Diagrams by Joe LeMonnier 



o o 



CENOZOIC 







'-QUATERNARY ^^^^ 



riARY 



CRETACEOUS 



CD 



C 

 



o 

 o 



■55 

 ■(1) 



a. 



o 

 c 

 o 

 o 

 o 



D. 







c 

 0) 



o 

 o 





 c 

 

 o 

 o 



'o 





 c 

 

 o 

 o 



UJ 





 c 

 

 o 

 o 

 



D 

 Q- 







1 ■■ ^ 





0.01 2 5 



24 



34 



55 



65 



llions of years ago 



like, nut-cracking premolars; and broad, 

 many-cusped molars, "multis" filled the 

 role later taken over by rodents. They 

 thrived in the Mesozoic and even persisted 

 in respectable numbers for some fifteen 

 million years after the demise of the di- 

 nosaurs. Their subsequent decrease in di- 

 versity and eventual extinction coincides 

 with the rise of the mouselike and squirrel- 

 like early rodents that were their main 

 competitors. 



While the most abundant skulls and 

 skeletons at Tugrug are those of the pos- 

 sibly egg-laying multis, a few fossils from 

 this site may belong to marsupials, or 

 pouched mammals. The rat-sized Delta- 

 theridiiim, for instance, had triangular- 

 shaped molars much like those of living 

 opossums. Deltatheridium and its close 

 relatives are known only from the Creta- 

 ceous of central Asia. A great variety of 

 Cretaceous marsupials inhabited North 

 America, but their record is largely one of 

 isolated teeth and partial jaws. Delta- 

 theridium is known from some excellent 

 skeletons; a nearly complete skeleton 

 found at Tugrug in 1 993 by American Mu- 

 seum preparator Amy Davidson may also 

 prove to be Deltatheridium. 



Of all the Mesozoic mammals from the 

 Gobi, ihe. piece de resistance is the placen- 

 tal group. These were among the prizes of 

 Roy Chapman Andrews's expeditions to 

 the Gobi for the Museum in the 1920s. In 

 the 1960s, joint Mongolian-Polish teams, 

 and later Mongolian-Soviet teams, re- 

 trieved an impressive suite of placental 

 skulls from new Gobi sites, including the 

 Tugrug beds. These rare skulls are among 

 the tiniest of Gobi fossils. They range 

 from less than an inch to two inches long. 



On the last day of our fieldwork at 

 Tugrug in 1991, Museum postdoctoral 

 research associate James Clark strolled 

 into camp; from his pocket he extracted a 

 small collecting bag containing a nodule 

 carefully wrapped in toilet paper. He un- 



42 Natural History 4/94 



raveled the paper to reveal a small skull, 

 crudely outlined in the matrix. The snout 

 region, however, was clearly delineated 

 and was that of a placental animal. Months 

 later, laboratory preparation confirmed our 

 impression in the field that this nearly per- 

 fect skull belonged to Zalambdcdestes, a 

 species whose relationship with more 

 modem placental orders greatly interests 

 us. Zalamhdalestes has long front incisors, 

 a gap between the incisors and the anterior 

 premolars, and long hind limbs, a combi- 

 nation of features reminiscent of rabbits. 

 Indeed, my colleague Malcolm McKenna 

 had a long-term hunch that Zalamh- 

 dalestes was a granddaddy rabbit — a 

 rather dramatic connection, since the first 

 undoubted lagomorphs (the order to which 

 rabbits and pikas belong) appear in the 

 fossil record some twenty million years 

 later (see "Early Relatives of Flopsy, 

 Mopsy, and Cottontail," page 56). I was 

 skeptical about Malcolm's idea, and we 

 had a running debate on the matter. The 

 Tugnig skull might determine the answer; 

 it is certainly the finest known skull of Za- 

 lamhdalestes, or indeed of any Mesozoic 

 mammal from Mongolia. At this early 

 stage of study, we have not resolved the 

 rabbit origin problem, but we have already 

 turned up some intriguing clues. 



In collaboration with Tim Rose, of the 

 University of Texas, we put our rat-sized 

 Zalamhdalestes skull under an industrial 

 strength CAT-scan. The machine made 

 1 ,600 high-resolution "slices" in cross sec- 

 tion, from which a computer program gen- 

 erated an animated sequence. Of course, 

 fossils do not preserve soft tissue such as 

 nerves and blood vessels, but various 

 holes and canals in the skull indicate the 

 pathways of these structures. From the 

 CAT-scanned images, we could tell that 

 the main pathway of the carotid artery ran 

 in two branches on either side of the mid- 

 line of the skull. This is a striking depar- 

 ture from the usual situation in placental 



mammals, in which the carotid crosses the 

 base of the skull away from the midline 

 and through the middle ear cavity. The 

 artery's position in Zalamhdalestes may 

 reflect the problem of packing a great deal 

 of equipment — in the form of nerves, 

 blood vessels, small ear muscles, and mid- 

 dle ear bones — into the diminutive skulls 

 of these mammals. 



The carotid arteries are also known to 

 take this middle route in some rabbits and 

 rodents. Could this indicate affinity? At 

 this stage the answer is not clear. The mid- 

 line route could be a very primitive condi- 

 tion merely retained in rabbits and some 

 rodents, but modified in most other mod- 

 em placentals. It might also occur in other 

 Mongolian species. We are eager to re- 

 solve this dilemma by casting a broader 

 net of comparisons over fossil and living 

 mammals and by CAT-scanning skulls of 

 other Mongolian animals, such as Ken- 

 nalestes and Asiorytes. These shrewlike 

 forms are even smaller than Zalamh- 

 dalestes, but we should be able to study 

 details of their skulls with the high-inten- 

 sity scanner. 



Anatomical data on Zalamhdalestes 

 and other Mesozoic creatures dispel some 

 myths about the roles of the earliest mam- 

 mals. The popular scenario depicts a 

 swarm of stealthy, sharp-toothed shrews 

 puncturing and consuming dinosaur eggs. 

 Doubtless some of these creatures were 

 capable of such habits, but a wide range of 

 feeding preferences existed, as demon- 

 strated by the seed-eating, nut-cracking 

 multis or the larger and possibly camivo- 

 rous beasts like Deltatheridium, which 

 could have devoured tiny Asiorytes or the 

 abundant lizards known from the Gobi's 

 Cretaceous period. The portrait of a shrew 

 that lived on and walked across the ground 

 also fails to describe adequately the vari- 

 ety of movements that different species 

 used in getting around their Mesozoic 

 habitats. Highly mobile ankle joints and 



