structure had ever been found in the entire order of copepods, and hence Dana must have been mistaken in what he thought he saw. Steenstrup and Lütken described and figured a similar structure in the maxillipeds of their new genus Perissopus (Kongelige Danske Vidensk. Selskabs Skrifter, ser. 5, vol. 5, 1861, pl. 12, fig. 25), and there is every reason for believing the structure in both genera to be genuine.

Absolutely hypothetical reasoning like that quoted above can have but little influence, and it certainly does not possess sufficient merit to prove or disprove the validity of any genus.

DEPARTMENT OF EDUCATION, WESTFIELD, MASS.

CHAS. B. WILSON

NEW DUST TREATMENTS FOR OATS SMUTS SINCE the introduction of copper carbonate for wheat bunt control (Darnell-Smith, and Ross, 1919)1 considerable interest has been shown in dust treatments for grain smuts. It was found by one of us (Thomas) in field tests in 1924 that copper carbonate alone was not effective in controlling oats smuts. However, when one part of either copper carbonate or copper sulfate was mixed with two parts of mercuric bichloride the dust was effective. These mixtures are too expensive for general use even though rapid and easy of application. Other tests showed that the mixture was less effective when inert fillers were added. In 1926 a mixture of one part of copper sulfate, one part of mercuric bichloride and one part of cresylic acid was found to control oats smuts. While the cost of this dust was only about half that of the copper sulfate-mercuric bichloride dust, yet it is also too expensive for general use.

None of these dusts, although they gave satisfactory control of oats smuts, was as cheap as the liquid formaldehyde. This liquid treatment is objectionable because of the difficulty in handling the wet grain and the possibility of seed injury. Since formaldehyde is so effective against smut, and the wet methods of grain treatment are objectionable, an attempt was made to put formaldehyde in a dust form. This was done by mixing 40 per cent. formaldehyde with either infusorial earth or charcoal. These dusts stick well and thoroughly coat the grains when mixed with them. In these tests dusts containing 9 per cent., 15 per cent. and 25 per cent. of 40 per cent. formaldehyde were used, each at the rate of 3 ounces per bushel

1 Darnell-Smith, G. P. and Ross, H. A dry method of treating seed wheat for bunt. Agr. Gaz. N. So. Wales 30: 685-692, 1919.

2 Thomas, Roy C. Dust treatment for smut in oats. SCIENCE, No. 1567, Vol. LXI: 47-48. January 9, 1925.

of grain. While the checks showed 47 per cent. smut the various formaldehyde dusts reduced smut to less than one per cent.

Another new treatment, iodine vapor dust, was tried in these same experiments. This dust was made by mixing finely ground solid iodine with infusorial earth. The iodine vaporizes readily at ordinary temperatures and diffuses through the infusorial earth giving it a light yellow-ochre color. This dust contained 5 per cent. by weight of iodine and was applied at the same rate as the formaldehyde dust. Only three smutted heads were found in three one-hundredth acre plots which were treated with this dust. It is possible that lower concentrations of iodine dust will also control the oats smuts. Further tests are under way. The cost of treating grain with these dusts is estimated at considerably less than 5 cents a bushel.

J. D. SAYRE R. C. THOMAS

OHIO AGRICULTURAL EXPERIMENT STATION, WOOSTER, OHIO

DO CATS SHARPEN THEIR CLAWS? LAST winter the family cat (castrated male, 3% years old) shed a number of claws in the house. These were found during January and February, some of them split lengthwise, the others intact. It struck the writer that the shedding of claws is probably a normal phenomenon with cats comparable to related phenomena, as that of the shedding of horns by deer. If this were true, it might be expected that some of the claws would be left in the bark of those trees which the cat used regularly for scratching. Upon investigation in April this bit of evidence was found in the form of two halves of a claw stuck into the bark of an elm and several halves lying under different trees used by the animal. The seetion of the bark was cut from the tree and with the pieces of claws has been mounted and placed in the college zoological museum.

