NOVEMBKR 5, 1915]

No originality is, of course, claimed for thissystem of labeling, but the writer does nothappen to know of its being used elsewhere forchemical specimens. Any one may easilydevise letters and numbers to fit his presentcollection as well as future additlons. Pro-vision may be made for alloys, commercialsamples, and the like, wherever neceessary.

It is convenient to write the letters andfigures in two lines, library style, as, " S 6 "is written large with the "th 2 " written small,beneath. Three figures, with the initial, wiflbe the maximum number of characters in thefirst line. For the lower line, the rule is notto use more than three letters, while twofigures will always be sufficient.The bottles used in the writer's collection

are one-pound and half-pound "salt mouth,"of uniform style, with "mushroom " stoppers.These are convenient sizes, the smaller sizebeing used mainly for the more costly sub-stances. There are two labels on each bottle.The larger labels are known to the trade asNo. 1006, about four by one and five eighthsinches. The names of the substances areprinted on these labels with rubber type incapital letters a quarter of an inch high.These labels are placed just below the shoulderof the bottle. The round labels used for theindex letters and numbers are known as "A88" or library labels. These are centeredunder the large label near the bottom of thebottle. A library assistant did the lettering ofthese labels, using india ink. The bottles aresealed with paraffin, and the labels coated withparaffin. The latter is necessary as the bottlesare kept on open shelves, and usually requirewiping with a damp cloth when they are to beshown. The paraffin further protects thelabels against accidental eontact with acids oralkalies.This system of labeiing is scarcely applica-

ble to organic compounds, unless one does notwish to keep them separate from inorganic.The writer, for present purposes, has made alist of the substances studied or referred to,in order, in the organic text used (one of themost complete published), and each substancegiven a number. This does not include sub-

stances that are impracticable to keep or pro-cure. The collection is so arranged that thesubstances mentioned in a given chapter arefound together, in numerical order, the miss-ing numbers to be supplied in the future. An-other possible arrangement is by classes ofcompounds. This has been used, but is ratherless convenient than the present arrangement.In the organic set, the half-pound bottle isthe maximum size for solids, the two-ouncethe minimum. For liquids, there are threesizes, from a half-liter down. The labels areNo. 1007 for the names, and No. 539 for thenumbers. The bottles are paraffined as in theinorganilc set. On account of the effect of lighton many organic compounds, the specimens arekept in a dark room, where the inorganic set isalso kept for convenience.The "looks" of one's teaching devices will

be certain to leave lasting impressions on theobserver. This is especially true in the chem-ical laboratory and lecture room. While it isnot claimed that the above-described methodsof labeling bottled chemical specimens are thebest that could be devised, they have servedthe writer very well and it is hoped that thedescription may interest others.

C. E. VAILCOLORADO AGRICULTURAL COLLEGE

REPORT OF THE SAN FRANCISCO MEETINGOF SECTION F OF THE AMERICAN AS-

SOCIATION FOR THE ADVANCE-MENT OF SCIENCE. II

Thursday. August 5Morning Session, Demonstrations

In charge of OLIVE SWEZY, University of CaliforniaEntameba Buecalis, Inez F. Smith, University

of California.Mitosis and Multiple Fission in the Flagellata,

Olive Swezy, University of California.Mitosis in Lamblia muris, Elizabeth Christian-

sen, University of California.Enflagellating and Exflagellating Soil Amaeba,

Charlie W. Wilson, University of California.Flagellates of Hemiptera, Irene MeCulloch, Uni-

versity of California.Drawings for Monograph on Dinoflagellata, C.

A. Kofoid and Mrs. Rigden-Michener, Universityof California.

SCIENCE 657

on April 23, 2021

http://science.sciencem

ag.org/D

ownloaded from

[N S. VOL. XLII. No. 1088

Drawings for Monograph on Pacific Tintin-noidea, C. A. Kofoid and Mrs. Elizabeth Puring-ton, University of California.

Balantidium from the Pig, J. D. MacDonald,University of California.

Drawings of Elasmobranchs, J. Frank Daniels,University of California.

