VI] GENERAL CHARACTERS OF BACTERIA loi 



element is noticed in a chain which is much larger than the 

 rest, and this larger element divides into two and four cocci 

 successively. So also those large elements described above 

 as occurring in the chains of streptococci show the suc- 

 cessive fission into two and four cocci. And it is this which 

 prompts him to say that these large elements found oc- 

 casionally in the chains or in the diplococci are not in- 

 volution forms, but are active elements which before 

 successively dividing grow up to large size. In staphylo- 

 coccus aureus liquescens, growing in gelatine, he has also 

 observed some of these large elements, though on the 

 whole they are not so numerous as in the streptococcus 

 growing in the same kind of medium. The normal mode 

 of division of a coccus is then (i) a slight enlargement 

 and division into two by transverse fissure, or (2) a coccus 

 enlarged to considerable size (four to six and more times) 

 and then successively divided into two and four and further 

 eight cocci of the normal size. 



' As regards bacilli all observations hitherto recorded agree 

 that a rod before dividing elongates sometimes more some- 

 times less, and then a transverse indentation appears about 

 midway, which ultimately becomes a fissure by which the 

 originally single rod divides into two ; according as the rod 

 was short or long, the resulting offsprings are more coccus- 

 like or more cyHndrical. Now, in the observations which 

 the writer has carried out as to the time of the division of 

 the different microbes mentioned above the writer has 

 repeatedly noticed that a single cylindrical bacillus not 

 infrequently divides almost simultaneously into three and 

 even four short rods. The writer has observed cylindrical 

 bacilli in preparations of bacillus anthracis, made directly 

 from the blood of guinea-pigs, which were uniform, and 

 there was no indication in the fresh specimen that they 