This is but an isolated observation. There are good grounds, however, for believing the conjectured explanation to be correct. Cats do not instinctively or from experience select good grinding surfaces, slightly rough and hard, such as a cement walk, the foundation stone or the corner boards of a house, or smooth hard posts. They make use of the rough bark of trees which is always much softer than their claws. Observations of their scratching movements show that the animals do not scrape downward over the surface of the object, but catch the claws into the surface and with a circular stroke pull first downward and then outward and slightly upward. Careful examination of the cat's paws each time when a

claw was found failed to reveal any sign of injury. It was impossible to identify the toe from which the claw had dropped. This strikes the writer as fair proof that the shedding of claws is a normal phenomenon. The claws of the rear feet are possibly lost as they become loosened, or they may be pulled out by the animal with his teeth. Cats are frequently seen to pull at their hind claws in a manner suggesting this.

The shedding of claws is most likely seasonal, as are the related phenomena in other animals. Why then should the cat carry on the scratching movements throughout the year? It is possible that a further function of the scratching may be that of keeping the claws from curving too much, consequently growing into and irritating the paw. The irritation caused by claws which are curved too much or by the itching or other annoyance of loose claws may be the stimulus that starts the scratching movements. In this connection a colleague, a zoologist, has called attention to a reaction of badgers. These animals frequently drop out of an intense fight, roll over on their backs and scrape the claws of their front paws by rapidly drawing the paws across each other, pads facing. In accounting for the continuation of the scratching activity throughout the year, however, the likelihood of this being a habit reaction must not be overlooked.

DEPARTMENT OF PSYCHOLOGY, BELOIT COLLEGE

OLE N. DE WEERDT

RECENT PUBLICATIONS OF THE NATIONAL RESEARCH COUNCIL

Two recent publications in the National Research Council's Bulletin Series should be of rather wide interest among scientific men. One (Bulletin 58) is entitled "Handbook of Scientific and Technical Societies and Institutions of the United States and Canada." The American section of this bulletin was compiled by Clarence J. West and Callie Hull, and the Canadian section by the National Research Council of Canada. The other (Bulletin 60) is entitled "Industrial Research Laboratories of the United States, including Consulting Research Laboratories, Third Edition." This bulletin was compiled by Clarence J. West and Ervye L. Risher. Both bulletins are the output of the National Research Council's Research Information Service, of which Dr. West is director.

The purpose of publication of the handbook is to present a ready guide to those scientific and technical societies, associations and institutions of the United States and Canada which contribute to scientific knowledge or further research through their activities,

publications or funds. Only those government institutions are included which administer private funds. Organizations directly controlled by universities or colleges have been omitted because it is expected that they will be covered by the forthcoming publication, "American Universities and Colleges," to be issued by the American Council on Education. Seven hundred and nine American organizations and seventy-four Canadian organizations are listed in the bulletin. The address of the secretary, the date of organization, the major object of the institution, the character of membership and amount of dues, time of meetings and information concerning publications are given for each institution.

The bulletin on Industrial Research Laboratories lists 999 such laboratories in the country, giving for each laboratory the name and address of the supporting industrial or commercial concern, the makeup of the research staff, and a list of special subjects to which the research activities of the laboratory are devoted. The first edition of this bulletin was published in 1920 and listed about 300 laboratories; a second edition (first revised edition) was issued in 1921 and listed about 600 laboratories. The present edition (1927) is the second revision of the bulletin.

The difficulties of compilation in connection with both of these publications make it inevitable that some errors, both of commission and omission, have been made by the compilers. The director of Research Information Service (National Research Council, Washington, D. C.) will be glad to have his attention called to any such errors noted by any who may have occasion to examine the bulletins.

NATIONAL RESEARCH COUNCIL, WASHINGTON, D. C.

