612 



ECOLOGY AND EVOLUTION 



be the result of fusion of once separated 

 small lakes. 



In each occupied African lake the pre- 

 daceous fishes of the genus Lates have 

 spht into two species or subspecies. Some 

 interbreeding occurs between these forms 

 in some lakes. In Lakes Albert and Ru- 

 dolph, one form lives in shallow waters, and 

 the other, with somewhat larger eyes, Uves 

 in deeper waters. Two species of Lates in 

 Lake Tanganyika and two large species of 

 Barilius in Lake Nyasa may also exhibit 

 such habitat isolation. 



Topographic isolation is probably the 

 main influence in speciation of fishes in 

 different lakes without adequate intercon- 

 nections, especially if the ecologic niche 

 occupied is the same. Speciation of fishes 

 in the same lake, with radiation into differ- 

 ent ecologic niches, is more probably the 

 result of habitat isolation. 



Remarkable speciation producing en- 

 demic snails has also taken place in Lake 

 Tanganyika (Yonge, 1938), which con- 

 tains sixty-eight endemic species of proso- 

 branchs out of eighty-four species of gas- 

 tropods. Yonge regards these prosobranchs 

 as arising "in the main from a common ori- 

 gin and gradually becoming adapted for 

 life in the variety of habitats which are 

 present in the wide and deep waters of 

 Lake Tanganyika." All the species have a 

 herbivorous diet. 



In many closely related species now 

 isolated through habitat separations, an- 

 cient topographic isolation may have ini- 

 tiated cessation of gene flow across the 

 population borders with subsequent diver- 

 gence and adaptation (Mayr, 1942, p. 

 215). Without knowing the detailed his- 

 tories of the populations involved, it is im- 

 possible to be sure that habitat isolation 

 and selection initiated the divergent evolu- 

 tion. 



Topographic barriers are unUkely in the 

 evolution of a bhnd Mexican cave fish 

 (Anoptichthys iordani) that interbreeds 

 with an eyed fish of the adjacent river 

 {Astyanax mexicanus) (see p. 674; Fig. 

 247). One may suppose that without the 

 habitat differences between the open 

 stream and the underground stream, these 

 ecologic subspecies (note that they have 

 been assigned to different genera) would 

 not have differentiated. Other isolating fac- 

 tors would seem to be secondary in this 



case, although it is possible that isolation 

 was greater at one time than it is at pres- 

 ent. Some authors would probably classify 

 these as allopatric species (p. 608); we 

 call them sympatric. Such difference ol 

 opinion revolves around the definition of 

 the categories. 



Flanders (1937) described the habitat 

 differences of three species of egg-parasitic 

 wasps, Trichogramma (Chalcidoidea), in 

 North America. Laboratory and field ob- 

 servations indicate that these species are 

 not host specific. T. embryophagum is 

 usually arboreal, T. evanescens is a field 

 form, and T. semblidis is commonly found 

 in marshes. When a species is liberated in 

 a habitat to which it is not adapted, it 

 soon disappears, even though suitable 

 hosts are present. The arboreal form typi- 

 cally crawls in search of eggs, while the 

 field form takes long ffights. Such differ- 

 ences seem to be adaptive to the normal 

 habitat of the species and may account for 

 the lack of success in cases of transfer. 



As might be expected, cases of isolation 

 in different biotic habitats are easier to dis- 

 cover than cases of isolation in different 

 physical habitats. Among the aphids there 

 seems to have been an evolution from aute- 

 cious types that complete the fife cycle of 

 the population on one species of host plant, 

 to heterecious forms that use two host spe- 

 cies ( Mordvilko, 1928). There is some in- 

 dication that the host inhabited by the fun- 

 datrices or stem-mothers is the original host. 

 Eriosoma rileyi, which develops on the 

 American elm, Ulmus americana, does not 

 emigrate to another host, while its close 

 relative, E. lanigerum, transfers from the 

 elm to some Pomoideae, such as the moun- 

 tain ash, Sorbiis americana, the hawthorn, 

 Crataegus, or the apple, Mains. In Europe, 

 where Ulmus americana is absent, E. lani- 

 gerum has been introduced, but cannot 

 complete its life cycle without its primary 

 host. The parthenogenetic forms are suc- 

 cessful on apples and other Pomoideae 

 (Mordvilko, 1935). 



There is presumably an advantage in 

 having two hosts. A great many heterecious 

 species have no close relatives that are aute- 

 cious: the primitive autecious cycle seems 

 to have been lost during evolution. The 

 secondary host, when the aphid invades its 

 roots, is not always the original host. These 

 aphids have many parthenogenetic genera- 



