Early  Journal  Content  on  JSTOR,  Free  to  Anyone  in  the  World  
This  article  is  one  of  nearly  500,000  scholarly  works  digitized  and  made  freely  available  to  everyone  in  
the  world  by  JSTOR.    

Known  as  the  Early  Journal  Content,  this  set  of  works  include  research  articles,  news,  letters,  and  other  
writings  published  in  more  than  200  of  the  oldest  leading  academic  journals.  The  works  date  from  the  
mid-­‐seventeenth  to  the  early  twentieth  centuries.    

 We  encourage  people  to  read  and  share  the  Early  Journal  Content  openly  and  to  tell  others  that  this  
resource  exists.    People  may  post  this  content  online  or  redistribute  in  any  way  for  non-­‐commercial  
purposes.  

Read  more  about  Early  Journal  Content  at  http://about.jstor.org/participate-­‐jstor/individuals/early-­‐

journal-­‐content.    

JSTOR  is  a  digital  library  of  academic  journals,  books,  and  primary  source  objects.  JSTOR  helps  people  
discover,  use,  and  build  upon  a  wide  range  of  content  through  a  powerful  research  and  teaching  
platform,  and  preserves  this  content  for  future  generations.  JSTOR  is  part  of  ITHAKA,  a  not-­‐for-­‐profit  
organization  that  also  includes  Ithaka  S+R  and  Portico.  For  more  information  about  JSTOR,  please  
contact  support@jstor.org.  
686 The TertiaryMarsupial. [July,

The best test of the relations in density between molten and

solidifiedrock is apparentlyto compare the densityof the rock
just beforefusion,or at least near that point,with the densityof
the same rock aftermelting. This would give a comparisonbe-
tweenthe crystallineand liquid states,while the usual method
only affordsa comparison between the liquid and the glassy or
semi-glassystates. It would also save any error arising from
cells in the cooled rock,if a solid mass was chosen in the first
place. Again the freshunalteredvarieties of a rock should be
chosen insteadof such old and alteredones as those usually ex-
perimentedupon.
In all discussions relating to the question of the liquidityof
the earth'sinterior,it is to be borne in mindthatthe chiefportion
of our knowledgeof the propertiesof liquids is derivedfromthe
study of water, a mobile liquid-while liquid rock,as lava or
meltediron,is viscous, and its laws and properties may on ex-
perimentbe found to differconsiderablyfrom those of watery
underlike conditions. Also in these solids the passage fromthe
solid to the liquid state or the reverse is not abrupt as is the
case withwater,for everygrade of viscosityexists between the
normal solidityand the approximatelyperfectliquid condition.
This is especially the case withironand seems to be so for the
common rocks.
(To be continued.)

.THE TERTIARY MARSUPIALIA.

BY E. D. COPE.

INCE Cuvier discoveredan opossum in the gypsum of Paris,

theknowledgeof the Marsupialsof the Tertiaryperiodsof
Europe and North America has been gradually extended. In
Europe theyhave been traced to the Middle Miocene, when they
disappear fromthat continent. In NorthAmerica we know them
fromOligocene beds (White River),when theydisappear,and are
only known as yet thereafteras membersof the existing fauna.
Descending the scale we have them in the Laramie in America
and Jurassicin Americaand Europe, and in the Trias in South
Africa. Whetherthe Triassic Mammalia of the northernhemi-
spherebelong to this order or not is uncertain. Under the head
of Creodonta' I have discussed the marsupial resemblancesof
I NATURALIST, March,April and May, i884.
x884.] The TertiaryMarsupialia. 687
thatdivisionof Eocene mammals,showingthat althoughtheir
dentitionis sometimes that of the carnivorousdivisionof the
marsupialorder,theycannotbe placedwiththem.
The extinctmarsupials belongto threetypes,as distinguished
by theformof theirsuperiormolarteeth. These are tritubercu-
late,quadrituberculate or multituberculate.To thefirstdivision
belong the carnivoroustypes,or Sarcophagaof Owen; to the
secondthekangaroosand thewombats, to which Owen's name
of Poephagamaybe applied.' The thirddivisionis entirely ex-
is
tinct,and characterized by havingat least threelongitudinal
seriesof tuberclesin itssuperiormolarteeth. To this suborder
I applythenameofMultituberculata. The suborderSarcophaga
includestheopossums,and in NorthAmericathe singleextinct
genusPeratherium 2Aymard.This is also the genus whichis
foundin the Oligo-
ceneand Mioceneof a
France. It differs -

