PROCEEDINGS OF THE AMERICAN ACADEMY OF ARTS AND SCIENCES. PROCEEDINGS OF THE AMERICAN ACADEMY OF ARTS AND SCIENCES. NEW SERIES. Vol. XXI. WHOLE SERIES. Vol. XXIX. FROM MAY, 1893, TO MAY, 1894. SELECTED FROM THE RECORDS. BOSTON: UNIVERSITY PRESS: JOHN WILSON AND SON 1894. , ibio a CONTENTS. Page I. On certain substituted Crotonolactones and Mucobromic Acid. By Henry B. Hill and Robert W. Cornelison . 1 II. A Revision of the Atomic Weight of Barium. Second Paper. The Analysis of Baric Chloride. By Theodore William Richards 55 III. New Genera and Species of Laboulbeniaceoz, with a Synopsis of the Known Species. By Roland Thaxter .... 92 IV. On the Formation of Volatile Compounds of Arsenic from Arsenical Wall Papers. By Charles Robert Sanger . 112 V. On Chronic Arsenical Poisoning from Wall Papers and Fabrics. By Charles Robert Sanger 148 VI. On the A isomorphic Linear Transformation of a Bilinear Form. By Henry Taber 178 VII. On some Laws of Cleavage in Limax. A Preliminary Notice. By C. A. Kofoid 180 VIII. Further Observations upon the Occurrence of Diamonds in Meteorites. By Oliver Whipple Huntington, Ph. D. 204 IX. Double Habrids of Antimony and Potassium. By Francis Gano Benedict 212 X. Certain Bromine Derivatives of Resorcine. By C. Loring Jackson and F. L. Dunlap 228 PROCEEDINGS OF THE AMERICAN ACADEMY OP ARTS AND SCIENCES. VOL. XXIX. PAPERS READ BEFORE THE ACADEMY. I. CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF HARVARD COLLEGE. ON CERTAIN SUBSTITUTED CROTONOLACTONES AND MUCOBROMIC ACID.* By Henry B. Hill and Robert W. Cornelison. Presented April 12, 1893. Several years ago a dichlorpyromucic acid was described by Hill and L. L. Jackson, f wbich differed from all the substituted pyromucic acids then known, in that it was readily decomposed by concentrated hy- drochloric acid at 100° at ordinary pressures. Carbonic dioxide was evolved and a neutral body was formed which melted at 52-53°, and contained a percentage of chlorine which corresponded to the formula C4H3C102. A body of similar properties melting at 77° had previ- ously been found in small quantity by Hill and Sanger $ among the products formed from pyromucic tetrabromide by the action of an alcoholic solution of sodic hydrate. A complete analysis of this body had shown that its formula was C4H3BrO;2, but a lack of material ren- * A part of the work described in the following paper was presented in the form of a thesis to the Faculty of Arts and Sciences of Harvard University in May, 1893, by Robert W. Cornelison, then candidate for the degree of Doctor of Philosophy. t These Proceedings, XXIV. 348. 1 Ibid., XXI. 158. VOL. XXIX. (N. S. XXI.) 1 2 PROCEEDINGS OP THE AMERICAN ACADEMY. dered a detailed study of it impossible. While the chlorine derivative was more accessible, it was not more carefully studied by Hill and L. L. Jackson since it was discovered only at the close of their inves- tigation of the chlorpyromucic acids. They were also but partially successful in determining the constitution of the ^-dichlorpyromucic acid from which it was so easily formed. They showed that one chlorine atom of this acid was in the 8 position, but could bring for- ward no definite facts to prove whether it was a /38-dichlorpyromucic acid stereometrically isomeric with the common form, or whether it was the third possible structural isomer, the -yS-dichlorpyromucic acid. In the former case its ready decomposition by aqueous hydrochloric acid might be due to its peculiar configuration, in the latter case this reaction might be conditioned by the simultaneous presence of halogens in the y and 8 positions. If the instability of the 'acid were due to the positions occupied by the halogen atoms, it seemed probable that the tri-substituted pyromucic acids which also contain halogen in the y and 8 places would show the same behavior. Although Hill and Sanger had observed no such decomposition in studying tribromjDyromucic acid, and Hill and L. L. Jackson noticed no such reaction with the tri- chlorpyromucic acid, in neither case had direct experiments in this direction been made, and it was quite possible that the decomposition in question had been overlooked. It was soon found, on trial, that tri- brompyromucic acid was decomposed in analogous fashion, but that the reaction was effected with much more difficulty, so that a temperature materially above 100° was necessary in order to bring it about. When heated to boiling with concentrated hydrobromic acid, carbonic dioxide was evolved and a body melting at 90-91° was formed, which showed a close resemblance to the monohalogenized bodies already known, and which had the similar formula C4H2Br202. From trichlorpyro- mucic acid the analogous body C4H2C1202, melting at 50-51°, could be made without difficulty. In studying the behavior of other substi- tuted pyromucic acids under the same conditions, it was found that the /38-dibrompyromucic acid could also be made to undergo the same decomposition, although but a comparatively small yield of a body C4H3Br02 melting at 58° could thus be obtained. This mode of decomposition was, therefore, not confined to those acids which con- tained halogen at the same time in the y and 8 positions. By follow- ing an entirely different method it was found that the /38-dibrompyro- mucic acid could be converted into an isomeric body, C4H3Br02, which melted at 77°, and was identical with that which had already been discovered by Hill and Sanger. The /38-dibrom pyromucic acid was HILL AND CORNELISON. — CROTONOLACTONES. 3 treated with bromine in aqueous solution, and the brommaleylbromide, C4HBr302, which, according to the observations of Hill and Sanger,* is thus formed, when carefully reduced by zinc dust and glacial acetic, acid, yielded the body melting at 77°. The nuicobromylbromide C4HBr302 described by Hill and 0. R. Jackson f was so simply related, as far as its empirical formula was concerned, to the body C4H2Br202 formed from tribrompyromucic acid, that it seemed pos- sible to establish experimentally a direct connection between the two. It was found that mucobromylbromide could readily be reduced by a variety of reducing agents, and that under certaiu conditions a nearly quantitative yield of the body C4ll2Br202 melting at 90-91° could be obtained from it. Moreover, this reaction could be reversed and muco- bromylbromide obtained by the action of bromine upon this reduction product. In the same way, by the reduction of mucochlorylbromide the body C4H2C1202 melting at 50-51°, which had been made from trichlorpyromucic acid, could readily be obtained in any desired quantity. Since the several bodies already mentioned were evidently perfectly similar in their structure, we chose for more detailed study the body C4H2Br202, which was the most readily accessible. The melting point of this substance, 90-91°, was not far removed from that assigned (88°) by Toennies t to a body of like formula, which he obtained by the oxidation of ^y-dibrompyromucic acid with bromine water, and it was not difficult to suppose the two bodies identical. Toennies pub- lished no analyses whatsoever of his product, and Hill and Sanger § were unable to obtain the substance in larger quantities; but according to their observations a preparation which gave the proper percentage of bromine melted at 89-90°. A careful comparison of the substance prepared according to Toennies with that obtained by the decomposi- tion of tribrompyromucic acid, or by the reduction of mucobromyl- bromide, showed the two bodies to be identical in every respect. The formation of dibrommaleic acid by the prolonged action of nitric acid upon this body sufficiently established the relative position of the two bromine atoms and left for its constitution the choice between the following formula? : — * These Proceedings, XXI. 166. t Ibid., XVI. 174. t Berichte d. deutsch. chem. Gesellsch., XII. 1202. § These Proceedings, XXI. 172. 4 PROCEEDINGS OF THE AMERICAN ACADEMY. H BrC - CO BrC - CH2 I I > BrC - CO BrC - CO II Toennies found that his product could be converted into mucobromic acid by oxidation with chromic acid, and therefore considered it to be the double aldehyde of dibromfumaric acid. We found, on the other hand, that the body showed none of the ordinary characters of an al- dehyde. On warming with chromic acid it was slowly oxidized, but mucobromic acid was not formed in quantity sufficient to enable us to identify it with precision. On long boiling with bromine and water mucobromic acid was formed. From concentrated nitric acid it could be recrystallized unchanged, and only after continued boiling was it oxidized to mucobromic and dibrommaleic acids. It did not combine with acid sodic sulphite, and did not react with hydroxylamine. Ani- line in alcoholic solution removed one atom of bromine and gave a phenylamido derivative of the form C4Il2(C6H5NH)BrOo, and phenyl- hydrazine also removed bromine. Aqueous alkalies dissolved it with the formation of a deep yellow solution, but decomposition soon ensued with the elimination of hydrobromic acid. The whole behavior of the body was in direct opposition to the assumption that it was a double aldehyde, and the unsyinmetrical structure was further established by the existence of two isomeric bodies of the same general constitution which contained a single bromine atom. The conclusion was inevita- ble that the body was a dibromcrotonolactone. The rigorous proof of its lactone structure, through its conversion into a salt of the corre- sponding oxy-ac:d was rendered difficult by the presence of the halogen since the halogen itself was rapidly removed in alkaline solution. By adding a decinormal solution of potassic hydrate to a cold dilute solu- tion of the substance it is true that it was easy to prove that very nearly two molecules of potassic hydrate were neutralized in the re- action, while but one molecule of potassic bromide was formed, but it was impossible to isolate definite products of the decomposition. On reducing the body with zinc and dilute sulphuric acid the bromine was completely removed, and on distillation a feebly acid solution was obtained, which on titration with decinormal potassic hydrate proved to contain a lactone. By the usual methods an amorphous barium salt was obtained, which when thoroughly dried had the percentage composition required by a baric oxycrotonate. The isolation of the HILL AND CORNELISON. — CROTONOLACTONES. 5 crotonolactone itself proved to be a matter of such difficulty that we attempted to find some derivative which would be more manageable. The bodies containing the aniline residue in place of one of the halo- gen atoms were found to be useless for our purpose, since they were decomposed by boiling in alkaline solution with the formation of phenylisocyanide, and the phenylamidocrotonolactone formed from them by reduction was also decomposed with the formation of aniline under the same conditions. It seemed probable that the corresponding derivatives containing the phenoxy group would prove to be much more stable. AYhile we have not yet succeeded in replacing the bro- mine of the body C4H2Br202 directly by the phenoxy group, we have had no difficulty in preparing such a body by the reduction of the bromanhydride of the mucophenoxybromic acid which was described by Hill and Stevens.* The body thus formed was easily shown to be a phenoxybromcrotonolactone, since it dissolved in hot alkaline solu- tions with the formation of the salts of the corresponding oxy-acid. On acidifying the well cooled akaline solution the phenoxybromoxy- crotonic acid was obtained, which was stable under ordinary conditions, but which was again converted into the lactone by heat. These results were fully confirmed by a study of the phenoxychlorcrotonolactone, which was prepared by the reduction of mucophenoxychlorylbromide. The position which the phenoxy group takes in entering mucobromic acid has already been shown with sufficient precision by Hill and Stevens, since a remarkably stable phenoxybromacrylic acid is readily formed from this product by the action of alkalies. The phenoxylac- tones containing halogen must therefore have the constitution BrC - CH2 C1C - CH2 V o CfiHBOC - CO C.H^OC - CO and the analogous bodies containing halogen alone the structure BrC - CH„ C1C - CH2 X0 X0 BrC - CO C1C - CO The formation of these bodies from tribrom- and trichlorpyromucic acids seems to us to be most readily explained by assuming that a tri- substituted furfuran is first formed with the loss of carbonic dioxide : * These Proceedings, XIX. 262. 6 PROCEEDINGS OP THE AMERICAN ACADEMY. xc = xc = C - COOH X0 / cx xc = xc = CH x o + co2 cx This then adds a molecule of haloid acid with the shifting of the double bond in a manner identical with that observed by v. Baeyer and Rupe* in the reduction of dichlormuconic acid : XC xc CH xo / cx + H X XC - CHo xo / XC - CX9 The addition product is then decomposed by water, giving the lac- tone. Possibly a molecule of water is directly added to the trisubsti- tuted furfuran, giving XC - CH2 \ XC o cx OH which would at once pass into the lactone by the loss of haloid acid. The conversion of the /3y-dibrompyromucic acid into the dibromcro- tonolactone, and of the /3y-dichlorpyromucic acid into the correspond- ing dichlorcrotonolactone by the action of bromine in aqueous solution likewise finds its explanation in the formation of a trisubstituted fur- furan. In aqueous acid or alkaline solution the carboxyl of the sub- stituted pyromucic acids is readily eliminated and replaced by bromine, as was first shown by Hill and Hartshorn f in the case of the 8-brom- pyromucic acid : XC = C - COOH \ 0 } = CH XC = CBr + Br2 = X0 + C02 + HBr XC = CH By the addition of water the substituted lactone is formed as before. We have shown that the bromanhydrides of mucobromic and muco- chloric acids are also formed in this same reaction. This formation is evidently due to the addition of bromine to the trisubstituted furfuran and the decomposition of this product by water : * Ann. Chem. u. Pharm., CCLVI. 