are usually finely interlaced. This interlocking probably accounts for the toughness of the nodules; the inner surfaces of broken nodules, however, can be disintegrated easily with the finger nail. The fibrous material is pure white and opaque, with a pearly lustre. The tiny prismatic crystals are lath-shaped and often longitudinally striated; they are colourless and transparent. When gently rubbed between two glass surfaces they show a perfect longitudinal cleavage and signs of a transverse one approximately at right angles to this. The specific gravity of the fibrous material is 2.302 (4), (specific gravity bottle and vacuum). The refractive index of the lathshaped crystals is 1-540 in one direction of extinction and 1-542 with the lower nicol rotated through 90°-determined in mixtures of clove and cinnamon oils at 30°C. in sodium light and controlled with an Abbe refractometer. They have apparently straight extinction, but individual spicules are so thin that an obliquity of several degrees would not be observable. The elongation is positive. The fibrous material has the following composition.1

30°

It contains no fluorine. The SiO2 is about 3 per cent. low, RO is about 12 per cent. high and H2O about 1 per cent. high compared with the theoretical composition of okenite of the accepted formula CaO, 2 SiO 2, 2 H2O, with 56-70 per cent. of SiO 2, 26-36 per cent. of CaO and 16.94 per cent. of H20.

The other minerals present are apophyllite and laumontite (see Pl. 13, fig. 2). The apophyllite is in well developed crystals up to 1.5 cm. square, showing (100), usually striated, (111) and (001). The laumontite is rather weathered and extremely friable. It has a good prismatic

1 I am indebted to Professor A. Lacroix for permission to make this analysis in his laboratory in the Muséum national d'Histoire naturelle, Paris.

cleavage and the the characteristic (201) cleavage. ẞ for sodium light is about 1.5151; the optical character is negative. The paragenesis is not easy to determine. In thin sections 2 the okenite fibres are seen penetrating the apophyllite and certainly seem to be replacing it. Again, idiomorphic apophyllites, jutting out of okenite nodules are seen, when the surrounding okenite is removed, to have lost their idiomorphic character wherever the two minerals are in contact, and often the contact persists to the centre of the okenite spherules. There is, however, just as definite evidence against growth by replacement; in a cavity protected from abrasion is a small apophyllite crystal with uncorroded prism, pyramid and basal plane faces, from which project scores of undoubted okenite spicules, firmly attached, showing at any rate that the growth of okenite was not incompatible with the persistence of unaltered apophyllite (see Pl. 13, fig. 3); while on another specimen from Nowroji Hill collected by Mr. J. Ribiero, a tiny, perfect crystal of apophyllite is seen poised on two needles of okenite. Probably conditions did not present an irreversible sequence. With minerals so similar in composition a slight change in temperature, in pressure, in concentration of any of the components of the system, might alter the necessary conditions for deposition of one mineral to those for re-solution of that and crystallization of the other, while a subsequent change allowed formation of the second without mutilation of the first or even, at a transition point, of simultaneous crystallization of both. Possibly the presence or absence of fluorine was one of the controlling factors. A careful search for fluorine on 2 g. of okenite by the amended method of Berzelius, showed its absence. 0.02 per cent. was found by the method of F. Pisani,1 but this not qualitatively confirmed. In apophyllite from the same specimen 0-94 per cent. of fluorine was found by the Berzelius method.

Laumontite appears to be the earliest of the three zeolites being found between trap and okenite and sometimes as a nucleus of the okenite spherules.

1 Cf. E. S. Larsen, Amer. Mineralogist, VI, 7, (1921).

2 Prepared at ordinary temperature by the vacuum method of E A. Wülfing, Cent, ralbl. Min. Geol. u. Pal., 1920, 317.

3 Cf. W. F. Hillebrand, Bull. U. S. Geol. Surv., DCC, 222, (1919).

4C. R. Ac. Sci., CLXII, 791, (1916),

A FRESHWATER FISH FROM THE OIL-MEASURES OF THE DAWNA HILLS. BY THE LATE N. ANNANDALE, C.I.E., D.SC., F.R.S., F.A.S.B., Director & SUNDAR LAL HORA, D.SC., Assistant Superintendent, Zoological Survey of India. (With Plate 14.)