fromDidelphys(the
trueopossum)inthe
of the
non-inflection
angle of themandi-
ble. Otherwisethe
twogeneraarevery FIG. I -Peratheriumfugax Cope,fromtheWhite
agreeingin River beds of Northeastern
similar, Colorado,twice natural
the number
thenu (eight)size. side;
ber (eight) Fig,a, anterior partof skullfrombelow; b,do.
Fig. c, partof rightmandibular
ramus, with
right
oftheinferior
incisor all themolarsortheiralveoli,fromabove; d, do.,from
teeth. Five species therightside.
have been foundin the White River beds of Colorado. The
largestof these,P.fugax Cope (Fig. i), had theskullas largeas
thatof a mole (Scalops aquauicus). The smallest,P. huntiiCope,
does notexceeda smallshrewin dimensions.
Of Poephaga no extinctformshave been foundin North
America.
The Multituberculata
includethreefamilies,which differ
as
follows:
Fourthsuperior
premolar (at least)liketruemolars..T.... Tritylodontida%.
Fourthpremolars(and probably others)moresimplethanfirst
truemolars
Po/yrnastodontida.
Fourthpremolars(and oftenothers)developedintoflatcutting
blades
Plagiaulacid&e.
1Owen places the Phascolomyidaein a distinctsuborder,but,as it appearsto me,
withoutsufficient
reason. sBulletinU. S. Geol. SurveyTerrs,V, No. 1, p. 45.
688 The TertiaryMarsupialia. [July,

In all threeof these families the incisor teethare in reduced

numbers,and are constructedon the rodent type,with an exter-
nal band of enamel. They thus approach the genus Phascolo-
mys (wombat), one of the Poephaga of the existing Australian
fauna. The genus Tritylodon (Fig. 2), recently described by
Owen, is fromthe South African Trias. It is a remarkablyspe-
cialized form,consideringits geological antiquity. Its formula
above is, I. 2; Pm. 4; M. 3. The lower jaw is unknown. The
median incisorsare developed at the expense of the laterals,and
are separated by a wide interspace. There is also a maxillary
diastema. The molars and last premolarsall support threerows
of shortlyconic tubercles(Fig. 7 h). The genus Stereognathus

b~~

FIG. 2.-Tritylodon longevus Owen, anterior part of skull,natural size; from

the Triassic beds of S. Africa. Fig. a, frombelow; b, fromabove. From Owen,
Quart.Journ.Geol. Society,i884, P. 146.

Charlesw.,includes species fromthe English Olite. The known

molars have threelongitudinal rows of crescentictubercles(Fig.
7 b). The species are no larger than a small shrew,while the
skull of Tritylodon
longavus is as large as that of a gray fox. In-
formationas to the structureof the skeletonof these remarkable
formshas not yet been obtained.
There is but one genus of the Polymastodontidc,the Polymas-
todon Cope. It is known fromthreespecies, all fromthe Puerco
Eocene of New Mexico. The largest,P. taoensisCope, has bones
equal in size to those of the large kangaroo, Macropus ma/or.
The jaw of the smallestspecies, P.foliatus (Fig. 5), is as large as
that of Hyrax capensis; that of the third species, P. fissidens
1884.] The TertiaryMarsupialia. 689

Cope, is intermediatein dimensions,and the dentitionhas some

well-markedpeculiarities. The charactersof the skeleton,so far
as known,are derivedfromthe P. taoensis (Figs. 3 and 4).
The angle of the lowerjaw is inflected,and the dental foramen
is at the anteriorapex of a large fossa,as in most marsupials.
There are but two true molars in each jaw, and a single simple
premolar,below. The condyle of the humeruspresentscharacters
shared by MeniscoEssusof the Laramie,whichare foundin lizards.
There is a strongand thickintertrochlear whichis so
ridgein front,
swollen at one side of the middle line as to resemblethe condyle

