Discussion 



Natural hybridizations are reportedly more 

 common among freshwater fishes than among 

 marine fishes (Hubbs 1955). Hubbs (1970) stated 

 that, "teleost hybrids are relatively easily pro- 

 duced and if the parental morphology is similar 

 the hybrids are easily reared." The results of 

 natural and artificial amphibian and teleost 

 crosses have been widely employed for estimating 

 degrees of phylogenetic divergence, revealing sys- 

 tematic patterns, and explaining mechanisms 

 controlling development and differentiation. 



Davidson (1968) reports that the closer the 

 phylogenetic relationship between the species 

 hybridized, the less likely the hybrid genome con- 

 trol will be displayed early in development. This is 

 because the mechanical aspects of early develop- 

 ment tend to be similar in closely related species 

 and may be primarily under the control of mater- 

 nal RNA accumulated in the egg prior to fertiliza- 

 tion. Davidson believes this, at least in part, ac- 

 counts for the commonly observed resemblance of 

 hybrids in early developmental stages to the mat- 

 ernal parent. The genetic influence of the paternal 

 genes in the hybrid genome may not be apparent 

 phenotypically until long after the onset of dif- 

 ferentiation (Davidson 1968). 



It is possible that such mechanisms account for 

 the maternal resemblance pattern observed in the 

 grunion hybrids as well. The hybrids resembled 

 the maternal parents in overall size and body 

 proportions, coloration, swimming ability, net- 

 escape capability, and dentition until long after 

 hatching. Only when the lateral scale rows were 

 counted at 141 days after hatching did the 

 influence of paternal genes become visibly and 

 quantitatively apparent. 



The numerous artificial and two natural hy- 

 bridizations (interspecific and intergeneric) re- 

 ported among the Atherinidae are reviewed by 

 Hubbs and Drewry (1959), Rubinoff (1961), and 

 Hubbs (1970). Natural hybrids reported between 

 Menidia menidia and M. beryllina along the At- 

 lantic coast of Florida (Gosline 1948) exhibit in- 

 termediate counts (i.e., scales and fin rays). Most 

 of the experimental crosses between these 

 atherinids resulted in low developmental success 

 and low survival rates except those of M. beryllina 

 9 X M. menidia 6 (Rubinoff 1961). Rubinoff did 

 not report whether any intermediate characteris- 

 tics existed in these hybrids nor was the reciprocal 

 cross attempted. 



Geographically isolated species forms adapt to 

 their respective environments by the evolution of 

 appropriate gene complexes. Then, if sympatry 

 reccurs and hybridization takes place, hybrid in- 

 dividuals will usually be selected against (Mayr 

 1963; Ford 1964). Hybrids not selected against 

 will usually be successful over only a narrow geo- 

 graphical range, since in animals, natural hybrid- 

 ization is commonly associated with environ- 

 mental perturbation (Mayr 1963; Manwell and 

 Baker 1970). 



The Menidia species are sympatric and hybrid- 

 ization does occur in northern Florida, a very nar- 

 row portion of the overlap in their ranges (Gosline 

 1948). Like these species, grunions are marine 

 fishes with similar, but not identical, ecological 

 preferences. However, the grunions are allopatric 

 and natural hybridization is not possible. 



According to Mayr (1963), some investigators 

 argue that renewed sympatry with hybridization 

 is required as a process of speciation in order to 

 "perfect isolating mechanisms," and, therefore, 

 unlike the Menidia species, the heterospecific 

 status of the grunions may be questioned, espe- 

 cially in light of the hybridization success reported 

 herein. We conclude that, despite our success at 

 hybridizing L. tenuis and L. sardina, the mor- 

 phological, physiological, and behavioral distinc- 

 tions between them warrant their continued rec- 

 ognition as separate species. 



Acknowledgments 



We thank O. M. Moffatt, V. J. Moffatt, and D. 

 Dutcher for accompanying and assisting the 

 senior author on the collecting excursions. We 

 especially thank O. M. Moffatt for saving the pre- 

 served milt of the Gulf grunion from being swept 

 away by an unusually high wave at La Jolla. We 

 acknowledge the time and efforts of P. C. Cook in 

 preparing the preliminary somatic-cell smears, 

 and we appreciate the recommendations concern- 

 ing the preparation of this manuscript given by W. 

 W. Reynolds, C. D. Ziebell, E. A. Stull, and J. S. 

 Frost. 



Literature Cited 



BRATANOV, C, and v. DIKOV. 



1961. Sur certaines particularites du sjserme chez les pois- 

 sons. Proc. Int. Congr. Anim. Reprod. 4:895-897. 



Davidson, E. H. 



1968. Gene activity in early development. Academic 

 Press, N.Y., 375 p. 



479 



