drawn with the camera. Of these 31 clones, 22 showed nine dyads or bivalents, which split into 9 + 9 at the first division (fig. 1); 3 clones had a total of 18 single chromosomes (or 9 bivalents), which divided at the first division into 8 + 10, or other unequal numbers, not commonly into 99; 5 clones were probably completely or nearly triploid, and irregular in their first division, resembling in this the triploid mulberries of Osawa, and like them showing a smaller number of chromosomes after the first division than the triple number (in these cases, 24 to 26 instead of 27); while one clone was regularly triploid, showing nine triads at the prophase and first metaphase, and a total of 27 chromosomes after the first division. This regularly triploid clone was obtained from Thorburn, New York, in 1920, under the name "Gladiator." It differed conspicuously from other Cannas with more than 18 single chromosomes, in its smaller flowers and lesser size, resembling in these the ordinary diploid Cannas. Thirtytwo pollen-mother-cells were drawn with the camera. In 18 of these pollen-mother-cells, the total number of chromosomes could be accurately counted, and was 27. In the 3 other cases where the total number could be counted, some chromosomes seemed to be missing, for the totals were 25, 25, and 24. (In 3 pollen-mother-cells showing the anaphase of the first division, or the metaphase of the second, only one group could be accurately counted, because of the slanting position of the other. In the 8 remaining prophase or metaphase figures, not all the trivalent chromosomes could be distinguised from the bivalents or univalents into which they had divided, or from which they were composed.)

Nine of the cells showed clearly how many chromosomes went to one pole and how many to the other after the first division.

(In the triploid Datura the chromosomes have been accurately counted in 64 pollen-mother-cells after the first division, and the distribution corresponds with a random one.)

Three of the early anaphase plates showed how the triads divided. Taking for "up" the upper side of the camera drawing, which was a random position, we have for the position of the two's with regard to the equator (fig. 2, Nos. 4 and 5):

In several prophases, one chromosome was disconnected from its corresponding dyad (fig. 2, Nos. 1 and 3). If this occurred more frequently it would lead to the condition found in the large-flowered Cannas, where triads, dyads, and monads are mingled at the metaphase; or to the situation of Osawa's mulberry clones, where triads are apparently less common; or finally to the triploid Oenothera lamarckiana, which, according to Geerts, shows only dyads and monads. In the triploid Canna, one complete count of both second divisions was made in a pollen-mother-cell. On one side of the cell-wall the numbers were 15 + 1 + 12, and on the other side, 13 + 13. The single chromosome seemed attached to the cellwall.

About half of the pollen-grains of this triploid Canna were nearly or quite empty; while the others were full or nearly full of cytoplasm, with one or more nuclei.

Summary.-(1) Most of the Cannas examined were diploid, showing nine dyads before the first division in the pollen-mother-cells, and these in most plants separated into 9 + 9.

(2) One of the triploid Cannas showed commonly nine triads, each of which separated into two and one on the spindle, in a random manner with regard to the two poles. (The same arrangement, though less easy to demonstrate, was also found in a triploid Datura.)

1 Belling, J., 1921, “On Counting Chromosomes in Pollen-Mother-Cells,” Am. Nat. 2 Blakeslee, A. F., J. Belling, and M. E. Farnham, 1920, "Chromosomal Duplication and Mendelian Phenomena in Datura Mutants," Science, 52, 388-390.

3 Geerts, J. M., 1911, "Cytologische Untersuchungen einiger Bastarde von Oenothera gigas," Ber. Deutsch. Bot. Ges., 29, 160-166.

* Müller, C., 1910, "Über karyokinetische Bilder in den Wurzelspitzen von Yucca," Pringsh. Jahrb. wiss. Bot., 47, 99–117.

Osawa, I., 1920, "Cytological and Experimental Studies in Morus, with Special Reference to Triploid Mutants," Bul. Imp. Sericult. Exp. Sta. Japan, 1, 317-369.

Rosenberg, O., 1918, "Chromosomenzahlen und Chromosomendimensionen in der Gattung Crepis," Arkiv Botanik, 15, No. 11, pp. 1-16.

7 Tahara, M., 1910, "Über die Kernteilung bei Morus," Bot. Mag. Tokyo, 24, 281–289. 8 Tischler, G., 1915, "Chromosomenzahl,-Form, und-Individualität im Pflanzenreiche," Prog. Rei. Bot., 5, 164–284.

• Wilson, E. B., 1910, "Studies on Chromosomes VI. A New Type of Chromosome combination in Metapodius," J. Exp. Zool., 9, 53–78.