C~~~~~~~~~~

FIG.3.-PolyrnaslodontaoinsisCope, jaws, two-thirdsnat. size; fromthe Lower

Puerco beds of New Mexico. Fig. a, rightmandibularramus rightside; b, do.,
internalside; c, do., fromabove; d, frombelow. Original,fromReportU. S. Geol.
SurveyTerrs.,F. V. Hayden in charge,Vol. in.
ofa femur. The trochleaforthe coronoid process of the ulna, on
the posteriorside, is narrowed so as to suggest a rotulargroove
(Fig. 4b'). The humeralcotylusof the ulna is adapted to thiscon-
dyle by a flareon each side (Fig. 4c). The astragalus is without
trochlea,as in most Puerco Mammalia, and the trochlearportion
is gentlyconvex anteroposteriorly.The head is much narrowed,
and has a narrownavicular face which is convex in only one, the
vertical,direction. On its outerside it bears a large flatfacetfor
the cuboid bone (Fig. 4d'). This form is much like that of the
kangaroos. It shows that the peculiar structureof the posterior
footof the Macropodida alreadyexisted at thisearlyday,though
 690 The TertiaryMarsupialia. [July,

perhaps in not quite so specialized a condition as at the present

time. The formof the astragalus shows that the internaldigits

b&

FIG. 4.-PolymastodontaobnsisCope; parts of individualrepresentedin Fig. 3,

two-thirds nat.size. Fig. 4a, caudal vertebra,front;a', do., leftside; b, humerus,
distal portion,front; 6', posteriorside. Fig. 4c, ulna, proximalpart,frontside; c',
fromabove. Fig. 4d, astragalus,fromabove; di, do., fromexternalside. Original,
fromReport U. S. G. SurveyTerrs., F. V. Hayden in charge,Vol. III.

are of very reduced size, the firstprobablywanting,in Polymas-

todon, and that the externaldigitswere large and constitutedthe
principal agent in progres-
C sion. That the animal had a
a large tail is provenby the few
_ caudal vertebrae preserved
(Fig. 4aa').
The inferiormolars have
but two rows of tubercles,

