143 



partially absorbed in most of the lot, but in those most advanced 

 was represented by an oil globule back of the head. The pectoral 

 fins were well developed, but no trace of the ventrals had as yet 

 appeared. The single median tin extended well in front of the vent, 

 and forwards on the back nearly to the head. The opercles did 

 not fully cover the gills. The most highly developed specimens — 

 those whose gill-sacs had nearly disappeared — had, at a short dis- 

 tance on either si-ie of the symphysis of the lower jaw, a sharp, 

 strong, raptatorial tooth, curved backwards and slightly inwards. 

 The base of this tooth was very broad, and the point acute and 

 slender. At a point behind each of these teeth about half their 

 distance from each other, was a second much smaller tooth, directed 

 almost exactly inwards. The upper jaw was, however, wholly tooth- 

 less. 



These fishes were all passed under the microscope, after having 

 been rendered transparent, but only four of them contained any- 

 thing whatever; three a little dirt, and the fourth a minute frag- 

 ment of the crust of the Gammarus, with which they had been 

 fed. 



Of one bundled and eleven specimens received February 17, 

 seventeen had taken food. I dissected nine of these and found 

 fragments of Gammarus and nothing else. iSinety specimens from 

 the same lot were examined February 25, and food was found in 

 fourteen. Four of these had eaten Gammarus fragments; two, 

 larvae of gnats; one, a small Cypris ; and eight contained small 

 fragments of the leaves and stems of vascular plants, including a 

 bit of netted-veined leaf and a little piece of pine wood. Thirty- 

 nine specimens, the labt of the lot, were received March 15, and 

 food was found in fourteen. I dissected nine of these, finding frag- 

 ments of Gammarus in four, a larva of a gnat, a Chironomus larva, 

 a larva of some undetermined fly, a minute vegetable fragment, a 

 Cyclops, a Cypris, and an undetermined Entomostracan each in one. 

 Thfee hundred and forty fry from the hatching house were examined 

 in all, in forty seven of which (fourteen per cent.) more or less food 

 was discernible. Of the thirty-five dissected, eighteen had eaten 

 Gammarus fragments; five, minute insect larvse ; four, Entomostraca, 

 and eight, small particles of vegetation. 



Only four lots were received from the spring, on the 9th, 14th, 

 17th and 25th of February, after which all died of starvation. In 

 the first hundred only one was found which had taken food, and 

 this had eaten a trace of filamentous Algie and a minute fragment 

 of the parenchyma of some higher plant, with a few diatoms. But 

 one of the second hundred contained even a trace of food, a minute 

 quantity of some thread-like Alga, the cells of which still contained 

 a little chlorophyll. In the third hundred, likewise, food was found 

 in but one. This consisted of a few particles of vegetable paren- 

 chyma, doubtless derived from the decaying plant structures in or 

 around the water. In the third lot of only forty-two specimens, six 

 showed traces of food, consisting almost entirely of a few filament- 

 ous Algae (including a fragment of Oscillatoria) and a little vegeta- 

 ble parenchyma. Pesmids and diatoms were observed in trivial 

 numbers. 