Papers: ProtozoologyCHARLES A. KOFOID, University of California,

presidingChromosomes in Protozoa: MAYNARD Mi. METCALF,

Oberlin, Ohio.This review has endeavored to show that in each

group of the Protozoa are found elongatedchromosomes which are linear aggregates of gran-ules and which split longitudinally in mitosis, giv-ing exact equivalence of the daughter nuclei. Inall major groups, except perhaps the 21astigo-phora, presexual and sexual phenomena, essentiallysimilar to those of Metazoa are known. The Meta-zoan mechanism of inheritance is therefore pres-ent, in some representatives at least, of all thegreat Protozoan groups. It has shown also that,in some representatives, at least, of the Plasmo-drome, during the vegetative phases of the life his-tory, the chromatin outside the caryosome, indeedapparently outside the caryoli, is thrown off astrophical chromidia. Such evolution of complex-ity of organization as has occurred in the Metazoaprobably could not occur until, as in the Ciliata, aconsiderable amount of chromatin is kept intactthroughout the whole life history, including itsvegetative as well as its presexual and sexualphases. The lower Protozoa have the Metazoantype of mechanism of inheritance in connectionwith their sexual phases but can not utilize it forthe development of a complex series of determin-ers for elaborate structural organization, for theydo not keep this mechanism intact during theirvegetative periods. The higher Protozoa appar-ently have the Metazoan type of mechanism of in-heritance and keep it intact through all their lifecycle. To what extent they have utilized its pres-ence to develop determiners, is an interesting ques-tion which does not seem beyond experimentalstudy with favorable material.

Problems on Bejuveneseence in Protozoa (illus-trated with lantern slides): LORANDE L. WOOD-RUFF, Yale University (read by title).

The Evolution of the Protozoan Nucleus and ItsExtranuclear Connections: CHARLES ATWOODKOFOID, University of California.

The essential and fundamental similarity ofProtozoan and metazoan nucleus is indicated bythe general trend of recent cytological investiga-tions among the Protozoa. The process of mitosishas the same sequence of phases, though thechronology of mitotic events and of the divisionof extranuclear structures varies from species tospecies and also among individuals within the spe-cies. In the early prophase in trichomonad flagel-lates and in Ncgleria the chromatin aggregates,presumably chromosomes each split longitudinally,or the chromatin network forms a split skein orthread which later fuses into one, or emerges onthe equatorial plate in chromosomes in which theprecocious splitting has entirely disappeared. Thechromosomes are definite in number and are dif-ferentiated among themselves as to size and be-havior in mitosis. There are suggestions of oddand even numbers (4, 5) there is (in Trichomonadflagellates in several genera and species) a smallchromosome lagging in the metaphase, and thereare instances of unequal division of chromosomes.The blepharoplast contains the centrosome in

trichomonad flagellates or is attached to it(Giardia-Lamblia). This organelle should not becalled a kinetonucleus; Schaudinn Is report of itsheteropole origin by mitosis is unsupported andprobably invalid. It arises in enflagellatingNcgleria (-Ammba) gruberi from the centrosomeor centriole in the central caryosome, which sendsout a radial fibril which enlarges at the peripheryinto a blepharoplast at the base of the twoflagellm which grow ouit from this enlargement.In exflagellation the flagella withdraw and to-gether with the blepharoplast retreat into thecaryosome. The blepharoplast is highly developedin parasitic flagellates and is directly connectedby a system of fibrils with nucleus, flagella, axo-style and parabasal body, the whole forming anintegrated structural unit which functions as aneuromotor apparatus comparable with that ofthe ciliate Diplodinium. It is a specialized struc-ture developed to the greatest extent in connec-tion with parasitic life demanding unusual ex-penditure of energy in locomotor activity. Theterm kinetonueleus should be abandoned as mis-leading. The Binuclearity Hypothesis in its nu-clear implications has no adequate foundations,and the order Binucleata should be discarded.The parabasal body of flagellates and Trichony-phida is perhaps the analogue of the macronucleusof the ciliates and is an extranuclear store of thechromatin related to the needs of flagellar andciliary activities.

658 SCIENCE

on April 23, 2021

http://science.sciencem

ag.org/D

ownloaded from

SCIENCE

Afternoon Session Papers. Geographic DistribuwtionJosEPs GRINNELL, University of California, pre-

sidingCalifornia as a Testing Ground for Theories of

Distributional Control: JOSEPH GRINNELL, Uni-versity of California.

