100 



Fishery Bulletin 88(1)^ 1990 



Caribbean for most of the time since its evolution in 

 the Eocene. 



Geologic events that cause fragmentation of the 

 contiguous, ancestral distribution are considered the 

 major means of distributional pattern formation 

 (Nelson and Platnick 1981). Although not all vicariant 

 events are identifiable at present, those known geologic 

 events that could have produced the present pattern 

 of hake distribution are explored below. 



Range expansion and vicariance of species B 



The closeness between fossil members of the genera 

 Palaeogadus and Merluccius has led Fedotov and Ban- 

 nikov (1988) to speculate that hake originated from a 

 species related to Palaeogadns intergerinus in the Mid- 

 dle Miocene. However, since fossil records indicate the 

 minimum age, the discovery of fossil Merluccius in- 

 feni:^ from the Tethys area of the Soviet Union in the 

 Middle Oligocene deposit may not necessarily signify 

 that Merluccius originated in the Middle Oligocene. The 

 genus could have made its appearance much earlier. 

 It is unclear when the hakes living on the western 

 North Atlantic seaboard (off North America) diverged 

 into two stocks, but species B must have expanded its 

 range southward and moved into the continental 

 shelves of low latitudes in the late Eocene (about 40 

 million years ago [MA]), an age characterized by a 

 major decrease in annual temperature (Frakes 1979, 

 Wolfe and Poore 1982). According to Savin et al. 

 (1975), equatorial temperatures may have been as low 

 as 20°C throughout the Oligocene (35-25 MA); conse- 

 quently, species B most likely migrated into Brazilian 

 waters during this long cooling period. 



In his discussions on Caribbean biogeogi'aphy, Rosen 

 (1975, 1978, 1985) concluded that there was an isth- 

 mian land bridge between North and South America 

 in the late Cretaceous or early Cenozoic. A similar in- 

 tercontinental link was also suggested by Savage (1966, 

 1982), Axelrod (1975), and White (1986) in their bio- 

 geographic studies on the herpetofauna, plants, and 

 silverside fishes, respectively. Alternatively, recent 

 geologic reports suggested that this isthmian link was 

 most developed sometime in the Eocene (Coney 1982, 

 Pindell and Dewey 1982). In that case during their 

 southward range expansion, hakes of the species B 

 would have been prevented from entering the Pacific. 



The area cladogram (Fig. 7) obtained from the 

 adopted phylogenetic hypothesis (Fig. 3) indicates that 

 species D had expanded across the Atlantic Ocean to 

 become the common ancestor of hakes in the eastern 

 Atlantic. The occurrence of such range expansion 

 would invoke a geologic event with a shallow-water 

 (<200 m) connection between South America and 

 Africa, because, from the parasitological point of view. 



it is the demersal adult hakes (and not the planktonic 

 larvae) that are responsible for this range expansion. 

 Since the fossil hake Me^iuccius inferus is known from 

 the middle Oligocene deposit in Europe (Svetovidov 

 1940), this eastward crossing of the Atlantic must have 

 occurred in the early Tertiary. However, no such ex- 

 tension of continental shelf in the Tertiary has been 

 proposed by geologists, except Vail et al. (1977) who 

 recognized five major lowstands in the Tertiary (Mid- 

 Paleocene, Early-Mid Eocene, middle Late Oligocene, 

 Late Miocene, and Late Pliocene-Early Pleistocene). 

 According to them, during these lowstands the sea level 

 fell below the edge of the continental shelf in most 

 regions; thus, the eastward expansion of species D 

 across the Atlantic Ocean in the low latitudes was possi- 

 ble. After reaching the eastern Atlantic, it spread north 

 and south along the coast of Africa. 



In the Early Miocene between 22 and 20 MA, the pro- 

 longed cooling trend initiated in the late Eocene was 

 reversed (Haq 1982). Temperatures in low latitudes 

 rose to modern values and might have exceeded 30°C 

 in some areas. This change in global climate, even if 

 it did not affect the adult hakes because of their pref- 

 erence for deep water (Grinols and Tillman 1970), could 

 have effectively prevented hatching of their eggs (nor- 

 mally at 11-14°C) and the normal growth of their 

 epipelagic larvae (found usually at 40-60 m); thus, the 

 hakes were prevented from occupying waters of low 

 latitudes. This unusual warming of the tropical waters 

 can be viewed as a vicariance event that split species 

 D into a northern segment which later developed into 

 another ancestral form (species E, Fig. 7), and a st)uth- 

 ern segment representing the common ancestor (spe- 

 cies F, Fig. 7) of the shallow-water and deep-water cape 

 hake. Species E must have invaded again the waters 

 of low latitudes during another cooling, which started 

 in early Middle Miocene (about 15 MA) when a major 

 enlargement of the East Antarctic ice-sheet developed 

 (Savin 1977, Woodruff et al. 1981). This would account 

 for the presence of Benguelan hake M. polli, one of the 

 decendants of species E, in the tropical waters of the 

 eastern Atlantic. However, further divergence of spe- 

 cies E is unclear given the present knowledge of geo- 

 logic history. 



The ancestor of New Zealand hake M. australis is 

 a sister group of species D. It must have occupied the 

 continental shelf off Argentina when the latter crossed 

 the Atlantic. The extinct hake Merluccius fimbriatus, 

 known from the Miocene deposit in Victoria, Australia 

 (Stinton 1958), may be a descendant of this stock since 

 the Drake Passage had remained open since the 

 Oligocene, between 29 and 28 MA (Haq 1984). Accord- 

 ing to Inada (1981), there are two distinct populations 

 of New Zealand hake: the New Zealand population 

 living in New Zealand waters, and the Patagonian 