25. t Berichte d. deutsch. chem. Gesellsch., XVIII. 448. HILL AND CORNELISON. CROTONOLACTONES. xc xc = xc xc CBr X0 / CH CBr2 X0 / CHBr + Br2 XC XC- XC CBr2 X0 / CHBr + H20 XC- CO xo / CHBr + 2HBr It has already been said that two isomeric monosubstituted crotono- lactones may be made by appropriate means from the disubstituted pyromucic acids containing the halogens in the /3 and S positions. By heating these substituted pyromucic acids with mineral acids the brom- and chlorcrotonolactones melting at 58° and 26° respectively may be made. The reaction is evidently perfectly analogous to the decompo- sition of the trisubstituted pyromucic acids under like conditions which has just been discussed. The carboxyl group is replaced by hydro- gen with the loss of carbonic dioxide and the substituted furfuran thus formed then passes into the lactone as before : + co2 xc == c - COOH s XC == CH H( xo / : = cx H( XC / 3= CX XC == CH XC -- CH2 H( > + H20 H( J -CO + HBr The ^-substituted lactones which are thus formed may be made much more conveniently by the partial reduction of the disubstituted lac- tones. In order to prepare the monohalogenized crotonolactones with the halogen in the a position from the /38-dichlor- and dibrompyromucic acids, derivatives of maleic acid must first be formed. The first step in the reaction is the formation of a trisubstituted furfuran through the replacement of the carboxyl by bromine : XC = HC = C - COOH XC == CBr )0 +Br2 CX H( XC / ; = cx 0+ C02 + HBr Through the addition of bromine with the shift of the double bond, as before, 8 PROCEEDINGS OF THE AMERICAN ACADEMY. xc = CBr \ HC cx xc- 0 + Br, = CBr2 \ HC- - CXBr a body is formed which through the action of water yields the deriva- tive of maleic acid : XC - CBr„ \ XC O + H20 HC - CXBr CO \ O + 2IIBr / HC - CXBr This in its turn gives the lactone by reduction : XC-CO xo / HC - CH2 In the case of the /38-dibrompyromucic acid the tribromfurfuran, which should appear as the intermediate product in this reaction, has been isolated in a pure condition by Mr. W. M. Booth in this Labora- tory, and been found to yield, when treated with bromine in aqueous solution, the brommaleylbromide described by Hill and Sanger. The a-bromcrotonolactone may also be formed by the action of bromine in aqueous solution upon /3-brompyromucic acid. The reaction is evidently identical with that through which the dibromcrotonolac- tone is formed from /?y-dibrompyromucic acid. A dibromfurfuran is first formed, which by the fixation of water and the elimination of hydrobromic acid passes into the lactone : BrC = C - COOH + Br2 BrC = HC O / CH CBr \ HC = / CH O + C02 + HBr BrC = HC CBr X0 / CH + H20 = BrC- CO \ O + HBr HC - CH, The isomerism observed in the monohalogenized crotonolactones must be due to the position of the halogen atoms, since, according to the views at present held concerning geometric isomerism, lactones can be formed from the maleinoid forms only. A definite proof that the HTLL AND CORNELISON. — CROTONOLACTONES. 9 halogen atoms in the a- and /?-bromcrotonolactones are attached to different carbon atoms is easily given. The o/3-dibromcrotonolactoue yields with hydriodic acid an iodine derivative which cau be reduced to /3-bromcrotonolactone. This same iodine compound gives with aniline a phenylamidobromcrotonolactone which on reduction is con- verted into the same phenylatnidocrotonolactone that may be made by the action of aniline upon the a-bromcrotonolactone : BrC — CH2 BrC — CH8 IC )o J -CO H( xo / )- CO BrC - CH2 HC - CHo C6H5NH( HC xo / ;-co ) — CHo S* \ / C6H5NHC - CO BrC / )-CO O It is evident that the acid described by Hill and L. L. Jackson under the provisional name of the ^-dichlorpyromucic acid, since it readily gives the a-chlorcrotonolactone on decomposition with mineral acids, must in fact be the •yS-dichlorpyrornucic acid, and its structure HC = C - COOH )o 310 = CC1 It is perhaps worthy of note that, although the $h- and the yS-di- chlorpyromucic acids give the two isomeric chlorcrotonolactones when heated with acids, they give the same a-chlorcrotonolactone when they are treated with bromine in aqueous solution, and the resulting product it reduced. Evidently two isomeric dichlorbromfurfurans are first formed, which yield the corresponding addition products with bromine, but these addition products are attacked by water in such a way that in each case the chlorine is left in the a position with regard to the oxidized carbon : C1C = CBr \ C1C- O HC = CC1 HC CBr. X0 / CClBr C1C- HC CO X0 / CClBr 10 PROCEEDINGS OF THE AMERICAN ACADEMY. HC C1C = CBr X0 / CC1 HC- C1C CBr2 / CClBr HC - CBr, \ / C1C - CO 0 The close relationship between mucobromic acid and bodies which have thus been shown to be derivatives of crotonolactone naturally recalled the suggestions which had already been made as to the constitu- tion of mucobromic acid itself. As early as 1882 Roser* pointed out that the so called fumaric aldehyde acid of Liinpricht might in reality be an oxylactone, HC H COH xO HC- CO and afterward in 1887 Anschiitz,t in his interesting and suggestive discussion of the constitution of maleic and fumaric acids, was led to the conclusion that mucobromic acid was an oxydibromcrotono- lactone : H BrC - COH V / BrC - CO O It was evident that the almost quantitative reduction of mucobromyl- bromide by such a reducing agent as stannous chloride in the cold was perfectly intelligible if the formula of this bromide were BrC- BrC- H CBr \ CO o while the formation of a lactone under these conditions from the bromanhydride derived from the aldehyde acid BrC-C II O BrC - COBr * Berichte d. deutsch. chem. Gesellsch., XV. 1523. t Ann. Chem. u. Pharm., CCXXXIX. 161. HILL AND CORNELISON. — CROTONOLACTONES. 11 could hardly be explained except by assuming that the aldehyde group was itself first attacked, and the lactone then formed by the elimination of hydrobromic acid from the body B^ - cgfe BrC - COBr This explanation seemed to us exceedingly improbable, and we fur- thermore showed by direct experiment that no perceptible amount of the dibromcrotonolactone could be formed by the action of stannous chloride and hydrochloric acid upon mucobromic acid itself, or by its reduction with zinc and acetic acid, while the alcohol acid Br{[ - COH BrC - COOH which would be formed if the aldehyde group were readily reduced would certainly yield the lactone with facility. The ready formation of mucobromylbromide by the action of bromine at 100° upon the dibromcrotonolactone seemed equally conclusive, and the lactone formula for mucobromylbromide established. Since phosphorous tri- bromide converts mucobromic acid almost quantitatively into its brom- anhydride, the lactone structure of mucobromic acid itself seemed to be proved. It can hardly be supposed that a bromide of the lactone formula could be formed from the aldehyde acid, for this would im- ply the replacement of the aldehyde oxygen through the phosphorous tribromide leaving the hydroxyl of the carboxyl group to form the lactone by the elimination of hydrobromic acid. Although Hill and O. R. Jackson * showed many years ago that mucobromic acid could readily be converted into dibrommaleic acid by means of argentic oxide, it was evident that even this mode of oxidation no longer war- ranted a definite conclusion as to its aldehyde character. We there- fore thought it necessary to take up the study of the behavior of mucobromic and mucochloric acids, as well as the related acids con- taining the phenoxy group, toward hydroxylamine. We found that these four acids gave oximes with ease, and in studying more in detail the compounds formed from mucobromic acid we found that they were perfectly analogous to the bodies formed by the action of hydroxyl- amine upon opianic, phthalaldehydic, and pseudopianic acids as * These Proceedings, XVI. 186. 12 . PROCEEDINGS OF THE AMERICAN ACADEMY. described by Lieberinan,* Allendorf,f Racine, X and Perkin.§ Muco- bromoxime is so unstable that it passes spontaneously into its anhy- dride, and can be formed only by the actiou of free hydroxylamine in aqueous solution. The anhydride is formed through the hydrochlorate in alcoholic solution, and by boiling this alcoholic solution or by heat alone is converted into dibrommaleiuimide. The oxime or its anhy- dride is further converted into the acid ammonium salt of dibrom- maleic acid by boiling it with water. The ready formation of oximes of perfectly normal character from mucobromic acid would in itself naturally be considered as establishing the aldehyde character of this acid. Still the difficulty in explaining the conversion of the broman- hydride of an aldehyde acid into a lactone by reduction with stannous chloride in the cold is so great that it seems to us more rational to suppose either that hydroxylamine acts directly upon an oxylactone of this type and that the ordinary oxime is then formed by molecular rearrangement, or that under the conditions of the reaction the hydrox- ylamine acts only through the fixation of a molecule of water, and the breaking of the lactone ring with formation of the aldehyde acid. In this connection we were interested to know whether the esters of mucobromic acid would enter into reaction as readily as the free acid. While it seemed probable that the ester of the aldehyde acid of the form BrC - Cq BrC-COOCH 3 would be attacked by hydroxylamine quite as readily as the free acid, the ester formed from the oxylactone. H BrC No / Br c-co might well prove more refractory. Several years ago Lieberman || tried the behavior of the ethylester of opianic acid with hydroxylamine, * Berichte A. deutsch. chera. Gesellsch., XIX. 2278, 2923. t Ibid., XXIV. 3264. } Ann. d. Chem. u. Pharm., CCXXXIX. 81. § Journ. Chem. Soc, LVII. 1069. || Berichte d. deutsch. chera. Gesellsch., XIX. 2926. HILL AND CORNELISON. — CROTONOLACTONES. 13 and obtained only the opianic oxime anhydride. We have been able to find described no other experiments with the esters of the acids in question. Our own experiments with mucobromic acid, and its methyl ester, showed that the ester was attacked with much more difficulty. While mucobromic acid is rapidly converted into its oxime and oxime anhydride, we failed to discover any appreciable action when hydroxylamine is added to a solution of methyl mucobromate in methyl alcohol at ordinary temperatures even after the lapse of several days. On boiling, the methyl ester of mucobromoxime was formed, identical with the body formed by heating mucobromoxime with methyl alcohol. Further experiments in this direction can alone show whether the difference in behavior between the acids of this series and their esters is sufficiently general to warrant definite conclusions as to their structure. The action of phenylhydraziue upon mucobromic and mucochloric acids and their esters, which we hoped to study in detail, proved to yield bodies which were not suitable for investigation. Many years ago Mr. C. W. Andrews, who was at that time an assistant in this Laboratory, made some experiments as to the action of aniline upon mucobromic acid and its ethyl ester, which he was unable to complete at the time, and which have never been published. He found that the reaction with mucobromic acid was complex, in that bromine was partially replaced by the aniline residue, and also that oxygen was eliminated by condensation. In marked contrast with this reaction was the behavior of ethyl mucobromate with aniline. One atom of bromine was here replaced, but the rest of the molecule remained unaltered, so that the body C4H2Br(NHC6H5)03C2H5 was formed in nearly theoretical quantity. It therefore seemed to us of interest to determine whether ethyl mucobromate would behave in an analogous way with ammonia. We found, however, in this case, that the reaction took a different course, and that the ethoxy group was first attacked. The body which is thus formed has the formula of the amide of mucobromic acid, but its behavior is in some respects anomalous. It dissolves readily in caustic alkalies, and may be reprecipitated unchanged by immediate acidification. It is but slowly converted into mucobromic acid by boiling with mineral acids, and with oxidizing agents it yields dibrommaleinimide. From muco- chloric acid a body of similar properties was obtained. o/3-DlBROMCROTONOLACTONE. Although tribrompyromucic acid is little affected by boiling hydro- chloric acid, it is quite readily decomposed, with the escape of carbonic 14 PROCEEDINGS OF THE AMERICAN ACADEMY. dioxide, when heated to 130° with diluted sulphuric acid (sp. gr. 1.43), or when boiled with concentrated hydrobrornic acid. The reaction seemed to be more neatly effected with the latter reagent, and we therefore heated tribrompyromucic acid with from four to five times its weight of concentrated hydrobrornic acid with a return condenser. When no further escape of carbonic dioxide could be detected, the somewhat dark colored clear solution was cooled, and diluted with water. A heavy crystalline precipitate was thus thrown down, which could easily be purified by recrystallization from alcohol or ligroin. The crude product obtained in this way amounted to about half the weight of the tribrompyromucic acid taken. Analyses of the body after several recrystallizations from alcohol showed that it had the formula C4H2Br20.2. For the analyses of this dibromcrotonolactone, and of several of its derivatives, which we publish, we are indebted to Mr. H. N. Herman, who made a preliminary study of this body seme two years ago, but was unable to continue the investigation. I. 0.2472 grm. substance gave 0.1788 grm. C02 and 0.0209 grm. H20. II. 0.2394 grm. substance gave 0.3711 grm. AgBr. i. ii. c Calculated for C4H,Br202. 