parieto quadrate arches; that all arches are absent from the Amphisbænian and Typhlophthalm suborders, and the zygomatic arch is incomplete in the Varanidæ. So this character has no ordinal signifi

cance.

Fourth.-Professor Owen opposes my statement that "there is no quadratojugal arch" by the observation that "in no reptile does the jugal or malar bone join the quadrate or tympanic bone". Professor Owen has here again fallen into error, since, in Hatteria (Sphenodon) and the order Rhynchocephalia, the malar does, according to Günther, articulate directly with the quadrate.* I cannot now refer to Professor Owen's early description of the same genus to see whether he has himself not pointed out this structure before Dr. Günther. Professor Owen knows also that the malar is connected with the quadrate in the Crocodilia by the mediation of a short quadrato-jugal bone, which fact is not directly contradicted in the sentence above quoted from his article. My object in citing this character was to show the distinction between the Pythonomorpha and the orders named.

Fifth, as to the form of the quadrate bone.-Like myself, Professor Owen finds it to differ from the corresponding elements in other orders. I have, however, not cited it in evidence of Ophidian affinity, although there is no propriety in Professor Owen's remark (p. 693) that "the tympanic (quadrate) bone alone suffices to refute the Ophidian hypothesis of the Mosasauroids". Comparing it with the quadrate of specialized snakes, he naturally finds differences; but he will find near resemblances if he will examine the same element in the Tortricine and the other low or generalized snakes which Müller combined under the name of Microstomata. Besides, great variations in the proportions of this and of various other elements are not inconsistent with coördinal affinity.

Sixth, as to the distinctness or coalescence of the nasal bones with surrounding parts.-Although this point is of no importance to the main question, I here observe that most of my specimens differ from the one figured and described by Professor Owen (fig. 14). He states that in the Mosasaurus missuriensis and Liodon anceps, the nasal bones are distinct; in various species of Clidastes and Platecarpus, they are coössified with other elements.

Seventh, as to the bony palate.-The partially free and dentigerous pterygoid bone is Ophidian as well as Lacertilian, but is not identical with the structure in the snakes, as I have pointed out. The supposed contact of these bones on the median line noted in Mosasaurus missuriensis is probably due to distortion, as it does not exist in most of the Pythonomorphous skulls which I have seen.

Eighth, as to the mandibular hinge.-I have not cited this in evidence. of any special affinity, for Pythonomorpha might be without it, and not lose their ordinal place. But there is a much greater resemblance be* On the Anatomy of Hatteria, in Trans. Royal Society, 1867, pl. i.

+ See my fig. of cranium of Cylindrophis, Proc. Am. Ass. Adv. Sci. xix. p. 217.

tween this part of the structure of these animals and some of the Erycid and Pythonid serpents than Professor Owen admits in his paper. Characters of vertebræ furnish the points of the essay from the ninth to the thirteenth. As I only cite a single vertebral character in my list of those of the order, most of Professor Owen's arguments on this head are irrelevant to my conclusion. I will, however, briefly review them. But firstly as to the one to which I have attached weight,-the absence of a sacrum. Professor Owen is unfortunate in his reasoning against the use of this feature as an ordinal definition. He says:-"The absence of a sacrum does not affect the mammalian grade of the Sirenia or Cetacea, so neither does it the lacertian nature of the Mosasaurians"! Here is committed the extraordinary oversight of comparing the rank of orders in a class with the rank of the subdivisions of an order among themselves. Professor Owen should have concluded the sentence with 66 so neither does it the reptilian nature of the Mosasaurians", in which case he would have been correct. The cases of the mammalian orders and that of the Pythonomorpha as orders of classes are indeed parallel. The absence of a sacrum is an important definition of the orders in the one case as in the other.

Of other vertebral characters I only mention two. Professor Owen cites the numerous hypapophyses of certain snakes as evidence against Ophidian affinities of Mosasauroids, but, as usual, selects those which have the largest numbers for comparison instead of those where the number is reduced. In the majority of non-venomous and Colubroid serpents, the hypapophyses are confined to the anterior part of the column, leaving the other vertebræ either smooth or not protuberant beyond the horizontal inferior line; e. g., Xenodon, Heterodon. The only exception to this rule is seen in the fresh-water snakes (Homalopsida), where the hypapophyses are numerous. The character is not, however, ordinal in any case.

