Lake Water and Sediment

Abstract

Procedures are given and field apparatus described for use of the millipore filter technique for lake water and sediment bacteria. Agar plate and filter procedures give the same counts using a supplemented sodium caseinate medium. Direct counts were higher than colony counts by factors from 59 to 233, with a mean of 111. Cultures made on lake sediment extract in agar were inferior to the caseinate medium in unproductive lakes, but when taken from productive lake soils they were adequate to produce colony growth. Several hundred quadruplicate tests were studied to establish errors, which had two components: the theoretical error of a Poisson series which depends on the number of colonies on a filter, and a handling error which is ±9%. Combining the above errors with the natural variability of profundal sediment from station to station and visit to visit, and expressing the standard error of the mean as a per cent, the expectations are: one visit to take one core at each of four profundal stations comprising the top two or three cm will yield, on quadruplicate plates, an error of ±130%. Hence, while a single visit may set a quite useful maximum it cannot be used to set a minimum. In the procedure adopted four field trips as above were made, by which the error was reduced to ±36%.Shallow cores, down to 15 cm or more, were sectioned. Surface scrapings did not yield the very high counts that others have reported. There was no significant decline in numbers down to about 5 cm, which comprises Naumann’s “actual” layer. Below this a decline sets in upon entering the “historical” layer. Deep cores were taken on two lakes down to base gravel presumed 15,000 yr old. There was a decline with depth which appeared to be logarithmic, although the deep counts were too close to zero to prove it.The authors’ counts, supplemented by those of Henrici and McCoy, are plotted against other measurements thought to be related to productivity, namely methyl orange alkalinity, conductivity, and (in paper V of the series) oxygen consumption over mud. A good regression is observed in all clear‐water lakes. However, brown‐water lakes, although low in alkalinity and conductivity, have high bacterial counts. In these lakes there is a good correlation of bacterial count with water color. This provides a possible basis for stating what part of the bacterial count of any lake is to be attributed to bog. By subtraction of the bog fraction the real or useful measurement as indicative of productivity is obtained.