Social control of sex change in fishes has been demonstrated experimentally only among harem-living species or those with a rigid dominance hierarchy.1-3 In such cases sex change usually occurs as a simple one-to-one replacement: loss of the dominant male or female induces sex change in the dominant fish of the opposite sex. in non-harem species with less rigid social and mating systems, one would predict socially mediated sex change to be under the control of a more flexible mechanism. We report that such a mechanism does indeed control sex change in Thalassoma duperrey, a reef-dwelling wrasse abundant throughout the Hawaiian archipelago. This species exhibits protogynous (female to male) hermaphroditism, lives in sexually integrated, overlapping home ranges, and mates promiscuously rather than in a harem4. Sex change in this species is not a function of paired replacement of dominant individuals. Rather, it is a function of the relative sizes of conspecifics in the social group. Their relative numbers on the reef may also be important. Experiments suggest that some threshold value of the proportion of larger or smaller fish within the home range probably triggers the initiation of sex change in individual females.

Fish were taken from coral reefs in Kaneohe Bay, Oahu, Hawaii, and brought immediately to the laboratory where they were held collectively in seawater tables for up to 2 days. During this period fish were sexed, weighed, measured. and placed individually in isolated seawater containers for 1 to 3 days before assignment and transfer to experimental pens. Pens, made of 12.7-mm (half-inch) wire mesh, measured 1m on each side5 and were submerged at fixed positions in a protected lagoon. There were no resident T. duperrey in the lagoon, which was an inappropriate habitat for these fish.

One to four adult wrasses were placed in each pen, with or without a barrier to separate individuals (Table 1). Small fish were 66 to 100 mm, standard length, and large fish, 101 to 135 mm. In the experiment with three fish, the entire size range was divided into three parts. When two or more wrasses were placed in a single pen, a minimum size difference of 10 mm was initially established. Tactile barriers were 12.7-mm wire mesh screens, and tactile-visual barriers were double-louvered panels that faced in opposite directions, allowing water flow but obstructing vision.

Experiments ran uninterrupted for 3 months, during which time fish were fed freely. Two replicates of each of the 12 experimental treatments (Table 1) were conducted simultaneously and repeated every 3 months until a sample size of five to nine individuals was reached for each treatment. At the end of each test, fish were killed and gonads examined histologically for evidence of sex change. Females were considered to have changed sex only if their gonads lacked intact oocytes and showed advanced stages of spermatogenesis. Treatment results were compared statistically by the Irwin-Fisher exact test, one-tailed6.

Experimental treatments 1 to 4 were designed to test the effects of absolute and relative size as well as social environment on the sex-change process (Table 1). Sex change occurred only in the larger individuals of female pairs (the number of changes in treatments 3 and 4 was significantly greater than that in treatments 1 and 2, P < 0.001). regardless of absolute size (treatment 3 versus 4, P = 0.73). It did not occur interspecifically (treatments 3 and 4 versus 5, P <0.01). These results show that sex change in one fish must be stimulated by the presence of at least one smaller conspecific.

The results of treatment 6 showed that the sex-change process is independent of the sex of the stimulus fish: smaller males were as effective as smaller females in inducing sex change (treatment 6 versus 3 and 4, P = 0.73). Adult wrasses normally undergo a color-phase conversion (initial phase to terminal phase) that is correlated with a change in spawning strategy4. However, treatments 7 and 8 suggest that sex change is independent of the coloration of the stimulus fish (P = 023). Tactile cues were not needed for sex change to occur (treatment 9 versus 3 and 4, P = 0.73), but visual cues were crucial (treatment 10 versus 9, P < 0.01). Placing females in an enclosure 12 times larger than the standard size showed that, at least within this range, sex change was independent of the amount of space available to females (treatment 11 versus 3 and 4, P = 0.37).

Treatment 12 was designed to test whether sex change can be inhibited as well as stimulated. A female wrasse potentially capable of change when placed next to a larger female, did not change sex even though a smaller female was also present. This showed that inhibition indeed occurs (treatment 12 versus 9, P <0.001).