19.83 Found. I. 19.73 H 0.83 0.94 Br 66.12 65.96 This same body may also be made directly from mucobromic acid by the action of phosphorous iodide. Phosphorus is dissolved in five times its weight of carbonic disulphide, equivalent weights of iodine and mucobromic acid well ground together are then added, and the nearly solid mass heated on the water bath with reverse cooler. The reaction sets in slowly, and frequently only after a part of the car- bonic disulphide has been allowed to escape through the cooler. With no solvent present the reaction is violent and yields little or no product. When the mass is completely liquefied, the rest of the car- bonic disulphide is distilled off, the flask well cooled, and cold water added in not too small quantity. On shaking, the crude lactone sepa- rates in a granular condition. In this way it is easy to obtain a product which amounts to two thirds of the weight of the mucobromic acid taken, but it contains a decided percentage of iodine, and the preparation of pure dibromcrotonolactone from it is a matter of great difficulty. By fractional crystallization from various solvents we found it impossible to eliminate the iodine, but on distillation with HILL AND CORNELISON. — CROTONOLACTONES. 15 steam the substance first carried over contained but little, and the percentage of iodine did not become large until about one half of the material taken had passed over. The material which had thus been partially purified by fractional distillation with steam, could not be further purified by recrystallization, but after boiling for some time with bromine water the iodine compound was oxidized, and a substance was then obtained which after several recrystallizations from alcohol possessed the properties and the composition of pure dibromcrotonolactone. 0.2439 grm. substance gave 0.3779 grm. Ag Br. Calculated for C4H,Br202. Found. Brm. 66.12 : 65. 93 The dibromcrotonolactone can much more readily be made from mucobromic acid by the reduction of its broman hydride. Muco- bromylbromide was described many years ago by Hill and O. R. Jackson,* who made it by the action of phosphoric pentabromide upon mucobromic acid. We have found that phosphorous tribromide is in many respects more advantageous for its preparation. The reaction runs somewhat slowly at 100°, but so smoothly that one molecule of the tribromide is sufficient for three molecules of mucobromic acid, and the yield is about 90 per cent of the theoretical amount. After the reaction is over the flask is well cooled, cold water is added, and the whole vigorously shaken until the oil which first separates solidi- fies in a granular form. The crude product melts at 54—55°, and by repeated recrystallization from small quantities of hot alcohol this melting point may be raised to 56-57°, one degree higher than the point given by Hill and O. R. Jackson. We attempted to purify the crude product by distillation in vacuo, but we found that the melting point was depressed rather than raised by this treatment. We have not further studied the change which is thus apparently brought about by distillation, for we soon found that a perfectly pure dibromcrotono- lactone could be made directly from the mucobromylbromide, as it was precipitated by water, while it was difficult, if not impossible, to do this with the distilled substance of low melting point. Hydriodic acid, or zinc dust with glacial acetic acid, reduces the mucobromyl- bromide to the lactone, although the former reagent yields a product which contains iodine. A far more efficient and convenient reducing * These Proceedings, XVI. 174. 16 PROCEEDINGS OP THE AMERICAN ACADEMY. a^ent we found to be stannous chloride with hydrochloric acid, since this gives at once an essentially pure product in satisfactory quantity. Mucobromylbromide is added to an equal weight of stannous chlo- ride dissolved in the same amount of concentrated hydrochloric acid. The reduction proceeds rapidly with the evolution of heat, although in working with small quantities the reaction is greatly facilitated by warming gently at first. "When the melting point of the mucobro- mylbromide is reached, the flask must be well shaken until the oil has completely disappeared. As the solution cools the dibromcrotonolac- tone crystallizes out, and still more separates on dilution. The yield which may be obtained in this way amounts to from 75 to 80 per cent of the weight calculated from the mucobromylbromide taken, or about 70 per cent of that theoretically required by the mucobromic acid employed. A single recrystallization from alcohol is sufficient for its complete purification. Dibromcrotonolactone is very sparingly soluble in cold water, and dissolves in from 30 to 40 times its weight of boiling water. As the hot aqueous solution cools, the lactone crystallizes in small six-sided plates, or in pointed prisms crossing at an angle of 60°. From alcohol, in which it is somewhat sparingly soluble at ordinary tem- peratures, although very readily soluble on heating, it crystallizes in bundles of long friable prisms. From concentrated nitric acid it crys- tallizes in large, clear, brilliant oblique prisms. In boiling ligroin it dissolves somewhat sparingly on boiling, and as the solution cools the greater part of it separates in finely felted needles. It is readily soluble in chloroform or benzol, more sparingly soluble in ether or carbonic bisulphide. It melts at 90-91°, and boils under a pressure of 18 mm. at 145°. With steam it volatilizes rapidly. Although it is remarkably stable in acid solutions, it is easily attacked by alkalies. The alkaline hydrates dissolve it, forming deep yellow solutions, and at the same time alkaline bromides are formed. In studying the action of decinormal potassic hydrate in the cold, we have found that approximately two molecules of potassic hydrate are neutralized by each molecule of the lactone taken, while but one molecule of potassic bromide is formed, but have been able to isolate no definite products of the reaction. The lactone was not attacked by hydroxylamine ; with aniline it yielded, with the elimination of bromine, the well crystallized aniline derivative, which will be described later. Phenyl- hydrazine also removed bromine, but gave no well characterized product. HILL AND CORNELISON. — CROTONOLACTONES. 17 Dibromfumaric Aldehyde of Toennies. The melting point and other properties of the clibromcrotonolactone recalled a body of like composition which was obtained in 1879 by Toennies,* through the action of bromine water upon 0-y-dibrornpyro- mucic acid. He gives no analytical data to support his formula, and apparently had but small quantities of material at his disposal. From its behavior he considered the body to be the aldehyde of mucobromic acid, or the double aldehyde of dibromfumaric acid. In 1866 Hill and Sangerf again prepared this body from /3y-dibrompyromucic acid, but were unable to obtain it in satisfactory quantities, and hoped to return to it at some future time. A further study of the reaction has shown us that it is by no means simple in its nature. While we have been unable to obtain a satisfactory yield of the desired product, we have succeeded in preparing an amount amply sufficient for its identifica- tion, and at the same time we have isolated a second product of the reaction. Preliminary experiments showed us that it was most advantageous to add about 20 per cent more than one molecule of bromine to the finely divided dibrompyromucic acid suspended in 20 times its weight of cold water. We therefore boiled 5 grm. of pure /3y-dibrompyro- mucic acid with 100 c.c. of water, and quickly cooled the solution in order to obtain the acid in a finely divided condition. "When the temperature reached 16°, 1.2 c.c. of bromine were added, and the whole well shaken. The color of the bromine gradually faded, but when it had completely disappeared a considerable amount of a well crystallized body remained undissolved. Instead of allowing the solu- tion to stand over night, as Hill and Sanger had done, we filtered out the insoluble substance at the end of three hours, and extracted the colorless aqueous solution thoroughly with ether. The ethereal ex- tract left on distillation a syrupy residue, which partially solidified on standing over night in vacuo over sulphuric acid. The crystals when thoroughly drained upon the pump and pressed weighed 0.85 grm. Two crystallizations from small quantities of hot alcohol yielded 0.65 grm. of substance, which melted at 90-91°, and showed the characteristic behavior of the dibromcrotonolactone. In order to identify it with- precision, we converted it into the a-iod-/3-bromcroto- nolactone, which is fully described later, by heating it with hydriodic * Berichte d. deutsch. cliem. Gesellsch., XII. 1202. t These Proceedings, XXI. 172. vol. xxix. (n. s. xxi.) 2 18 PROCEEDINGS OF THE AMERICAN ACADEMY. acid, and found the product to melt at the proper point, 118-119°. Ou reduction with zinc and acetic acid it yielded the /3-bromcrotono- lactone melting at 57-58°, and with aniline the a-phenylamido-/3- bromcrotonolactone with its characteristic properties, both of which bodies, as we shall afterwards show, may be made in the same way from the dibromcrotonolactone. As we shall show more fully later, we were unable to confirm the statement of Toennies that the dibrom- crotonolactone could easily be converted into mucobromic acid by oxidation with chromic acid. By the action of bromine in aqueous solution a part of the /3y-dibrompyrornucic acid had been converted into dibromcrotonolactone, according to the equation given by Toennies, C6H4Br2Oa + Br2 + H20 = C4H2Br202 + C02 + 2 HBr. The viscous oil which had been drained from the crystalline product deposited a few more crystals of the lactone on long standing, but we have not yet examined it further. The insoluble matter, which had been removed by filtration before the extraction with ether, was washed with a dilute solution of sodic carbonate. But a small amount dissolved, and upon acidification 0.2 grm. of a sparingly soluble acid was precipitated, which melted at 189-190°, and was evidently unal- tered dibrompyromucic acid. The residue, insoluble in the alkaline solution, weighed 1 grm., and consisted chiefly of mucobromylbromide. It gave in alcoholic solution a deep blue evanescent color on the ad- dition of sodic carbonate, and one crystallization from alcohol was sufficient to raise the melting point to 50-57°. A part of the di- bromp3'romucic acid had therefore been decomposed according to the equation C5H4Br,03 + 2 Br2 + H20 = C4HBr302 + C02 + 3 HBr. Mucobromylbromide is so slowly attacked by water that little muco- bromic acid could have been formed under the conditions chosen ; but it is found in abundance when the product of the reaction stands for a long time in dilute solution. The alcoholic mother liquors obtained from the recrystallization of the mucobromyi bromide were evaporated, and the residue boiled with water under a reverse cooler until the bromanhydride had been converted into mucobromic acid. A small amount of insoluble material was then left, which had the odor of tetrabromfurfuran, but we were unable to identify it with precision. A preliminary experiment under essentially the same conditions had already yielded us a crystalline body, which was proved by its melting HILL AND CORNELISON. — CR0T0N0LACT0NE3. 19 point (G4°), and other characters, to be tetrabromfurfuran. Hill and Sanger * had previously found that this substance was formed in considerable quantities when bromine was gradually added to the dibrompyromucic acid suspended in water. Action of Oxidizing Agents. As we already have said, dibromcrotonolactone may be crystallized without alteration from concentrated nitric acid, but on long boiling it is slowly oxidized, and mucobromic acid or dibrommaleic acid formed. Even fuming nitric acid attacks it with difficulty, and after boiling for half an hour the greater part of the lactone taken may be recovered unchanged. We dissolved the lactone in 8 times its weight of concentrated nitric acid (sp. gr. 1.42), and boiled the solution for three hours under a reverse condenser. The unaltered lactone was then driven off with steam, and the acid solution evaporated to small volume. On cooling, mucobromic acid separated in abundance, which melted at 119-120°, after recrystallization from hot water. Even after boiling for six hours with concentrated nitric acid the oxidation was far from complete, but it was then easy to establish the formation of dibrommaleic acid through the melting point (114-115°) of its anhydride. On boiling the lactone with bromine water, it is slowly converted into mucobromic acid. The reaction may be hastened greatly by using concentrated hydrobromic acid as a solvent, and the oxidation is then completed in a comparatively short time. The mucobromic acid which we obtained, when recrystallized from water, melted at 120-121°. With chromic acid the action is also very slow, and we have been unable to identify any products of the oxidation except carbonic dioxide. If an amount of chromic acid was used which corresponded to one atom of oxygen for each molecule of the lactone, several hours at 100° were needed for the complete reduction of the chromic acid. On distillation large quantities of the unaltered lactone were obtained (melting point 90-91°), and the retort residue, when evaporated to small volume, deposited nothing on cooling. This residue was therefore extracted with ether, the ethereal extract shaken with a dilute solution of sodic carbonate, and this alkaline solution immediately acidified with hydrochloric acid, and again extracted with ether. The ether then left upon evaporation a small syrupy residue, which was strongly acid to test paper. After standing for several days, several minute clusters of rhombic plates could be seen under * These Proceedings, XXI. 172. 