sustain whether of Darwin, or of Archbishop Ussher — he seeks only to know the truth, whatever may be the consequences. Perhaps the points mentioned in this paper further along have not had sufficient attention hitherto. "Come, let us reason together."

The first extract above sets the close of the "Ice Age" entirely too far back. One of the objects of this paper is to make good this assertion. From the facts set forth below, it is reasonable to conclude that even on Croll's theory alone the close was not over 40,000 years ago, and possibly not over 35,000. If the views of Professor LeConte, given in his paper of January, 1891, have the weight and influence which their importance demands, it seems to me that there need no longer be any contest between gacialists who reject the astronomical theory, by reason of the remoteness of the time, and those who refer the ice age to terrestrial causes alone. Professor LeConte's theory is so clearly and tersely set forth that it is best to quote it entire, as given by Professor Wright on pp. 618-9, including the figure used in illustration:

3. That the removal of the ice-load by melting was the cause of the re-elevation to the present condition.

"4. That all these effects lagged far behind their causes. "This lagging of effects behind their causes is seen in all cases where effects are cumulative. For example, the sun's heating power is greatest at midday, but the temperature of the earth and air is greatest two or three hours later; the summer solstice is in June, but the hottest month is July; and in some cases the lagging is much greater. The cause of sea-breezes,―i.e., the heating power of the sun, — culminates at midday, but the effects in producing air-currents culminate late in the afternoon and continue far into the night, long after the reverse cause, i.e., the more rapid cooling of the land, has commenced.

"Now, in the case under consideration, it is probable that the lagging would be enormous in consequence of the reluctant yielding of the crust and the capacity of ice to produce the conditions of its own accumulation. Although the elevation produced the cold, and therefore the ice accumulation, yet the latter culminated long after the former had ceased, and even after a contrary movement had commenced."

So far LeConte.

The close of the glacial epoch as above given -70,000 years is wholly arbitrary, and is evidently very far from the truth, as will be shown. At that time the eccentricity of the earth's orbit was nearly twice as great as it is now, and the consequent excess of the sun's time on one side of the equator above that on the other side (depending on the longitude of the perihelion) was

about fourteen days. It had decreased from thirty-five, when the difference was greatest. But this difference of fourteen days would work in the direction of great difference of climate between the hemispheres, and would so continue to work as long as there was any difference at all. And not only so, but the effect would continue and accumulate according to the universal law of nature in the cases above cited, long after the smallest eccentricity had been reached. And that smallest eccentricity occurred less than 45,000 years ago, whether the computation be made by the formula of LeVerrier or by that of Stockwell.

The last sentence of the extract from LeConte is significant: "Yet the latter culminated long after the former had ceased and even after a contrary movement had commenced." In this latitude the usual temperature for the first week in December is not very different from that of the first week in March. Yet the sun in the first case is more than twenty-two degrees south of the equator, and at the latter date is scarcely five degrees. In like manner, and in accordance with the law above named, suppose the intense cold resulting from the wide glaciation of the northern parts of this continent, to have continued long after the eccentricity had reached its minimum, then it is not only possible, but even probable, that the close of the ice age was not more than 35,000 years ago, even if 30,000 would not be a more accurate designation. In which event, the objection to the astronomical theory, by reason of the long time elapsed since the days of high eccentricity, is wholly removed. And Professor Wright himself, although long favoring the short period of 10,000 years, has lately seen cause to doubt whether this is not too small. In a letter to the New York Nation, under date of Sept. 15, 1892, in view of his recent investigation of the old northern outlet of the great lakes into the Ottawa River, he says the facts " will... considerably lengthen our computation."

This throwing back of the close of the ice age by the glacialist, and the preceding shortening of the period in accordance with a universal law of nature, may both serve to strengthen the hypothesis of LeConte, and commend it to all interested in these questions, as the explanation which best accounts for the admitted facts.

CURRENT NOTES ON CHEMISTRY.-I. [Edited by Charles Platt, Ph.D., F.C.S.]

Properties of Diamonds.