VERNON KELLOGG

SCIENTIFIC APPARATUS AND

LABORATORY METHODS

PREPARATIONS OF STAINED DECALCIFIED BONE WHICH RIVAL GROUND SECTIONS

GROUND Sections of bone, besides being difficult to prepare, are often unsatisfactory for student use either on account of their thickness or due to the fact that they have been mounted in thin xylolbalsam, resulting in the displacement of the air from the lacunar and canalicular spaces of the tissue. It is, however, possible to prepare decalcified bone in such a way that all the advantages of canalicular detail are obtained. Two methods by Schmorl,1 the picro-thionin and the thionin-phosphotungstic acid

1 1909. Schmorl, G. "Die pathologisch-histologischen Untersuchungenmethoden." Vogel, Leipzig.

methods, give excellent results and the detail demonstrated surpasses that observed in ground sections. With the exception of a few departments of dental histology, neither of these methods is in general use in American laboratories. I have been unable to find Schmorl's original description of his methods but they are repeated in a more recent work of 1909. An excellent discussion of the methods is also found in a paper on the structure of bone of Fasoli2 and adequate directions for the successful use of these methods are given by Carleton3 in his recent book on histological technique. References to Schmorl's methods may also be found in the works of Lange and Fischer. It seems unnecessary to completely outline the method since it can be readily obtained in English in a modern text-book on histological technique. Formol, Orth's, Müller's or Regaud's fluids may be used for fixing. Fluids containing mercuric chloride should be avoided. Best results are obtained with celloidin or frozen sections. If nuclear patterns are desired, the tissue should be first stained in alumcarmine or hemalum, as the success of the picrothionin method depends entirely on the precipitation of the thionin in the lacunae and canaliculi. The picro-thionin method is best adapted to work with old bone, while the phosphotungstic acid method is more useful for demonstrating the histology of young bone and the process of ossification.

DEPARTMENT OF BIOLOGY,

NEW YORK UNIVERSITY

ALDEN B. DAWSON

SOME FIXATIVES FOR BOTH NUCLEI AND MITOCHONDRIA

A 2.5 per cent. solution of copper bichromate C. P. (Eimer and Amend) has a pH of 2.0. When root tips of Zea are fixed in it the fixation image is that of chromic acid, i.e., the nucleolus appears as a spherical, darkly staining body in a hollow nucleus whose surface is composed of the chromatin reticulum. The mitochondria are either dissolved by the fixative or by the dehydrating alcohol. If, however, a slight excess of cupric oxide is added to the solution, the pH is altered to about 4.6 and the fixation image is greatly changed. There is here no hollow space around the nucleolus; the nucleus is a solid body, and in the resting stages the chromatin reticulum is much

2 1905. Fasoli, G. "Ueber die feinere Struktur des Knochengewebes." Arch. mikr. Anat., Bd. 66, S, 471. 3 1926. Carleton, H. M. "Histological Technique.'' Oxford University Press.

4 1913. Lange, W. "Histologische Zahnärzte." Springer, Berlin.

Technik

für

5 1910. Fischer, Bau und Entwicklung der Mundhöhle.

höhle.

less distinct. In the dividing nucleus the spireme shows up distinctly and the chromosomes are well preserved. While the spindle fibers are not distinguishable individually, collectively they are well delineated. The mitochondria are well fixed and mordanted and can be followed through each of the mitotic stages. This fixative has the following faults: the resting nuclei show little detail, the cytoplasm is somewhat distorted and the outer layer of cells is generally over fixed. The addition of .05 per cent. acetic acid causes the resting nuclei to show more detail, though one must be cautious in the use of this acid, for a slight excess of copper acetate will dissolve the mitochondria. The most successful formula for the fixative is:

The material should be left in the solution for from 36 hours to six days, and when thus fixed both chromosomes and mitochondria are well stained with Heidenhain's haematoxylin. Destaining should not proceed as far as is usual for an examination of the nuclei, for the mitochondria do not hold the stain as well as the chromosomes and can be completely decolorized before the chromosomes have started to fade.