THE DESTRUCTION OF PHOSPHORESCENT ZINC SULFIDES BY ULTRA VIOLET LIGHT

BY LEONARD B. LOEB AND LLOYD SCHMIEDESKAMP

RYERSON PHYSICAL LABORATORY, UNIVERSITY OF CHICAGO

Communicated by R. A. Millikan, June 12, 1921

It has long been known that the phosphorescence exhibited by certain alkaline earth sulfides is gradually destroyed when they are subjected to the action of alpha particles from radio-active matter. It has also been found that a permanent destruction of the phosphorescent properties of such sulfides is caused by the bombardment of the sulfides with canal3 rays or with cathode2 and beta1 rays. In this process of destruction, the sulfides are caused to phosphoresce brilliantly by the destroying agents, and the destruction is accompanied by a change in color of the sulfide, generally a darkening.

In 1919 J. Perrin1 and more recently R. W. Woods have shown that in the process of fluorescence, a process assumed to be somewhat similar to phosphorescence, the fluorescing molecules undergo chemical changes which result in the loss of the fluorescent power. In other words light of a given wave-length falling upon a molecule of fluorescent substance causes it to emit light of a different wave-length, and in this process the molecule is chemically changed so that it can no longer fluoresce.

The detection of the destruction of the fluorescent properties of certain substances is complicated by the fact that the fraction of molecules that are destroyed per second is relatively so minute that the time of exposure required to produce a measurable effect must be very long. As Perrin showed, this time may be much reduced by using very thin films of the solutions of fluorescing molecules. Wood has shown that it may also be much reduced by using intense sources of light. In view of the destruction of the phosphorescence of zinc sulfides by other agents which cause them to phosphoresce, and in view of the powerful technique developed by Perrin and Wood for the destruction of the fluorescent property by light, it occurred to one of the writers to attempt to detect a destruction of the phosphorescent property of the zinc sulfides through the action of light. The only reference to observations on such a destruction yielded by a survey of the literature consisted of the following quotation taken from a long paper by Baerwald3 on the effect of canal rays on phosphorescence. Baerwald states "It is remarkable that intense ultra violet light from a mercury arc also destroys the phosphorescence of zinc sulfide. As it becomes dark in the process we may suspect a chemical change of the sulfide into an allotropic form,....

For the preliminary experiments which are being reported here, three

different samples of phosphorescent zinc sulfides were used. One was a sample obtained from the Radium Standard Chemical Company in 1918, the others were two samples obtained in France from MM. Guntz and Ziegler of the powder works at Bouchet in 1918. The specimens to be studied were prepared by sifting the fine crystals upon small circular plane glass plates, 7 mm. in diameter which had been moistened with a very dilute solution of collodion in amyl acetate (about 3 drops of an ether collodion solution of the consistency of "New Skin" in 100 cc. of amyl acetate). Microscopic examination of the specimens showed that they were covered with a uniform layer of the sulfide occupying about 99% of the surface. The glass plate was fastened to a plate of thin transparent quartz, with the layer of sulfide pressed firmly against the quartz. The back side of the glass plate was then painted with optical black to permit the photometry of the sample. At the same time that the specimen was mounted on the quartz plate a control specimen as nearly the same as possible was similarly mounted on a glass microscope slide.

The quartz-mounted specimen was then placed in the image of a quartz Hareus mercury arc thrown upon the wall by a quartz lens of 3 cm. aperture and 15 cm. focal length. The arc operated continuously at about 2.5 amperes in a light-tight box. A blackened thermometer bulb placed in the image of the arc showed a rise in temperature of but 0.5° C. above that of the room, so that the heating of the specimen produced by the arc was not of importance. The glass-mounted control sample was kept in a light-tight box and was removed from this only at the time of photometric measurement.

The phosphorescent intensities of the quartz-mounted specimen and the control were compared by photometric measurements on the phosphorescence exhibited by them when studied in a simple phosphoroscope. The specimens were mounted on the window of a small photometer with the blackened side towards the window. This window was diffusely illuminated from behind by a small automobile headlight lamp. The intensity of illumination of the window was varied, and hence the photometry was accomplished, by changing the distance of the lamp from the window until the phosphorescent specimen faded against the window which served as its background. A color match between the lamp and the specimen was obtained by the use of absorbing screens of colored glass. In the photometry the specimens were excited to phosphorescence by the light from a 75-watt lamp focussed on them. The exciting light was put on and cut off in the usual way between the intervals of observation by means of slits in a metal disc rotated at a high speed, the procedure being in all respects the standard procedure used in phosphorescent measurements. After the quartz sample had been exposed to ultra violet light for a suitable interval of time, its intensity was compared with the

control as above. The results obtained on three samples are tabulated below.

In the first column are given the intensities of the specimens compared to the controls corresponding to the number of hours exposure to the ultra violet light recorded in the second column. In the third column are given the dates on which the measurements were made; while the fourth column is reserved for comments. In every case it may be seen that there is a marked reduction in the phosphorescent intensity with