20 PROCEEDINGS OF THE AMERICAN ACADEMY. the microscope. In appearance they closely resembled mucobromic acid, but they were insufficient in quantity even for a melting point determination. Since the dibromcrotonolactone reduced an ammoni- acal solution of argentic nitrate on heating, we made one attempt to effect its oxidation with argentic oxide. On boiling an aqueous solu- tion of the lactone with a large excess of well washed argentic oxide, metallic silver was formed, but at the same time argentic bromide in large quantity. An approximate quantitative determination showed that 87.0 per cent of the total bromine contained in the lactone had been converted into argentic bromide. After a careful search we failed to find any products of the oxidation except an amount of dibrommaleic acid, which was just sufficient for its complete identifi- cation through the microscopic appearance of its barium salt and the melting point of its anhydride, 114-115°. Action of Bromine. Dibromcrotonolactone is not attacked by bromine at ordinary tem- peratures, but at 100° substitution is quite rapidly effected. If equal molecules of bromine and the lactone are taken, a colorless oil is ob- tained after several hours' heating, which on standing gradually solidi- fies, and the crystalline solid may be purified by recrystallization from small quantities of alcohol. The substance which was thus obtained melted at 56-57°, and gave with alkalies in alcoholic solution the characteristic blue color described by Hill and O. R. Jackson.* Mucobromylbromide had therefore btjen formed from the dibrom- crotonolactone according to the equation C4H2Br202 + Br2 = C4HBr302 + HBr. When heated with an excess of bromine further substitution is effected and a product is obtained in which both the hydrogen atoms of the dibromcrotonolactone are replaced. For the preparation of this body we heated equal molecules of mucobromylbromide and bromine in sealed tube at 125-130°. At this temperature the reaction proceeds rapidly, but it is also completed at 100° on longer heating. The col- orless oil which we obtained gave no blue color in alcoholic solution with sodic carbonate, and on standing gradually solidified. The carefully pressed solid proved to be extremely soluble in alcohol, ether, chloroform, benzol, or carbonic disulphide, but it was somewhat more sparingly soluble in ligroin and could be purified by recrystalli- * These Proceedings, XVI. 175. HILL AND CORNELISON. — CROTONOLACTOXES. 21 zation from this solvent. The percentage of bromine which the sub- stance contained corresponded to the formula C4Br402. I. 0.1 G85 grm. substance gave 0.3177 grm. AgBr. II. 0.1884 grm. substance gave 0.3552 grm. AgBr. Calculated for Found. C4Br402. I. II. 80.00 80.23 80.20 Br This substance crystallizes from ligroin in clustered leaflets which melt at 58-59°. It has a strong suffocating odor like that of the acid bromauhydrides. When heated with water it is slowly dissolved and the solution then contains dibrommaleic acid. The acid was as usual identified by the crystalline form of the barium salt, and by the melt- ing point (114-115°) of its anhydride. On treating the body with stannous chloride and hydrochloric acid the dibromcrotonolactone is again formed by reduction. The mode of formation, the composition, and the behavior of this body justify the conclusion that it is the un- symmetrical form of dibrommaleylbromide. Its behavior will be fur- ther studied in this Laboratory. We have made many unsuccessful attempts to prepare an identical or an isomeric body directly from dibrommaleic acid. We have tried the action of phosphoric penta- bromide upon dibrommaleic anhydride at temperatures which varied from 100° to 225°, and either obtained the unaltered anhydride, or else carbonization ensued. We were equally unsuccessful in our attempts to prepare such a product through the salts of the acid. Action of Aniline. Aniline reacts upon the a/3-dibromcrotonolactone at ordinary tem- peratures and forms a-phenylamido-/3-bromcrotonolactone. This sub- stance is most readily prepared by dissolving the lactone in twenty parts of alcohol, diluting this solution with an equal weight of water and then adding somewhat more than two molecules of aniline. On standing long needles of the aniline derivative separate in abundance, which may be purified by recrystallization from sixty per cent acetic acid, and afterward from alcohol. I. 0.2570 grm. substance gave 0.4426grm. C02 and 0.0763 grm. H20. II. 0.2393 grm. substance gave 0.1775 grm. AgBr. III. 0.2727 grm. substance gave 13.4 c.c. of moist nitrogen at 21° and under a pressure of 763 mm. 22 PROCEEDINGS OF THE AMERICAN ACADEMY. c Calculated for C10H8BrNO2 47.24 I. 46.97 Found. II. H 3.15 3.30 Br N 31.50 5.51 31.56 III. 5.63 a-Phenylamido-/3-bromcrotonolactone is almost insoluble in ether carbonic disulphide, or ligroin, but dissolves somewhat more readily in boiling benzol or chloroform. It is somewhat sparingly soluble even in boiling alcohol, still less freely in boiling water, and as the solutions cool is deposited in each case in long brilliant needles. In hot glacial acetic acid the substance dissolves readily ; from chloroform it crystallizes in thin transparent plates. When slowly heated it melts with decomposition at about 165°, but if the capillary tube containing the substance is plunged into the heated bath it melts promptly at 186-187° and immediately decomposes. In alkaline solutions it dis- solves readily on warming, and if the solution be quickly cooled it crystallizes out apparently unchanged, but on heating for a longer time decomposition ensues, as the strong odor of pheuylisocyanide shows. By the action of sodium amalgam bromine is removed and a-phenylamidocrotonolactone, which we shall describe later, is formed. We have been unable to replace the second atom of bromine by heat- inff with an excess of aniline. Action of Hydriodic Acid. When o/3-dibromcrotonolactone is dissolved in ordinary distillable hydriodic acid and the solution is boiled for a short time, or heated for a longer time at 100°, one atom of bromine is replaced by iodine, and as the solution cools the a-iod-/3-bromcrotonolactone separates in long prisms. This same body is also formed by the action of phosphorous iodide upon mucobromic acid or when hydriodic acid acts upon muco- bromylbromide, and it may easily be made the chief product of the latter reaction. For analysis it was recrystallized from alcohol. I. 0.2393 grm. substance gave 0.1453 grm. C02 and 0.0229 grm. H20. II. 0.1874 grm. substance gave 0.2747 grm. AgBr + 'AgI. III. 0.2031 grm. substance gave 0.2978 grm. AgBr + Agl. C H Br + I Calculated for C4II2BrI02. 16.61 0.69 71.63 I. 16.56 1.05 Found. II. 71.76 in. 71.78 HILL AND CORNELISON. — CROTONOLACTONES. 23 The a-iod-/3-bromcrotonolactoue is quite readily soluble in benzol or chloroform, more sparingly soluble in ether, carbonic disulphide, or lio-roin. It dissolves freely in boiling alcohol, and as the solution cools it is deposited in quite large, oblique, colorless prisms, which gradually change color on exposure. With steam it distils with difficulty. The melting point of most preparations of this substance we have found to be constant at 118-119°. Still we were frequently unable to raise by recrystallization alone the melting point of material which melted at too low a temperature, and on one occasion a preparation made in the usual way melted one degree higher at 119-120°. Toward aqueous alkalies this substance behaves like the corresponding bromine deriva- tive, an alkaline iodide being formed in the decomposition. Concen- trated nitric acid or bromine water on boiling liberates iodine. When heated with one molecule of dry bromine in sealed tube at 100°, iodine is also liberated, and but a small quantity of hydrobromic acid is formed. With aniline in dilute alcoholic solution it yields the a-phe- nylamido-/S-bromcrotonolactoue, which has already been described, al- though not quite as smoothly as the dibromlactone, since dark-colored viscous products, which we have not further examined, are formed at the same time. On reduction with zinc dust and glacial acetic acid, the /3-bromcrotonolactone melting at 58°, which we shall presently describe, was formed. After dilution with water the acid solution was extracted with ether, the ethereal extract washed with a dilute solution of sodic carbonate, and the crystalline residue obtained by the evapo- ration of the ether recrj^stallized several times from small quantities of alcohol. The body thus obtained had the properties of the /3-brom- crotonolactone, melted at 57-58°, contained but an unweighable trace of iodine, and gave on analysis the required percentage of bromine. 0.2078 grm. substance gave 0.2407 grm. AgBr. Calculated for C4H3Br02. Found. 49.08 49.27 Br When the iodbromcrotonolactone was boiled with hydriodic acid the separation of iodine soon ensued. After long boiling with the addi- tion of red phosphorus, although traces of volatile products, which had the odor of fat acids, had been formed, the main product of the reduction was a viscous oily body, which could be extracted by ether from the diluted solution, but from which we were unable to prepare any material suitable for analysis. With distillable hydriodic acid and red phosphorus in sealed tubes at temperatures below 180° we 24 PROCEEDINGS OP THE AMERICAN ACADEMY. obtained substantially the same result, while prolonged heating at 200° brought about decomposition. We had no better success when we employed hydriodic acid saturated at 0°, although in this case decom- position set in at a lower temperature. /?-Bromcrotonolactone. When /38-dibrompyromucic acid is boiled with concentrated hydro- bromic acid, it is slowly decomposed, carbonic dioxide is evolved, and /3-bromcrotonolactone is formed. The reaction does not run as smoothly as it does with tribrompyromucic acid, and more or less car- bonization ensues. When the reaction appears to be completed, the dilution of the dark brown solution usually precipitates a small amount of dark-colored unaltered acid, but no lactone is thus thrown down. The filtered solution is then extracted with ether, the ethereal extract washed with a dilute solution of sodic carbonate, and dried with calcic chloride. The residue left after distilling off the ether gradually solidifies on standing, and the crude product, which amounts to about one quarter of the dibrompyromucic acid taken, may be recrystallized from small quantities of alcohol or from ether. The same body may be prepared much more conveniently by the reduction of the a/3-di- bromcrotonolactone with zinc dust and acetic acid. The dibromcro- tonolactone is suspended in its own weight of 80 per cent acetic acid, and somewhat more than the calculated weight of zinc dust is then added with careful cooling. At first the reduction proceeds rap- idly with the evolution of heat, but several hours at ordinary tempera- tures are necessary for its completion. When the zinc has nearly disappeared the viscous solution is warmed, filtered, and the clear fil- trate cautiously diluted with water. The bromlactone is thus precipi- tated as an oil, which after cooling and shaking soon solidifies in the form of colorless feathery crystals. The precipitated lactone amounts to 45 per cent of the weight of the dibromcrotonolactone taken, and somewhat more may be obtained by extracting the mother liquor with ether. T. 0.2417 grm. substance gave 0.2609 grm.C02and 0.0440 grm. H20. II. 0.1913 grm. substance gave 0.2214 grm. AgBr. Found. II. 49.24 c Calculated for C4H3Br02. 29.45 i. 29.44 H 1.86 2.02 Br 49.08 HILL AND CORNELISON. — CROTONOLACTONES. 25 The /3-bromcrotonolactone is readily soluble in alcohol, chloroform, benzol or carbonic disulphide. It is decidedly less soluble in ether, and very sparingly soluble in ligroin. From small quantities of alcohol it crystallizes in colorless clustered prisms, from ether by slow evaporation in lai'ge transparent six-sided plates. It is quite readily soluble in hot water, and as the hot aqueous solution cools it separates in clear obliquely truncated prisms. It melts at 58°, and boils under a pressure of 18 mm. at 140°. It distils with steam, although with some difficulty. Aqueous alkalies dissolve it with the formation of a deep yellow color, and the solution then contains an alkaline bromide. In dilute alcoholic solution aniline also removes bromine, but forms at the same time a dark-colored viscous oil, from which we have been able to isolate no crystalline product. Action of Bromine. /3-Bromcrotonolactone is but slowly attacked by bromine in the cold. If one molecule of bromine is added to the powdered lactone, a clear deep red solution is soon obtained, but the color fades so slowly that several days are required to complete the reaction at ordinary temper- atures. On opening the tube a small quantity of hydrobromic acid escaped, but the weight of the crystalline product was substantially the same as that of the materials employed. After several recrystalliza- tions from alcohol an analysis also showed that the substance had been formed by the addition of bromine. 