Sex change is socially controlled in T. duperrey and is initiated by visual stimuli. As long as the fish is reproductively mature, there is no apparent critical size at which sex change normally occurs. In our experiments, sex change in one individual was stimulated by some aspect of the presence of one or more smaller conspecifics. Though it did not require removal of a male or a larger individual, it apparently was inhibited by some aspect of the presence of one or more larger conspecifics. Relative size was critical to both stimulation and inhibition, while coloration and sex apparently were not.

Other hypotheses proposed to explain the proximate causes of sex change in various fishes are those dealing with suppression1, priming2,and a sex-ratio threshold2.For T. duperrey. we suggest that size ratios, rather than sex ratios or loss of dominant individuals, may be important proximate cues to individual sex-change candidates. Thalassoma duperrey interact with relatively large numbers of conspecifics on a regular basis. The relative numbers of larger and smaller fish may be the best proximate indication of the chances for reproductive success. If. for example. there are many larger fish (usually male) and few smaller fish (usually female), there would be too few females with which a new male could mate. There would also be too many larger males competing for those mates. If, on the other hand, the proportion of larger fish is low, then sex change would be advantageous.

Thus in social systems such as that of T. duperrey, sex change in one-to-one correspondence with male loss would be too rigid a strategy to follow. The size-ratio mechanism provides maximum flexibility to individual sex-change candidates in socially variable environments. It would appear to operate efficiently only in fishes with large social units, however. Congeneric labrids and other labroids that live in overlapping home ranges at relatively high population densities may well exhibit such a mechanism. The precise role of behavior and the specific visual cues involved in this mechanism have yet to be determined.

Table 1
Results of sex-change experiments with T. duperrey. Vertical bars represent the placement of barriers by fish size and sex. Data on sample size and sex change refer to the largest female in each treatment (to the larger female of the pair of females opposite the male in treatments 7 and 8 and to the large and medium females in treatment 12).
Treatment
Fish/pen
 Barrier Fish size Sex
Sample
size
Sex change
N
Per
cent
1
1
 None Small F
5
0
0
2
1
 None Large F
6
0
0
3
2
 None Small F F
7
6
86
4
2
 None Large F F
6
5
83
5
2
 None Varied F + T. ballieui
8
1
12
6
2
 None Varied F M
7
6
86
7
4
 Tactile Varied | smaller F F | M F*
6
6
100
8
4
 Tactile Varied | smaller F F | M F*†
6
4
68
9
2
 Tactile Varied | smaller F | F
7
6
86
10
2
 Tactile-visual Varied | smaller F | F
9
1
11
11
2‡
 None Varied F F
6
4
68
12
3
 Tactile Large | medium | small F | F | F
8 | 9
6 | 0
75 | 0
* The male fish was at least 10 mm smaller than the larger female across the barrier.
† Differs from treatment 7 in that the male is terminal color phase.
‡ Pen size was 12 times larger than in other treatments.

 

REFERENCES AND END NOTES

1. L. Fishelson. Nature (London) 227, 90 (1970): D. R. Robertson. Science 177, 1007 (1972); H. W. Fricke and S. Fricke. Nature (London) 266, 830 (1977).

2. D. Y. Shapiro, Adr. Study Behav. 10, 43 (1979); J. Theor. Biol. 82, 411(1980).

3. ―, Science 209, 1136(1980).

4. R. M. Ross. Thesis, University of Hawaii, Honolulu (1982); Proc. IV Int. Coral Reef Symp. 2, 575 (1981).

5. An extra large pen (treatment 11 in Table 1) was 12 times larger (4 by 3 by 1 m).

6. L. A. Marascuilo and M. McSweeney, Non-parametric and Distribution-free Methods for the Social Sciences (Brooks-Cole, Monterey, Calif., 1977), p. 96.

7. We thank E. S. Reese and S. R. Haley for discussions, as well as J. S. Stimson, D. Y. Shapiro, and R. R. Warner for comments on the manuscript. Supported in part by a research assistantship from the Hawaii Institute of Marine Biology to R.M.R. Contribution 661 of the Hawaii Institute of Marine Biology.

Back to top