But since then, they haven't changed much 

 compared with the evolutionary changes 

 that must have occurred earher. Paleontol- 

 ogists have had only a few glimpses of the 

 jaws and teeth, often of just a few isolated 

 teeth, of earlier specimens. Thus we have 

 had no clear picture of what whole skulls 

 or whole skeletons looked like in those 

 early days of lagomorph history. Palaeo- 

 lagiis is much more like modern lago- 

 morphs than like the animals we began to 

 find in 1991 near Naran Bulak. For ex- 

 ample, since about thirty-five million 

 years ago, rabbits, hares, and pikas have 

 had high, prismatic, rootless cheek-teeth, 

 used for grinding up vegetation with a 

 side-to-side motion quite different from 

 that of primitive mammals or rodents. 



The lagomorph pattern of folded 

 enamel and dentine on the tops of the high 

 cheek-tooth crowns is unique, and its ori- 

 gin has puzzled generations of paleontolo- 

 gists. Although theories abound, no one 

 has been able to figure out exactly how the 

 lagomorphs' pattern of cusps and valleys 

 originated from the simpler triangular 

 cusp pattern of more primitive mammals. 

 But the teeth of our Naran Bulak speci- 

 mens are not high-crowned, folded, or 

 rootless like the cheek-teeth of advanced 

 lagomorphs. Rather, the Naran Bulak ani- 

 mals have cheek-teeth that are rooted and 

 low-crowned, with a triangular cusp pat- 

 tern that is little modified, even though the 

 enamel on the inner side of the upper teeth 

 sometimes enters partway into a tooth 

 socket. Their low enamel crowns have a 



clear dental pattern that can be related to 

 that of many primitive mammals, as well 

 as to the highly modified design of ad- 

 vanced lagomorphs. We now know how 

 the dentition of lagomorphs has changed 

 from a structure like that of primitive 

 mammals to the unique pattern shown by 

 modem representatives. 



Other lagomorph features are shared by 

 our Naran Bulak finds. For instance, the 

 joint between the jaw and the skull is high 

 on the side of the skull, as in later lago- 

 morphs. Another feature shared with later 

 lagomorphs is the projection of a sliver of 

 the frontal bone of the skull roof forward 

 onto the side of the snout, between the 

 main bones on the face (maxillary) and 

 snout (premaxillary). The incisive foram- 

 ina, the holes in the front of the palate be- 

 hind the two pairs of upper incisors, are 

 very elongated, another telltale clue of 

 linkage with more modem lagomorphs. In 

 still another traditionally lucky feature, the 

 rabbit's foot, the anatomy of the ankle in 

 our specimens is far more lagomorphlike 

 than rodentlike, although we don't know 

 whether the Naran Bulak animals hopped. 



Other characteristics of our Naran 

 Bulak fossil lagomorphs are primitive, not 

 yet modified from features shared with 

 other (nonlagomorph) mammals of the 

 time. The typical flexure of the snout and 

 the shortening of the palate of modem 

 lagomorphs' skulls are not present in our 

 specimens, nor are certain changes in the 

 bony parts of the ear region that took place 

 closer to thirty-five miUion years ago. The 



lacy filigree of bone on the sides of the 

 snout in modern rabbits is only faintly 

 suggested by .some tiny openings in the 

 Naran Bulak lagomorphs. The upper 

 cheek-teeth in our creatures still had fairly 

 large roots but these are much reduced in 

 size in some and wholly lost by other, later 

 lagomorphs. These technical anatomical 

 details help to establish our Naran Bulak 

 fossils as primitive members of the mam- 

 malian order Lagomorpha, and they also 

 show that different parts of organisms can 

 evolve at different rates. Thus the long in- 

 cisive foramina and upper incisor distribu- 

 tion evolved long before the palate short- 

 ened or the molars became prismatic. 



But the evolutionary trail does not end 

 here. The teeth in our Naran Bulak skulls 

 are closely similar to those of the Mimo- 

 tonidae, an extinct family of lagomorph- 

 like mammals known from snouts and 

 jaws but not from well-preserved com- 

 plete skulls. Mimotonids occur mostly in 

 southem China in rocks about sixty mil- 

 lion years old. Our colleague Li Chuan- 

 kuei in Beijing has been amassing a large 

 and important collection of mimotonids 

 for years and has recognized their affini- 

 ties with Lagomorpha. His fossils, as well 

 as our more completely preserved ones 

 from Naran Bulak, suggest to me that a 

 Mongolian late Cretaceous mammal 

 known as Banmlestes may also be related 

 to lagomorphs. Their teeth share some fea- 

 tures: an enlarged pair of anterior incisors 

 accompanied by other, smaller rear in- 

 cisors, a developing gap between the in- 

 cisors and the cheek-teeth, and inner 

 enamel of the upper cheek-teeth that 

 sometimes enters the tooth sockets. The 

 lower incisor of Banmlestes extends far 

 back in the jaw, beneath the anterior mo- 

 lars, and its enamel is restricted to an outer 

 U-shaped band of single-layered enamel, 

 like that of later lagomorphs. However, the 

 creature does not have the pecuhar lago- 

 morphlike forward-extending sliver of 

 frontal bone that our Naran Bulak speci- 

 mens share with lagomorphs. 



Banmlestes, in tum, is closely related to 

 the enigmatic Zalambdalestes, one of the 

 Mongohan late Cretaceous mammals first 

 collected by the American Museum's 

 Central Asiatic Expeditions in the 1920s 



An anist 's inteifretation of 



Palaeolagus, an early rabbit that 



lived in Nonh America some thirt}'- 



five million years ago 



Drawing by Frank Ippolito 



57 



