May, 1903.] 



KNOWLEDGE. 



101 



With regard to mammals the evidence is practically 

 decisive and complete as to an intimate connection with (and 

 therefore their proljable descent from ) reptiles. And the 

 same is true with regard to the affinity Ijetween reptiles 

 and salamanders. So far as a direct connecting chain is 

 concerned, salamanders indeed cannot l)e affiliated to fishes, 

 but the collateral, or what may be termed circumstantial, 

 evidence points most strongly to the former existence of 

 such connecting links, and consequently to the evolution 

 of the former from the latter group. Very much th« same 

 may be said concerning the resemblances between birds 

 and reptiles. 



Although the gaps in the fish-salamander series are still 

 very large, and we are unable to show at present any inter- 

 mediate form between reptiles and birds, it is important to 

 notice that in not one single instance is there a scrap of 

 evidence which could be construed as detrimental to the 

 evolution doctrine. Such evidence as there is, be it full or 

 be it sparse, in every instance points, so far as it goes, to 

 the derivation of the newer from the older, of the higher 

 from the lower, of the specialised from the generalised. 



To revert to mammals, it was mentioned in the course of 

 our remarks on the relationship of that group to the 

 auomodout reptiles that some at least of the small mam- 

 malian jaws from the Stonesfield slate and other Jurassic 

 formations, as well as those from the Cretaceous, not 

 improbably indicate the ancestors of mammals other than 

 the modern mouotremes. It has been very generally 

 considered that many or all of these Mesozoic mammals 

 were marsupials, but recent changes in our conception of 

 the nature and origin of the latter group have tended to 

 discredit this idea. And the probability is that none of 

 the Mesozoic mammals were marsupials, or at all events 

 not marsupials as we now know them. Far more likely is 

 it that these early mammals— the presumed direct descen- 

 dants of the auomodouts — laid eggs and approximated in 

 the characters of their skeletons to the modern mouotremes. 

 Together with the direct ancestors of the latter (which, as 

 already mentioned, there is some reason to believe may 

 have had a peculiar type of dentition inherited from a 

 special branch of anomodonts) they not improbably formed 

 a primitive group for which Huxley's name of Prototheria 

 is available. 



All this is, however, more or less vague conjecture, and 

 it is time to return to facts. Apart, then, from these small 

 Mesozoic forms of which the afliuities are unfortunately 

 so uncertain, the most primitive, and, at the same time, 

 some of the earliest (if not actually the earliest) Tertiary 

 mammals with which we are acquainted are the so-called 

 Creodontia, or primitive C.irnivora. These creodouts, which 

 varied in size from that of a small fox to that of a bear, 

 were long-jawed mammals, with a dentition of a carnivorous 

 type, but lacking the differentiation of a pair of teeth in 

 each jaw into special cutting instruments — the carnassial. 

 or flesh, teeth of cats and dogs and most other niodera 

 landCarnivora. That the creodouts were the direct parents 

 of the latter is now generally admitted by palaeontologists,* 

 and it is likewise highly probable that they also gave rise 

 to the Insectivora (shrews, hedgehogs, teurecs, etc.). 



That they were not marsupials (that is to say in the 

 (jrdinary restricted acceptation of that term) may be 

 regarded as well established; but certain more or less 

 nearly allied forms from the Tertiary deposits of South 

 America appear to indicate such a complete transition 

 from the typical creodouts to the carnivorous marsujiials 

 as to render it seemingly very difficult, if not indeed im- 

 possible, to draw any satisfactory distinction tictwcou the 

 two groups. 



* "See J. 1.. Wortmau, American Journal o/ Science, 1901-1'. 



And here a few words may be devoted to certain 

 peculiarities in the organization of modem marsupials. 

 In common with the great majority of mammals other 

 than marsupials, the creodouts develop two series of teeth 

 — a small milk, or baby, series, and a larger jiermanent 

 series, the anterior members of which vertically replace the 

 former as they are shed. Marsupials, on the other hand, 

 if they change any teeth at all, change only a single pair 

 in each jaw, and considerations into which we need not 

 now enter render it probable that their restricted tooth- 

 change is what naturalists call a specialised, and not a 

 primitive feature. In other words, it appears that mar- 

 supials have lost the complete tooth-change chai-acteristic of 

 most other mammals, and not that they have only commenced 

 to develop the same. Similarly, recent investigations tend to 

 show that marsupials, like the higher mammals, formerly 

 produced their young with the aid of the structure known 

 as the placenta, of which some of them still retain a 

 vestige. Fui'ther, modern marsupials are characterized 

 by the presence of unossified spaces in the bony palate of 

 the skull, and likewise by a peculiar bending-in of the 

 hinder part of the inferior border of the lower jaw — the 

 inflection of the angle of the lower jaw, as it is technically 

 called. 



Now the extinct South American mammals refen'ed to 

 above differ from the typical creodouts in replacing a 

 smaller number of teeth, showing in this respect a complete 

 transition from the former to the true marsupials, in which, 

 as already mentioned, only a single pair in each jaw is thus 

 replaced. They show, moreover, the marsupial inflection 

 of the lower jaw, although they lack, in most cases at any 

 rate, vacuities in the palate. As to the presence or absence 

 of a placenta, nothing can of course be said, as indeed is 

 the case with the typical creodonts. 



So far as the available evidence goes, these South 

 American sparassodonts, as they are called, seem to justify 

 the statement that between the creodonts on the one hand 

 and the carnivorous marsupials on the other there is such 

 a close connection that (here and elsewhere on the 

 assumption that evoluiion is the true explanation of the 

 resemblances of animals to one another) there seems every 

 reason for regarding the one group as descended from the 

 other, or both as divergent branches from a common 

 ancestor. 



This being so, the question narrows itself as to whether 

 creodonts are more primitive than marsupials. Priinii 

 facie, the inflection of the lower jaw and the presence of 

 vacuities in the palate, might apparently be just as well 

 acquired as primitive features ; but since the former 

 feature occurs in some of the Jurassic mammals it would 

 seem to be primitive. Both characters may, however, 

 have been lost in the more typical creodonts with which 

 we are acquainted. On the other hand, if, as is generally 

 believed, the reduced tooth-replacement in marsupials is 

 due to degeneration it is manifest that in no sense can the 

 latter group bo ancestral to creodouts. And the same 

 seema to be demonstrated by the vestigial marsupial 

 placenta, even if creodouts were non-placentals. The 

 balance of evidence thus seems to be in favour of the 

 derivation of marsupials from creodonts.* It should be 

 added, however, that this is not the view of Dr. Woi-tman, 

 who has made a special study of the latter group. In his 

 opinion marsupials are not the descendants of placentals, 

 but, together with creodonts, are the derivatives of a non- 

 placeutal stock. This leaves unexplained the origin of the 

 marsupial vestigial placenta. 



* It hus been stat«d tliat one of tlio Juiussic mammals had » 

 reduced tooth-replacement like tliat of modex-n marsupials, but the 

 minute size of the specimen readers the evidence indecisive. 