THE

HE fish described in this note was collected by Professor J. W. Gregory, F.R.S., at Mepale in the Dawna Hills, Tenasserim. Its remains are preserved in stiff clay evidently of lacustrine origin and associated with the limestone in which the shells 1 from the same locality which one of us has recently described were obtained. The type-specimen will be returned to Glasgow University.

The species evidently belongs to the family Cyprinidæ and we believe to the subfamily Cyprininæ, but its characters are so distinct that a new genus must be set up for it. We propose for it the name :—

DAUNICHTHYS, gen. nov.

The head is large and about as deep as the body. It was apparently flattened above with the eyes in its upper half. The jaws are not suctorial. The body is short and moderately deep.

10

a

TEXT-FIG. 1.-Dorsal and anal fing.

(a) Dorsal fin;

(b) Anal fin.

1, 2, 3 bony spines of the dorsal; 1′-12′ branched and flexible rays of the dorsal.

1 Annandale, Rec. Geol. Surv. Ind., LV. pp. 97-104, (1923).

There are at least 31 vertebrae, of which 13 appear to be caudal and the caudal and trunk regions are about equal in length. The lateral line runs along the tail below the vertebral column.

The dorsal fin is situated near the middle of the body and is of moderate length. There are at least 15 rays, of which about 12 are branched. The last bony ray is stout and strongly serrated in its upper two-thirds. The caudal fin is long and deeply notched, with the two halves pointed and equal. The ventral lies below the dorsal and has more than six rays, none of which is strongly developed. The anal, which is situated behind the dorsal, is of moderate length and contains two unbranched and 9 branched rays. There is no trace of scales in the specimen.

DAUNICHTHYS GREGORIANUS, sp. nov.

D. 3/12; A. 2/9 ; P. 7+; V. 6+; C. 30.

The length of the head is contained 3 times in the total length without the caudal fin. It is as deep as the body. The greatest depth of the body is contained a little over 3 times in the total length without the caudal. The dorsal fin was probably as high as the depth of the body below it. Its first bony ray is short, the second of moderate length and the third much longer and deeply grooved throughout its length. The branched rays of this fin and certain rays of other fins of other fins are longitudinally grooved. The pectorals and the ventrals are widely separated and cannot have overlapped. The commencement of the anal is nearer to the ventrals than to the base of the caudal.

The total length of our example is 56 mm., that of the head 14-2 mm., the greatest depth of the body 14 mm., and the length of the caudal fin 13 mm.

Having thus described the genus and the species we will now proceed to examine the specimen in greater detail.

Skull and associated structures-In the region of the head the jaw-bones, the opercular bones, the secondary pectoral arch and the bones of the brain-case can be distinguished after careful examination, but the other bones have been completely broken up. It seems, however, quite probable that there was a complete circumorbital ring for traces of it can still be made out. Of the jaw-bones, the lower jaw is broken in the middle longitudinally, while the dentary of the unexposed side, which is also visible, is further broken

into two pieces. The maxillaries are represented by nodule-like bones at the top of the premaxillary, which is closely approximated to the dentary. From the direction of the jaws, which are directed

=

TEXT-FIG. 2.-Bones of the jaws, operculum etc., and the anterior modified vertebrae. d=dentary; mx-maxilla; p. mx-premaxilla; p. o. preoperculum; i. .interoperculum; 8.0-suboperculum; o. operculum; p. f-pectoral fin; p. c. = post clavicular process; c. cleithra; tp1=transverse process of first vertebra; tp2=transverse process of second vertebra; tp-transverse process of fourth vertebra; r1rib of first vertebra; r3-rib of fifth vertebra; r = rib of sixth vertebra; s=scaphium; i. 7.-interossicular ligament; f = frontal; p=parietal; 8.0= supraoccipital.

almost vertically upwards in the specimen, it is evident that the mouth-opening must have been directed obliquely upwards as in the living Catla. The four opercular bones are quite clear and are well developed. It seems to be quite clear from the position of these bones that they have been detached from the jaw-bones post mortem and have been pushed backwards and downwards by external pressure. The secondary pectoral arch is complete and well developed. It is not emarginate anteriorly, but exhibits a somewhat primitive form of the cleithra. The rib-shaped post-clavicular process of the secondary arch is also well marked. Lying alongside the posterior border of the cleithra is a rib-shaped structure, which in all probability represents a rib of the first vertebra, for we know of no other similar structure in this position in the living fishes.

Regan, Ann. Mag. Nat. Hist., (8) VIII, p. 28 (1911),