DECEMBER 31, 1897.] 
the original transparent colonies grown in 
the incubator. The final aspect of the cul- 
ture is compared to a miniature archipelago 
in which the islets would be represented by 
the colonies first developed, and the surface 
of the water by the layer subsequently 
formed at the lower temperature. 
Blood-serum and potato proved them- - 
selves to be rather unsuitable media for B. 
acteroides. In milk the germ grows readily, 
but without producing coagulation. The 
most favorable fluid medium tested by San- 
arelli was beef broth containing lactose and 
calcium carbonate. 
The germ is pathogenic for most of the 
domestic animals. Mice, guinea-pigs and 
rabbits succumb readily to inoculation with 
a pure culture. The rabbit is considered as 
the most favorable subject for experimental 
inoculation, and possesses notable advan- 
tages over the guinea-pig both in suscepti- 
bility and in the regularity and constancy 
of symptoms and death. The dog, how- 
ever, presents the most instructive instance 
of close analogy with the disease yellow 
fever as it manifests itself in man. Both 
in the symptoms and in the anatomical 
lesions Sanarelli was able to trace a cor- 
respondence at once constant and precise. 
As is the case in man, the liver and the 
kidneys are the organs especially attacked ; 
secondary infections with the streptococcus 
and colon bacillus sometimes occur. 
In the second memoir Sanarelli details 
the results of his experiments with the yel- 
low fever toxin. Cultures of the germ 15-20 
days old, made in ordinary peptonized meat 
broth and filtered through a Pasteur-Cham- 
berland tube, afforded him a potent toxin. 
The toxin thus prepared, when injected 
into the bodies of susceptible animals, pro- 
duced substantially the same symptoms as 
inoculation with the specific bacillus. In the 
dog, particularly, inoculation with the germ- 
free toxin set in motion the same train of 
specific symptoms and caused the same 
SCIENCE. 
983 
pathological changes in the tissues. ‘‘ The 
toxin of yellow fever is an exceedingly 
powerful cellular poison comparable solely, 
in some points, to the diphtheria toxin. Its 
contact with the tissue elements of the ani- 
mal organism, especially the higher species, 
determines, like that of the diphtheria toxin, 
a violent irritation, followed by retrogressive 
processes which always end in the necrosis 
and fatty degeneration of the protoplasm.” 
Some very interesting experiments bear- 
ing on the question of mixed infection are 
next described. When B. icteroides in sown 
upon culture media on which, respectively, 
the colon bacillus, streptococcus and proteus 
have been previously grown, it is found 
that the growth of the former is distinctly 
inhibited by the presence of the soluble 
products of the other microbes. The latter, 
on the contrary, grow excellently in a me- 
dium previously inhabited by B. icteroides, 
and are only slightly incommoded by the 
presence of the soluble products of one an- 
other, the products of the proteus bacillus 
seeming mostinjurious to all concerned. A 
similar result was revealed by a study of 
the ‘vital antagonism’ of B. icteroides and the 
microbes concerned in the secondary infec- 
tions. Both streptococcus pyog. and staphylococ- 
cus pyog. au. speedily gain the upper hand 
over B. icteroides, and a similar, though less 
marked superiority, is manifested by the 
colon bacillus. These facts certainly shed 
much light on the difficulty of demonstra= 
ting the presence of the yellow fever germ 
in the bodies of victims of the disease, and 
go far to explain the negative result reached 
by many observers. 
In an attempt to account for the impor- 
tant part played by maritime commerce in 
the diffusion of yellow fever Sanarelli re- 
cords a curiously significant observation. It 
was noticed that gelatin plates sown with 
B. icteroides sometimes remained without de- 
velopment, although agar plates sown at the 
same time evinced abundant growth. But 