In discussing the other vertebral character, the structure of the atlas and axis, I am charged with the failure to recognize the homology of the odontoid process with the centrum of the atlas. There is no ground for this charge; and as Professor Owen finds no characters which distinguish these parts from the corresponding ones in Colubroid snakes, I leave it.

As the fourteenth point, the significance of the structure of the teeth may be considered. I have already adverted to the wide difference in the mode of support of the crowns by the jaws from that which is universal in the Lacertilia. Professor Owen repeats a former dictum, that this kind of attachment "is a feature of resemblance to the lacertians called acrodont". Now even the term "resemblance" can hardly be admitted; and as to homology between the two kinds of dental attachment, there is none. Says Professor Gervais, in the Zoologie et Paléontologie Françaises, tome i., page 262, in describing some teeth which he refers to Liodon, in a note,-"C'est à tort que l'on décrit les dents des Mosasaures

comme réellement acrodont à la manière de celles de beaucoup de Sauriens actuels." Professor Owen goes on to say,-"The enamel develops a pair of opposite low ridges which are minutely crenate; the crenation becomes abraded at the apical part of used teeth, but is demonstrated in unworn and unextricated crowns. Many saurians, both Crocodilian and Lacertiau, show the creno-bicarinate character, but no Ophidian does." If the characters here mentioned were as universally present in the types to which Professor Owen refers as he seems to suppose them to be, they would have less significance than he attaches to them; but the variety presented by all the orders of reptiles is such as to render the above remarks quite irrelevant. Moreover, the statements are inaccurate. Teeth with two cutting edges are not uncommon in the Ophidia (e. g., genus Ophibolus, the posterior maxillaries), and are far from universal among Pythonomorpha. The teeth of Platecarpus are characterized by the absence of cutting edges, having a subcircular section.* In Clidastes, they are not crenate.

Fifteenth.-The presence of osseous dermal scuta is cited in evidence of the Lacertilian relationship of the order. Should such scuta have existed, it would not make the Mosasaurida Lacertilians, since they characterize other orders much more generally; but I am safe in saying that such structures had no existence in the known genera of Pythono. morpha. I have recently received large accessions of material belong. ing to these reptiles in admirable preservation, and have found no dermal bones. I have observed certain osseous segments arranged in lines, whose character I have not yet determined. Their form is rectangular, their tissue spongy, and their surfaces without sculpture.

Sixteenth. The presence of the columella is rightly regarded by Professor Owen as evidence of Lacertilian relationship. But this character is not a crucial test, since the lizards of the suborder Rhiptoglossa are without it, and the Rhynchocephalia and various turtles possess it.

IV. CONCLUSIONS.

I now recur to the propositions which I endeavored in the work already cited to demonstrate, and which have not been admitted by Professor Owen. They are expressed in the following language:-"That these repliles... constitute a distinct order of the Streptostylicate group; ... that they present more points of affinity to the Serpents than does any other order." My conclusions that they are not nearly related to the Varanidæ, and that the order is nearer to the Lacertilia than to any other, being sustained by Professor Owen, are not further considered.

As regards the claim of the Mosasauroids to position in an order distinct from Lacertilia, I do not enumerate a large number of subordinate characters, in which they differ from all known Lacertilia, because such * Report U. S. Geol. Sur. Terrs. ii. p. 141. + Report U. S. Geol. Surv. Terrs. ii. p. 126.

#

are not of ordinal value.* They might be wanting from Pythonomorpha and present in Lacertilia without violating their ordinal boundaries. I enumerate those which appear to be essential only. They are the fol lowing:

Subclass STREPTOSTYLICA.

Order PYTHONOMORPHA.

1. The parietal bones are decurved on the sides of the cranium, and are continuous with the alisphenoid and proötic elements.