0.2965 grm. substance gave 0.5175 grin. AgBr. Calculated for C4H3Br30,. Found. Br 74.30 74.27 This body, which from the mode of its formation must be the a/3o-tri- brombutyrolactone, is readily soluble in alcohol, ether, benzol, carbonic disulphide, or chloroform, but is more sparingly soluble in ligroin. From alcohol it is deposited in large, well formed, brilliant prisms, which melt at 63-64°. On boiling with water a part distils un- changed, but decomposition soon sets in, carbonic dioxide is evolved, and hydrobromic acid is formed. If the /3-bromcrotonolactone is heated to 100° with one molecule of bromine, the color of the bromine soon disappears, but at the same time hydrobromic acid is formed in considerable quantities. With an excess of bromine at 100° mucobro- mylbromide is formed in abundance. The product which we obtained melted at 56-57°, gave in alcoholic solution a deep blue color with so- dic carbonate, and when heated with water yielded mucobromic acid. 26 PROCEEDINGS OF THE AMERICAN ACADEMY. Action of Oxidizing Agents. We have been unable to obtain any characteristic products by the oxidation of the /3-bromcrotonolactone. On boiling with concentrated nitric acid oxidation takes place, but carbonic dioxide is evolved in abundance, and no other definite products were isolated. When heated with bromine in aqueous solution the addition product is ap- parently formed at first, but on boiling this is soon broken up with the evolution of carbonic dioxide. We also noticed in this case the forma- tion of a small amount of a highly crystalline body which on recrystal- lization from alcohol formed long lustrous prisms. The melting point (54° ), the camphor-like odor, and other physical properties, render it extremely probable that this substance was pentabromethan, but it was insufficient in quantity for analysis. The aqueous solution upon evaporation gave a small viscous residue, from which on long standing a few microscopic rhombic plates separated which appeared to be mucobromic acid. o-Bromcrotonolactone. Since the yb dibrompyromucic acid is as yet unknown, and the a bromine atom of the a/3-dibromcrotonolactone seems always to be first attacked by reagents, it was necessary to try some new method for the preparation of the a-bromcrotonolactone. It occurred to us that it was by no means impossible that it might be made from brom- maleylbromide, since the chloranhydrides of similar dibasic acids were known in several instances to yield lactones on reduction. We soon found on trial that our conjecture was correct. Hill and Sanger * had already shown that a crystalline body having the formula C4HBr30o, and yielding monobrommaleic acid when decomposed by water, could be made from /3S-dibrompyromucic acid by the action of bromine in aqueous solution. This brommaleylbromide we therefore prepared, although we modified slightly the method of preparation in that we dissolved the /3S-dibrompyromucic acid in the requisite amount of a di- lute solution of sodic carbonate, and added to this feebly alkaline so- lution slightly more than two molecules of bromine. We attempted to purify the product by distillation in vacuo, and found that it boiled without noticeable decomposition at 124-125° under a pressure of 17 mm. The distillate remained liquid for a long time, and then but * These Proceedings, XXI. 166. HILL AND CORNELISON. — CROTONOLACTONES. 27 partially solidified with the separation of the original substance. The solid portion melted at 54-55°, while the brommaleyl bromide, accord- ing to Hill and Sanger, melts at 55-56°. The liquid portion was much more readily attacked by cold water than the original substance, and gradually dissolved, leaving about one third its weight of the crystalline solid, which had been held in solution. In the aqueous solution was found monobrommaleic acid melting at 130°, and by evaporation monobromfiimaric acid melting at 178° was obtained. These facts seem to us sufficient to show that we had in our hands two different bodies, each of which showed the properties of a brom- maleylbromide. The relation between these bodies will be more fully investigated. For the preparation of the a-bromcrotouolactone we found that no purification of the crude product was necessary. The brommaleylbromide is readily reduced by a solution of stannous chloride in strong hydrochloric acid, but the isolation of the lactone which is formed is difficult. Although it is volatile with steam, it dis- tils slowly, remains dissolved in the distillate, and must be extracted from the dilute solution by ether. We have found it more convenient to use zinc and acetic acid for the reduction. The brommaleylbromide is suspended in one and a half times its weight of 80 per cent acetic acid, and somewhat more than the calculated amount of zinc dust added with careful cooling. The reaction is at first attended with great evolution of heat, and the zinc dust must be slowly added. After the completion of the reaction the addition of water to the filtered solution precipitates but a small part of the lactone, and the solution must be thoroughly extracted with ether. The ethereal solu- tion, when washed with a dilute solution of sodic carbonate and dried with calcic chloride, leaves upon distillation the beautifully crystalline a-bromcrotonolactone contaminated with but small amounts of oily impurities. The thoroughly pressed substance may most readily be purified by recrystallization from ether. This same a-bromcrotono- lactone is also formed by the action of bromine in aqueous solution upon /3-brompyromucic acid. As is the case with the analogous reaction by which n/3-dibromcrotonolactone is formed from /3y-dibrom- pyromucic acid, the yield is small, and we have as yet merely satis- fied ourselves of the identity of the product through its physical properties. 0.2000 grm. substance gave 0.2304 grm. AgBr. Calculated for C4H3Br02. Found. Br 49.08 49.01 28 PROCEEDINGS OF THE AMERICAN ACADEMY. The a-brorncrotonolactone is readily soluble in alcohol or chloroform, somewhat more sparingly soluble iii ether or benzol, sparingly soluble in cold carbonic disulphide, more readily in hot, and sparingly soluble in ligroin. It dissolves quite freely in boiling water and distils slowly with steam. It crystallizes in long transparent prisms of adamantine lustre, which melt at 77°. It dissolves in alkalies with a deep yel- low color and the formation of an alkaline bromide. It also reduces argentic nitrate in ammoniacal solution on boiling. With aniline, when dissolved in dilute alcohol, it gives the well crystallized a-phenyl- amidocrotonolactoue. The a-bromcrotonolactone appeared to be iden- tical with the body of like formula and the same melting point which Hill and Sanger * obtained in small quantity as a by-product in the preparation of the dibrompyromucic acids by the action of alcoholic sodic hydrate upon pyromucic tetrabromide. Still their statements concerning the behavior of their substance toward bromine were at variance with our own observation, and a more vigorous proof of the identity of the two bodies was necessary. Fortunately a small amount (0.125 grm.) of their old preparation still remained, and we were able to establish its identity with our own product, since it gave with aniline the same a-phenylamidocrotonolactone melting at 217-218°, | and was attacked by brornine and water in the same way. Action of Bromine. a-Bromcrotonolactone is very slowly attacked by dry bromine at ordinary temperatures. If one molecule is added, the lactone does not dissolve at first, and the color of the bromine remains apparently unchanged for several days. The reaction then seems to proceed, more rapidly, and at the end of the sixth day we found that the color had faded completely. The tube then opened with pressure, and a large amount of hydrobromic acid escaped. Substitution had evi- dently taken place, and the weight of the liquid product corresponded precisely with that which would theoretically be required by the replacement of one hydrogen atom by bromine. The product of the reaction remained for a long time liquid, and its alcoholic solution gave but a greenish color on the addition of sodic carbonate. After long standing in desiccator a few large clear crystals appeared, which from their melting point (90-91°) and other characters were shown to be the a/3-dibromcrotonolactone. After distilling with steam, in order to remove the dibromcrotonolactone, the aqueous solution left on * These Proceedings, XXI. 158. t See page 31. HILL AND CORXELISON. — CROTOXOLACTONES. 29 evaporation a viscous residue. After long standing in desiccator a few microscopic rhombic plates appeared, which were doubtless muco- bromic acid. On repeating this experiment later, when the mean temperature of the room was much higher, the reaction was com- pleted in about half the time, the product gave a decided blue color in alcoholic solution with sodic carbonate, showing the formation of mucobromylbromide, but on long standing crystals of dibromcrotono- lactone were also formed. On heating with two molecules of bromine for several hours at 100°, mucobromylbromide is formed in quantity. The product gradually solidified on standing, gave an intense blue color with sodic carbonate, and after recrystallization from alcohol melted at 56-57°. Action of Oxidizing Agents. Concentrated nitric acid oxidizes the a-bromcrotonolactone without much difficulty at 100°. Carbonic dioxide is freely evolved, and a small quantity of a volatile oil is formed, which has a suffocating, sharp odor, and contains both nitrogen and bromine. We were un- able to isolate other characteristic products of the reaction. With bromine in aqueous solution we had somewhat better success. The reaction progressed slowly, but after boiling for several hours with an excess of bromine a small amount of an oily product had been formed which was not further examined. The aqueous solution upon evapo- ration gave characteristic crystals of mucobromic acid, which, after re- recrystallization from water, melted at 120-121°. The weight of mucobromic acid thus obtained was, however, barely 20 per cent of the theoretical amount, and carbonic dioxide was also formed in the reaction. Hill and Sanger had found that the body which they de- scribed as having the composition of a bromcrotonolactone and melt- ing at 77° was converted by bromine and water into an amorphous substance insoluble or sparingly soluble in all common solvents. Two experiments which we tried with small quantities of their old prepa- ration failed to confirm their statements. It gave precisely the same results which we had already obtained with the a-bromcrotonolactone. The yield of mucobromic acid given by the reaction is so small, and the quantity of the old material at our disposal was so limited, that we were unable to purify our product sufficiently for a sharp deter- mination of the melting point. Its appearance and behavior left no doubt as to its identity. 30 PROCEEDINGS OF THE AMERICAN ACADEMY. Action of Aniline. If aniline is added to a solution of the a-bronicrotonolactone in dilute alcohol, the bromine is gradually displaced at ordinary tempera- tures, and the corresponding phenylamidocrotonolactone is formed. The somewhat dark-colored crystalline mass, which separates after the lapse of twenty-four hours, may be purified by recrystallization from glacial acetic acid, and afterward from alcohol. The a-chlor- lactone which was described by Hill and L. L. Jackson,* and of which we shall speak later, also reacts as readily with aniline, and naturally gives the same product. The same body may furthermore be made, although with more difficulty, by the reduction of the a-phenylarnido-/3-bromcrotonolactone. The elimination of the bromine takes place so slowly in acid solution that we have found it more ad- vantageous to employ sodium amalgam with dilute alcohol, and to allow the reduction to proceed in alkaline solution at ordinary temper- atures. Although more or less decomposition of the phenylamido- bromcrotonolactone takes place, as the strong odor of phenylisocyanide which is developed shows, a satisfactory product is obtained in this way. The finely powdered a-phenylamido-/3-bromcrotonolactone is suspended in 50 times its weight of about GO per cent alcohol, and an excess of sodium amalgam containing 2 per cent of metallic sodium is added. The substituted lactone gradually goes into solution, and we usually have found the reduction complete at the end of four or five hours. The clear solution is then acidified with acetic acid, the alco- hol driven off upon the water bath, and the crystalline product which separates recrystallized from glacial acetic acid. Occasionally we have found it necessary to treat the product a second time with sodium amalgam in order to remove the last traces of bromine. I. 0.1991 grm. substance gave 0.4996 grm. C02 and 0.09G2 grm. H„0. II. 0.2114 grm. substance gave 14.9 c.c. of moist nitrogen at 21°. 3 and under a pressure of 754 mm. Calculated for Found. CIOH9N02. I. II. c 68.57 68.42 H 5.14 5.36 N 8.00 7.94 * These Proceedings, XXIV. 353. HILL AND COUNELISON. — CROTONOLACTONES. 