RIVALRY OF MALES AND SECONDARY SEXUAL DIFFERENCES. 93 



tation, the oldest was a gastrula covering but one-third of the yolk. The eggs had 

 not been developing more than 5 days, probably not more than 2 at the utmost, 

 and decided beyond a doubt that these fishes are oviparous and not viviparous. 

 In one individual 61 eggs were found, in another 70. The exact number in the 

 other two, I can not give, but the number does not differ greatly from the above. 

 From one side of one I took 35 eggs, from another individual an uncertain number. 

 The remaining eggs were left in the gills to develop, but all that were not subse- 

 quently preserved finally died. 



The female with eggs can readily be distinguished by her distended gills, and 

 since dead eggs become opaque, such can readily be distinguished through the 

 translucent operclcs and branchiostegal membrane. Dead eggs are retained in the 

 gill cavity till they disintegrate. 



I have never secured as many young from any female as the eggs enumerated 

 above. This may have been either on account of the dying of many eggs or the 

 liberation of the young during the struggle of capture. 



Emphasis need be laid on the fact that A-wblyopsis is not- "viviparous and that 

 its breeding period extends at least from the first of March to November and 

 probably throughout the year. A female with nearly ripe eggs was secured on 

 September 9, and since these would have been carried either as eggs or young for 

 about 2 months longer, November is a safe limit. During March the spawning 

 season is evidently at its beginning, and it is during this month and April and May 

 that the early stages may be looked for with the greatest confidence. 



No eggs were deposited in the laboratory. Females with eggs in the gill 

 cavities had to be sought for in the caves. To secure embryological material when 

 a female containing favorable stages was captured, she was isolated in a small 

 aquarium and the number of eggs needed freed from the gill cavity by gently 

 raising the edge of the operculum. The rest of the eggs were permitted to remain 

 in their natural surroundings until another lot was wanted. During the early 

 stages of development the edges of the operculum are closely pressed to the neck 

 and there is no danger of freeing more eggs than are wanted unless the fish is 

 roughly handled. During the later stages of development the tension of the oper- 

 culum is relaxed and eggs or larvae can be much more easily removed, but there 

 is a correspondingly greater danger of liberating more young than are wanted. 

 If the female is disturbed or confined during the latest stages of brooding, some 

 or all of the young will escape. The eggs freed from the gill cavity will continue 

 their development uninterruptedly, but the gill cavity of the female offers such a 

 unique and self -regulated hatchery that they were usually left in it. 



RIVALRY OF MALES AND SECONDARY SEXUAL DIFFERENCES. 



In an aquarium containing six specimens of Amblyopsis, two took a great 

 antipathy to each other and engaged in vigorous contests whenever they came in 

 contact. Frequently they came to have a position with broadside to broadside, 

 their heads pointing in opposite directions. The fight consists in quick lateral 

 thrusts toward the antagonist to seize him with the mouth. The motion is in- 

 stantly parried by a similar move by the antagonist. This blind punching may 

 be kept up for a few seconds, when, by their vigorous motions, they lose each 

 other and jerk themselves through the water from side to side, apparently hunt- 

 ing for each other. At this time they are very agile and move with precision. 