THE experiments of M. Moissan in the production of artificial diamonds and other precious stones, his remarkable results in the reduction of the most refractory oxides and his whole line of work at high temperatures, are well known to readers of the scientific magazines. Some of the more recent investigations have been of the properties of the diamond when heated in oxygen, hydrogen, chlorine, etc. When the temperature is raised slowly the combustion is correspondingly slow and without production of light, being recognized solely by the action of the gas evolved on baryta solution. At 40°-50° above the point at which this slow combustion takes place the combustion becomes more rapid, producing a visible flame. Yellowish-brown carbonado burned with a flame at 690°; black carbonado with a flame at 710°-720°; transparent Brazilian diamond without a flame at 710°-720°; transparent crystallized Brazilian diamond without a flame at 760°-770°; cut diamond from the Cape without a flame at 780°-790°; Brazilian bort and Cape bort without a flame at 790°, and with a flame at 840°; very hard bort without a flame at 800°, and with a flame at 875o. As a rule, the harder the diamond the higher its point of ignition.

When heated to 1200° in hydrogen the Cape diamond loses nothing in weight, but becomes lighter in color and less transparent; dry chlorine and dry hydrogen fluoride have no action at 1100-1200°. Sulphur attacks diamonds at 1000°, but with carbonado carbon bisulphide is readily produced at 900". Sodium vapor has no action at 600°. Iron at its melting point attacks the diamond with the production of graphite on cooling; melted platinum also combines readily. Fused potassium hydrogen sulphate and alkali sulphates, potassium chlorate and nitrate are all without action on the diamond, but, according to Damour, attack

carbonado. The diamond is rapidly dissolved when heated to a high temperature with carbonates of the alkalies, carbonic acid being given off, but no hydrogen, and hence M. Moissan concludes that diamonds contain no hydrogen or hydrocarbons.

Treated with hydrofluoric acid, and then with aqua regia and finally washed, dried and burnt in oxygen, the diamonds yielded an ash consisting in all cases but one chiefly of ferric oxide. Cape bort contained also silica, calcium and magnesium, and Brazilian carbonado, silica and calcium, with a trace of magnesium. One specimen of green transparent bort from Brazil left a minute quantity of ash, which contained silica, but no iron.

Preparation of Pure Alumina.

The preparation of pure alumina from bauxite, which is always accompanied by more or less silica and oxide of iron, has commonly been carried out as follows: Taking advantage of that property of alumina, which enables it to act as either base or acid, according to its environment, the bauxite is fused with sodium carbonate, the resulting products being sodium aluminate

and sodium silicate. The mass is then extracted with water and the sodium aluminate passed into solution. The silicate of soda, owing to a deficiency of base, is but little acted upon by the water and with the ferric oxide is left in the residue. From the solution of the sodium salt the alumina is precipitated by passing carbonic acid gas, carbonate of soda being regenerated at the same time.

This process has lately been improved by first precipitating a portion of the solution of aluminate by the gas in the cold, producing a small quantity of crystallized alumina hydrate of the same composition as Gibbsite, Al,O, 3 H,O. This, then, is added to the main bulk of solution, and a complete separation of the whole is secured, the soda being left in the caustic state. The reaction which takes place has been investigated by M. A. Ditte, and is explained as follows: A solution of sodium aluminate may be regarded as amorphous hydrated alumina dissolved in caustic soda. The form in which a body crystallizes from a solution is largely determined by the character of the crystal introduced to start crystallization. Hence in the process described above the tendency of the whole is to crystallize in the form of the several crystals first introduced, and as the crystalline form of alumina is less soluble than the amorphous in alkaline solutions, there is a gradual complete precipitation. Stirring facilitates the separation of the crystals by bringing those already formed into contact with fresh portions of solution. There is finally left only that proportion of alumina corresponding to the solubility of Gibbsite in caustic soda under the conditions existing.

Silk from Wood.

At the Paris Exposition in 1889 M. de Chardonnet gave to the world his process for the manufacture of silk from wood and received the highest honors from the jury of award. Since that time the process has been further developed and has presumably attained a practical success; silk is being manufactured at Besançon from wood pulp such as is used in the fabrication of certin kinds of paper According to F. B. Loomis, U. S. Consul at St. Etienne, the following process is used: The pulp is carefully dried in an oven and plunged into a mixture of sulphuric and nitric acids, then washed in several water-baths and dried by alcohol. The product thus prepared is dissolved in ether and alcohol with the production of collodion similar to that used in photography. This collodion, which is sticky and viscous, is enclosed in a solid receptacle furnished with a filter in the lower end. An air-pump supplies air to the receptacle, and by its pressure forces the collodion through the filter, removing all impurities. The collodion flows into a horizontal tube armed with three hundred cocks having glass spouts pierced by a small hole of the diameter of the thread of a cocoon as it is spun by the silkworm. The spinner opens the cock and the collodion issues in a thread of extreme delicacy. This thread, however, is not yet fit to be rolled upon spools on account of its viscosity and softness, being still collodion and not "silk." To obtain the necessary consistency the thread as it issues is passed through water, by which means the ether and alcohol are washed out and the collo lion solidified

