DISCUSSION

The fossils from Pakistan and Uganda broadly overlap in age, although there are localities in Pakistan that are significantly older than those from East Africa and the Zinda Pir sequence as a whole represents a longer period of time. There are no bovids in the Ugandan sequence, and the family apparently does not appear in sub-Saharan Africa until about 16 Ma. In Pakistan true bovids may appear as early as 19.5 Ma, and the family is certainly present at 18.5 Ma. Since they are not known earlier elsewhere, their appearance in southern Asia may document the beginning of this very important radiation. The density of appropriate-sized remains at about 23 Ma strongly suggests that bovids did not occur in Pakistan at that time, but it is possible that they made an as yet undetected appearance between 23 and 19.5 Ma because the remains of medium-sized pecorans are sparse during this interval.

The large bodied Progiraffa exigua may be an early representative of the Giraffidae in southern Asia at circa 20 Ma. Other primitive giraffoids that are approximate contemporaries are present in East Africa and Namibia, but they are smaller and as far as known lack ossicones. Large, ossicone-bearing giraffoids are not present in Africa until slightly later at circa 18 Ma.

Other than for the lophiomerycid (and tragulids) the remaining ruminants discussed in this paper represent early pecorans that are not clearly related to the modern families. They are generally similar in both size and morphology to a variety of Late Oligocene and Early Miocene taxa of Europe and central Asia. Nevertheless, none of the East African or southern Asian species is particularly close to the Eurasian taxa. The species of Pakistan in particular give evidence of turnover in the ruminants that is analogous to what is known in the Late Oligocene and Early Miocene of Eurasia.

While we have not treated the tragulids here, they are also diverse and common in both East Africa and Pakistan. In East Africa their record extends to 20 Ma, at which time there are at least three species. While the Pakistani tragulid material is as yet unstudied, the first occurrence is older than in Africa, which is expected because there are older fossiliferous deposits in Southern Asia. In the Zinda Pir area the oldest tragulid is from locality Z 150 at approximately 23 Ma. The slightly younger site Z 127 has three species associated with the primitive pecoran Bugtimeryx pilgrimi. In contrast, tragulids apparently do not make an appearance in Europe until MN3 or even MN 4 at about 18 Ma.

The apparent differences between East Africa and southern Asia in ruminant diversity is in part due to the longer period of time represented in the Pakistani fossil sequences and in part due to the lack of a recent systematic treatment of the African Early Miocene ruminants. It is our opinion that revision of the wealth of material collected in Kenya and Uganda during the course of the past 50 years is likely to reveal a much richer record of ruminants, one well worth the effort of studying.