cycle generally consisted of (1) general epeirogenic uplift of a central region for a long time and the shedding from it of large amounts of debris to partly surrounding basins; (2) compressional deformation in the partly surrounding basins to form a belt of major orogeny; and (3) subsidence of the former region of uplift, with the creation of a basin having deep water in places. The marginal areas of the new basin soon became sites of heavy sedimentation. The Gulf of Mexico is believed to have begun to subside in the Permian period, the western Caribbean Basin to be due to Cretaceous and early Tertiary subsidence, and the eastern Caribbean Basin is thought to have resulted from middle and late Tertiary subsidence.

According to Chace and Hobbs (1969), most of the present islands of the Antilles are not known to be older than Oligocene or early Miocene. Consequently, the ancestors of the present fresh-water and terrestrial faunas on them could not have been established earlier.

Belkin (1962) believed that this "American Mediterranean" region has been an important center of evolution of new types of mosquitoes and probably other terrestrial and fresh-water organisms rather than a barrier to their dispersal. In the course of island formation through fragmentation of a land area, great environmental stress would be placed on surviving populations and these would be greatly reduced and isolated. Under these conditions there would be an ideal opportunity for quick fixation of new adaptive types.

These would have a chance to become established and better adapted and they could then disperse when the isolated areas became approximated again with one or more of the adjoining continental masses. Limited hybridization would probably also occur when contact was reestablished and would contribute to further evolution. Belkin believed that this hypothesis helped to explain several controversial problems as enumerated below:

(1) The presence of equally primitive and distinct members of a phylad both in the Northern and Southern Hemispheres. Usually it is interpreted that the phylad originated in a continental area in one hemisphere and then dispersed to another. According to Belkin as exemplified by the subgenus Ochlerotatus of Aedes, the situation is easily explained if the phylad originated on an intercontinental island and dispersed to both hemispheres where taxa evolved independently. Bram (1967) reported that 47 of the 61 American species of the subgenus Culex (Culex) are found in the American Mediterranean, which apparently was their primary center of origin.

(2) The presence of annectant types on the larger islands and on the periphery of intercontinental areas. Belkin considered that these might represent relatively unsuccessful types evolved in isolation that had not been able to radiate and are confined to the immediate vicinity of their place of origin.

(3) The successive replacement of phylads during evolution of a minor

group. This could be accomplished by the origin of new dominant types in isolation of the intercontinental area and their subsequent invasions of continental areas.

The Culicoides furens group may illustrate Belkin's third type of situation. This coastal salt-marsh group comprises five species-C. alahialinus Barbosa, ranging from Panama to Ecuador; C. barbosai (West Indies distribution (p. 26)); C. furens (p. 43); C. gorgasi Wirth and Blanton, a Panama endemic; and C. cancer Hogue and Wirth, a Costa Rican crab-hole endemic. C. furens is by far the most dominant, wideranging, and successful species of this group (fig. 5) and may actually be a sister species of the similarly dominant C. schultzei (Enderlein) of the Ethiopian-Oriental fauna. Since no Afro-Asian species appears to be closely related to C. schultzei, it seems on morphological grounds that the origin of this species was with the American C. furens group. At the same time that C. furens was achieving dominance, it seems that C. cancer and C. gorgasi evolved as ecologically restricted endemics, and C. barbosai followed C. furens part way in the western Caribbean to Florida (fig. 5) with a much restricted ecological niche located a little closer to the actual seashore. C. schultzei evolved from this group and is even more fresh water tolerant than C. furens, breeding in a wide spectrum of wet soil habitats.

The West Indian Culicoides exhibit an unusual degree of endemism, possibly because of their low ecological valence in the sense of

Belkin's second situation. Three of the nine endemic species belong to the C. nigrigenus group of the subgenus Anilomyia, a group of seven circum-Caribbean species that breed, so far as is known, in the leaf axils of bromeliads (fig. 2). This group has one widespread mainland species, C. nigrigenus Wirth and Blanton, ranging from Mexico to Colombia and Trinidad; two Lesser Antillean endemics, C. decor and C. dominicanus; C. hayesi Matta from Mexico and Honduras; C. lutealaris Wirth and Blanton and C. chrysonotus Wirth and Blanton from El Salvador, Costa Rica, and Panama; and C. farri endemic to Jamaica. We suspect that through isolation the Antillean endemics evolved from mainland ancestors, C. decor and C. dominicanus through Venezuelan stocks, with the ancestor of C. farri coming across the Nicaraguan Swell. Cuba and Hispaniola are terra incognita in this distribution pattern and should yield valuable collection data in this group especially. Absence of Anilomyia species in Puerto Rico, where many insect collections have been made, suggests that this group has invaded the Antilles independently from the west and from the southeast.

The subgenus Drymodesmyia contains two groups. The C. copiosus group contains 17 species that breed primarily in rotting cacti in the Southwestern United States and in Mexico, with a few species ranging through the West Indies to Florida and Puerto Rico and through Central America to Panama, Venezuela, and Trinidad. The three West Indian

species each have a different distribution pattern (fig. 4), indicating for each a different ancestral pathway to the islands. C. jamaicensis is widely distributed on the mainland from Mexico to Panama, Venezuela, and Trinidad and is found on the Greater Antilles and Bahamas. Its pathway was probably across the Nicaraguan Swell. C. loughnani is probably a Greater Antillean species that reinvaded the mainland through Florida and possibly followed the Gulf Coast to Texas. More collecting is necessary in Mexico to be sure it does not range through Yucatan to proximity with western Cuba. C. panamensis is a Central American species that has crossed the Nicaraguan Swell to Jamaica and the Cayman Islands. It is reasonable to deduce that C. jamaicensis evolved on the mainland and migrated to the Greater Antilles; C. loughnani probably traveled from the Antilles to the mainland; and the ancestral home of C. panamensis may have been either place, but more likely the mainland.

The second group of Drymodesmyia is the C. daedalus group, which contains two West Indian species breeding in tree holes, C. borinqueni and C. bredini (fig. 4). Both are West Indian endemics; C. borinqueni occurs in Jamaica and Puerto Rico and C. bredini only on Dominica. The former is closely related to C. hinmani Khalaf from the Southern United States, whereas the second has many similarities to C. daedaloides Wirth and Blanton from Panama. Evolutionary pathways are rather speculative in this group,

but we would surmise that the ancestor of C. hinmani, C. borinqueni, and C. bredini came to the Greater Antilles across the Nicaraguan Swell from Central America, C. bredini evolved through isolation in the east, and C. hinmani developed similarly on the Florida mainland, leaving C. borinqueni in the Antillean heartland.

The distributional picture of West Indian species of the C. debilipalpis group is complex (fig. 6). C. archboldi is a Dominican endemic related to C. eublepharus Macfie. It probably came to the Lesser Antilles by way of Venezuela and Trinidad, where related species still occur. C. guadeloupensis and C. trilineatus are closely related West Indian endemics, the first species endemic to Guadeloupe Island and the second ranging from Puerto Rico through the Lesser Antilles to Grenada and Barbados. The most interesting distribution is shown by C. hoffmani, a West Indian endemic ranging from Jamaica and the Cayman Islands in the west to Barbados and Trinidad in the east. This species is closely related to C. debilipalpis Lutz, a widespread neotropical species ranging from Maryland to Florida and Texas in the United States and from Honduras in Central America to Trinidad and to Brazil and Argentina in South America (see Wirth and Blanton, 1971). The closely related C. eadsi Wirth and Blanton is allopatric with C. debilipalpis in southern Texas and in Mexico, whereas in the West Indies C. hoffmani replaces it except in Trinidad where both species occur.