DAHLBERG and CONYERS: ECOLOGICAL STUDY OF GOBIOSOMA 



County, Georgia. An abundance of shell on bot- 

 tom probably would provide a suitable spawning 

 site. 



EGG SIZE AND FECUNDITY 



To analyze fecundity, eggs were counted in 

 nests and ovaries. Eggs from ovaries of three 

 G. hosci and one G. ginsburgi were counted and 

 measured (Table 4) . Thirty-five to 50 randomly 

 selected eggs from both ovaries were measured 

 with a filar micrometer eyepiece adapted to a 

 20 X dissecting scope. Random egg diameter 

 measurements were made as in Springer and 

 McErlean (1961). 



Table 4. — Sizes and numbers of eggs from ovaries of 

 Gobiosoma bosci and G. ginsburgi in Georgia. Two size 

 classes were found in G. ginsburgi. 



In one ripe G. ginsburgi, ova of two size clas- 

 ses were equally distributed throughout the 

 ovary. Large ova were 0.55 to 0.68 mm, and 

 small ova were 0.11 to 0.21 in diameter. In G. 

 robustum the size range was somewhat greater 

 and ripe ova were 0.476 to 0.782 mm (Springer 

 and McErlean, 1961). The immature eggs may 

 be spawned the following year or not at all. 

 Springer and McErlean concluded that it is pos- 

 sible that sufficient time remains in the fall for 

 the small egg class to be spawned; however, they 

 also observed that the late spawners (mostly in 

 fall) were entirely or mostly of a difl!"erent year 

 class than the early spawners (mostly in spring) . 

 In the fall, females matured at small sizes, 14.6 

 mm in G. robustum and 12.9 in G. longipala 

 (Dawson, 1966). In three gravid G. bosci we 

 found none of the small size class of eggs in 

 the ovaries. 



Springer and McErlean (1961) reported up 

 to 402 eggs in a single ovary and approximately 

 the same number in left and right ovaries of 

 G. robustum. Our counts are based on both 

 ovaries. Four nests of G. ginsburgi contained 

 354, 400, 790, and 1,884 eggs. Distinctly dif- 

 ferent stages of development were apparent in 

 the largest nest. One fem.ale G. ginsburgi (25 

 mm) had 435 ripe eggs plus a similar number 

 of immature ova. Fourteen G. bosci nests con- 

 tained 332 to 2,000 eggs and three others had 

 3,933, 8,000, and over 9,000 eggs. The fact that 

 three gravid females contained only 701 to 1,382 

 eggs points to a polygamous nature of the males. 

 Further evidence of polygamy is the diff"erence 

 in developmental stages of adjacent egg masses 

 in nests of G. bosci and G. ginsburgi. This 

 polygamous behavior occurs in a closely related 

 group, Gobius, and in an oyster associate, the 

 oyster toadfish, Opsanus tau (Breder and Rosen, 

 1966). However, Runyan (1961) concluded that 

 variation of developmental stages within Gobie- 

 sox egg clusters probably resulted from deposi- 

 tion on successive days, and she found no evi- 

 dence for polygamy. 



Fishes characteristic of oyster reefs have 

 evolved parallel patterns of reproductive be- 

 havior. Adhesive eggs are incubated and guard- 

 ed by the males. There are various degrees of 

 reliance on oyster shells for nests. The extent 

 of polygamy is not known. Low fecundity is 

 characteristic of fishes that exhibit parental 

 care. We found that egg counts in oyster shells 

 were 319, 1,001, and 2,302 for Gobiesox and 

 1,058, 1,502, and 3,856 for Chasmodes. Runyan 

 (1961) found 1,600 eggs in an average-sized 

 female Gobiesox and 300 to 2,500 eggs in shells. 



ACKNOWLEDGMENTS 



Thanks are extended to Dr. V. G. Springer, 

 U.S. National Museum, Dr. W. S. Woolcott, Uni- 

 versity of Richmond, and Dr. H. D. Hoese, Uni- 

 versity of Southwestern Louisiana, for many 

 valuable comments on this manuscript. We 

 thank Richard W. Heard, Jr., for providing the 

 dredge collections. 



285 



