annual fluctuation than in the present study area.

It was suggested earlier that improved productivity and thus food availability at times of high rainfall may account for breeding during this period and indeed some studies on primary productivity suggest that productivity and biomass increase with increasing rainfall (Alias, 1980; Md. Rashid, 1980).

Figure 1 shows the first and most intense breeding peak of the year (March, April, May) occurs exactly at the time of the peak rainfall. However, the second peak of breeding precedes the rainfall peak although juveniles would be in this

REPRODUCTION I N T PECTORALIS

300 I

165

280

g 200 .- 180

I60

I40

J J A S O - D J F . M A

90 f

70 2 e

50

40

30

(L

M FIG. 5 . Rainfall (histogram) and rainfall duration (dots), mean data for 1968-1979.

population at the time of the peak rainfall. This is the main rainfall period of the year in terms of both total amount of rain and duration of rainfall (Fig. 5) . This raises two questions: (1) why is there a difference in timing of the two breeding peaks relative to the rainfall peaks and, related to this, (2) which environmental factor(s) trigger the breeding peaks?

In the floodplain rivers of Africa, fish are thought to be sensitive to the rising water levels (Lowe-McConnell, 1975). In the present study, the periods of heavy rain do not precede either of these breeding peaks and could not therefore be the cue for gonadal development. Two other abiotic factors, maximum-minimum temperature changes and occurrence of dry spells have been examined from Malaysian climatological data summarized by Dale (1 960, 1963). Although both factors showed clear annual cycles, neither could be linked to the breeding pattern. Many other abiotic and biotic factors may be important in determining the timing of reproduction, e.g. Kramer (1978) suggests a number of possible biotic hypotheses in discussing some tropical fish breeding cycles.

OVARIAN CYCLE Some fish species fall neatly into one of the three spawning type categories

described by Hickling & Rutenberg (1 936) and Qasim & Qayyum (1 96 1) on the basis of oocyte size distribution (e.g. the single batch total spawners, Mackay & Mann, 1969; Ruelle, 1977) the multiple batch spawners with two or more distinct batches of ripening eggs (Qasim, 1956; MacKay & Mann, 1969); and the multiple batch spawners without clear distinction of batches (Macer, 1974). However, some others have found exceptions to these types and, in such cases, additional information on spawning behaviour has been required to confirm the spawning type (de Silva, 1973; Staples, 1975; Wootton & Mills, 1979).

From the oocyte size distribution in Fig. 4, it seems that T. pectoralis conforms to the typical pattern of a single batch total spawner since a clearly defined group of oocytes was seen even at early stages of ovarian development, this batch disappearing in a spent fish. However, the existence of partially spent fish as evidenced by the losses from the largest groups of oocytes (Table 11) suggests that there is more than one spawning act for this batch of oocytes. Bagenal (1969) states that fish designated as single batch spawners release eggs in batches but the

166 A. J . HAILS A N D Z. ABDULLAH

intervals between shedding are small. In T. pectoralis it seems unlikely that the interval between batches is small since in some months partly spent fish formed a significant proportion of the population (e.g. 40% in April). If they only remained partly spent for a very brief period, they would be unlikely to be caught in large numbers on a single monthly sampling trip.

Two additional pieces of evidence suggest that this species may spawn several times in shedding a batch of eggs. Firstly, in one of the Fisheries stations in Malaysia which farms T. pectoralis, it has been reported that in the spawning raceways it is standard practice to keep males and females in the proportion 2 : 1. Secondly, a small number of egg counts from collected bubble nests indicated that the number of eggs spawned in a nest is quite small (2460,2440, 1584) again suggesting that a female may shed eggs over an extended period.

Thus it seems reasonable to propose that T. pectoralis is a total spawner as only one batch of oocytes ripens and is spawned, the batch being spawned over an extended period. (It is possible, since there is no clear separation of the ripening batch from the smaller oocytes, that some oocytes in the 14-16 micrometre unit size group may continuously move into the ripening batch but this seems unlikely.) It is clear that some caution should be exercised in applying too rigidly the definition of spawning types on the basis of oocyte size distribution.

