THE DISTRIBUTION OF BACTERIA IN i.AKES 



61 



numbers of bacteria found in it by Miss 

 Bere, tlie only observation approaching 

 tliose of the Russian workers at Lake Glubo- 

 koje. Perhaps "saprobic" would describe 

 the lake better than any other term. 



Boulder Lake is also a special case. The 

 lake is broader, but rather shallow (5 

 meters). It occurs in the course of the 

 Maiiitowish River, and receives dark brown 

 bog Avater from the upper reaches of that 

 stream. The organic matter in this water 

 is dejjosited in tlie lake; the bottom deposit 

 is "dy"-like. The effluent stream is clear. 

 Apparently the lake is taking care of the 

 organic matter brought to it. It is a good 

 fishing lake, and probably should be consid- 

 ered as eutrophic, although it has some dys- 

 trophic characters. The remaining eu- 

 trophic lakes named in Table XI are typical 

 enough, and have been described elsewhere. 



Of the oligotrophic lakes. Crystal and 

 Weber deserve special mention. They are 

 very similar in size, though Crystal Lake is 

 somewhat deeper. They are seepage lakes 

 occurring close together, with unusually 

 clear water. But Weber Lake had been sub- 

 jected to fertilization exiDeriments for sev- 

 eral years before the observations here 

 recorded were made. This perhaps ac- 

 counts for the higher numbers of bacteria. 



The two dystrophic lakes are typical 

 sphagnum bog pools occurring in a spruce 

 forest, the water still and cold. Mary Lake 

 is much deeper than Helmet Lake. 



The data presented in Table XI concern 

 only observations from stations in the open 

 lake, since there are no data available for 

 littoral stations save in Lake INIendota and 

 Lake Alexander. Only summer data are 

 included. The data from the various lakes 

 are therefore comparable from the stand- 

 point of seasonal and horizontal distribu- 

 tion. The counts of periphytic bacteria, 

 and the water plate counts, represent aver- 

 ages of samples from all depths (1, 2. or 3 

 meter intervals) from top to bottom. 

 Where the bacteria on slides were too few 

 to count, they have been considered as zero. 

 The counts therefore represent averages of 

 a considerable, but varying, number of indi- 

 vidual counts. The microscopic counts of 



bacteria in evaporated samples of water, 

 taken from Miss Bere's paper, are in some 

 cases based upon a single sample; where 

 samples were taken at dift'erent depths or 

 on different occasions, I have averaged the 

 results. But the number of counts averaged 

 is considerably less than in the otiier three 

 columns. Partly for this reason, and partly 

 because of the nature of the technique, I 

 consider them less accurate than the other 

 data. 



One of the most striking features of the 

 data in Table XI is the close correspondence 

 of the plate counts of lake water and the 

 counts from submerged slides. The number 

 of bacteria per square centimeter per day 

 deposited upon the slides is almost directly 

 convertible to bacteria per cubic centimeter 

 which can be cultivated on agar plates. 

 Such is the magic of averages 1 The num- 

 bers of bacteria in the bottom deposit, and 

 those computed from evaporated samples do 

 not correlate so closely. 



It will be seen that the eutrophic lakes 

 contain much larger numbers of bacteria 

 than the oligotrophic ones, no matter which 

 of the four methods is used, when averages 

 are considered, but that the figures derived 

 from the submerged slides and from the 

 plate counts of water are more uniformly 

 consistent in classifying the lakes than are 

 the other two sets of figures; i.e.. the lowest 

 counts for the eutrophic lakes are higher 

 than the highest counts for the oligotrophic 

 lakes in the first two columns, but this is 

 not true of the third and fourth. The in- 

 consistencies in the microscopic counts from 

 evaporated samples are perhaps to be ex- 

 plained by inaccuracy of the data. 



The two dystrophic lakes are widely dif- 

 ferent. Tlie vertical distribution of bacte- 

 ria on submerged slides indicates that in 

 these lakes bacterial activity is largely con- 

 fined to the very surface of the water. It 

 follows that, the deeper the lake, the smaller 

 will be the counts when observations from 

 all levels are averaged. It is perhaps on 

 this basis that we may explain the higher 

 counts from Helmet Lake (of the magni- 

 tude of the eutrophic lakes) and the lower 

 counts from Mary Lake (of the magnitude 