b D
I^,.,.
o
Al andthepenultimate
is much
^ thanthelast one. The
~~~~~larger
formertooth looks a good
deal like a reduced last infe-
riormolar of some species of
Mastodon. The familywas
FIG. 5.-Polymnastodonfoliatus Cope; part
of rightmandibularramus,naturalsize, Fig. Of herbivoroushabits,and is
5a, rightside; 6, left do.; c, fromabove; d, probably the ancestral type
posteriorview,showingmassetericand dental
fosse. From the Lower Puerco of New of the kangaroos. The dis-
Mexico. Original, fromthe Report of the covery of this remarkable
U. S. Geol. Survey Terrs., F. V. Hayden,geu inorLwrEc e
Vol.Ill. genus in our Lower Eocene
beds marksan importantadvance in the knowledge of the origin
1884.] The TertiaryMarsupiaclia. 691

of one of the most interestingof living forms. On the other

hand, the Polymastodontidaemay well have derived theirorigin
fromthe Tritylodontidae,which were also of herbivorous or
granivoroushabits. The familyof the Plagiaulacidal is one of
the most peculiar among those of the Mammalia, whetherwe
considerits structureor its relations to geological time. Com-
mencing in the Jurassicperiod, it persisted throughthe Creta-
ceous to the Eocene. It then disappeared fromview to remind
us once more of its existenceby its probable descendant,the ex-
traordinarypouched lion of the Pliocene period of Australia,
Thy/acoleocatnifexOwen (Fig. 9). The familyexhibitsthe usual
successional relation of its component genera. In this respect
it repeatswhat I have already pointed out as a law of succession
in placental Mamnialia,2a reductionin the number of premolar
teeth. The followingtable exhibitsthese relations:
I. Tubercles of superiormolarcrescentic.
Fourthpreniolarserrate,not ridged............................... ,eniscoessus.
II. Tuberclesof molarssubconic.
a. Four'compressedpremolarsbelow.
Premolarsserrate,notridged . . .lenacodon.
Premolarsridgedand serrate....................... Plioprion.5
aa. Three compressedpremolars.
Premolarsridged................................................ fPlagiaulax.
aaa. Two premolars.
Fourthpremolarridged ... Pilodus,
aaaa. One premolar.4
Fourthpremolarridged . .. ............... IVeoplagiaulax.
Fourthpremolarsmooth . .L iolonus.5

Of these seven genera but nine species are thus farknown.

Ptilodus and Ctenacodon have two species each; and each of the
othersbut one. Ctenacodon,Plagiaulax and Plioprionare Juras-
sic; Meniscoessus is Cretaceous,and the remainingthreegenera
are Eocene Tertiary. The American genera are Ctenacodon
Marsh (Fig. 7 g), Meniscoessus Cope, and Ptilodus Cope. The
firstnamed is the most generalizedof the family. The Menisco-
essusconquislus Cope, has the distinctionof being the onlyknown
mammalof the Cretaceous period.
1 This familyis the equivalentof Marsh's " order" Allotherias
2 BulletinU. S. Geol. SurveyTerrs.,VI, p. i68.
S Gen. nov.,typePlagiaulax minorFalconer.
4 Numberunknownin Liotomus.
5 Gen. nov., type zVeoplagiaulax mnarshii
Lemoine.
VOL. XVIII.-NO. YVI. 44
692 Marsupialia.
TheTertiary [July,
The genusfirstdiscoveredwas the .Plagiaulaxof Owen, of
whichthetypicalspecieswas foundin the Purbeckbed of the
Isle of Wight,England. It was made the subjectof a memoir
by Falconer. Ctenacodonwas nextdiscoveredin theJurassicof
Wyomingby Marshin i879. In i 88o Dr. Lemoine discovered
the first
Tertiaryrepresentative and in i882 named
of thefamily,
it Neop/agiaulax eocanus(Fig. 6). In November, i88i, I de-
scribedthefirstAmericanTertiary whichwas discoveredin
form,
the Puerco beds of New Mexico,underthe name of Ptilodus

FIG. 6.-Neofpagiaulax eocanus Lemoine, from the Cernaysianbeds of Reims,

France; mandibularramiand teethof threeindividuals which are representedby
the lettersa, b and c. Figs. a, b, and c, much enlarged. Figs. a-, 6t and c', natural
size. Fig. 6'", fromabove. From Lemoine,Bulletinde Soc. Geol. de France, 1882,
p. 249.

medicevus(Fig. 8). Its presencein thatformation, with

together
variousotherassociatedtypes,proved the nearhomotaxy the
of
beds exploredby Dr. Lemoine near Reims with thoseof New
Mexico.'
Up to thistimethe great Cretaceousperiod had remained a
blankin the historyof Mammalianlife. Europeanpaleontolo-
gistshad examinedthefresh-waterbeds of this periodformam-
malianremains withoutsuccess.Amongthem,thegifted Kowalev-
I See the NATURALIST, i883, p. 870.
1884.] The TertaryMarsupialia. 693