31 a-phenylamidocrotonolactone is very sparingly soluble even in boil- ing water. It dissolves quite readily in boiling alcohol, more spar- ingly in cold, and very sparingly in hot benzol or chloroform ; in ether, ligroin, or carbonic disulphide, it is nearly insoluble. It is readily soluble in hot glacial acetic acid, and crystallizes on cooling in feather-formed aggregations of branching needles. The melting point of the substance varies somewhat according to the rapidity with which it is heated ; but with proper care to avoid long heating it was found to melt quite sharply at 217-218°, although decomposition ensued on further heating. Material made from the a-chlor- and a-bromlactones, as well as by the reduction of the phenylamidobromlactone, showed identically the same behavior on heating, and when tested side by side in the same bath melted at the same temperature. The body dis- solves readily in hot concentrated hydrochloric acid, but aniline seems to be formed at once. In hot alkaline solutions it dissolves somewhat more readily than in water, and if such a hot solution is quickly cooled it separates apparently unchanged. On longer heating decom- position sets in. A solution of the body in hot dilute baric hydrate was boiled for a short time and distilled. The distillate contained aniline in abundance, as was shown by the formation of tribromaniline melting at 118-119°. At the same time baric carbonate had been precipitated, and in solution was found a barium salt which was ex- tremely soluble in water, was precipitated by alcohol from the con- centrated aqueous solution in an amorphous form, and was left as a brownish varnish upon evaporation. Reduction of the o/3-Dibromcrotonolactone to Crotonolactone. When dibromcrotonolactone is heated to 100° with dilute sulphuric acid and granulated zinc, it is rapidly reduced, and at the end of sev- eral hours the reduction is complete. It was easy to prove by quan- titative determinations that all the bromine of the original substance had been converted into hydrobromic acid, and the behavior of the solution toward decinormal potassic hydrate also showed that it con- tained a lactone in nearly theoretical quantity. AVe found it practi- cally impossible to extract the lactone from aqueous solution, and we therefore distilled this solution in a current of steam. The distillate which we obtained was feebly acid, but it also contained a small quan- tity of a lactone, which, however, we found by titration to be relatively large in proportion to the free acid. The amount of free acid varied 32 PROCEEDINGS OF THE AMERICAN ACADEMY. very much in our preparations, and apparently disappeared entirely, if during the reduction too large an excess of sulphuric acid was not used. With ordinary care the free acid did not correspond to as much as one tenth of the total alkali which was neutralized on boil- ing, and we were unable to increase that amount beyond one fifth by any modification of the mode of reduction. Our determinations fur- thermore showed that each litre of distillate could contain but 1.25 grm. of crotonolactone if this body had really been formed by redaction. We made many fruitless attempts to isolate the crotonolactone from this dilute aqueous solution, and since our main object was to establish beyond question the lactone nature of the dibromcrotonolactone from which it had been formed, we proceeded at last to prepare from the solution a salt of the corresponding oxycrotonic acid. To the dilute lactone solution we added slightly more than the required amount of baric hydrate, heated the alkaline solution to boiling, precipitated after the lapse of some time the slight excess of baric hydrate by car- bonic dioxide, and evaporated to small volume the filtered solution. We obtained in this way a somewhat brown syrupy residue, which on drying yielded a hard varnish. For its purification it was dissolved in a small amount of water, and alcohol added until a portion of the salt had been precipitated. The filtered solution was then concen- trated by evaporation, a trace of a haloid barium salt removed by the cautious addition of argentic carbonate, and the clear solution evapo- rated upon the water bath. After drying for some time at 100°, the gummy salt became sufficiently friable to enable us to powder it, and it was then thoroughly dried at 100°. A complete analysis showed that this salt had the composition of a baric oxycrotonate. The sec- ond determination of carbonate was made in the wet way by the method of Messinger,* the first by the more usual method with potas- sic dichromate in open tube. I. 0.19G1 grm. salt gave 0.2044 grm. C02 and 0.0694 grm. H20. II. 0.1432 grm. salt gave 0.1503 grm. C02. III. 0.2414 grm. salt gave 0.1655 grm. BaS04. m Ba 40.44 40.33 * Berichte d. deutsch. chem. Gesellsch., XXI. 2910. Calculated for Found Ba(C4H,03)2. i. II. c 28.30 28.42 28.62 H 2.95 3.93 HILL AND CORxNELISON. — CROTONOLACTONES. 33 The formation of the crotonolactone through the reduction of the dibromcrotonolactoiie was thus established. Since the salts of the oxycrotonic acid all seemed to possess the same uninviting properties, and our main end was already reached, we made no further study of them. a/3-DlCHLORCROTONOLACTONE. The a/3-dichlorcrotonolactone may be prepared most easily by the reduction of mucochlorylbromide. This body, which has not yet been described, was readily made by the action of phosphorous tribromide upon mucochloric acid. The reaction ran perfectly smoothly, and but little more than one molecule of phosphorous tribromide was needed for complete reaction with three molecules of mucochloric acid. After heating at 100° until the liquefied product grew turbid with the separa- tion of phosphorous acid, the flask was well cooled, and cold water added with constant shaking. The oil which separated gradually so- lidified, and the granular product which was obtained could readily be collected by filtration. As in the preparation of mucobromylbromide we have found it easy to obtain 88-90 per cent of the theoretical yield, and the crude product was sufficiently pure for further use. An analysis of a sample recrystallized from ligroin gave the required per- centage of halogen. 0.2721 grm. substance gave 0.5556 grm. AgCl + AgBr. Calculated for C4HCI202Br. Found. CI + Br 65.09 64.91 Mucochlorylbromide is very readily soluble in alcohol, ether, chloro- form, carbonic disulphide, or benzol ; in ligroin it is somewhat more sparingly soluble. From a solution in dilute alcohol it may be ob- tained by spontaneous evaporation in large transparent plates which melt at 36°. In alcoholic solution it gives with dilute alkalies a deep purple color, which rapidly passes into a wine-red and finally to yellow. With sodic carbonate we have also frequently noticed the formation of a deep green color after the wine-red. The reduction of the mucochlorylbromide is easily effected by means of stannous chloride and hydrochloric acid. The reaction proceeds slowly in the cold, but may be carried to the end at ordinary tempera- tures. On warming gently it runs rapidly with the evolution of heat, and in a short time, if the reaction is promoted by vigorous shaking, the mucochlorylbromide completely disappears. On cooling and add- ing water the dichlorlactone separates in long slender needles, while vol. xxix. (n. s. xxi.) 3 34 PROCEEDINGS OF THE AMERICAN ACADEMY. the mother liquor yields on distillation with steam a small quantity of the same body. The yield amounts to 78 per cent of the weight which could theoretically be obtained from the mucochlorylbromide, or, taking both reactions into account, 70 per cent of the amount which the mucochloric acid employed should give. The substance recrystallized from ligroin gave on analysis the following results : — I. 0.4362 grm. substance gave 0.4955 grm. C02 and 0.0574 grm. H20. II. 0.3097 grm. substance gave 0.5789 grm. AgCl. c Calculated for C4H2C1202- 31.37 Found. I. 30.97 H 1.31 1.46 CI 46.40 II. 46.21 This dichlorcrotonolactone we have also made by the decomposition of trichlorpyromucic acid, and by the action of bromine water upon /3y-dichlorpyromucic acid. Trichlorpyromucic acid appears to be more stable than tribrompyromucic acid. Still it is slowly decomposed with the evolution of carbonic dioxide, when heated to boiling with 50 per cent sulphuric acid, and the dichlorlactone is formed. We have made by this method only a sufficient quantity of material to enable us to identify it with precision. In studying the action of bromine upon /3y-dichlorpyromucic acid we followed the same method that we had employed with the /3y-dibrompyroraucic acid, and found that the re- action followed precisely the same course. In this case, however, we proved the presence of mucochlorylbromide amoDg the insoluble pro- ducts of the reaction only through the characteristic color which was developed upon the addition of sodic carbonate to the alcoholic solution. The dichlorlactone was isolated from the aqueous solution by distilla- tion with steam, as well as by extraction with ether. Its identity was proved by its melting point, and by its conversion into a-phenylamido- /3-chlorcrotonolactone melting at 183°.* a/3-Dichlorcrotonolactone is very readily soluble in benzol or chloro- form, readily in alcohol or ether, somewhat more sparingly soluble in carbonic disulphide, and sparingly soluble in cold ligroin, although more freely soluble in hot. It dissolves quite readily in boiling water, and volatilizes freely with steam. The substance crystallizes ordinarily in long silky needles, but on slow evaporation of the ethereal solution it is deposited in clear six-sided plates. It melts at 50-51°, and boils * See page 36. HILL AND CORXELISON. — CROTONOLACTONES. 35 under a pressure of 18 mm. at 114—115° without uoticeable decompo- sition. With aqueous alkalies it behaves like the corresponding bro- mine compound, and gives a bright yellow solution. Hydroxvlamine has no action upon it, and phenylbydrazine removes chlorine without forming any crystalline product. Aniline when added to its solution in dilute alcohol gives the highly crystalline a-phenylamido-/3-chlorcro- tonolactone. Hydriodic acid gives at 100° the a-iod-/3-chlorcrotono- lactone. Action of Oxidizing Agents. a-/3-Dichlorcrotonolactone is oxidized with even more difficulty than the corresponding dibromlactone. After boiling for several hours with eight times its weight of concentrated nitric acid (sp. gr. 1.42), a large part of the lactone remained unaltered, and could be recovered by distillation with steam. The aqueous solution was then evaporated to dryness on the water bath, and the crystalline residue treated with small quantities of cold water. Mucochloric acid was then left behind, which after recrystallization from hot water melted at 124-125°. The cold water had taken up a small quantity of a readily soluble acid, which undoubtedly was dichlormaleic acid. Since mucochloric acid yields dichlormaleic acid on oxidation with nitric acid, although it is attacked with more difficulty than mucobromic acid, it did not seem to us worth while to take further steps for its preparation in larger quan- tity and identification. Bromine dissolved in concentrated hydro- chloric acid oxidizes the dichlorlactone, although in this case also Ions continued boiling is required to complete the react'on. Action of Aniline. Aniline readily reacts upon the a/3-dichlorcrotonolactone at ordi- nary temperatures, and forms the a-phenylainido-/3-ohlorcrotonolactone. The reaction runs best with a dilute solution in 50 per cent alcohol, and on standing the product separates in broad glistening leaflets, which are more or less highly colored. This color may be removed by recrystallization from glacial acetic acid or from alcohol. 0.1491 grm. substance gave 0.1011 grm. AgCl. Calculated for CinII8ClNO,. Found. CI 16.91 16.76 o-Phenylamido-/3-chlorcrotonolactone is quite readily soluble in boil- ing alcohol, and but sparingly soluble in cold. It also dissolves quite freely in boiling chloroform or benzol, more sparingly in ether or 86 PROCEEDINGS OF THE AMERICAN ACADEMY. carbonic disulphide, and is almost insoluble in ligroin. It is very sparingly soluble in boiling water, and is deposited in the form of fine needles as the solution cools. From chloroform it crystallizes in clear flat prisms, which melt quite sharply at 183°, although decomposition sets in at a somewhat higher temperature. Toward alkalies it behaves like the corresponding bromine derivative. It dissolves more freely in dilute sodic hydrate on heating than in water, and, if this solution is quickly cooled, it crystallizes out apparently unchanged. On longer heating decomposition ensues with the formation of phenylisocyanide. Action of Hydriodic Acid. The action of hydriodic acid upon the aj3-dichlorcrotouolactone we have studied only so far as to prove that a-iod-/3-chlorcrotonolactone is formed. This body we have also made directly from mucochloryl- bromide by the action of hydriodic acid at 100°. In order to keep up the strength of the hydriodic acid we found it advantageous to add at the same time potassic iodide and red phosphorus. The product was precipitated by the addition of water decolorized with sulphurous acid and recrystallized from alcohol. 0.2015 grm. substance gave 0.3116 grm. AgCl + Agl. Calculated for C4H2C1I02. Found. Cl+I 66.45 66.36 a-Iod-/3-chlorcrotonolactone dissolves very readily in benzol, is freely soluble in ether, but is more sparingly soluble in chloroform or carbonic disulphide. It dissolves easily in hot alcohol, but as this solution cools the greater part separates in long flattened needles which melt at 108-109°. In its behavior toward reagents it closely resembles the bodies of similar constitution already described. /3-Chlorcrotonolactone. By the action of zinc dust and glacial acetic acid the a/3-dichlor- crotonolactone may be reduced without difficulty to the /3-chlorcro- tonolactone, but the physical properties of the latter body make its isolation a little more troublesome than that of the corresponding body containing bromine. We dissolved the dichlorcrotonolactone in nearly twice its weight of 80 per cent acetic acid, and slowly added with cooling somewhat more than the theoretical quantity of zinc dust. The reduction was at first attended with the evolution of heat, but afterward became sluggish. After standing over night, the HILL AND CORNELISON. — CROTONOLACTONES. 37 viscous solution, in which some metallic zinc was still suspended, was warmed, filtered, and the clear filtrate diluted with water. As the oil which was then precipitated showed no tendency to crystallize when cooled and scratched, the whole was extracted with ether, the ethereal solution washed with a dilute solution of sodic carbonate, dried with calcic chloride, and the ether removed by distillation. The liquid residue left by the ether was then distilled under diminished pressure. After two distillations the greater part of the product was collected between 128° and 130°, under a pressure of 22 mm., and on cooling this fraction to 10° the larger portion of it solidified. The solid was drained on the pump in a Gooch crucible, and carefully pressed with filter paper. This perfectly dry material was then melted, cooled, and the crystalline solid again pressed, and these operations repeated for a second time without raising perceptibly the melting point. An analysis showed that the body was a chlorcrotonolactoue. 