418 E. J. ALLEN. 



natural sea-water from the Laboratory tanks better cultures result than 

 have so far been got in any medium which has natural instead of artificial 

 sea-water as a basis. 



THE DIATOM CULTURE USED. 



A culture of the diatom Thalassiosim gravida Cleve, isolated some 

 years ago,* w^hich has been kept since then by successive inoculations in 

 fresh culture medium, has been used almost entirely for these experi- 

 ments. This species is especially useful owing to the fact that in healthy 

 cultures the cells hang together in long chains, whereas when the culture 

 is unhealthy or becoming exhausted the chains break up. This is a most 

 useful guide when watching the progress of an experiment. 



The Purity of the Culture. — The culture contains no other diatom except 

 T. gravida and no other organisms except bacteria. It would of course be 

 preferable, if it were possible, to remove all the bacteria, so as to deal with 

 a perfectly pure culture of the diatom. Many attempts have been made 

 to attain this end, but so far without complete success, though it has been 

 possible to carry the process of purification so far that only one species 

 of bacterium capable of forming colonies on a peptone-agar platef was at 

 all abundant. The method adopted for purifying the culture was that of 

 differential poisoning, a suitable poison being added to a number of culture 

 flasks in a series of gradually diminishing strengths, in the hope that one 

 strength might be found which would kill the bacteria without kilUng the 

 diatom. 



A measure of success was obtained with Copper sulphate in this way. 

 In the most successful case a solution of the salt was added to 100 c.c. of 

 culture medium containing Thalassiosira gravida in such proportion that 



* Allen and Nelson, loc. cit., p. 460. [Q.J. M.S., p. 412.] The species was then 

 thought to be a variety of Thalassiosira decipiens. Subsequent examination by Mr. 

 Nelson has convinced him that it is really Th. gravida. The extreme delicacy of the 

 siliceous skeleton of these diatoms makes the determination of species founded chiefly on 

 valve structure very difficult. The species was formerly thought to be a variety of 

 Thalassiosira decipiens Grun. since tlie only markings that were observed were cliaracter- 

 istic of this species, altliough no markings at all could l)e resolved with the great majority 

 of valves. Examination of the present cultures by Mr. Nelson witli more perfect apparatus 

 has shown the typical 2'h. gravida deve valve structure to which species this form is now 

 referred. It is not unlikely that the older cultures were a mixture of Th. decipiens and 

 gravida from which the decipiens have died out. 



t It should be remembered that possibly the piesence of some bacteria in the cultures 

 is necessary for their success, tliough iMiquel (Le Dlatomiste, I, 1890-3, pp. 153-6) states 

 definitely tliat he obtained cultures of fresh water diatoms which were entirely free from 

 bacteria, and Richter {Ber. dcut. hot. Gesell, XXI, 1903 and later papers) also succeeded in 

 obtaining such bacteria-free cultures on solid culture media. 