Insect Transmission of Swamp Fever: J. W. SCOTT,University of Wyoming.During the summer of 1914, the writer obtained

experimentally a well-defined ease of swamp fever,and the conditions of the experiment leave nodoubt that the disease was contracted through theagency of certain biting insects. This disease, fre-quently known as "'infectious'' or perniciousanemia, is a serious and destructive blood diseaseof the horse. It has been reported from France,Germa-ny and Japan, and is widely distributed inNorth America, from Texas to the Northwest Ter-ritories, and from the states of Washington andNevada to the Province of Ontario, Canada. Itbas an altitudinal distribution from near sea-levelto at least 9,000 feet. It is usually found inswampy regions, but has also been reported fromrolling, wooded countries. Wherever found, it maybecome epidemic and ca-use the loss of a large per-centage of a given herd of horses. The diseaseshows a seasonal distribution, reaching a maxi-mum number of cases in late summer or earlyautumn. The disease itself is characterized byprogressive emaciation and an intermittent rise intemperature; it is frequently accompanied byanemia, and the rises in temperature are some-times at quite regular intervals. The organism isfilterable, for the disease is transmitted by theinjection of blood serum after it has been passedthrough a Berkfeld filter.

In France, Vallei and Carri concluded that nat-ural transmisson took place through drinkingwater contaminated with urine or feces from aninfected horse. Van Es, of North Dakota, in1911, after a study of several years, thought thisthe mest probable explanation. Swingle, how-ever, in Wyoming, in 1912 and 1913, showed thatit was an extremely difficult matter to secure in-fection by way of the alimentary canal, his nu-merous experiments with urine and feces all re-sulting negatively.At this point in 1913 the writer began the

problem. Since internal transmission was a diffi-eult matter, only one or two cases being known(Van Es, Schlatholter), and sinee the contamina-tion of drinking water could hardly explain epidem-ies, it was believed that natural transmission mustbe by means of some external agent. Accordingly

659

in the spring of 1914 a screened cage was erectedcapable of holding five horses. The cage had anentryway 10 feet long, each end closed by a doorand the screen had 16 meshes to the inch.The first experiment in this cage was with vari-

ous kinds of mosquitoes and resulted negatively.A longer and more conclusive experiment withmosquitoes the present summer has had a similarresult. There was next introduced into this cage aconsiderable number of flies; these were house-flies, stable-flies and a few other wild flies, in-cluding one of the smaller species of Tabanids.The house-flies and stable flies thrived and in-creased rapidly in numbers between the first andthe twenty-fifth of August; the other flies soondied and were not observed to attack the horses.The stable-flies were observed to feed in largenumbers on both infected and well horses. Ex-cept for two or three days infeeted horses werekept continuously in the cage from July 27 to Au-gust 28. Three well horses were exposed in thecage during this time. On August 28 horse No.22, a healthy strong animal, showed a temperatureof 102.8. After two more fever periods this horsedied October 5. Subinjection of his blood hasproduced typical cases, showing that he bad thedisease. The temperature of this horse had beennormal from June 9 to August 28, and he had notbeen outside of the cage for 25 days, while 10-14days is the ordinary incubation period of the dis-ease. Under these conditions there appears to beno escape from the conclusion that insects trans-mitted the disease. It is believed further that thestable-fly was responsible for the transmission forthe following reasons: (1) These flies were ob-served to attack the horses viciously. (2) Nega-tive results of two other experiments show thatthe mosquitoes were not responsible, even thougha few were still in the cage during August. (3)Houseflies do not bite, and the other flies presentin the cage did not attack the horses, so far as onecould observe, and soon died.The experiment, with the stable flies alone, is

being repeated, and another experiment involvingsome of the Tabanidie is also in progress.

The Big Bears of Western North America, withSpecial Reference to their Distribution: C. HARTMERRIAM, United States Biological Survey.The bears are the largest of living carnivores

and are widely distributed, being found in bothAmericas and in Eurasia. The typical genus'Ursus occurs in both eastern and western hemis-pheres. South America is the home of Tremaretos,

NOVEMBIR 5, 1915]

on April 23, 2021

http://science.sciencem

ag.org/D

ownloaded from

[N. S. VoL. XLII. No. 1088

a short-faced bear resembling Arctotherium, ofthe American Pleistocene, while the north polarregions are inhabited by the polar bear, Thal-arctos. The ancestry of the bears is still clouded,none being know-n below the Pliocene.