0.3071 grm. substance gave 0.3720 grm. AgCl. Calculated for C4H3C102. Found. CI 29.95 29.95 The /3-chlorcrotonolactone is very readily soluble in alcohol, ether, chloroform, or benzol; less soluble in carbonic disulphide, and spar- ingly soluble in ligroin. It may be crystallized most conveniently by strongly cooling the ethereal solution after the addition of ligroin. It crystallizes in well-formed flat prisms, which melt at 25-26°. The boiling point of the pure substance was found to be 124-125° under a pressure of 18 mm. It is quite freely soluble in hot water, and distils slowly with steam. Toward the alkaline hydrates and aniline it behaves like the /3-bromcrotonolactone. We have not yet studied its behavior toward bromine or oxidizing agents. Although the /35-dichlorpyromucic acid is an extremely expensive substance on account of the difficulties which lie in the way of its preparation, and the small yield which can be obtained, its conversion into this same /3-chlorcrotonolactone by heating with acids seemed to us of sufficient importance to warrant an experiment in this direction. "We therefore heated to boiling 0.7 grm. of the acid with 7 grm. of sulphuric acid of sp. gr. 1.43. Carbonic dioxide was slowly evolved, and the sulphuric acid became strongly colored. After boiling for several hours the solution was cooled, diluted, and extracted with ether. The ethereal extract was then washed with a dilute solution of sodic carbonate, which took up a small amount of the unaltered dichlorpyromucic acid, and dried with calcic chloride. The greater part of the ether 38 PROCEEDINGS OF THE AMERICAN ACADEMY. was removed by distillation, the residue transferred to a weighing tube and heated to 60-80° in vacuo. The liquid residue which remained weighed 0.13 grm., and on cooling to 13° it almost com- pletely solidified. The well pressed crystalline solid remained friable at the temperature of the room (20°) and melted at 25-26°. Ee- crystallization from a mixture of ether and ligroiu failed to raise this melting point. It was thus proved that the /3-chlorcrotonolactone had been formed by the decomposition of the /38-dichlorpyroruucic acid. a-CHLORCROTONOLACTONE. The a-chlorcrotonolactone melting at 52-53° was prepared several years ago by Hill and L. L. Jackson * through the decomposition of the yS(x)-dichlorpyromucic acid. The close relationship between this substance and a body of similar composition containing bromine, which had been discovered by Hill and Sanger,f was recognized at the time, but no attempt was made to establish experimentally this connection. The body containing bromine, which was described by Hill and Sanger, has already been shown to be the a-bromcrotono- lactone, and it is easy to establish the identity in structure of the bromine and chlorine derivatives through the reaction with aniline. The a-chlorcrotonolactone when dissolved in dilute alcohol yields with aniline the a-phenylamidocrotonolactone melting at 217-218°, which we have already described as prepared in the same way from the a-bromcrotonolactone. Since the a-bromcrotonolactone had been made from the /3S-dibrom- pyromucic acid through the reduction of the brommaleylbromide, which is formed from it by the action of bromine, it seemed to us of interest to prepare the a-chlorcrotonolactone by the same method from the /38-dichlorpyromucic acid. Pure /38-dichlorpyromucic acid was suspended in 25 times its weight of cold water, and dissolved by the cautious addition of sodic carbonate. Two molecules of bromine were then added to the feebly alkaline solution, and the whole al- lowed to stand over night. As the oil which had separated would not crystallize when strongly cooled and scratched, it was taken up with ether, and the residue left on the evaporation of the ether dis- solved in 80 per cent acetic acid. Zinc dust was then added to this solution with careful cooling. When the reaction appeared to be complete, the filtered solution was diluted, extracted with ether, the ethereal solution washed with sodic carbonate and distilled. The * These Proceedings, XXIV. 353. t Ibid., XXI. 158. HILL AND CORNELISON. — CROTONOLACTONES. 39 residue soon crystallized in long slender prisms, which when recrys- tallized from ligroin melted at 52-53°, and in other respects showed the behavior of the a-chlorcrotouolactone. The /38-dichlorpyromucic acid, like the corresponding acid containing bromine, may, therefore, be made to yield either the a- or the /3-chlorcrotouolactone. While we had no doubt that the •yS-dichlorpyromucic acid would show an essentially different behavior, and would yield, when treated in this way, the same n-chlorcrotonolactone which it gives when heated with acids, we thought it worth while to establish the fact by experiment. The yS-dichlorpyromucic acid was therefore treated with bromine in a feebly alkaline solution, and the oil, which was thus formed, was then reduced with zinc dust and glacial acetic acid. It was easy to isolate, as before, the a-chlorcrotonolactone melting at 52-53° with its characteristic properties. a-PHENOXY-/3-BROMCROTONOLACTONE. Many years ago Hill and Stevens * prepared from mucobromic acid, by the action of potassic phenylate, a derivative of mucobromic acid, in which one atom of bromine was replaced by the phenoxy group, and to which they gave the name mucophenoxybromic acid.f They showed that this acid had the same general structure as the mucobromic acid, in that it could be converted by oxidation into a phenoxybrommaleic acid, and that it yielded on decomposition with alkalies a phenoxybromacrylic acid. The stability of the latter acid in alkaline solution led them to the conclusion that it contained the phenoxy group in the a position. Although none of the substituted crotonolactones which we had studied could directly be converted into the salts of the oxy-acids, it seemed to us probable that the corre- sponding bodies containing the phenoxy group would react more smoothly. As the preliminary experiments which we made on the action of sodic or potassic phenylate upon the dibromcrotonolactone gave us little hope that the desired bodies could be made in this way, we turned to the mucophenoxybromic acid, and found that the a-phe- * These Proceedings, XIX. 262. t Beilstein in the second edition of his " Handbuch der organisehen Chemie " (II. 429) has seen fit to change this name to phenoxymucobromic acid. It is quite possible that we might have selected a better name for our new body, but it certainly would be difficult to find one more misleading than that which Beilstein has chosen to give it. Whatever else it may be, it certainly is not a phenoxymucobromic acid. H. B. H. 40 PROCEEDINGS OF THE AMERICAN ACADEMY. noxy-/3-bromcrotonolactone could easily be made by the reduction of its bromanhydride. The mucophenoxybromylbromide was not described by Hill and Stevens, but we found that it could be made without difficulty, in nearly theoretical quantity, by the action of phosphorous tribromide. We used but little more than one molecule of the tribromide to three molecules of the acid, and heated at 100°. When the liquid product became turbid with the separation of phosphorous acid, we added cold water with constant shaking. The bromide immediately separated in a granular form, and could be recrystallized from small quantities of alcohol. It formed fine concentrically grouped needles, which melted at 95-96°. 0.2363 grm. substance gave 0.2656 grm. AgBr. Calculated for C4H(0C6Hs)Br202 Found. Br 47.90 47.82 Mucophenoxybromylbromide is readily soluble in ether, chloroform, or benzol, readily soluble in hot alcohol, more sparingly in cold, and sparingly soluble in ligroin. With alkalies in alcoholic solution it gives a yellow color. The reduction of this body to the correspond- ing lactone is slowly but completely effected by stannous chloride and hydrochloric acid at ordinary temperatures, if a little alcohol be added at the same time to facilitate solution. We have found it more con- venient, however, to use zinc dust and acetic acid. Since the lactone is almost insoluble in dilute acetic acid, and an excess of zinc dust apparently does no harm, we suspended the bromide in twice its weight of glacial acetic acid (99.5 per cent), and slowly added with careful cooling an excess of zinc dust. After standing; for several hours, the reduction is complete, and the addition of water to the fil- tered solution throws down a beautifully crystalline substance, which may be recrystallized from alcohol. The product thus obtained proved on analysis to have the composition of a phenoxybromcrotonolactoue. The weight of the lactone as precipitated by water was 80 per cent of the theoretical amount. T. 0.3929 grm. substance gave 0.6796 grm. C02and 0.1008 grm. H20. II. 0.1906 grm. substance gave 0.1410 grm. AgBr. Found. I. II. 47.18 2.85 31.48 Calculated for C4H,(0C6H5)Br02 c 47.07 H 2.75 Br 31.37 HILL AND CORNELISON. — CROTONOLACTONES. 41 The a-phenoxy-/3-bromcrotonolactone is readily soluble in ether, chloroform, benzol, glacial acetic acid, or in hot alcohol, and but spar- ingly soluble in cold alcohol, or in ligroin. It crystallizes in six-sided plates which melt at 72°. It dissolves slightly in boiling water, and volatilizes slowly with steam. In alcoholic solution it is not attacked by aniline at ordinary temperatures. With aqueous alkalies it is slowly carried into solution, on heating, with the formation of the corresponding salts of the oxy-acid. a-Phenoxy-$-brom-y-oxycroto7iic Acid. If the a-phenoxy-/3-bromcro- tonolactone is dissolved in a hot solution of potassic hydrate, and after thorough cooling a slight excess of hydrochloric acid is added, the solu- tion soon becomes filled with colorless pearly scales of the phenoxybrom- oxycrotonic acid. For analysis the substance was merely washed well with cold water, and dried over sulphuric acid. I. 0.2782 grm. substance gave 0.4484 grm. C02 and 0.0857 grm. H20. II. 0.1588 grm. substance gave 0.1092 grm. AgBr. c Calculated for C4H4(OC6HB)Br03. 43.95 Found. I. 43.96 ii. H 3.30 3.08 Br 29.30 29.25 The a-phenoxy-/3-brom-y-oxycrotonic acid is readily soluble in al- cohol or ether, quite readily soluble in hot chloroform or benzol, and almost insoluble in ligroin. From chloroform it crystallizes in bril- liant flat rectangularly terminated prisms. It is very sparingly solu- ble in cold water, and when warmed it dissolves somewhat more freely, but the solution immediately grows turbid with the formation of the still more sparingly soluble lactone. The identity of the phenoxy- bromcrotonolactone formed in this way was established by analysis. 0.2696 grm. substance gave 0.2007 grm. AgBr. Calculated for C4II2(OC6HE)Br02. Found. Br 31.37 31.66 The melting point of the acid cannot be determined with precision, since it is converted into the .lactone by heat, but under ordinary con- ditions it is found to be about 98°. The loss of weight which we observed on melting the acid by short exposure to a temperature of 110°, and allowing the fused mass to stand over sulphuric acid, corre- sponded to a little more than one molecule of water. 42 PROCEEDINGS OP THE AMERICAN ACADEMY. 0.6318 grin, substance lost 0.0450 grm. H20. Calculated for C4U4(OC0H6)BrO3. Found. H20 6.59 7.12 Baric a-phenoxy-fS-brom-y-oxycrotonate, Ba(C4H3(OC6H5)Br03)2 . 3 H20. — The barium salt is readily soluble even in cold water, and crystallizes in flat prisms with rectangular terminations. It apparently contains three molecules of water, almost the whole of which it readily loses over sulphuric acid. 0.5891 grm. air-dried salt lost over sulphuric acid 0.0430 grm. H20, and at 100° it lost 0.0052 grm. H20 in addition. Calculated for Ba(C4H3(OC6H5)Br03)2 . 3 H20. Found. H20 7.35 8.18 0.4939 grm. anhydrous salt gave 0.1677 grm. BaS04. Calculated lor Ba(C4H3(0C6H5)Br03)2. Found. Ba 20.11 19.97 The anhydrous salt dissolves without difficulty in strong alcohol, but in a few moments the solution turns solid with the separation of finely felted needles of a salt containing alcohol of crystallization. The same salt may also be obtained by dissolving the hydrous salt in hot alcohol. We have been unable to obtain satisfactory determinations of the amount of alcohol which is thus taken up, since the salt efflo- resces very rapidly on exposure to the air, and loses the last part of its alcohol with great difficulty. a-PHENOXY-0-CHLORCROTONOLACTONE. It seemed to us desirable to confirm the results which we had ob- tained with the phenoxybromcrotonolactone by preparing also the corresponding bodies containing chlorine. Hill and Stevens had made no experiments as to the action of potassic phenylate upon mucochloric acid, but we found that mucophenoxychlonc acid could readily be made in this way. As we were interested only in the preparation of the phenoxychlorcrotonolactoue, Mr. H. P. Nash undertook a more careful study of this acid, and will present his results in a separate paper. From the mucophenoxychlonc acid we prepared without diffi- culty the corresponding bromanhydride by the action of phosphorous tribromide. The reaction ran smoothly, and gave us, as in the pre- vious cases, which we have described more in detail, about 90 per cent of the theoretical yield. The body was readily soluble in ether, chloroform, benzol, or glacial acetic acid. It was very readily soluble HILL AND CORNELISON. — CHOTONOLACTONE3. 43 in hot alcohol, more sparingly in cold, very sparingly soluble in cold ligroin, though more easily soluble in hot. When recrystallized from small quantities of alcohol or from ligroin, it was obtained in radiating needles which melted at 89-90°. 0.2320 grm. substance gave 0.2631 grm. AgCl + AgBr. Calculated for C4H(0CuHj)C102Br. Found. CI + Br 39.89 39.53 In the reduction of the mucophenoxychlorylbromide to the corre- sponding lactone we followed the method which had already proved convenient with the mucophenoxybroniylbromide. Zinc dust was slowly added with careful cooling to the bromide dissolved in glacial acetic acid. When the reduction was complete, the filtered solution was diluted with water, and the crystalline body which was thus thrown down was recrystallized from a small amount of alcohol. The lactone which was precipitated by water amounted to 76 per cent of the theoretical yield. I. 0.3243 grm. substance gave 0.6765 grm. C02 and 0.0995 grm. H20. II. 0.2256 grm. substance gave 0.1541 grm. AgCl. Calculated for Found. C4H,(0C0H5)ClO,. I. II. C 57.02 56.89 H 3.33 3.40 CI 16.87 16.89 The a-pheuoxy-/3-chlorcrotouolactone is readily soluble in ether, chloroform, benzol, carbonic disulphide, and in hot alcohol or ligroin, although but sparingly soluble in these solvents when cold, and almost insoluble in water. From alcohol it crystallizes in six-sided plates, or flattened prisms, which melt at 67-68°. In alcoholic solution it is not affected by aniline at ordinary temperatures. When heated with aqueous alkalies it is gradually carried into solution with the formation of the salt of the corresponding oxy-acid. a-Phenoxy-fi-chlor-y-orycrotonic Acid, C4H4(OC6H5)C103. — If the a-phenoxy-0-chlorcrotonolactone is dissolved in a moderately dilute solution of potassic hydrate by the aid of heat, and to the well cooled solution a slight excess of hydrochloric acid is added, the phenoxyehlor- oxycrotonic acid separates almost immediately in colorless pearly scales. The crystals were removed by filtration, thoroughly washed with cold water and dried over sulphuric acid. The substance was then analyzed without further purification. 