and transformed into an elastic thread as brilliant and resisting as ordinary silk. The dangerous inflammability of this substance, as prepared above (two centimetres per second), has been reduced, according to the inventor, by passing the spun thread through a solution of ammonia, thus rendering it as slow of combustion as any other like dress material.

Up to January of this year none of the more important silk manufacturers of St. Etienne or Lyons had invested heavily in this enterprise, but all express confidence in the process and believe it is destined to figure largely in the commercial world.

New Method for Melting Points.

A. Potilitzin is the author of a new method for the determination of the melting points for substances melting below 450°, this being the highest temperature which a nitrogen-filled mercury thermometer can indicate. One end of a hard-glass tube, 5 mm. bore and 500-600 mm. in length, is drawn out to capillary fineness and the other is bent at right angles. The capillary is dipped into the molten substance, the melting point of which is to be determined, so that on cooling the tube is closed by a solid plug of the substance 3-4 mm. long. The other end is connected with a manometer by means of which a pressure exceeding that of the atmosphere is maintained within the tube. The tube, along with the principal thermometer and also one for stem correction, is inserted into a wide test-tube, which is then immersed in a bath of fusible metal. When the melting-point is reached the plug softens and is expelled by the internal pressure, so that the sudden equalizing of the pressure in the manometer indicates the moment when the substance melts, the thermometers being then read off. Potassium nitrate was found by this method to melt at 336.57° (mean of eight experiments); by immersion of the thermometer direct into a large mass of the salt the melting point was found to be 336°.

Pigments Used in Some India-rubber Toys. India-rubber has been generally, and correctly, accepted as a suitable material for children's toys; but investigation into the manufacture of the latter reveals the fact that many as placed upon the market contain harmful ingredients. A. Bulowsky has recently called attention to several dangerous ingredients as, for instance, in black dolls, which are often colored "in the mass with lead pigments. Red articles are also most usually colored in mass, the pigment being antimony sulphide, which, however, being unattacked by the saliva may be considered innocuous. Grey rubber goods generally contain zinc oxide, and hence particularly when, as is sure to be the case, the toy is brought to the child's mouth, an element of danger is introduced. Superficial coloring is frequently accomplished by means of poisonous pigments. These remarks are applied in particular to foreign manufactures, and though, doubtless, the same coloring matters are used in this country, I have yet to learn of a case of poisoning from coloring in mass. Superficial pigments, from their disposition to flake and from the greater quantity brought into contact with the mouth, are certainly to be avoided. It is difficult, moreover, to estimate the amount of damage done by these toys owing to the many petty ills and derangements of infancy, the poison received by the child very likely is insufficient to develop well-defined symptoms or to direct suspicion, but at the same time may be the cause of an indisposition which itself brings on crying, wakefulness, and general wear on the little body struggling for existence.

SCIENCE:

PUBLISHED BY N. D. C. HODGES, 874 BROADWAY, NEW YORK.

SUBSCRIPTIONS TO ANY PART OF THE WORLD, $3.50 A YEAR.

To any contributor, on request in advance, one hundred copies of the issue containing his article will be sent without charge. More copies will be supplied at about cost, also if ordered in advance. Reprints are not supplied, as for obvious reasons we desire to circulate as many copies of Science as possible. Authors are, however, at perfect liberty to have their articles reprinted elsewhere. For illustrations, drawings in black and white suitable for photoengraving should be supplied by the contributor. Rejected manuscripts will be returned to the authors only when the requisite amount of postage accompanies the manuscript. Whatever is intended for insertion must be authenticated by the name and address of the writer; not necessarily for publication, but as a guaranty of good faith. We do not hold ourselves responsible for any view or opinions expressed in the communications of our correspondents. Attention is called to the "Wants" column. It is invaluable to those who use it in soliciting information or seeking new positions. The name and address of applicants should be given in full, so that answers will go direct to them. The Exchange column is likewise open.