It is very important to make up the fixative at least 24 hours before it is to be used. It must be shaken frequently in the interval and the excess copper oxide allowed to settle. If it is used too soon the fixation image will be that of chromic acid. It is best to wash out the fixative with 70 per cent. alcohol. If the dehydration is too prolonged the mitochondria will be dissolved out of the peripheral cells. A half hour in each of 70 per cent., 85 per cent. and 95 per cent. alcohol, and an hour in each of two changes of absolute, are sufficient for the dehydration.

Another solution which fixes both chromosomes and mitochondria is:

If calcium carbonate is used the pH is about 2.2; if lithium carbonate is used it is about 4.8.

When this fixative is washed out with water and the dehydration proceeds slowly, the dividing nuclei and the cytoplasm appear beautifully fixed. The chromosomes are a trifle shrunken so that in the metaphase they show the split very clearly. The cytoplasm appears quite smooth with sharply delineated vacuoles. In the root tip the growth of the vacuoles and their behavior during mitosis can be easily followed. Unfortunately the mitochondria are dissolved out of the epidermis and cortex and remain only in the central cylinder. If the fixative is washed out with

70 per cent. alcohol and the dehydration is relatively rapid, the cytoplasm appears more granular and the mitochondria are preserved in nearly the whole tissue.

It is evident that there is an important relation between the pH of a bichromate and its fixation image. If it is too acid it will fix the chromatin but not the mitochondria, if it is too basic it will fix the mitochondria but not the chromatin. Certain bichromates in the presence of an oxide or a carbonate of the element which furnishes the cation will buffer

at a point where they will fix both nuclear and cytoplasmic elements. Others, as their pH number is raised, suddenly change from nuclear to cytoplasmic fixatives. The pH of this point of change shows quite a range for the various bichromates. Thus ammonium bichromate pH 4.2 and potassium bichromate pH 4.4 are much too basic to fix the chromatin, while lithium bichromate pH 4.6 has the fixation image of chromic acid. Zinc bichromate pH 5.2 will fix both chromosomes and mitochondria with its characteristic fixation image.

A detailed description of the fixation images of various bichromates is being prepared.

BUSSEY INSTITUTION,

HARVARD UNIVERSITY

CONWAY ZIRKLE

SPECIAL ARTICLES

THE MnII SPECTRUM EXCITED BY RARE

GAS IONS

THE MnII spectrum was excited by the method recently described by Duffendack and Smith1 and tested by the writers2 on the CuII spectrum. In this

1 Phys. Rev. 29, 914, 1927; Nature, May 21, 1927. 2 Phys. Rev. 29, 925, 1927.

them and simultaneously excite them to the degree that the ionizing potential of the rare gas exceeds that of manganese, 7.4 volts.

An argon ion on contact with a Mn atom can ionize it and excite the resulting Mn+ ion to the extent of 15.4-7.4 8.0 volts. In the process the argon ion is neutralized by combination with an electron taken from the Mn atom and energy to the amount of 15.4 equivalent volts is made available. 7.4 equivalent volts of this is expended in extracting the electron from the Mn atom, leaving eight equivalent volts to be accounted for. Smyth and Harnwell and Hogness and Lunn3 have demonstrated by positive ray analyses that ionization may occur upon contact between an ion and a molecule. In the investigations cited above1,2 it has been demonstrated that the excess energy may go toward exciting the ion formed. Hence, when argon ions are used, one may expect to

produce by this process Mn+ ions excited to states

whose levels are less than eight volts or 84,800 cm-1 above the normal state of Mn+ but none excited to a higher degree. Consequently, lines of the MnII spectrum whose initial states are below 64,800 cm-1 should appear and lines originating in higher states should be absent from the spectrum thus excited. If, however, neon (ionizing potential 21.5 volts) is substituted for argon, Mn+ ions excited to states whose levels are less than 14.1 volts or 114,210 cm-1 are produced and lines from these levels should appear in the spectrum.

