21 
Carp from Havana and points downstream showed no abnormal- 
ity, except for a few individuals, perhaps one in a hundred, which 
supposedly had straggled downstream from Feoria Lake. At points 
between Feoria and Havana, intermediate percentages of "knothead" carp 
were found. At Fekin, for example, about 30 per cent of the knothead 
abnormality occurred. These percentages were based on carp averaging 
about 5 years of age, hence it seemed clear that carp did not migrate 
any considerable distances in the five years from the time the knot- 
head characteristic was determined in the fry stage until they were 
caught 5 years later, else the dividing line between knothead and 
normal carp would not be so clear-cut as it is between Feoria and 
Havana, Table I shows that the calculated average migration of a 
carp in five years is 17.9 miles. This figure was rather accurately 
predicted from a consideration of the distribution of knothead and 
normal carp in the Illinois River. 
The data on tagged carp have been arranged in Table Ix, 
p. 22, to show the rates of upstream and downstream migration. The 
rate of upstream migration shows a calculated migration constant of 
0.24 miles in one day, while the downstream rate is 0.57 miles per 
aay. i regard these rates as significantly different. The greater 
downstream rate may be a reflection of the effects of drouth since 
fishes commonly seek an optimal stream size and move downstream in 
times of drouth. The data on tagged "fine fish" show upstream move- 
ments almost ten times as rapid as downstream. The reason for this 
difference from the behavior of carp is not clear. It may be supposed 
that "fine fish," most of which are found in bottomland lakes and 
other quiet waters, travel upstream willy-nilly when they find them- 
selves in a current. Their rate of migration in lakes is only one- 
fourth their upstream rate. It must be remembered that any fish which 
shows a preferential downstream movement under all conditions cannot 
long remain an Illinois fish and leave descendants to inhabit Illinois 
waters. Since these "fine fish" are not typical stream fishes, their 
reactions may be imperfectly adapted to stream life. This view seems 
to be supported by the low rates of migration found for typical stream 
fishes such as channel cat, common redhorse, black sucker, carp, and 
quillback. It aprears that typical stream fishes are those which 
have a set of reactions which allow them to hold their place ina 
stream without crowding into the headwaters or being swept down to sea. 
While the evidence obtained by analysis of statistics 
presented here shows that the movements of fishes are random in 
nature, it is not incompatible with many kinds of observations which 
show that the movements of fishes are directed, i.e., not random. 
These seemingly opposite views may be readily harmonized by supposing 
that this randomness of fish movement appears only as a gross rela- 
tionship covering long periods of time, while those observations 
tending to show that fish movements are directed commonly arply to 
much shorter periods of time. A great number of physical variables in 
water affect the movements of fishes, and these are interrelated in so 