FECUNDITY In broad terms, fecundity may be defined as the number of eggs produced by an

individual in its lifetime (Lowe-McConnell, 1975). Bagenal (1978) finds this definition unsatisfactory because of the difficulties involved in determining such a figure but nonetheless, for the puposes of examining reproductive strategies in fish, it would be the most meaningful. However, most authors define fecundity in more practical terms as the number of ripening eggs in the ovary just before spawning and even here problems arise when studying multiple spawners (e.g. Macer, 1974; De Silva, 1975) and total spawners which may spawn more than once in a season. From this study, it would be feasible to use the equations relating length and weight to egg number or the relative fecundity (39 1 5- 12.7 eggs g-l body weight) estimate for comparison with other populations. Tan et al. (1973) have reported that the production of T. pectoralis in Malaysian paddy-fields has declined by 60% in recent years and suggests that this may be partly due to the effects ofpesticides.

It is useful to look at the fecundity of this species in relation to that of other species. By expressing the fecundity of each species as a power function of total length (T.L.) in Table IV fecundity data has been brought together for comparison. Where authors gave regression equations relating egg number to T.L. these equations were used to calculate a predicted egg number given a T.L. This egg number was then expressed as a power function of T.L. In some cases, notably those cited by Lowe-McConnell (1975), no equations were available and the length and corresponding egg number were used directly to calculate the fecundity figure. A few generalizations can be made from the data in this Table. As suggested by Lagler et al. (1977), marine fish tend to be more fecund than freshwater fish. Lowe-McConnell (1975) has made the generalization that in tropical fish, multiple spawners produce fewer eggs at a time than total spawners, while Lowe (1955) and Welcomme (1967) suggest that species exhibiting parental

TABLE IV. Fecundity of a range of fish species

Total Location/Habitat/Species Spawning length Fecundity

type (cm) (T.L.) Authority

Temperate marine Hippoglossoides platessoides

Limanda ferruginea

Clupea pallas

Arnoglossus laterna

Sprattus sprattus*

Trachurus trachurus*

Temperate freshwater Leuciscus leuciscus

Rutilus rutilus

Salmo trutta Salmo gairdneri

Gasterosteus aculeatus

Tropical freshwater Alesles leuciscus A . nuase A . dentex Labea victorianus Marcusenius victoriae

Hippopotamyus grahami

Trichogaster pectoralis

Hypostomous plecostomus Tilapia zillii

Tropical freshwater T. leucostica

T. esculenta

T. nilotica

Total

Total

Total

Multiple

Multiple

Multiple

Total

Total

Total Total

Multiple

Total Total Total Total Total

Total

Total

Multiple Multiple

Multiple

Multiple

Multiple

15 3.686 30 3.659 38 3.772 54 3.853 17 3.343 22 3.36 1 8 4.27 1

10 4.210 10 4.000 15 3.954 25 3.737 38 3.756

Bagenal(l957)

Pitt (1971)

Nagasaki (1958)

Gibson & Ezzi

De Silva (1973

Macer (1974)

13 2-769 20 2.934 10.6 2.949 22.6 3.272 27.4 2.073 28.5t 1.879 36.7t 2.172 5.0

8 16 22 18 1 1 19 7

13 16 20 12 14 22

17 24 17 36 15 25

2-882$ 3.5655

3.958 3.513 3.274 3.667 2.81 1 3.303 2.833 3.338 3.579 3.636 1.915 3.123 2.900

2.025 2-034 2.040 2.07 1 2.409 2.347

1980)

Mann ( 1974)

Papageorgiou (1979)

Bagenal(l969) Scott (1962)

Wootten (1973)

In Lowe-McConnell(1975) In Lowe-McConnell(l975) In Lowe-McConnell(1975) In Lowe-McConnell(l975) In Lowe-McConnell(1975)

In Lowe-McConnell(l975)

Present study

In Lowe-McConnell(1975) Siddiqui (1 979)

Welcomme ( I 967)

Lowe(1955)

Babiker & Ibrahim (1975)

*Egg count for one spawning cycle, i.e. all batches ofeggs counted. ?Specimens from two different lakes. $Egg count for one spawning act. §Egg count assuming three spawning acts in a season.

A. J . HAILS A N D Z. A B D U L L A H 168

care show further reductions in fecundity as the extent 61 parental care‘increases. From Table IV, the tropical total spawners do appear to be more fecund than multiplp +wxwrs. B ~ n r If .rePva!.d s,rw~arniqg. ax Iirnl~&?d ( , m q r XikpG spam at least six times per season), the fecundity is less than the cube of the lenglh in most cases. However, the situation is confused as many tropical mulCple spawners also exhibit parental care and this is true of the Tilapias’. Welcomme (1967) has drawn attention to the comparison between T. zillii, which wardq eggs and fry and t k mmifh. b c d ~ q T . Zmmttjrja, ewdm.ra md nilotica.