In an article which will shortly appear in the Medical News, entitled "Investigation of the Outbreak of Beriberi on Board the Bark Pax' from Ceylon with a Cargo of Graphite," I show that, from the deficient packing of 1,200 tons of graphite, the cargo was exposed to the moist air encountered on a tropical voyage, all but six of a crew of nineteen were stricken with beriberi.

The bark J. C. Warns," from Java and Macassar, with a cargo of green coffee, arrived in New York June 23, 1890. The captain and three men had died of beriberi. The coffee had been picked and shipped too green. Mr. Tobias, consignee of the cargo, showed me a sample of it; it was charred, carbonized, and almost destroyed. The coffee had fermented. The outbreak of beriberi on a ship from Java, where the coffee has been carbonized, is a regular occurrence. Java coffee owes its value in our market to its color; in order to obtain this color, the captains take their cargoes quite green, which favors a slight fermentation during the trip. Sometimes they go too far; the coffee is too green, and the fermentation too violent; in such cases there is always carbonization; the grains stick together in great masses, and abundant fumes (carbonic gases) fill the ship. The iron ship Glenmorag, Captain Currie, 133 days from Colombo, Ceylon, with 1,100 tons of graphite on board, 800 tons of cocoa-nut oil, etc., arrived July 17 in New York. This ship, loaded in Colombo alongside the "Pax" (mentioned above), travelled the same course, at an interval of two weeks. She lies now at the Atlantic dock, in Brooklyn, again alongside the "Pax." She had no beriberi outbreak. From her first mate I have the following information:

Crew, 28 men; captain's wife and two children on board; or all, 31.

She is a Scotch Glasgow boat, and the crew is English and Scotch. Before taking this cargo, this ship had carried from Barry Dock, near Cardiff, a cargo of coal to Buenos Ayres, South America, and taken a ballast of sand to Colombo, Ceylon. Before these trips she had been in the wheat trade from Tacoma, Washington, to Havre, France. She is remarkably dry, and the cleanest ship one would wish to see. I went down her hold and examined every part of it; there is not a smut nor a stain anywhere about it. The iron part is especially clean: no trace of incrustation of carburetted iron, which might have indicated the action of hot moist

air on the carbon.

None of the barrels containing the graphite

was broken. The packing was exceedingly good, the dunnage consisted of sticks and cocoa-nut bulls, so that it was impossible for the barrels to roll and break, and thus expose their contents to the action of the air.

The diet bill was about the same, or even poorer, than that of the "Pax." Nine casks of salted beef and seven barrels of pork were consumed during the trip. Fresh beef (tinned) three times a week, one-half pound to a man, and a half pound of salt meat on the same days; other days a full pound of salt meat a day. One-half pound of rice for each man on Saturdays; no vegetables except onions with the soup three, times a week. The ship being Scotch, oatmeal made part of its fare for two and a half months after starting, when it ran out. No sickness whatever during the voyage. One death by accident. The captain attributes the good condition of his cargo and his crew to the change of winds and cooler weather which he enjoyed from the Cape of Good Hope to the North Atlantic. His log is indeed very different from that of the "Pax."

In Science., vol. xxii., No. 545, p. 16, Venable states that "the metallic carbides are usually formed by the action of intense heat upon the metal in the presence of carbon. The form of this carbon is capable of being greatly varied. Graphite, amorphous carbon, and many hydrocarbons, may be used. The heat of the ordinary furnace is sufficient to form the carbides of the metals already mentioned, zinc, copper, and notably iron. All of these carbides, under certain conditions, give off their carbon in the form of hydrocarbons. The same smell can be detected in all during their decomposition. In some cases, as iron and zinc, the decomposition is caused by the action of an acid. The carbides of the earths (of which graphite is one, in conjunction with iron) decompose in moist air, and more rapidly in water.”

I may point again here to those broken barrels of the "Pax," which exposed the carbon to the influence of tropical air.

I have examined microscopically the blood of four of the sufferers of the "Pax," and obtained the following results: Captain Geeseicke, sick since May 16 with beriberi; 800 diameters, inch oil-immersion objective; red discs, irregular in outline, congregated in masses, with ragged edges, not inclined to form rouleaux; quite plastic; colored streaks or rays of pink and red, showing the presence of biliary matter, biliverdin crystals; black spores, not free but entangled in the hummocks of corpuscles. It may be noted that the oedema of this patient's legs only left him two days before this examination.

