248 BULLETIN OF THE UNITED STATES FISH COMMISSION. 



than one per cent. The net result of two sendings was to get 73 per 

 cent, of the reports desired : 



I should have preferred a third sending at the expiration of two hun- 

 dred days from the second. I estimate that a third call on the remaining 

 11,133 offices might have produced about G,000 more replies, and brought 

 the per cent, of results up to 86 per cent, of what was desired. 



I have taken two hundred days as the limit of returns. The number 

 that arrive after that is very insignificant. But it will be valuable to 

 know what results to expect earlier. Taking the total of 1G,996 replies 

 which came in the first two hundred days, the percentage which came 

 in each of the ten twenty-day periods was as follows : 



The return within one hundred days of 97i per cent, of all that would 

 arrive within two hundred days, shows that when subsequent calls are 

 to be made one hundred days is a very good limit to fix at which to 

 send again. The receipts under the second call were even more pre- 

 cipitate than the first, being augmented slightly by the driblets from 

 the first call. It was as follows : 



