BY ARTIFICIAL MEANS. 239 



Frelinmiary Experimerit. — The diatomaceous sediment is agi- 

 tated in a small bulk of sterilised water, and this water — still turbid, 

 but deprived of the larger impurities by decantation — is placed in 

 a new flask, or one that has to be heated almost to redness. 



A drop of this liquid is placed on a slip of glass, whose upper 

 surface is divided into squares of one-tenth of a millimetre. The 

 water is allowed to evaporate, and you count under the microscope 

 how many diatoms are held in suspension in each evaporated drop, 

 noting especially the number of the frustules of the species that 

 you desire to insulate. 



Let us suppose that the drop has deposited five hundred frus- 

 tules on the surface of the lined piece of glass — four hundred of 

 various species and one hundred individuals of the species you 

 desire to insulate. Hence, it is nearly certain that one drop of 

 this liquid would introduce into a litre of water five hundred frus- 

 tules — one for 2 ccm. 



The result stated : — Thus, by introducing one drop of this 

 tested liquid into a litre of water, and distributing half-a-ccm. of 

 this Avater (after well shaking it) into twenty new macerations, five 

 among them will have been fertilised with one diatom and fifteen 

 will remain sterile. In the five cases of fecundation, the operator 

 may hope for a pure culture of the diatom he seeks to insulate. 



Results. — -After having exposed to the light of day the macera- 

 tions thus treated, it will be found that things are as above stated, 

 that twenty-five for every hundred cultures will give positive results ; 

 in a word, experiment will prove the theory. But the observer 

 may simplify this work, and put on his side all the chances of 

 success. He should endeavour to secure in the cultures the pre- 

 dominance of that diatom that he desires to obtain in a state of 

 purity. In point of fact, if this diatom is found in the primitive 

 mixture in the proportion of four hundred to five hundred ; in five 

 cases of fecundation, four will be determined by the species that 

 he desires to insulate, and then in place of twenty macerations it 

 will sutifice to put a dozen to work. 



The common risk in this method is the fear of not pushing the 

 dilution to a degree sufificiently advanced ; and on the other side 

 it is hardly necessary to add that if the dilution be too great the 

 results will be negative, whence the need for the prelimi7ia7-y 

 experiment. 



