On Certain Bacteria from the Air of New York City. 331 



the resulting cultures tend to exhibit the same characters as the 

 parent culture. When, however, cultures are made from separate 

 colonies from a plate, the individual A'ariations tend to become 

 much more apparent, as the above shows. 



To proceed with the experiments. Having obtained cultures 

 derived from a single one but with slightl}' different characters, j 

 proceeded to test these further, with the application of selection, 

 as follows : 



(1) No. 6 was selected as showing the least power of liquefying 

 gelatin and a series of plates were made from it b^' dilution. Of 

 the resulting colonies, nine w^ere planted on gelatin and of these 

 nine I selected the one which liquefied gelatin most rapidly and 

 the one which liquefied the least so (for there was a considerable 

 difference, as in the first instance) and planted them side by side 

 with the original No. 6. Here is the result in tabular form as 

 before : 



It shows that instead of holding true to the character acquired 

 by No. 6 of a slower rate of liquefaction, there was a marked ten- 

 denc}' to return to the original quick liquefaction, which could 

 probabh' alone be overcome \iy a longcourse of selection. Notice 

 also the difference in the rate of growth as shown by the fact that 

 the No. 6 finally catches up with the new No. 8, perhaps owing to 

 the approaching exhaustion of the medium. 



(2) No. 10 was selected as showing the best effect of reduction 

 of nitrate, being much better than the original culture. It was 

 hoped, by selection, to produce a form which would reduce nitrate 

 as well as normal B. lactis erythrogenes. Notice that there was at 

 first a marked tendencj' to approach the normal tj'pe of B. lactis 

 erythrogenes. The original culture corresponded in its reducing 

 effect to No. 8 above, whereas eight out of nine of its progeny 

 reduced nitrate better than it did. It was thouo;ht that this 

 function might be easil}- further increased. 



Ten colonies were tubed from gelatin plates made from No. 10, 

 and planted in nitrate solution at the same time. Of these it was 



