236 BULLETIN OF THE BUREAU OF FISHERIES. 
successes and in the effort to obtain results on a scale large enough and with cost 
small enough to deserve the adjective “‘practical.’”’ The total outputs for the 
years are: 
T89822 5-825 2. cape (2) VOOR! Seen sane ae 27; 300" 19062552 - S526 =---- 28 C189, 384 
SO Gane eee eee (©) WGI eres eee nee ae £37,500) 1907- aa d2s= = ose ana 4294, 896 
TQ00s ena BREA) [ai oy le Se ae SS See ofhiye Ipuishhe oa eae €322, 672 
OO ee SiG 7a GOS nee 103, 572 
The rearing of considerably over 300,000 lobsters in the small plant at 
Wickford recalls the confession of faith written ten years ago, at the conclu- 
sion of the first season’s work: 
We know perfectly well that many others have failed in doing what we attempt, 
but until we are thoroughly convinced that the young lobster can not be ‘“‘brooded”’ 
we propose to continue our work./ 
Manner of determining output.—It was early realized that “estimates” 
of the number of lobsters in experimental work are practically worthless and 
therefore all the fourth-stage lobsters which are taken account of at all (many 
thousands of others have accidentally escaped) have been individually counted. 
Within the last few years, when the numbers have run up into hundreds of 
thousands, the operation of counting individuals has consumed much time, but 
the satisfaction of accuracy in results has been sufficient compensation. A 
comparatively easy and very accurate method of counting is now in use. The 
“Jobsterlings”’ are dipped out of the hatching boxes with flat woven-wire strainers 
which take up from one to twenty at a sweep; these are recorded on an auto- 
matic counting register held in the hand. The count at each sitting is then 
posted (fig. 7 pl. x). ; 
It is of little use to estimate the number of a lot of first-stage fry. More 
than once the lots so estimated, even by experts, have yielded not more than 
10 per cent of the estimated number; sometimes, no doubt, they would run 
considerably over. For this reason, in order to ascertain the proportion of 
newly hatched fry to the fourth stage, the individuals must be counted both 
before and after the experiment. This is a rather tedious process, but it is war- 
ranted and necessary when new methods or new devices of construction are 
tested for their relative efficiency. 
Tested by this method both the large canvas bags used until this year and 
the present boxes have yielded on several tests 40 per cent of fourth-stage lob- 
sters from lots of 20,000 newly hatched fry. In one test of the canvas bags 48.2 
per cent were obtained in a 20,000 lot. In testing for relative value of foods in 
a “Several.” ¢ 24,800 to fifth stage. €5,481 to fifth stage. 
b “Several hundred.” @ 4,900 to fifth stage. J Bumpus, op. cit. 