18

Henry Oelrichs (second mate), German, 27 years old. Has been fourteen days sick with beriberi. Examination of the blood: 500 diameters, inch objective; red corpuscles, very plastic, aggregated in hummocks. Many black spores are seen floating about, free in motion. Fibrine in excess, light in texture, and lumpy. Blood very thick, syrupy, and plastic. No motion, showing want of circulation. Excess in the coloring matter. This same case examined: 900 diameters, inch objective immersion lens, shows excess of fibrine in ropes, biliary matter in great excess; no crystalline formations; blood quickly oxidizes and forms a solid mass. The black spores above mentioned are quickly held by the fibrine; the red discs are distended, bladder-shaped, and have very ragged edges. The meniscus-shape is lacking, there being great irregularity in outline and color, some are even squareshaped. Some discs have an excess of color, some are very pale. On the edges of the corpuscular mass the color quickly disappears, in consequence of rapid coagulation.

Isaac Hegglund, a Swede, 27 years old, has had beriberi since crossing the equator, six weeks ago. Legs are now very thin, but still some soreness remains; knee reflexes still lost. First sound of heart prolonged. Microscopical examination of the blood, 900 diameters, inch objective, shows rouleaux well formed, no spores, no filaments, slight feverish condition shown by spiculated outlines of some of the red corpuscles. Fibrine is assuming a normal form, showing meshes very regular; no distension of red corpuscles.

Emil Jensen, a German, 19 years old, sixteen days sick with beriberi. Black spores in active motion and very plentiful; freely scattered in the field of observation; circulation very torpid; fibrine

very irregular, light in texture; biliary matter freely scattered; blood discs distended and with ragged edges; red corpuscles congregated in masses; fibrine forming in heavy clots; blood rapidly coagulating; black spores are quickly fastening in the fibrine.

We have here, in the 14th day sick and the 16th day sick cases, black spores in active motion and biliary matter in both cases, and the corpuscles distended bladder-shaped, in ragged-edged condition; the fibrine quickly clotting. And in the captain's case, which was the worst of all, we have still black spores, biliary matter, and ragged-edged corpuscles.

In the 6th week case, a much milder case, moreover, than any of the others, it is reasonable to assume that in some way the patient has quickly eliminated the poison. There is no biliary matter in his blood, no black spores, no abnormal fibrine, no distension of red corpuscles; the latter are perfectly formed in rouleaux.

Examinatian of urine of Henry Oelrichs (second mate, bark "Pax"), July 17th, 1893 (14th day of beriberi): —

Odor, light, aromatic, and feverish.

Color, light (yellow) amber.

Reaction, excessively acid.

Appearance, transparent.

Specific gravity, 1.032 +.

Weight of a fluid ounce, 470.27 grains.

6.605 grains. 1.404

Traces of sugar and carbonic acid gas are commonly observeď in the urine of beriberi patients.

Dr. Wallace Taylor, Osaka, Japan, sends me three interesting tables, which he made from examinations of 134 cases of beriberi. These examinations were made with Hayem's hæmatometer and Gower's hæmacytometer. The average corpuscular richness for the 134 cases is 94 per cent. This, he says, corresponds to the clinical experience in cases of beriberi. Most of the cases of beriberi seen by the general practitioners are well-fed, wellnourished, full-blooded appearing men. The ill-fed, poorly-nourished, weak constitution cases are the exception. During the past few years he has kept a record of the physical condition of the beriberi patients, and he gives this record, together with another record, of a beriberi hospital in Tokio:

cent.

In his table No 3 there is shown a general diminution of the hæmoglobine. The average hæmoglobine in 101 cases is 81 per In some of these cases the amount is very low, being below 65 per cent, and with but few exceptions the per cent of hæmoglobine is below the per cent of corpuscles, showing a deficiency of the individual corpuscles in hæmoglobine.

The appearance of biliary matters, which I have shown in my analyses of the four cases of the bark "Pax," would show by itself a deficiency of hæmoglobine.

In the Tribune Mèdicale, Sept. 10, 1891, Messrs. Bertin-Sans and Moitessier show that it is the presence of hydrogen and carbonic acid in oxicarbonized blood that prevents the total destruction of hæmoglobine.