2. The ophisthotic is largely developed, and extends upward and forward to the walls of the brain-case.

3. A distinct element connects the squamosal with the parietal bone above the opisthotic.

4. The teeth have no roots.

5. There is no sacrum.

6. There is no sternum.

7. The bones of the limbs possess no condylar articular surfaces. Of the preceding seven characters, the decurvature of the borders of the parietal bones at the margins, and their continuity with the margins of the proötic bone, is of importance as a character not found in the Lacertilia and universal among Ophidia. Even in Aniellidæ† and in the Amphisbania, the most snake-like of lizards, the lateral borders of the parietals are free, and are separated by a fissure from the greater portion of the proötic.‡

The opisthotic has a greater development than in lizards, where it does not reach the brain-case upward. In the serpents, its contact with the brain-case is well known. The existence of another element lying on the opisthotic, first pointed out by Marsh, is an important character. The anterior extremity of this bone enters into the side-wall of the cranium below the parietal, occupying much the position of the pterotic, and resembling, even more than the opisthotic, the suspensorium of the Ophidia. Should this be a true homology, the affinity to the Ophidia is strengthened; and should it prove to be a distinct element, not found in either Ophidia or Lacertilia, the claims of the new order to existence are maintained. In either case it is clear that the Ophidian suspensorium is not the squamosal bone.

The demonstration of my second assertion, i. e., that the Pythonomorphous order presents more points of affinity to the serpents than does any other order, may be seen in the above list of characters. Professor Owen doubtless believes with me that the Lacertilia are more nearly allied to the Ophidia than is any other order, so that I only need to show that the *I have enumerated eleven subordinate characters on pp. 125-126 of my report, Hayden's Series, vol. ii.

+ See Proc. Acad. Phila. 1864, p. 230, for the osteology of this family.

Compare Professor Owen's figures of crania of Liodon, fig. 15, with Monitor, fig. 7, and Python, fig. 13.

Pythonomorpha are nearer to the Ophidia than are the Lacertilia to establish the truth of my position. Five of the seven characters enumerated above are so clearly of this nature that my statement is abundantly justi fied. And it may be true without necessarily implying close affinity with 'the typical serpents. Of course, the points of approximation in Ophidia are to those which do not present the extreme of modification of the order, but to such more generalized forms as the Tortricidæ, Erycidæ, Scolecophi dia, etc., which are also nearest the lizards. Had Professor Owen desired a character in addition to the numerous ones which I have cited, in which they do not resemble the Ophidia, he might have added the absence of the trabecular grooves of the basi- and presphenoid, noticed by Huxley as distinguishing the serpents from the Lacertilia. But this interesting feature does not characterize the order Ophidia. The groove is reduced in Xenopeltis, and is wanting in the Typhlopida.

It only remains to show the inexact nature of the comparison which Professor Owen draws between the relations of the seals to other Carnivora, and those existing between the Pythonomorpha and Lacertilia. These relations he considers to be similar; that is, that as the seals are an aquatic form of Carnivora, so the Pythonomorpha are an aquatic form of Lacertilia. I affirm, in opposition to this view, that the relations in the two cases are totally distinct.

The seals agree with the Carnivora in all those important respects in which I have shown the Pythonomorpha to differ from the lizards. The seals possess a sternum and sacrum like other Carnivora; neither do they differ in the structure of the brain-case nor otic region from the same order. The teeth have dentinal roots like other Carnivora; and although the limbs are adapted for aquatic use, and formed superficially like those of Pythonomorpha, their bones are like those of Carnivora in all important respects. They possess the usual condylar articular faces, even to the phalanges; they have ungues also; so that all the parts common to the limbs of Carnivora may be found in the seals. The difference between the limbs of Lacertilia and Pythonomorpha is radical in general and in particular.

Professor Owen objects to the name which I have given to the order, and seems to think it conveys an erroneous impression. Such an impression as to my meaning appears to have been made upon my critic: what I mean to convey by it can be readily understood by reference to my definitions. The name would not be erroneous even if applied to an eel or other serpent like animal without the least affinity to Ophidia, and is rather more appropriate than the names Ichthyopterygia for reptiles whose fins are not truly like those of fishes, or Dinosauria, some of which are small and weak. As to the use of the term sea-serpent, since I have not referred these reptiles to the Ophidia, the term involves no I have used the same expression in writing of the contemporary Elasmosauri, of totally distinct affinities. As the first name proposed for these reptiles as a natural group, with a definition, the name I have given will stand in accordance with all the rules of nomenclature.