Chapter IV ■ — 53 — Methods of Enumeration 



rarely contains this many), it must be concentrated before the bacteria 

 can be counted by the direct microscopic method. It is not practical to 

 concentrate sea water by evaporation at reduced pressures, as advocated 

 by Bere (1933) ^"tl others for fresh water, because of the high salt con- 

 tent of sea water. With very little concentration, certain salts commence 

 to precipitate from sea water, carrying many bacteria with the crystals. 

 Moreover, the higher osmotic pressure of the concentrated sea water has 

 undesirable elTccts on the morphology of microorganisms. By evaporat- 

 ing lake water at a reduced pressure at 35° C. to concentrate the bacteria, 

 KusNETZOW and Karzinkin (1931) microscopically counted 2000 to 4000 

 times as many bacteria in water from Lake Glubokoje as were demon- 

 strated by plate counts. They preserved the water with formalin, used 

 agar as a fixative, and stained the smears with carbol-erythrosin. 



Centrifugation is not a feasible method of concentrating the bacteria in 

 sea water because the specific gravity of most of the bacteria is near that 

 of the sea water, 1.024 to 1.030. The force of a high-speed angle cen- 

 trifuge precipitates some of the bacteria with the sediment while others 

 are concentrated at the top of the liquid, but the majority remain sus- 

 pended in the sea water or become attached to the glass. Some of the 

 problems of concentrating marine microplankton and nannoplankton by 

 centrifuging are discussed by Lebour (191 7), Lohmann (1922), and oth- 

 ers. Most of these methods detect only clumps of bacteria, zooglea, bac- 

 teria which are attached to larger organisms, and bacteria which exceed 

 10 fj. in diameter. 



Snow and Fred (1926) treated water with aluminum hydroxide to 

 facilitate the fiocculation of bacteria. They reported the recovery of 

 95 per cent of the known number of suspended bacteria in the stained 

 sediment after centrifuging. The direct microscopic examination of 

 water samples from Lake Mendota treated in this way revealed the pres- 

 ence of from 7 to 14 times as many bacteria as did plate counts. 



MtJLLER (1912a) concentrated bacteria by centrifuging loo-ml. sam- 

 ples of water treated with ferric perchlorate. A measured quantity of 

 the sediment was transferred to a known area of glass slide and stained 

 with gentian violet for microscopic examination. He believed that the 

 direct counts, which were 25 times higher than plate counts, detected 97 

 to 99 per cent of the bacteria in water. 



Various modifications of the centrifuge-fiocculation method used by 

 Snow and Fred (1926) have been applied to sea water with inconclusive 

 results. The principal difficulty has been in the precipitation by the 

 flocculating agents of various salts from sea water, thereby obliterating 

 some of the bacteria from view. From 10 to 200 times more bacteria were 

 found in sea water by its direct microscopic examination than by plate 

 counts. The statistical treatment of the data from the examination of 

 100 samples shows that whereas the deviation of duplicate plate counts 

 was 2.3 to 21.7 per cent, the duplicate successive dilution method counts 

 (using ten tubes for each effective dilution) deviated from the mean by 

 19.4 to 106.7 per cent and duplicate direct microscopic counts deviated 

 from the mean by 32.8 to 790 per cent. In other words, plate counts are 

 much more reproducible than direct microscopic counts on samples of sea 

 water, although the latter detect more microorganisms. 



Provisions are made for the direct microscopic observation of chem- 

 ically flocculated ultraplankton by Utermohl (1931) who observed the 

 stained organisms from below by inverting the microscope. His inverted 



