[3] PROPAGATION OF SCIIOODIC SALMON IN 1881-'82. 109S 



tlie stream were higiier than aiij' year since 1875. xV sudden rise of 

 water, owing to copious rains in August, 1880, had been followed by an 

 abundance of fish in the stream early in September. The high water of 

 1881 did not have the same effect on the fish, scarcely any salmon en- 

 tering the stream till after the middle of September. The inference 

 naturally suggested is that the condition of the stream favorable to a 

 late summer or early fall run of salmon is not so much a high stage of 

 water as a sudden rise; but the phenomena observed are hardly suffi- 

 cient for confident generalization. Moreover, during ten days in August 

 the gates at the dam were closed for certain repairs on a dam at Calais, 

 and meanwhile the flow of water was confined to that entering the canal. 

 From August 3 to September 10 there was a fall of 5 J inches, and from 

 September 10 to October 29 a further fall of 15 J inches ; November 5, a 

 rise of 2 inches, owing to rains on the two preceding days; and after 

 that date there was neither rise nor fall until December. 



The usual nets were placed across the stream and canal about the 

 middle of September, but no preparations for the capture of the salmon 

 were made until October 29, when it was observed that the most for- 

 ward of them had begun to form their ridds above our nets. On the 

 night of October 31 the capture of fish began. The manipulation was 

 delayed until IS'ovember 8, when some hundreds of salmon had been 

 collected, and a part of them exhibited great uneasiness, a few actually 

 beginning to spawn in the inclosures. The work proceeded as usual 

 until November 19, when all the salmon taken had been deprived of 

 their spawn, and the almost entire cessation of the catch told that the 

 season was at a close. 



An accident during the work of spawning confused the different lots 

 of fish so that the number taken from day to day cannot be stated with 

 the usual accuracy, but the tally-book shows exactly the number of 

 females that were manipulated, and enables me to make an estimate of 

 the total number of males, which, I am very confident, is within 15 of 

 the true number. According to these estimates there were taken 652 

 female salmon, 370 males, and one of unknown sex — total, 1,023. There 

 were 621 females that yielded spawn, and the eggs obtained from them 

 are estimated at 947,000, being an average of 1,525 eggs from each 

 female. 



3. — Shipment of spawn. 



The development of the eggs intended for earliest shipment was car- 

 ried on in hatchery No. 3, the remainder being kept in the colder water 

 of No. 2. It is from the latter that the reserve is always selected, since 

 the retardation of their development will bring them out in the spring 

 much nearer the natural date than if developed in the warmer water. 



The shipment of eggs began .January 12 and closed March 1. The 

 losses up to the time of the division of the eggs, when they were either 

 shipped or set aside for the reserve, aggregated 87,091, of which 62,159 

 are known to have been uuimpregnated eggs. From this we may fix 



