The sera of Nos. 5 and 55 were tested with spores of two other molds, Absidia glauca and a species of the Mucor racemosus type. There seemed to be a slight agglutination with both species. It is possible that the rabbits tested had received so many spore injections that their sera had largely lost their specificity and so were reactive to spores of any species of this mold group. The results do not seem to warrant many definite conclusions in regard to the agglutination reaction but suggest the need of further investigation.

The results outlined in the preceding pages show that repeated intravenous injections into rabbits of fungus spores fail to cause the development in their blood of cytolytic substances capable of dissolving such spores, but apparently stimulate the formation of antibodies that cause them to agglutinate.

LITERATURE CITED

I. BLAKESLEE and GoRTNER. On the occurrence of a toxin in juice expressed from the bread mold-Rhizopus nigricans. BIOCHEMICAL BULLETIN, 1913, ii, p. 542.

2. GORTNER and BLAKESLEE. Observations on the toxin of Rhizopus nigricans. American Journal of Physiology, 1914, xxxiv, P. 353.

3. BLAKESLEE. A possible means of identifying the sex of (+) and (-) races in the mucors. Science, 1913 (n. s.), xxxvii, p. 880.

4. BLAKESLEE. Lindner's roll-tube method of separation cultures. Phytopathology, 1915, v, p. 68.

5. BLAKESLEE and GoRTNER. See C. B. Davenport, "The Department of Experimental Evolution," Year Book of the Carnegie Institution of Washington, 1913, xii, p. 99.

STUDIES ON THE PHYSICO-CHEMICAL PROPERTIES OF VEGETABLE SAPS

3. A comparison of the physico-chemical constants of the juices expressed from the wall with those from the included carpellary whorl in proliferous fruits of Passiflora gracilis

J. ARTHUR HARRIS AND ROSS AIKEN GORTNER,
WITH THE COLLABORATION OF JOHN V. LAWRENCE

(Carnegie Institution of Washington, Station for Experimental Evolution, Cold Spring Harbor, N. Y.)

(WITH PLATE 3)

1. Introduction. In this paper we present a limited portion of the data which we have gathered concerning the physico-chemical properties of the expressed juices of morphologically differentiated fruits of Passiflora gracilis.

Our purpose in carrying out this work has been to ascertain whether there may not be chemical differences underlying, or at least associated with, morphological differences in plant organs. The fruits of Passiflora gracilis furnish excellent material for such a study. One of us (Harris, 1906) has described some of the more common abnormalities in these fruits. Later (Gortner and Harris, 1913) we attempted on a small scale to ascertain whether there is any consistent chemical difference between normal and teratological fruits. The results of this investigation led to work with improved methods of analysis during the summer of 1913, when some 800 individual samples (400 teratological and an equal number of normal controls) of fruits were examined. All of these data have not as yet been assembled; results from another season's cultures will be secured before conclusions are drawn concerning certain of the relationships.

This paper is limited to a comparison of the properties of the juices expressed from the wall of the abnormal fruit with those from the mass of abnormal tissue which it contains.

The proliferous fruit of P. gracilis is characterized in the simplest cases by the production, from the base of the fruit, of a stalked whorl or series of whorls of incompletely closed carpels. In such cases the prolification must be regarded as morphologically a continuation of the axis, whose activity generally closes with the formation of the carpels constituting the normal ovary. The included or abnormal mass, as we shall sometimes designate this accessory carpellary whorl, is generally large, green and turgid, but it may be small and shrivelled. It may become so large as to rupture the normal ovary wall. The variation in structural details is very great and will form the subject of a forthcoming memoir by one of us.

The fruits which have furnished the materials for this paper have been selected with great care as to morphological type. They comprise exclusively those which are trimerous (the normal condition, i. e., with six external sutures and three placentae) in the organization of their wall, and have a large living basal prolification with an external whorl of four carpels. The diagrams (Plate 3) make clear the type of fruit dealt with. Numerous detail drawings will be published when the morphological problems are taken up.

2. Material. The materials upon which the constants were determined were drawn from five cultures of P. gracilis grown under slightly varying conditions on the grounds of the Station for Experimental Evolution in the summer of 1913. All the plants were typical and (relatively) normally grown. These five series do not include all of the fruits taken into consideration in our work, but since the other experiments were made under highly abnormal conditions of growth for special purposes, it seems quite legitimate to leave them out of account here.

Three of the cultures represent a strain which has been under cultivation at the Station since the summer of 1908, when the seed was secured from an American dealer. The second lot of seed was a commercial sample bought from Haage und Schmidt, of Erfurt. The series considered2 are:

CULTURE A'. American seeds. The young plants were transferred to large pots of garden soil, which were plunged in the garden early in May. Collections were made from 24 plants.

CULTURE B. This was carried out in precisely the same manner as Culture A, with the exception that 175 gm. of bone meal was thoroughly mixed with the soil of each pot. Collections were made from 25 plants.

CULTURE C. This and the following culture was made from the Haage und Schmidt seed, which was germinated considerably later than the lot that gave rise to the preceding cultures. The young plants were transferred directly from 8 cm. pots into the garden. They were divided into two lots, the first of which (Culture D) was placed in a garden somewhat separated from that in which Cultures A to C were grown. Altogether there were 43 plants.

CULTURE D. The plants (30) were identical with those of Culture C but grown in the garden separate from that in which Cultures A-C were grown. (The gardens were about 200 yards

apart.)

CULTURE E. A number (53) of plants from the same sowing as experiments A and B were potted up in 15 cm. pots, in soil which was considerably better than that in the garden. After a few days in the greenhouse to allow the plants to become established, these pots were plunged in the garden adjoining Cultures A and B.*

The distribution of the collections from which the abnormal fruits were drawn is shown in Table 1.

The entries in this table show that in exp. C and D the materials were gathered from well matured vines late in the fall, whereas in the other three series the collections were made at intervals throughout the season. Thus, the entire materials of Culture C were taken for analysis from Sept. 29 to Oct. 3, inclusive. All the