Of the 339 young salmon taken in the trap, 322 were under 2 inches long, and 17 over 4 inches. 178 of the smaller specimens were measured accurately with the following results:
The smallest had just absorbed the yolk-sac, and in many there was yet a small quantity of yolk remaining. Those 1.6 inches long had not been feeding over three weeks, as other observations show that they grow at the rate of .3 inch a month. That we found young salmon that had barely absorbed the yolk-sac indicates that they begin their migration as soon as they are able to swim. That they remained of the same size from October 8 to November 30 indicates that practically all start down stream at the same age. If any of the same age as those taken October 8 had been taken during the last days of observation, the fact would have been indicated by their increased size. In 53 days they would have grown over half an inch, and would have been at least two inches long. There were no intermediate sizes
between 1.8 inches and 4.0 inches. The larger individuals had remained over from the previous year. The trap was so set that they could not have avoided it during the day. The fact that none was taken during the day indicates that they travel more or altogether at night. The last record shows that 24 were taken from 8 to 9 a. m., which is the only day record. This is accounted for by high and muddy water, caused by a heavy rain during the previous night. This and other observations indicate that high and muddy water, especially the latter, is an incentive to day travel.
The two months' observations with the trap at Battle Creek were deemed of such importance that two observation stations were established on the river the first of January, 1899, and equipped with similar traps. The upper station was established at Balls Ferry, some three miles above the mouth of Battle Creek; the other at Walnut Grove on the lower river. The diagrams indicate the number of salmon fry taken in the traps at the two stations: