The paper contains a discussion of the above results. It is shown that probably Lord Rayleigh's value of r may be too high by as much as 0'0002, in consequence of the fact that the mercury in his terminal caps was 5° or 6° C., but no complete explanation of the differences between his result and those of Rowland, Kohlrausch, and ourselves, has been found. The difficulty of working with tubes such as those used by the Lorentz, 1-2 metres in length, and 1, 2, and 3 cm. in diameter, may perhaps account for his value for the ohm, viz., 105·93.

XI. "Researches on the Structure, Organisation, and Classification of the Fossil Reptilia. VI. On the Anomodont Reptilia and their Allies." By H. G. SEELEY, F.R.S. Received June 20, 1888.

(Abstract.)

The author examines the structure of the skull in the Dicynodontia, and discusses the interpretations of its elements and affinities given by Sir Richard Owen, Professor Huxley, and Professor Cope, and arrives at the conclusion that the interpretation of the bones of the. palate may be varied. The quadrate bone is found, though it is absent from many specimens owing to loose articulation, and the malleus is recognised as a normal element in the skull, which articulates with the quadrate and is free, except at its extremities. The palatine bones are internal to the pterygoids, and the pterygoids extend forward to the maxillary. The columella is found in more than one specimen. Many new specimens are described which further elucidate the structure of the skull. The first of these shows that the upper part of the foramen magnum is formed

by the supra-occipital bone, and that the element which has appeared to be a supra-occipital is the inter-parietal. Evidence is given of the form of the brain case, which is found to be high and narrow. Details are given of the structure of the squamosal bone, and of its relation to the quadrate and other cranial elements; and it appears that the squamosal usually embraces the quadrate, so as to extend in front of it, and sometimes to hide it, so that both the quadrate and squamosal sometimes contribute to form the articulation for the lower jaw. Evidence is offered of the sutures which divide the bones of the skull from each other. The sub-nasal element, found in Pareiasaurus, is met with in Dicynodonts, sometimes below the narine, and sometimes within its floor in the position of a turbinal. A new type of quadrate bone, which is regarded as Anomodont, is described, and found to differ from the usual form in being perforated in the antero-posterior direction. A summary of the structure of the skull is illustrated by a restoration showing its sutures.

Further contributions are made to a knowledge of the vertebral column. The cervical vertebræ are described, the atlas and axis are regarded as anchylosed, and succeeded by an intercentrum which has no neural arch. The cervical ribs are comparatively long, and articulate by a long fork with the neural arch, as well as with the centrum. Further evidence is given of the structure of dorsal vertebræ, showing that the rib is attached to a single transverse process of the neural arch. The caudal vertebræ of Platypodosaurus, eleven in number as preserved, are described; and some observations are made on the mode of ossification of the intervertebral substance. Additional materials further elucidate the Anomodont scapular arch, and examples of scapula and coracoid are described; but the only additional pelvic bone described is the pubis of Titanosuchus.

An account is given of the limb bones, which are elucidated by large bones associated with the skull fragments described by Sir R. Owen as Titanosuchus ferox. They contribute to a knowledge of the femur, humerus, and fibula in that type, and are associated with small bones of the extremities which are probably metacarpals. The ulna is described, which was referred by Sir R. Owen to Pareiasaurus, and evidence is given that it possessed terminal epiphyses of different form to any which are known in fossil reptiles, the proximal epiphysis having much the character of the olecranon of a mammal. A massive Anomodont tibia, also referred by Sir R. Owen to Pareiasaurus, is described, and found to possess a distal talon of mammalian pattern.

Further observations are made upon the Theriodontia, as restricted to the genus Galesaurus, the skull of which is further elucidated. The author also describes new material, making known the structure of the skull, palate, and scapular arch of Procolophon; from which it appears that the pre-coracoid is exceptionally well developed, and

united by suture to the coracoid. The inter-clavicle had the slender T-shaped form of the bone in Ichthyosaurus.

Procolophon has teeth on the vomera and pterygoid bones, and the structure of the palate and the post-orbital region show that the Procolophonia forms a distinct division of the Anomodontia. Observations are made on the relations of the European and South African Anomodonts, and on the relation of the Anomodontia to the Pelycosauria and to Cotylosauria. Comparison is made with Placodus, which genus has two exoccipital condyles, comparable to those of mammals, and appears to have lost the basi-occipital condyle. Comparisons are made with other extinct reptilia to show the relation of the Anomodonts to the Saurischia, and other reptilian types. Observations are offered on the theory of the Anomodont skull, and on the effect of the articulation of the lower jaw with the squamosal in causing a diminished growth of the malleus and quadrate, converting them into the malleus and tympanic.

The larger groups included in the Anomodont alliance are regarded as the Pareiasauria and Procolophonia; Dicynodontia, Gennetotheria, and Pelycosauria; the Theriodontia, Cotylosauria, and Placodontia are regarded as coming under the same sub-class, which at one end of the series exhibits characters which link reptiles with amphibians, and at the other end of the series link reptiles with mammals.

XII. "A new Form of Eudiometer." By WILLIAM MARCET, M.D., F.R.S. Received June 20, 1888.

[PLATE 14.]

