Ecology of the burrowing amphipod, Pontoporeia affinis, in Lake Michigan / by Wayne Paul Alley.

113

the exception). In the two dimensional presentation of data, such as
Pontoporeia density versus distance from shore, it is impossible to
determine how much of the apparent relationship is actually attributable
to depth, bottom type or other environmental parameters. Multiple linear
regression is a useful statistical tool in the study of biological systems
because it has the ability of mathematically removing the effects of
confounding environmental factors.
A remarkable aspect of the multiple regression analysis was the
high multiple correlation coefficients that were obtained at the four depth
zones: 0.756 for 10-35 m, 0.599 for 36-65 m, 0.677 for 66-105 m, and
0. 826 for greater than 105 m. The high multiple correlation coefficients
imply that the physical and biological parameters "explain" a large
proportion of the changes in amphipod counts.
Within the 10-35 m depth region of the lake, the multiple regression
equation showed a significant interaction of depth, sediment type, and
distance from shore which indicated that inshore regions of the lake present an unsuitable environment for Pontoporeia. The turbulence attributable to waves and water currents of this region have the ability of
resuspending and redepositing the finer sediments and detritus into other
areas that are less mixed. These shallow areas contain surficial sediments of coarse beach sand and hard bottom types, and fewer amphipods
are found there.
The necessity of including the depth squared term in the multiple