By sweeping their solution of oxicarbonized blood and water, with a current of hydrogen and carbonic acid gas, and an addition of sulphide of ammonia, they obtained the spectrum of reduced hæmoglobine. They thus show that oxicarbonized hæmoglobine can be readily transformed into a mixture of methæmoglobine and oxide of carbon. It is therefore reasonable to suppose that in an outbreak of beriberi where we have the presence of oxides of carbon and a deficiency of hæmoglobine (observable in all cases of beriberi) the latter is the effect of the former.

In Japan, the universal burning of charcoal produces the oxides, which held down in the low places by the moist atmosphere of the beriberi season, there is produced on a large scale and continually during the moist season what happens on board of each of those ships which come to us from the East with carbonized cargoes and beriberi-sick crews.

THE STRUCTURE AND AFFINITY OF THE PUERCO UN

GULATES.

BY CHARLES EARLE, B. SC. (PRINCETON).

The discovery in 1880 by Baldwin of the presence of mammalian remains in the Puerco beds of New Mexico, was one of the most important in the history of vertebrate paleontology in this country. This rich mammalian fauna has been wholly described by that able investigator, Professor E. D. Cope, and to him we are indebted for having made known to the scientific world the interesting mammals which are imbedded in this formation. As I have lately been studying a collection in the American Museum of Natural History from the Puerco, I propose in this paper to sum up some of the results of my investigations as relating in particular to the primitive ungulates of this formation, and especially to attempt to place some of these forms in or near their proper phylogenetic positions in the system.

As a word of introduction I would remark that most of the remains from the Puerco are in a poor state of preservation, and this applies particularly to the skeleton. The teeth are often well preserved, so that in working out the affinities of these mammals we are generally dependant upon the character of the teeth to discover their relationship to other forms. A very striking peculiarity in the dentition of the Puerco mammals, as pointed out by Professor Cope, is the fact that their superior molars are generally of the tritubercular pattern, and these upper teeth are associated with inferior molars, which are tubercularsectorial, or a modification of the latter. In the tubercular-sectorial tooth the anterior portion is raised above the posterior or talon, and consists of three elevated cusps. By the modification of the latter pattern of molar, both the highly specialized sectorial teeth of the Carnivora and the quadritubercular teeth of the Ungulates have been derived.

In general we may say that, besides the characters of the teeth, the mammals of the Puerco epoch, in their skeletal structures, as far as known, are of a decidedly primitive type. The skull is short and heavy, with the orbit well forward over the teeth; the various processes of the skull for muscular attachment are prominent. Correlated with their low structure in general was the exceedingly small brain, as illustrated by the genus Periptychus. As the structure of the skeleton in the latter genus is the best known, I will enumerate some of its characters. The feet of Periptychus were plantigrade. The hind foot had five toes, the external one being not much shortened. The structure of the astragalus is well known in Periptychus and important, as teaching us one of the characters of the structure of the primitive foot in general. This bone is short and strongly depressed; the neck of the same is short and heavy. In all modern mammals which are digitigrade the trochlear surface of the astragalus, articulating with the tibia, is deeply grooved, whereas in Periptychus this surface is plane and flat. Another very important and primitive character of the astragalus in Periptychus is that it is perforated by a well-marked foramen. I am not aware that this perforation of the astragalus occurs in any recent Ungulate. The astragular foramen is of constant occurrence in Puerco mammals and also is present in some of their descendants in the Wasatch (Coryphodon).

In one respect the foot of Periptychus is more advanced than that of the genus Phenacodus, which is from a later formation (Wasatch); I refer to the articulation of the cuboid bone with the astragalus, but in general the foot structure of Phenacodus is far advanced in its evolution over that of Periptychus. Phenacodus was a digitigrade mammal, with the outer toes much shorter than the median. The long bones of the skeleton in Periptychus are short and heavy; this applies particularly to the humerus, which is an exceedingly heavy bone; its distal extremity is perforated by an entepicondylar foramen, another primitive character of this genus. The character of the humeral condyles in Periptychus is peculiar and different from all modern Ungulates. In the latter group the trochlear surface of the humerus is interrupted by a strong ridge separating the external from the internal articular surface. Now in Periptychus, as well as in Phenacodus, there is no such interruption of the condylar surface of the humerus, and it has the same character as in the modern Carnivora, thus showing how these two widely separated orders at the present time approach each other in their osteological characters in the Eocene. The ulna and fibula are large in Periptychus and more nearly approach the size of the bones of the preaxial side of the limbs than in modern forms.