In broad terms, Trichuguster pectoralis is within the range for tropical total spawners and may be identified as a very fecund fish, particularly when it is remembered that in this species, the male guards the eggs until they hatch and, from the breeding cycle data, it is likely that an individual female may spawn twice a year. It is a very successful species in Malaysian standing waters and constitutes 72% of the catch (by weight) in paddy-field drainage systems (Tan et al., 1973). The fecundity is similar to the indigenous Trichogaster trichupterus, although only limited information is available; from one fish measured in this study and from the work of Tan (1969), the fecundity of T. trichupterus is around

To summarize the life history of this species it is a fast growing fish, attaining sexual maturity in less than six months (Fishery Department, Sarawak, internal report 1969; Seow, 1980). At the present study site, all fish were sexually mature at a S.L. length of 120 mm for females and 1 15 mm for males. It has been shown to be highly fecund and a total spawner. Although no information is available on life span it seems unlikely that lifespan is long since the maximum S.L. recorded in this study was 159 mm.

3.4-3.6.

We gratefully acknowledge the facilities and technical assistance provided by the Biology Department of the Agricultural University of Malaysia. We thank Dr C. J. Hails, University of Malaya, for the rainfall data shown in Fig. 5 .

References

Alias, J. (1980). Population ecology of the benthic macrofauna of a (UPM) pond. Hons. Thesis, Agricultural University of Malaysia.

Ang, Kok Jee (1971). The reproductive biology of some Malaysian Anabantidae with special reference to Betta pugnax Cantor. MSc. Thesis, University of Malaya.

Babiker, M. M. & Ibrahim, H. (1979). Studies on the biology of reproduction in the cichlid Tilapia nilotica (L.): gonadal maturation and fecundity. J. Fish Biol. 14,437-448.

Bagenal, T. B. (1957). The breeding and fecundity of the Long Rough Dab Hippo- glossoides platessoides (FABR) and the associated cycle in condition. J. mar. biol.

Bagenal, T. B. (1969). The relationship between food supply and fecundity in Brown trout,

Bagenal, T. B. (1978). Fish fecundity. In Ecology of Freshwater Fish Production (S. D.

Berry, A. J. (1975). Patterns of breeding activity in West Malaysian gastropod molluscs.

Berry, P. Y. & Varughese, G. (1968). Reproductive variation in the Torrent Frog Amolops

ASS. U.K. 36,339-375.

Salmo trutta. J. Fish Biol. 1, 167-182.

Gerking, ed.), 520 pp. New York: Wiley.

Malay. J. Sci. 3(A), 49-59.

larutensis (Boulenger). J. Linn. Soc. (Zool.) 47, 547-559.

169 R E P R O D U C T I O N I N T. PECTORALIS

Dale, W. L. (1960). The rainfall of Malaya, Part 111. J . trop. Geogr. 14, 1 1-28. Dale, W. L. (1963). Surface temperatures in Malaya. J. trop. Geogr. 17,57-7 1 . De Silva, S. S. (1973). Aspects of the reproductive biology of the sprat Sprattus sprattus

(L.) in inshore waters ofthe West coast ofscotland. J . Fish Biol. 5,689-705. Delahunty, G . & De Vlaming, V. L. (1980). Seasonal relationships of ovary weight, liver

weight and fat stores with body weight in the goldfish, Carassius auratus (L.). J. Fish Biol. 16,5-13.

Gibson, R. N. & Ezzi, I. A. (1979). Aspects of the biology of the Spotted Dragnet Callionymus maculatus Rafinesque Schmaltz from the West Coast of Scotland. J . Fish Biol. 15,546-553.

Hickling, C. F. & Rutenberg, E. (1956). The ovary as an indicator of spawning period of fishes. J. mar. biol. Ass. U.K. 21,3 11-377.

Kramer, D. L. (1978). Reproductive seasonality in the fishes ofa tropical stream. Ecology

Lagler, K. F. (1977). Ichthyology, 2nd edn, 506 pp. New York Wiley. Langham, N. (1980). Breeding biology of the edible nest swiftlet Aerodramus fuciphagus.

Ibis 122,447-46 1 . Le Cren, E. D. (1951). The length-weight relationship and seasonal cycle in gonad

weight and condition in the perch. J. Anim. Ecol. 20,201-2 19. Lenton, G . M. (1980). The ecology of barn owls (Tyto a h a ) in the Malay Peninsula with

reference to their use in rodent control. Ph.D. Thesis, University of Malaya. Lowe, R. H. (1955). The fecundity ofTilapia species. E. Afvi. Agric. J . 21,45-52. Lowe-McConnell, R. H. (1975). Fish Communities in Tropical Freshwaters, 337 pp.