Systematically the bears are a compact group,yet within the genus Ursus several subgroups may berecognized; the subgenus Euarctos, containing theblack or cinnamon bears and the subgenus Ursw,containing the grizzly and brown bears. Of thelatter (grizzly and brown bears) two, three andeven four species, representing different species-assemblages, may be found in a single locality, asin the fossil deposits of Rancho La Brea or inYellowstone National Park. The characters usedin identification are chiefly cranial and dental.Skulls of males of the grizzly group are two, andin some cases, three times the bulk of those offemales of the same species. At present about 40species of grizzlies and 10 of brown bears arerecognized, where formerly but one of each wasknown. For example, California once containedrepresentatives of five different groups of grizzlies.The recognition of this large number of specieshas been made possible by the extensive collectionsof skulls in the United States National Museumand in the California MIuseum of Vertebrate Zool-ogy.

Fossil Tertiary Mollusca of the Rocky MountainRegion: T. D. A. COCKERELL, University of Col-orado, Boulder, Colorado.The recent expeditions of the American Museum

of Natural History, primarily for the discovery ofmammalian remains, have brought to light somevery interesting Tertiary land and freshwatershells, principally in Wyoming and New Mexico.From a study of these we are led to the followingconclusions:

1. Certain of the most characteristic genera ofthe Rocky Mountain land shells living to-day areapparently very ancient inhabitants of the samegeneral region, and have, perhaps, at no time ex-tended very mueh beyond it. The most note-worthy example is Oreohelix, represented in theEocene and Paleocene by large species, some ofthe specimens showing the sculpture of the apicalwhorls, which agrees with that of the groupcalled Radiocentrum by Pilsbry. A species ofHolospira from the New Mexico Paleoeene is ex-tremely like a living species of Arizona.

2. Ashmunella, one of the most characteristicendemic genera of the southwest, is not at presentknown below the Pleistocene. It is not possibleto be sure, at present, wxhether all the peculiar

genera of ;the Rocky Mountain region and south-west are very ancient inhabitants of that country,but it seems very likely that they will all be found,sooner or later, in the early tertiaries.

3. Various eircumpolar genera of small snails,such as Pupilla and Cochlicopa, represented in themodern fauna by species identical or nearly iden-tical with those of Europe and northern Asia, areapparently lacking in the Tertiary, or at any ratein the Eocene. They may have been overlooked,but it is probable that they have reached our re-gion in much more recent times, from Eurasia, astheir small amount of modification, or total lackof it, would suggest the Eoeene collections do con-tain small species of more characteristicallyAmerican types as Vitrea and Thysanophora.

4. The Paleocene and Eoce-ne faunas included aseries of genera entirely different from anythingnow living in the same region, but evidently re-lated, at least in part, to the present Central Amer-ican and West Indian faunas. Thus we have fromthe Eocene of Wyoming species apparently refer-able to PIeurodonte and Eucalodium. It is notclear, at the present time, whether this CentralAmerican fauna originated northward, or whetherit had its main center in the region where it nowexists, merely extending northward during a timewhen the country Inow represented by Wyomingand Colorado was low, moist and warm.

5. The most remarkable discoveries have beenof small shells belonging to the Bulimulide, hav-ing the aperture or last whorl upturned, repre-senting at least two genera (Protoboysia andGras gerefla) and four species. These, while be-longing to extinct genera, show evident relation-ship with some of the South American snails, andat the same time a remarkable resemblance to theIndian genus Boysia, which is supposed to belongto the Pupillida, though the anatomy is unknown.

6. Among the fresh-water shells, the Unionidieare the most interesting. About the end of theCretaceous and beginning of the Tertiary we findin the Roeky Mountain region a large and variedseries, resembling the types which now inhabit theMississippi Valley. At about the same time thatthe dinosaurs disappeared, these mussels also de-parted, leaving an impoverished Unionid fauna inthe Eocene, which in its turn eventually died outaltogether. It seems probable that this change wasconnected with earth movements.

Isolation as a Factor in the Evolution of Thaislaimellosa (illustrated with specimens and lan-tern slides): TREvon KINCAID, University ofWashington.

660 SCIENCE

on April 23, 2021

http://science.sciencem

ag.org/D

ownloaded from

NovEmBim 5, 1915]

Friday, August 6

Morning Session, DemonstrationsIn charge of F. W. WEYMOUTH

Living Ova of Rat in Serum, J. A. Long, Uni-versity of California.

