62 BOTANICAL GAZETTE [JANUARY 
and they enter the embryo sac through the micropyle. In seeds they occur 
constantly between the embryo and the endosperm. They are seen in young 
seedlings and in all later stages. 
Careful study was made of the bacteria, both in the host plants and in 
artificial cultures. They were found to bear a close resemblance to the tubercle 
bacillus, both in structure and in behavior. Consequently von FABER regards 
these organisms as members of the Mycobacteria, and he gives them the name 
Mycobacterium Rubiacearum, nov. sp. Of large interest also was the dis- 
covery that nitrogen is fixed in the artificial cultures of these organisms. 
Von FABER succeeded also in getting pure cultures of the Rubiaceae, the 
seeds being sterilized by placing them for 25 minutes in hot water at a tempera- 
ture of 50°. The pure cultures grew far less vigorously than biotic cultures, 
and the leaves were lighter in color. It was observed also that nitrogen is 
fixed in the symbiotic cultures but not in the pure cultures, so that in nature 
these Rubiaceae can get their nitrogen supply directly from the air. Still 
further recalling the nutritive relations between the Leguminosae and their 
root bacteria is the fact that the bacterial tissue of the leaves of the Rubiaceae 
is rich in starch, which may serve the bacteria for food; there is evidence also of 
bacterial decadence, involution forms, and eventual phagocytosis. Pure 
cultures seem to indicate that each host species has its own bacterial “‘adapta- 
tion form.” 
MieHE has also followed brief preliminary reports ® of his studies of 
hereditary symbiosis by detailed accounts." * Ardisia crispa, one of the 
Javanese Myrsinaceae, has glandular thickenings on the leaf margins. These 
represent modified hydathodes and have commonly been regarded as albumen 
ds. It is now shown that these structures resemble ZmmMERMANN’S 
bacterial knots and are caused by bacteria. Furthermore the bacteria are 
present throughout the life history of Ardisia, and the details of bacterial 
entrance and subsequent behavior are astonishingly like those reported simul- 
taneously in the Rubiaceae by von FaBer. The microorganisms enter the 
leaves through stomata, which later become closed. They were observed by 
MteHE in the embryo sac, in the seed, and in the vegetation point of the 
seedling. Two species of bacteria have been isolated and are called Bacillus 
foliicola and B. repens. In old cultures there occur curved and branched 
involution forms. In most respects these organisms resemble vON FaBER’S 
9 Miexe, H., Die sogenannten etree! an mie Blattern von Ardisia crispa 
A.DC. Ber. ‘Dench. Bot. Gesells. 29: 156-157. 
= , Uber Symbiose von Bakterien mit eases Biol. Centralbl. 32:46-59- 
IQI2. 
™———. Die Bacterienknoten an den Blattrindern der Ardisia crispa. In 
Javanische Stantien: Abhandl. Kénigl. Sachs. Gesells. Wiss. 32:398-431. tort. 
uu ————, Weitere Untersuchungen iiber die Bakteriensymbiose bei Ardisia crispa 
I. Die Mikroorganismen. Jahrb. Wiss. Bot. 53:1-54. pls. 2. 1913. 