London & New York: Longman. Macer, C. T. (1974). The reproductive biology of the horse mackerel Truchurus trachurus

(L.) in the North Sea and English Channel. J. Fish Biol. 6,415438. Mackay, I . & Mann, K. H. (1968). Fecundity of two cyprinids in the River Thames,

Reading, England. J. Fish. Res. Bd Can. 26(1 I ) , 2795-2805. Mann, R. H. K. (1974). Observations on the age, growth, reproduction and food of the

dace Leuciscus leuciscus (L.) in two rivers in southern England. J. Fish Biol. 6,

Md. Rashid. A. (1980). Preliminary studies on the ecology and primary production in a fish pond. Hons. Thesis, Agricultural University of Malaysia.

Medway, Lord & Wells, D. R. (1976). The Birds of the Malay Peninsula, Volume I/: Conclusion and Survey of Every Species, 448 pp. London: H. F. 8r G . Witherby. Kuala Lumpur: University of Malaya Press.

Nagasaki, Fuzuko (1958). The fecundity of Pacific Herring (Clupea pallasi) in British Columbia coastal waters. J. Fish. Res. Bd Can. 15,3 13-330.

Papageorgiou, N. K. ( I 979). The length-weight relationship, age, growth and reproduction ofthe roach Rutilus rutilus (L.) in Lake Volvi. J . Fish Biol. 14,529-538.

Pitt, T. K. (1 97 1). Fecundity of the Yellowtail Flounder (Limandu ferruginea) from the Grank Bank, Newfoundland. J . Fish. Res. Bd Can. 28,456477.

Qasim, S . Z. (1956). The spawning habits and embryonic development of the Shanny (Blenniuspholis (L.)). Proc. zool. SOC. Lond. 127,79-93.

Qasim, S. Z. & Qayyum, A. (1961). Spawning frequencies and breeding season of some freshwater fishes with special reference to those occurring in the plains of Northern India. Indian J . Fish. 8 , 2 4 4 3 .

Ricker, W. E. (1973). Linear regressions in fishery research. J . Fish. Res. Bd Can. 30,

Ruelle, R. (1977). Reproductive cycle and fecundity of White Bass in Lewis and Clark

Scott, D. P. (1962). Effect of food quantity on fecundity of Rainbow Trout, Salmo

Seow, R. ( 1 980). Culture Dotential of Trichogaster pectoralis (Anabantidae) in

Siddiqui, A. Q. (1979). Reproductive biology of Tilapia zillii (Gervais) in Lake Naivasha,

59,976-985.

237-2 53.

409-434.

Lake. Trans. Am. Fish. SOC. 106,61-76.

gairdneri. J . Fish. Res. Bd Can. 19,7 15-73 1.

stabilization ponds, 142 pp. M.Sc. Thesis, University of Malaya.

Kenya. Envir. Biol. Fish. 4,257-262.

170 A . J . H A I L S A N D Z. A B D U L L A H

Soong, Ming Kong (1 948). Fishes of the Malayan Padi-fields: I-Sepat Siam. Malay. Nut.

Staples, D. J. (1975). Production biology of the upland bully Philypnodon hrevicep,s Stoke11 in a small New Zealand lake. I. Life history, food, feeding and activity rhythms. J . Fish Biol. 7, 1-24.

Tan, C. E., Chong, B. J., Sier, H. K. & Moulton, T. (1973). A report on paddy and paddy-field fish production in Krian, Perak. Bull. Minisf. Agric. Fish. Muluysiu No. 128.

Tan, Cheng Kiat (1969). Maturity and spawning of Trichogustrr trichopterus. Unpublished. M.Sc. Thesis, University of Singapore.

Welcomme, R. J. (1967). The relationship between fecundity and fertility in the mouth- brooding cichlid fish Tilupia leucostictu. J . Zool. Lond. 151,453-465.

Wootton, R. J. (1973). The effect of size of food ration on egg production in the female 3-spined stickleback, Gusterosteus aculeatus L. J. Fish Biol. 5,89-96.

Wootton, R.. J. & Mills, L. A. ( I 979). Annual cycle in female minnows Phoxinus phoxinus from an upland Welsh Lake. J. Fish Biol. 14,607-6 18.

Yap, S. Y. (1978). The biology and ecology of Osteochilus husselti with particular reference to its fisheries in Subang Reservoir, Malaysia. M.Sc. Thesis, University of Malaya.

J . 3,87-90.