The quantitative determination of oxygen, simple as it appears at first sight, is found in practice beset with many difficulties. Liebig's method with pyrogallic acid and potassium hydrate, though considered as yielding correct results, takes too much time, and is unsatisfactory in some respects, so that the eudiometer has become of general use for the estimation of oxygen. I shall not attempt to describe the various forms of eudiometer, but it may be assumed that Regnault, so well known for the care he bestowed on his investigations, had adopted a very correct kind of eudiometer in the researches he undertook with Reiset on the chemical phenomena of respiration.* Other eudiometers have been made since then, such as the ingenious instrument of Dr. Frankland for gas analysis, which has proved most serviceable. I claim for the present form of eudiometer that it is correct and reliable in its working, simple in construction, and easy of manipulation. The main objects of an eudiometer must be the easy introduction of the air to be analysed, the ready mixture of that air with a known volume of pure hydrogen gas, and the correct reading

* ' Annales de Chimie et de Physique,' 3rd Series, vol. 26, 1849.

of the volume after explosion. It will be seen that these conditions are entirely fulfilled in the present instrument; and it has, moreover, the advantage of being available in conjunction with Pettenkofer's method for the determination of carbonic acid in atmospheric &ir.

The eudiometer as figured in the accompanying Plate has the form of a T-piece, the vertical limb of which is a straight tube about 60 cm. in length and 12 cm. in diameter; it is divided into 50 or 60 c.c. and tenths of cc., like a common burette. The upper end of this tube is closed air-tight with a steel cap, from which lateral tubes project right and left; these tubes are bent V-shaped, or rather in the form of a lyre. At the junction of the lateral tubes with the cap, there is a three-way stop-cock allowing of the passage of air or gas in four different directions, viz., first through the tubes cut off from the body of the eudiometer; secondly, into the eudiometer, which is done by raising it in the mercury trough; thirdly, out of the endiometer, on the side opposite that from which it was introduced, which is effected by depressing the tube in the mercury; fourthly, through the tubes and eudiometer simultaneously. The eudiometer is held tightly by two claws projecting at different heights from a vertical iron rod connected with a rack and pinion movement. The iron rod, together with the eudiometer, is immersed in mercury contained in a straight cylindrical glass vessel.

The hydrogen used for the explosion is prepared for that special object from zinc and sulphuric acid in the ordinary way, and washed through an alkaline solution, rather than obtained condensed in iron bottles from the manufacturers, and it is collected in a bell-jar suspended over water. The bell-jar I use holds 11 litres of gas; it is balanced by a counterpoise, and its weight, as it moves up and down in water, is regulated by another counterpoise hanging from a cycloid, so that the gas in the holder is always under atmospheric pressure; an oil-gauge fixed to the holder shows at any time the pressure in the bell-jar. Should the gas fail to be absolutely under atmospheric pressure, the equality of pressures may be ensured by the use of the adjusting instrument I have described in a former communication. It consists of a clamp fixed to the rim of the tank, and made to grasp at will the cord holding the counterpoise; a screw in connexion with the clamp enables the cord, and consequently the bell-jar, to be drawn up or down. For the actual requirements of the analysis, a receiver for the hydrogen holding only one litre of gas would suffice, but it is better to have a larger gas-holder in which to store up the hydrogen for future determinations.

Moreover, the cycloid arrangement for regulating the weight of the bell-jar, though very convenient, may be dispensed with, as the gas in the receiver can be brought approximately under atmospheric pressure

by means of weights, while the adjusting screw will enable its being accurately placed under atmospheric pressure.

The analysis is made as follows:

We suppose that air for analysis has been shaken with barium hydrate in a glass jar of a capacity of about 10 litres, and made according to the form adopted by Dr. Angus Smith* for the determination of carbonic acid in air by Pettenkofer's method. This jar is closed by a tight-fitting india-rubber cap, which I cover with several coats of copal varnish; from this cap two short india-rubber tubes project, each of these tubes being clamped by a pinch-cock. After the agitation is over, and when all the carbonic acid is taken up by the alkaline solution, the fluid is poured out from the jar into a glass-stoppered bottle holding about 100 c.c. This can be done easily without letting any air into the jar, as the india-rubber cap will collapse somewhat while the fluid is allowed to run out through one of the india-rubber tubes in the cap, a very small quantity of fluid only being left in the jar. The india-rubber tube is again clamped, and the bottle holding the barium hydrate is sealed with paraffine and left undisturbed for the precipitation of the carbonate and subsequent analysis.

The glass jar full of air free from carbonic acid, and absolutely saturated with moisture, is placed under a funnel supported on a filter stand, and the funnel is connected with one of the india-rubber tubes projecting from the cap, while the other tube has a short piece of glass tubing inserted into it, to which a longer india-rubber tube is fixed.

Everything is now ready for the determination of the oxygen of the air contained in the glass jar. After turning the stop-cock in the cap of the eudiometer, so as to allow the hydrogen gas to wash out the steel tubes and top of the eudiometer, the latter is lowered in the cylinder until the mercury is in contact with the cap, and therefore very near to the stop-cock. The eudiometer is next connected by narrow india-rubber tubing with the hydrogen receiver on which a weight has been placed, and on opening the receiver hydrogen rushes out, washing thoroughly the passage through which it will have to reach the eudiometer, and driving out the very small quantity of air contained in the steel cap between the mercury and the stop-cock. I found it convenient to stop the end of the V-shaped tube letting out the gas with short india-rubber tubing and a pinch-cock. When a few hundred cubic centimetres of gas have gone through, the three-way tap is turned by one-quarter of a turn, so as to place the tube in communication with the hydrogen; it is now easy to rinse the eudiometer with that gas, by raising the eudiometer, and then giving the three-way cock half a turn, so as to bring the instrument in communi*Air and Rain,' 1872.