Papers: Marine ZoologyF. M. McFARLAND, Stanford University, presiding

The Occurrence and Possible Causes of PeriodicVertical Movements of Aquatic Organisms: C.0. ESTERLY, Occidental College, Los Angeles,California.It is well known that many aquatic organisms,

particularly those of the plankton, are moreabundant at or near the surface by day than bynight. This is because they ascend from deeperwater in the early part of the night and descendfrom the surface later.

In order to understand the fundamental causesof this phenomenon it is necessary to have accu-rate knowledge of the field conditions under whichcollections are made, particularly of the tempera-ture and salinity of the water and of the light in-tensity as represented by the time of day at least.It is highly important that field observations besupplemented by laboratory experiment to deter-mine to what sorts and degrees of stimuli the or-ganisms respond. Furthermore, each species mustbe studied by itself.The explanation of the diurnal movement that

at present is most satisfactory is the one based onresponses to different factors in the environment,but our knowledge in this respect is very incom-plete. The "mechanical explanation" whichtakes into account mainly the changes in viscosityof the water does not satisfy. The suggestionthat the alternating rise and fall is due to meta-bolic rhythms has received so little attention thatits worth is not apparent.

Some Physiological Characters of Marine Animalsfrom Different Depths: V. E. SHELFORD, Uni-versity of Illinois, Urbana.

Field Study of Animal Behavior as contrastedwith Laboratory Study: ELLIs L. MICHAEL,Scripps Institute, La Jolla, California.The object of this paper is to emphasize the

necessity of employing tw-o essentially distinct butmutually helpful methods of research in anystrictly scientific study of animal behavior. Allbiologists readily admit that an important func-tion of the well-known laboratory method is that

of analyzing the mechanism involved in an organ-ism 's response. Few, however, seem to realizethat this method is incapable of revealing howany particular species is related to its environ-mental complex.

Yet, the minute a definite answer to such a ques-tion is sought, a little thought clearly shows thatrecourse must be had to some method of field ob-servation. For instance, could any amount oflaboratory experimenting reveal the fact thatSagitta bipunctata is more abundant between 15and 25 fathoms than at any other depth, or thatit increases in abundance as the distance from thecoast decreasesi

In order to demonstrate the indispensability of afield method three aspects of the behavior of S.bipunctata are illustrated in some detail. First,the variation in abundance at all depths at differ-ent times of day is considered, showing how thespecies migrates vertically. Second, the fact thatit accumulates in all depths in greatest numberswhen the temperature lies between 13° and 160 C.is revealed. Third, data are presented showingthat the species is more abundant, on the surfaceat least, when the salinity lies between 33.55 and33.70.

Finally, the question of the interpretation of theresults is discussed. Attention is called to the factthat an adequate interpretation necessitates re-course to the laboratory method. Obviously thefield method can not, except by interference, as-certain the nature of response involved, i. e., asto whether the demonstrated relations are due to atropism, a taxis, a metabolic reaction, or to adirect physical effect of changes in viscosity anda specific gravity.

The Influence of Chance on the Number of Or-ganisms Collected in Plankton Nets: GEORGE F.MCEWEN, Scripps Institute for Biological Re-search, La Jolla, California.Measurements of a number of hydrographic ele-

ments and the corresponding abundance, or num-ber of a species per unit volume of water in theocean supply the data for determining the way inwhich the distribution of the species is related tothese elements.

Practically, the distribution of the smaller or-ganisms must be inferred from the number ofeach species collected per haul made with a plank-ton net. A planikton net filters rather variablefraction of the water that would pass through thenet rim, if unobstructed; also the estimate of thedistance hauled is usually subject to accidentalerrors of importance.

SCIENCE 661

on April 23, 2021

http://science.sciencem

ag.org/D

ownloaded from

[N. S. VOL. XLII. No. 1089

Grouping values of the abundance thus ob-tained with respect to each of a series of magni-tudes of one of the environmental elements ex-hibits in tabular form the relation of the distri-bution to that element and the others linked with it.The errors in estimating the abundance arisingfrom the use of a plankton net, the variability ofthe elements other than the one in question, andthe fact that only a relatively small sample of thetotal population is examined have an importantaccidental influenee on the character of the rela-tion.

Illustrations are given of a method of testingthe significance of the difference between two val-ues of the abundance, which depends upon theprobability that a difference in the same directionwould arise by chance if the number of hauls wereindefinitely inereased. A plan is briefly outlinedfor eliminating the effects of all but one of theelements and testing the significance of the rela-tion indicated by the corrected values of theabundance. The same methods are also applicableto statistical treatment of other kinds of quantita-tive data.