sky,too soon lostto science,spentmuchtimein thesouthof France

among the formations which most nearlyrepresentthe American
Laramie formation,but found no Mammalia. It remained for
Mr. Wortman to crown a
series of successful expedi- a
tions by the discoveryof the a
Meniscoissusconquistusin the
Laramie formationof. Da-
kota, its loose teeth being .
found mixed with the teeth in
of dinosaurs and scales of c b
gar-fishes.The charactersof
the molar teeth are highly
appropriateto the geological
age of the genus, the supe- 4
rior molar resemblinhg bot
that of the JurassicStereog-
nathusand the Eocene Poly- FIG. 7.-Fig. a, Ctenacodon serratusMarsh,
I nat. size, fromMarsh. Fig. b, molar of
mastodon. Stereognathus 1, fromOwen. Fig c,
oofiticus,
I have shown that the cut- right
fourth uppermolarof Tritylodon long,
vus,fromOwen, .. Figs. d-f, Meniscoissus
ting tooth of the lower jaw conquistusCope, I nat. size. Fig. d, supe-
rior molar; Figs. e-f, superior fourthpre-
in the genera with but one molar1; g, humeral condyles of a smaller
such tooth, as Ptilodus and species withjaw, foundwiththe MAniscogssus.
Thylacoleo, is the fourthprenmolar;2while the similartooth in
the existingkangaroo-rats,withwhich it has been compared,is
a third premolar. In the living genera the fourthpremolar
resemblesa true molar. It is necessaryto rememberthis factin
the attemptto ascertain the phylogenyof the Multituberculata.
This is not an entirelyeasy task, owing to the questions which
arise as to the origin of the cutting premolars themselves. In
general it is true of Mammalia that simple premolarsprecede the
complex in time; butan exceptionto thisrule is to be seen in the
tritubercular superiorsectorialtooth of some Creodontaand Car-
nivora. Whether the premolarsof this familyare primitiveor
derivativeis not as yet known. If they be primitivethey may
be direct modificationsof the serrate teeth of the herbivorous
Dinosauria or Theromorpha. The complex characterof the pre-
molars in the older Tritylodon suggests the possibilityof the
otheralternative. The general historyof the Plagiaulacidaecon-
1 This toothmay possiblybelong to a Saurian.
2NATURALIST, I882, P. 521.
694 The TertiaryMarsupialia. [July,

firmsthe theoryof derivationfrom complex premolars,and we

observe in the laterform,Thylacoleo, a simplification of the true
molarsalso. The molariform fourth premolars in the existing
Macropodidaeconfirmthis view. In
orderto connectthese latterwiththe
ancestral formof the Plagiaulacidoe,
on a former occasion, I posited a
theoreticalform which should com-
_ bine threeof the cutting premolars
_ h of the extinctfamilywiththe molar-
\ iformfourthpremolar of the Macro-
podidae. This I named Tritomodon.'
The discoveryofTritylodonhas added
+ A_ confirmationto thisview,at leastas re-
gardsthepriorexistenceofthemolari-
FiG. 8. .-J'tilodus mediavus formpremolars.The Polymastodon-
Cope left mandibularramus,
nat. size; fromthe Upper Puerco tidaywere probablyderived fromthe
beds of New Mexico. Fig. at Tritylodontidaby the usual process
externalside; b, internaldo.; c, b
view, I nat.size. Orig-
superior of numberof teethand
of reduction
specialization of those that remain.
inal, from Report U. S. Geol.
SurveyTerrs.,III, F. V. Hayden T
incharge. Thylacoleo' mustbe regardedas the
type of a familydistinctfromthe Plagiaulacida, since it has but
one true molar tooth in the upperjaw. The compositionof that
tooth is unknown,so that it is not certain whetherthe family
Thylacoleontidaemust be placed in the Multituberculataor Sar-
cophaga. That it is a direct descendantof the Plagiaulacidx I
thinkthereis no doubt. The following phylogeneticscheme is
similar to one I published in the NATURALIST, i882, P. 521, with
some addition, and the removal of Polymastodon (Catopsalis)
fromthe Plagiaulacide:

1 NATURALIST, I 882, P. 5 2 1.
2Owen, Quar. Journ.Geol. Society,London, i883.
I1884r.] The TerliaryMarsutpialii. 695

Tritylodon

"Tritomodon"

Ctenacodon

Meniscoessus Plioprion Phascolomys

Plagiaulax

Polymastodon Ptilodus Hypsiprymnus

Neoplagiaulax Liotomus Macropus

Thylacoleo

It appears fromthe precedingconsiderationsthatthe dentition

of the implacental Mammalia has had a historyindependentof
that of the placental series so far as regards the herbivorous
typesat least. I have shown thatthe primitivetypesof the pla-
