236 SULPHUR BACTERIA 



Methods of Investigation. 



1. A sample of pond water containing purple bacteria in 

 sufficient numbers to colour the water may be used. Light 

 is passed through a thin layer of fluid, and examined in the 

 spectroscope. This was the method of the earlier investigators, 

 for example, Lankester and Engelmann. The obvious dis- 

 advantage of the method consists in the impurity of the fluid. 

 Green Algae are frequently present and even mere traces of 

 chlorophyll affect absorption spectra. 



2. The colouring matter may be extracted by solvents, 

 and the absorption spectra of the coloured fluids examined. 

 This method has the same disadvantage as the first. 



3. The most reliable results are obtained when the colouring 

 matter is extracted from pure cultures of the organisms. 

 Only absorption bands of the non-sulphur purple bacteria 

 have liitherto been examined by the third method, because 

 until quite recently pure cultures of the sulphur bacteria have 

 not been obtained. This method, however, is not free from 

 objections, which are those that are inherent in all spectro- 

 scopic methods of examination. The chief objection is the 

 extreme unreliability of the evidence obtained in this way 

 when applied to the investigation of complex organic matters. 

 Such evidence must be limited to very wide generalisation, 

 for the following reasons : — ■ 



(i) Increase in the concentration of a fluid giving absorption 

 bands leads to a widening of the bands, not to an 

 intensification of their opacity. 



The range of wave-lengths occupied by the bands of 

 strong solutions may be considerably greater than those 

 occupied by the bands of weaker solutions, and may 

 extend to several colours of the spectrum. The difference 

 in the range may be so great that whilst four or five 

 dift'erent colours are affected by the bands of strong 

 solutions, only two, or at most three, are affected by 

 those of the weaker solutions. It is therefore not possible 

 to fix a definite range of wave-length as characteristic 

 of the bands of any particular coloured fluid. 

 (ii) This difficulty would not be so great if it were always 