The Boring Mollusca of the Pacific Coast: Mas.IDA S. OLDROYD, Long Beach, California.

The Life History of the Pacific Herring: C. Mc-LEAN FRASER.The Pacific herring appear in large schools all

along the Pacific coast from California to Alaska.The average weight of those caught in purse seinesis about 3.6 oz. and the length 8 inches, not in-cluding caudal finrays. They wander about insearch of food, which consists mainly of copepods.There is but one spawning season in one locality.Near the biological station, Nanaimo, this is inFebruary and March. Spawning takes place inshallow water, but this may be incidental to therequirement of barnacle larvmn for food at thattime. Both females and males rub against the seaweed while spawning. The spawn adheres, batchesout in 14 or 16 days and the yolk lasts another sixdays. First spawning takes place at the age ofthree or four years. Most herring caught are fromfour to eight years; some were found ten yearsold. The scale increases in size in approximnatelythe same ratio as the other parts of the body andthe different year growths are marked off bywinter checks or rings. Those who have calcu-lated the growth of the fish in each year from therate of grow-th of the scales have failed to takeinto account that the scale does not start to growwhen the fish does. The herring is 3.5 cm. long

before the appearance of the scale. This shouldbe taken into consideration when the length of thefish is divided accordiing to the divisions of thescale as shown by winter checks.

The Nuclear Phenomena in Paramecium: R. T.YOUNG, University of North Dakota.During the San Francisco meetings on Thursday,

August 5, of the American Association for theAdvaancement of Science, there was formed a Pa-cific Branch of the American Society of Zoologists.The officers elected at this meeting were:

President, V. L. Kellogg, Stanford University,Palo Alto, California.

Vice-president, R. M. Yerkes, Santa Barbara,California.

Secretary and Treasurer, Joseph Grinnell, Uni-versity of California, Berkeley, California.

Executive Committee, C. 0. Esterly, OccidentalCollege, Los Angeles, California; Barton W. Ever-mann, California Academy of Science, San Fran-cisco, California; Charles L. Edwards, Los Angeles,California; J. Frank Daniel, University of Cali-fornia, Berkeley, California; Harold Heath, Stan-ford UTniversity, Palo Alto, California.On Thursday, August 5, there was formed a

Pacific Coast Branch of the American Society ofNaturalists with the following organization:

President, Barton W. Evermann, CaliforniaAcademy of Sciences, San Francisco, California.

Vice-president, John F. Bovard, University ofOregon, Eugene, Oregon.

Secretary, Ellis Leroy Michael, Scripps Institutefor Biological Researel, La Jolla, California.

Treasurer, L. L. Burlingame, Stanford Univer-sity, Palo Alto, California.

Executive Committee, Trevor Kincaid, Univer-sity of Washington, Seattle, Wash.; Harry B.Torrey, Reed College, Portland, Oregon; FrankM. McFarland, Stanford University, Palo Alto,California.The society will take the place of the local bio-

logical societies of the Pacific coast.The Biological Society of the Pacific met at the

Hotel Sutter, San Francisco, August 4, for theirannual meeting. The address of the evening wasgiven by Dr. Harry Beal Torrey, of Reed College,on "I Research and the Elementary Student ofScience." At this meeting the Biological Societyvoted to drop its organization in favor of thenewly organized Pacific Branch of the AmericanSociety of Naturalists.

H. V. NEAL,Secretary

662 SCIENCE

on April 23, 2021

http://science.sciencem

ag.org/D

ownloaded from

ASSOCIATION FOR THE ADVANCEMENT OF SCIENCE. IIREPORT OF THE SAN FRANCISCO MEETING OF SECTION F OF THE AMERICAN

H. V. Neal

DOI: 10.1126/science.42.1088.657 (1088), 657-662.42Science 

ARTICLE TOOLS http://science.sciencemag.org/content/42/1088/657.citation

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

registered trademark of AAAS. is aScienceAdvancement of Science, 1200 New York Avenue NW, Washington, DC 20005. The title

(print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for theScience

Advancement of Science. No claim to original U.S. Government Works.Copyright © 1915 The Authors, some rights reserved; exclusive licensee American Association for the

on April 23, 2021

http://science.sciencem

ag.org/D

ownloaded from