cental series were tritubercular, and then quadritubercular, and
thencrested. In the herbivorousmarsupials,on the contrary,we
commencewith multitubercular forms,and it is yet an open ques-
tion whetherthese have had a quadri- and tritubercularancestry
or not.
The Plagiaulacida of the Jurassic period are of very small
size, none of them exceeding in dimensions the house mouse.
The same is true of the species of the Eocene period hitherto
foundin Europe. The Americanspecies are larger,the Plilodus
medicevusequaling probablythe Norway rat (Fig. 8), while the
P. trovessartianus is one-thirdsmaller. The Meniscoessusconquis-
tus is still larger,equaling about the Polymastodon folia/us(Fig.
7). The arrangementof the crestsof the fourthpremolarin the
species of Plagiaulacidx differsas follows: In Plagiaulax, Plio-
prionand Neoplagiaulax this tooth is grooved. In Ptilodus the
grooves have become so wide (Fig. 8) that the wide intervening
ribs have become narrowkeels. In Meniscoessus there are no
keels, but the marginof the crownis serrate(Fig. 7 a c).
696 The TertiaryMarsupialia. July,

The Plilodus mediavus furtherdiffersfromthe Neoplagzaulax

eocwnusin the more rodent-likecharacterof its incisorteeth. In
the latterspecies these teeth resemble more those of the kanga-
roos in theiranteriordirection. The diastema is longer in Ptil-
odus, thus increasing the rodent resemblance. The fourthpre-
molar is stronglyserratein the Neoplagiaulax, resemblingin this
also the Mesozoic types.
The discussion between ProfessorOwen on the one side, and

FIG. 9.-Thylacoleo carnifexOwen; skull frombelow, one-halfnat. size; from

thePliocene beds of Australia. From restorationby ProfessorOwen in Geological
Magazine, 1883,p. 289.
Messrs. Falconer,Ktefftand Flower on the other,as to the nature
of the food of Thylacoleo, is known to paleontologists. From
the formof the teethalone, ProfessorOwen inferredthe carniv-
orous natureof the food of this genus, while his opponents in-
ferreda herbivorousdiet fromthe resemblance betweenthe den-
titionand that of the herbivorousHypsiprymnus. I have pointed
out (1.c.) that the comparisonof Thylacoleo with Hypsiprimnus
1884.] Vestigesof Glacial Man in Minnesota. 697

is weakened by two considerations: First, the cutting teeth in

the two genera are not homologous; second, the grindingseries
of molars,completein Hypsiprymnus,is almost wantingin Thy-
lacoleo. It evidentlydoes not followthatbecause Hypsiprymnus
is herbivorousThylacoleo is so also. ProfessorFlower refersto
the reductionof the molars in Thylacoleo as slightly compli-
catingthe problem,and concludes that the food of that animal
may have been fruitor juicy roots,or even meat. It is difficult
to imaginewhat kind of vegetable food could have been appro-
priatedby such a dentitionas that of Ptilodus and Thylacoleo.
The sharp,thin,serrateor smooth edges are adapted for making
cuts and dividingfood into pieces. That these pieces were swal-
lowed whole is indicatedby the small size and weak structureof
the molar teeth,which are not adapted for crushingor grinding
anythingbut verysmall and softbodies, It is not necessaryto
suppose that the dentitionwas used on the same kind of food in
the large and the small species. In P/ilodus medkevusthe diet
may have consistedof small eggs which were picked up by the
incisorsand cut by the fourthpretnolars. In Thylacoleocarnifex
it might have been larger eggs, as those of the crocodiles, or
even the weaker livinganimals. The objectionto the supposition
thatthe food consistedof vegetables,is foundin the necessityof
swallowing the pieces without mastication. In case it should
have been of a vegetable character the peculiar premolar teeth
would cut offpieces of fruitsand other soft partsas suggested
by Professor Flower, but that these genera could have been
herbivorousin the mannerof the existing kangaroos,with their
fullseries of molars in both jaws, is clearlyan inadmissiblesup-
position.

VESTIGES OF GLACIAL MAN IN MINNESOTA.

BY MISS FRANC E. BABBITT.

(Continuedfrompage 605, YJune

numberr)
THE notch quartzes hithertoexamined have been differentiated
by marked peculiaritiesof distribution,
worth while to enu-
meratein this place.
First: as therewere originallyno quartzesaffordedby the soil
above the stratum,so therewere none yielded by that lyingim-
mediatelybelow; although it would naturallybe expected that