ROOT-NODULES OF LEGUMINOSAE. 125 



unknown substances together cause plasmolysis and death of the bacteria when they reach a certain 

 concentration and this prior to the extrusion of the pectin-like substance. 



The most important difference between the ordinary nodule bacteria and the bacteroids is not 

 in form, since there are many unbranched bacteroids, but in their plasma contents which is vacuolate 

 and otherwise unlike that of the ordinary rods. The real difference, therefore, is in the processes 

 through which the plasma passes. This phenomenon was observed in solutions containing i per cent 

 grape-sugar. The bacteria from soy-beans gave striking results. Almost without exception the 

 slender, feebly staining bacteroids showed in one week lateral, generally spherical but sometimes broad- 

 based, projections consisting of strongly staining plasma. Occasionally these projections occurred at 

 both ends. The projection or outgrowth was strongly differentiated from the rod by tincture of 

 iodine : The rod stained bright yellow, the projecting plasma red brown. In living uncolored prepa- 

 rations these projections were highly refractive bodies. They were not destroyed by dilute potash 

 solution or dilute sulphuric acid, nor influenced in their ability to stain with carbol fuchsin. Staining 

 alive in neutral red and eosin gave unsatisfactory results. With iodine potassium-iodide the projec- 

 tions stained mostly yellow while the rodlets remained unstained. To this projecting plasma Hiltner 

 gives the name nuclear plasma in contrast to the part remaining within the rod which stains weakly 

 with carbol fuchsin and which he calls cell-plasma or nutrient plasma. Similar observations were 

 made on lupin bacteria. 



After 2 weeks' growth the appearance of the soja bacteria was yet more striking. The two parts 

 were still present and stained as before. From all appearances these bacteroids, except for a scarcely 

 visible layer lining the walls, consisted of the same albuminous substances as is found collected in the 

 vacuoles of the normal bacteroids. 



After 49 days growth the nuclear plasma of the bacteroids from the bacteria of Trifolium incar- 

 natum had broken up into normal bacteria. The same thing had occurred with those from Trifolium 

 hybridum and T. pratense, and new bacteroid formation had begun. This process held true for all 

 the bacteria of the Vicia group. 



The bacteria of Phaseolus vulgaris showed, especially clearly after 3 weeks' growth, the breaking 

 up of the much enlarged, strongly staining bodies into normal bacteria. 

 The following conclusions are drawn from these observations : 



In i per cent grape-sugar solution bacteroid formation is rapid. They are distinguished from 

 the normal rods by larger size and by a sharp differentiation in the plasma; one part, the germ-plasm 

 or nuclear plasm, either pushing out as a spherical projection or gathering in definite spots within 

 the bacteroid. Two groups are distinguishable, those which retain the form of rodlets when enlarging 

 into bacteroids and on which the projections appear at the ends, and those which enlarge chiefly in 

 breadth becoming spherical or pear shaped, and almost completely filled with the strong staining 

 nuclear plasm : The rapidity of this process varies with the various nodule bacteria. 



Cultures in solutions containing grape-sugar in concentrations of from o.or per cent to 5 per 

 cent showed that, in general, growth is the stronger as the solution is weaker. The 2 per cent and 

 5 per cent solutions remained almost clear 4 days after inoculation. A concentration lying between 

 o. i per cent and i per cent causes the most rapid bacteroid formation and plasma differentiation. 



In saltpeter solutions the nodule bacteria from the pea made most vigorous growth in the stronger 

 concentrations (i and 2 per cent) but bacteroid formation was greater in the weakest solutions, i. e., 

 o. i per cent and 0.05 per cent. In these solutions after 3 days' growth a few isolated rods remained 

 almost unaltered, while the bacteroids which were branched, and 10 times as long as broad, were 

 collected in groups of a hundred or more. They stained uniformly, and while vacuoles were present 

 they were not so highly differentiated as in the grape-sugar solutions. The soy-bean organism grew 

 badly in saltpeter solutions and otherwise differently from the pea bacteria but with the formation 

 of bacteroids. These were also obtained in asparagin water. 



In o.oi per cent to 5 per cent peptone solutions, the bacteria from soy-bean and Vicia saliva 

 were totally unchanged after 4 weeks' growth, showing that this substance does not induce bacteroid 

 formation. Growth occurred in the 5 per cent solution but was best in the i per cent solution. Soy- 

 bean bacteria which grew poorly in saltpeter solutions did better than the vetch bacteria in peptone 

 solutions. 



In i per cent grape-sugar solutions which contained also 0.5 per cent to 2 per cent saltpeter the 

 bacteroids of the pea showed a slight differentiation in their contents. Small, strongly staining 

 granules appeared which at times seemed to lie on the exterior of the bacteroid. In general, also 

 in solutions containing grape-sugar, the saltpeter caused very large, branching bacteroids with a net- 

 like plasma. As the saltpeter in the weaker solutions was used up by the organism the effect of the 

 grape-sugar became evident in the pushing out of the plasma masses which stained red brown with 

 iodine and which were much larger than in pure grape-sugar solutions. 



A solution containing o. i per cent asparagin showed astonishing results with the vetch bacteria, 