Now the question arises, what great groups of mammals of later epochs than the Puerco are represented in this formation. I think that we may safely say that there were only a few main stems of Puerco mammals which persisted until later periods, and I shall endeavor to show that these stem forms were the direct ancestors of later types. As in so many cases, in seeking to determine phylogenetic relationships, we must turn to the investigations of Professor Cope to decide this question in part at least. He has described mammals from the Puerco which he considers to be Ungulates in their affinity, others to be related to the Carnivora, and still other types which resemble the Lemu

roidea in the structure of their teeth.

As I am only dealing with the Ungulates in this paper I shall speak of certain genera which Professor Cope and other paleontologists have determined to be closely related to this group.

The group of primitive Ungulates which Professor Cope has designated the Condylarthra is not a very homogeneous one, it appears to me, and perhaps with Schlosser we had better speak of a condylarthrous stage, through which all Ungulates are supposed to have passed rather than to attempt to confine these early forms all in the suborder Condylarthra. At least as shown by Professor Osborn, the characters laid down by Professor Cope as limiting the Condylarthra, would not include some of the forms (Periptychus) which Professor Cope has embraced in this suborder.

When we attempt to separate the Ungulates from the Unguiculates of the Puerco we are met with the obstacle that in most cases the distal phalanges of the feet have not been preserved. Accordingly we are dependant upon the character of the dentition to diagnose and separate these two groups. However, so low down geologically speaking as the Puerco, the different groups of Ungulates are not supposed to be distinctly differentiated, and then again in most cases the structure of the skeleton, and especially of the carpus and tarsus of these forms, is totally unknown. I believe, however, that the stems leading to the main types of the Ungulates which we meet with in the Wasatch, are fairly well separated in the Puerco, and more so than has been generally accepted. For example, when we consider another group other than the Ungulata, the Creodonta, we find a number of well-marked families of this order in the Puerco, which are distinct and lead up in some cases to types of later epochs. The Creodonta, with low-crowned, purely bunodont teeth, such as are included in the Triisodontidæ, the more specialized and trenchant dentition of the Provivirrida (Deltatherium), and again the low-crowned and nearly quadritubercular lower molars of the Arctocyonidæ (Claenodon, Scott). The lastnamed genus is very likely the ancestor of the Wasatch (Anacodon).

Turning again to the Ungulates, what are the types of this order which we can distinguish in the Puerco? To attempt to decide this question we must rely on the characters of the teeth in nearly all cases. To ascend to the mammals of the Wasatch period for a moment we observe in that formation the Perissodactyles are distinct from the Artiodactyles. The former group has superior molars with six cusps, which may be either distinct or fused; the lower molars are quadritubercular. In the Artiodactyles of the Wasatch the superior molars are of the tritubercular pattern and the lower teeth are sexitubercular, or more nearly of the primitive tubercular-sectorial type already mentioned. Again, the premolars of the Perissodactyles are more complex than these of the Artiodactyles. Returning to the Puerco we find the same state of things well foreshadowed, although these two stems may have not passed the condylarthrous stage. In the genus Euprotogonia (= Protogonia), we have the supposed condylarthrous representative of the Perissodactyles, and in the genus Protogonodon of the Puerco I believe we are dealing with an ancestral Artiodactyle. I am aware of the fact that the skeletons of these two genera are totally unknown, so until they are discovered we will be unable to say whether these two forms were true Condylarthra or if they had assumed more of the characters which are typical of the two great divisions of the Diplarthra. I think that from a study of the teeth of the above genera that the two lines of the Diplarthra were fairly well separated even in the Puerco.

The upper true molars of Euprotogonia in the typical form, E. puercensis, consist of six tubercles. The superior premolars are simpler than in Phenacodus. A character of the upper molars of Euprotogonia, and separating it well from Phenacodus, is the absence of the parastyle and mesostyle. When we study the structure of the lower teeth in Euprotogonia, we are surprised to find them so highly developed for a Puerco form. The last lower premolar is nearly as complex as it is in the Wasatch Phenacodus, and in the typical species the crescents of the inferior true molars are as plainly marked as in the last-named genus. In