406 REPORT OF THE HARVARD AFRICAN EXPEDITION 
this last hypothesis. Cowdry ! in his study of the microchemistry of the nuclear 
and cellular inclusions in virus diseases, points out that these are occasionally 
feebly basophilic, but that the majority are strongly acidophilic (or oxyphilic). 
We found nothing to suggest the intracellular bodies of Holmes and other 
authors, but, on the other hand, we did find, on several occasions, a fungus in- 
fection in the latex. 
It must be admitted that the spores of fungi which are present in the tissues 
of several of the plants affected with the African mandioca disease, are some- 
what similar in appearance to the inclusion bodies of Hippeastrum mosaic, of 
wheat rosette, and of the Fiji disease, as illustrated by Kunkel? in his recent 
article on the Virus Diseases of Plants. They resemble more the inclusion bodies 
of Hippeastrum mosaic, since an appearance somewhat resembling budding 
may be observed in Kunkel’s illustration. However, in the sections of the 
African mandioca disease they can be definitely identified as elements of a 
fungus, which is not the case in Kunkel’s sections. 
While it seems probable that this extensive fungus infection of the latex of 
two of the mandioca plants examined must play some role in the pathological 
condition encountered, it is also realized that since the fungus infection was not 
demonstrated in the majority of the diseased plants, it probably is merely a 
secondary invader which has been imposed upon a primary virus infection. 
Obviously, the infection in Liberia should be carefully studied and inocu- 
lation experiments made, both with filtrates of the latex and with the fungus 
isolated. Such studies were impracticable for us to carry out on the present 
Expedition, especially from the standpoint of the time they would have con- 
sumed. Our object is particularly to call attention to the wide prevalence of the 
infection in Liberia and to the necessity for further studies regarding its etiology. 
No observations suggesting insect transmission of this mandioca disease were 
made by us in Africa. Aphis maidis has been shown to transmit other forms of 
mosaic disease in various parts of the world. It, however, was not seen on any 
of the mandioca plants in Africa. In fact, the only arthropod observed on a few 
examples of diseased mandioca in Africa was a species of mite collected by Be- 
quaert. Kenneth Smith * has pointed out that in the mosaic group of diseases 
it is fairly obvious that the insect is nothing more than a mechanical carrier, but 
that in certain other plant diseases the virus is spread only by one particular 
insect. . Steiner * has reported experiments which at least suggest that the 
root-knot nematode Caconema radicicola Cobb ( = Heterodera radicicola) is able 
to transmit tomato mosaic. Whether the transmission of the mandioca infection 
in Africa may occur (as elsewhere in other definite forms of mosaic disease) both 
through cuttings and through insect transmission will require further investiga- 
tion. 
We did not observe in Liberia, the disease of mandioca caused by Cecido- 
myiidae that we met with in Amazonia, particularly along the banks of the 
1 Cowdry: Science (1928), LXVIIT, 40. 
* Kunkel: Filterable Viruses, edited by Thos. Rivers (1928), p. 364. 
’ Kenneth Smith: Brit. Med. Jour. (1929), Part I, p. 448. 
4 Steiner: Loc. cit. 