The experimental procedure consisted in photo

graphing the spectra of low voltage ares in mixtures

of argon and Mn vapor and neon and Mn vapor in a tungsten furnace apparatus. The manganese was put into a molybdenum trough mounted inside a cylinder of thin sheet tungsten and insulated from it. This trough constituted the anode of the arc, and the cathode was a tungsten filament mounted inside the cylinder and parallel to its axis. The tungsten cylinder was itself mounted inside a metal water-cooled vacuum chamber, filled to the desired pressure with argon or neon, and was heated by passing a sufficiently large current through it. The manganese in the trough was thus vaporized and any desired vapor pressure could be maintained inside the cylinder. The filament was then heated and a low voltage arc maintained in the mixture of argon or neon and manganese vapor within the furnace. The spectrum of the arc was photographed through quartz windows sealed onto the vacuum chamber.

The results support the hypothesis outlined above. When argon was used, lines originating in the 'P and

3 Nature, Jan. 15, 1927; Phys. Rev. 29, 830, 1927; ibid., 30, 26, 1927.

SP levels (Fig. 1) appeared in the spectrum but none from the "D level. When neon was substituted for argon and all other conditions kept the same as before, the "P-D lines came out strongly.

This work led us naturally to the analysis of the MnII spectrum. In 1923 Catalan published four multiplet arrangements in the spark spectrum of manganese which can easily be recognized as 'S- 'P, "P-D, "S-5P, and 5D-5P. The lowest term of the MnII spectrum may be expected to be the 'S term and so the levels of the septet terms were immediately established, as shown in Fig. 1. Catalan's multiplets cm2 70000

enabled us to determine the relative levels of the quintet system and so the first problem was to find intercombination lines which would fix the positions of the two systems with respect to each other. The difference 5S-7S can be estimated from convergence limits in the MnI spectrum.5 Dr. O. Laporte, who has given us valuable suggestions on the nature of the MnII spectrum, had recently calculated this difference and furnished us his result, 9,477 cm-1. Using this value, intercombination lines were quickly found which fixed the difference at 9,474.3 cm-1 and established the relative positions of the two systems. Lines have been found for the transitions indicated in Fig. 1, and the work of completing the analysis of the spectrum is in progress. The similarity of this spec

4 Phil. Trans. Roy. Soc. 223, 127, 1923.

5 McLennan and McLay: Trans. Roy. Soc. Canada 20, 15, 1926.

FOG PRECIPITATED BY TREES

THE collection by vegetation of moisture from fog has interested me for a long time. I recently found an opportunity to approximately measure the amount collected by trees.

During the summer west winds blow the moistureladen air from the Pacific Ocean up and over the hills back of Berkeley, California. Nearly every afternoon fog collects on the hills at elevations above 800 feet and stays until the morning sun dissipates it. Occasionally it remains the entire day.

About twenty-five years ago pine and eucalyptus trees were planted on the sides and tops of the hills over large areas which prior to that time were bare of all but grass. Trees were found only in canyons, while brush covered many of the slopes, particularly those sloping to the north. These trees grew slowly for a number of years but have made very rapid growth in the dry years since 1917. The summers here are nearly rainless and all vegetation on the hills usually dries up during this rainless season, except in protected spots and in canyons where moisture is more plentiful.

I have long noticed that the soil beneath trees is more moist than elsewhere, the additional moisture coming from the collection of water from the fog dripping to the ground. I recently (July 31) collected samples of soil from beneath trees and from ten feet from trees, where soil and other conditions were identical. Samples were collected from surface to 12 inches depth and the moisture determined. Here are the results:

Assuming the weight of soil as 90 lbs. per cubic foot, these differences in percentage are equivalent to the following in inches of rainfall. Pine, elevation 1,500 feet, 2.87 inches. Pine, elevation 1,600 feet, 3.60 inches. Eucalyptus, elevation 1,650 feet, 2.33 inches. The soil was moist much deeper than 12 inches, so the total difference in inches of water collected is much above that shown.

The area of ground covered by trees, where the 6 Catalan: Anales Soc. Esp. de Fis. y Quim. 21, 84,

1923.