44 PROCEEDINGS OP THE AMERICAN ACADEMY I. 0.4291 grm. substance gave 0.8242 grm. C02 and 0.1541 grm. H20. II. 0.2160 grm. substance gave 0.1342 grm. AgCl. Found. n. c Calculated for C4H4(0CGH5)C103. 52.51 i. 52.37 H 3.94 3.99 CI 15.52 15.36 a-Phenoxy-/3-chlor-y-oxycrotonic acid crystallizes in flat rectangu- larly truncated prisms which melt at about 76°, but the melting point varies with the rapidity of the heating. It is very soluble iu alcohol or ether, quite readily soluble in chloroform or benzol, less soluble in carbonic disulphide or ligroin, and sparingly soluble in water. On heating with water it behaves like the compound containing bromine, which has already been described, and is rapidly converted into the lactone. This conversion is immediately effected when the acid is heated to its melting point. The acid was melted by short exposure to a temperature of 110°, and the fused mass thoroughly dried over sulphuric acid. The loss in weight corresponded almost exactly to that recpaired by one molecule of water. 0.4068 grm. substance lost 0.0333 grm. H20. Calculated for C^HjCOCoH^ClOg. Found. H20 7.88 8.11 The acid may also be converted into the lactone by longer con- tinued heating at a temperature below its melting point, and this change seems to take place slowly even at ordinary temperatures. After the lapse of many weeks a combustion of the material which had previously given us the proper percentage of chlorine showed that the acid had in part been converted into tlie lactone. 0.2004 grm. substance gave 0.4018 grm. C02 and 0.0676 grm. H20. Calculated for Calculated for C4II4(0Cf,H5)C103. C4H2(OC6H5)C102. Found. c 52.51 57.02 54.69 H 3.94 3.33 3.75 Baric a-phenoxy-fi-chlor-y-oxycrotonate, Ba(C4H3(OCeH5)C103)2 • 3 H20. — The barium salt of the acid is very soluble even in cold water, and crystallizes in long colorless prisms. The air-dried salt contains three molecules of water, all of which it loses over sulphuric acid. HILL AND CORNELISON. — CROTONOLACTONES. 45 0.5906 grm. air-dried salt lost over sulphuric acid 0.0492 grin. H20, and wheu heated at 100° it lost in addition 0.0002 grm. II20. ' Calculated for Ba(C4iI3(0C0Hr))Cl03)2 . 3 H20. Found. H20 8.36 8.36 0.4814 grm. anhydrous salt gave 0.1869 BaS04. Calculated for Ba(C4H3(006H5,C103)2. Found. Ba 23.19 22.82 This salt behaves toward alcohol precisely like the baric phenoxy- bromoxycrotonate, and forms the same finely felted needles. Mucobromic Acid. It was shown many years ago by Hill and O. R. Jackson,* that mucobromic acid could be converted into dibrommaleic acid by oxida- tion with argentic oxide. When their investigation was completed (1880), it was already known that phenylhydrazine reacted upon cer- tain aldehydes, but experimental evidence as to the generality of the reaction had not accumulated, and the importance of this body as a reagent for aldehydes and ketones was not urged by E. Fischer f until several years later. The characteristic behavior of hydroxylamine with aldehydes and ketones was also not discovered by V. Meyer t until two years later. Since the facts which we have described in the preceding pages, tended to show that mucobromic acid was an oxylactone and not an aldehyde acid, it was evidently necessary for us to study its behavior toward phenylhydrazine and hydroxylamine. Phenylhydrazine has unfortunately given us no bodies the physical properties of which invited further investigation, but with hydroxyla- mine we have been more successful. We have also studied the action of ammonia upon ethylmucobromate, and found that mucobromamide is thus formed. Mucobromoxime, C4H3Br2N03. — Hydroxylamine acts with great ease upon mucobromic acid, but we have been able to isolate the oxime only when the reaction takes place in an alkaline aqueous solution. The decomposition is almost instantaneous, so that after the lapse of a few minutes the solution deposits upon acidification the * These Proceedings, XVI. 186. t Berichte d. deutsch. chem. Gesellsch., XVII. 572. t Ibid , XV. 1165, 1324, 1527, 2778, 2783; XVI. 822. 46 PROCEEDINGS OF THE AMERICAN ACADEMY. oxime in a granular condition. This granular precipitate was at once removed by filtration, well washed with cold water, and dried over sulphuric acid. I. 0.2397 grm. substance gave 0.1 575 grm. C02 and 0.0415 grm.H20. II. 0.3133 grm. substance gave 0.2043 grm. C02.* III. 0.2524 grm. substance gave 0.3481 grm. AgBr. d. in. c Calculated for C4H3Br2N03. 17.58 i. 17.92 Found. II. 17.58 H 1.10 1.92 Br 58.61 58.68 This oxime is readily soluble in alcohol or ether, and but sparingly soluble in chloroform or benzol. In cold water it is also sparingly soluble, but dissolves readily on boiling. If the hot solution is imme- diately cooled the oxime anhydride separates, but otherwise the acid ammonium salt of dibrommaleic acid is formed, and remains in solution. The oxime has no definite melting point, but is evidently converted into the anhydride by heat. It usually melts, at least in part, at about 90°, but this point varies with the rate of heating. Mucobromoxime Anhydride, C4HBr2N02. — Under ordinary condi- tions the only product formed by the action of hydroxylamine or its hydrochlorate upon mucobromic acid is the anhydride of the oxime. Moreover, if in the preparation of the oxime by the method just described the precipitated substance is allowed to stand for several hours in the mother liquor, the anhydride is formed. The material used in Analysis III. was made by the action of hydroxylamine hydro- chlorate upon a solution of mucobromic acid in strong methyl alcohol ; for the other determinations, preparations were employed which had been made in a solution in dilute methyl alcohol with the addition of one equivalent of sodic carbonate. In each case the reaction ran rapidly at ordinary temperatures, and the anhydride which separated was washed well with cold water, and dried over sulphuric acid. I. 0.2927 grm. substance gave 0.4334 grm. AgBr. II. 0.2393 grm. substance gave 0.3547 grm. AgBr. in. 0.1801 grm. substance gave 0.2670 grm. AgBr. IV. 0.2014 grm. substance gave 10.0 c.c. of moist nitrogen at 19° and under a pressure of 753 mm. V. 0.3127 grm. substance gave 14.5 c.c. of moist nitrogen at 15° and under a pressure of 764 mm. * The hydrogen in this combustion was lost. HILL AND CORNELISON. — CROTONOLACTONES. 47 Calculated for Found. C4HBr,N02. I. II. HI. IV. V. Br 62.75 63.00 63.07 63.08 N 5.49 5.65 5.45 Mucobromoxirae anhydride dissolves quite readily in alcohol, chloro- form, or benzol ; with somewhat more difficulty in ether or carbonic disulphide; and is almost insoluble in ligroin. It crystallizes in den- dritic needles, which melt at about 117-118°, but this melting point varies with the conditions under which it is taken. If the melted sub- stance be heated to a higher temperature, it solidifies again below 140° and melts then for a second time at 218°. This behavior is due to the conversion of the anhydride into the isomeric dibrommalein- imide, which melts, according to Ciamician and Silber,* at 225°. If this isomerization is brought about by plunging the tube containing a considerable quantity of the substance into a bath heated to 115°, the reaction is violent, but also attended with decomposition. The anhydride is sparingly soluble in cold water, although more readily in hot. If the hot solution is quickly cooled the anhydride separates, but on prolonged heating it is completely converted into the acid ammo- nium salt of dibrommaleic acid. This acid was identified through its characteristic barium salt and the melting point (114°) of its anhydride. Methyl Ester of Mucobromoxime, C4H2Br2N03CH3. — Although hydroxylamine hydrochlorate reacts at ordinary temperatures upon mucobromic acid when dissolved in strong methyl alcohol, and forms the mucobromoxime anhydride, if the reaction takes place at the boil- ing point of the methyl alcohol the methyl ester of the oxime results. The same body is also formed when the anhydride is boiled for a short time with methyl alcohol. For its preparation we found it ad- vantageous to dissolve mucobromic acid in rather less than twice its weight of methyl alcohol, to add a little more than one molecule of hydroxylamine hydrochlorate dissolved in the smallest possible quan- tity of water, and to boil with reverse cooler for twenty minutes. On cooling, the methyl ester then separated in abundance. L 0.4336 grm. substance gave 0.3370 grm. CO., and 0.0915 grm. H20. II. 0.2946 grm. substance gave 0.3899 grm. AgBr. Pound. II 56.31 * Berichte d. deutsch. cliem. Gesellsch., XVII. 556. c Calculated for C4H,Br2N03CH3. 20.90 i. 21.19 H 1.74 2.34 Br 55.75 48 PROCEEDINGS OP THE AMERICAN ACADEMY. The methyl ester of mucobromoxime is readily soluble in alcohol, somewhat more sparingly soluble in ether, and very sparingly soluble in chloroform or benzol. It is also very sparingly soluble in cold water, but dissolves freely on heating, and, if the solution is quickly cooled, it separates in finely felted needles. On continued boiling it is converted into dibrommaleinimide, and the same change is also effected by longer boiling with methyl alcohol. The ester melts at 146-147°, and at a slightly higher temperature solidifies with the formation of dibrommaleinimide, which again melts at 218°. The conversion of the ester into the imide may also be brought about by dissolving it in sodic carbonate, and acidifying the solution with hydro- chloric acid. The imide melting at 218° is thus thrown down. If the alkaline hydrates are used to dissolve the ester, great care must be used to avoid an excess. This same ester may also be made by the action of hydroxylamine hydrochlorate upon methyl mucobromate dis- solved in methyl alcohol. We found, however, that substantially no reaction took place when a ten per cent solution of the ester stood at ordinary temperatures for several days, either with free hydroxyl- amine or with its hydrochlorate. The solutions when cooled with ice and salt deposited no crystals in material amount, and on dilution with water the original ester was recovered essentially unchanged. With free hydroxylamine there seemed to be but slow action even on boiling; but on boiling the solution containing the hydrochlorate for half an hour, a copious separation of the methyl ester of the muco- bromoxime melting at 146-147° was obtained on cooling. On boil- ing for a longer time this ester was almost wholly converted into dibrommaleinimide melting at 218°. Although it had been found that the methyl ester of mucobrom- oxime could not be dissolved in alkalies without the formation of dibrommaleinimide, it seemed to us possible that we might form a silver salt. On the addition of an alcoholic solution of argentic nitrate to an alcoholic solution of the ester, no salt was precipitated, and no crystals separated on standing. Alcoholic ammonia was then cau- tiously added, and the first drop brought do*vn a heavy yellow pul- verulent precipitate, although the addition of two molecules was necessary to complete the precipitation. The salt was insoluble in water or alcohol, and was not affected by an excess of ammonia. On aualysis it proved to be an ammonio-silver salt of dibrommaleinimide, corresponding to the ammonio-silver compound of dichlormaleinimide described by Ciamician and Silber.* Nitric acid decomposed it with the liberation of dibrommaleinimide. * Benclite d. deutscli chem. Gesellsch., XVI. 2394. HILL AND CORNELISOX. — CROTONOLACTONES. 49 I. 0.2916 grm. substance gave with HBr 0.1440 grm. AgBr. II. 0.2384 grm. substance gave with HN03 and AgN03 0.2347 grm. AgBr. III. 0.3502 grm. substance gave 22.2 c.c. of moist nitrogen at 19° and under a pressure of 754 mm. Found. III. Ag Calculated for C4Br2N02AgNH3. 28.51 i. 28.37 Found. II. Br 42.23 41.8$ N 7.39 7.37 A salt which closely resembles that which we have just described is precipitated when an alcoholic sokition of argentic nitrate is added to an alcoholic solution of dibrommaleinimide. It can contain no am- monia, however, and is the silver salt of the imide. 0.3153 grm. substance gave 10.7 c.c. of moist nitrogen at 17° and under a pressure of 777 mm. Calculated for 04Br2NO,Ag Found. N 3.87 4.02 The strongly acid filtrate from this salt gave with alcoholic ammo- nia an ammonio-silver salt identical with that prepared from the methyl ester of mucobromoxime. 0.3553 grm. substance gave 21.9 c.c. of moist nitrogen at 17° and under a pressure of 775 mm. Calculated for C4Br,N02AgNH3 Found. N 7.39 7.30 In the preparation of the mucobromoxime methyl ester directly from mucobromic acid, there is also formed an oil, which, although some- what soluble in water, is precipitated by the addition of water to the methyl-alcoholic mother liquors. We have not obtained it in a condi- tion fit for analysis, but as it is readily decomposed by water with the formation of dibrommaleic acid it is doubtless an acid ester of this acid. 3fucobromamide, C4TIBro02NH2 — When ammonia gas is passed into a solution of ethyl mucobromate in dry ether, reaction immediately takes place, and mueobromamide separates. This body may also be pre- pared by the action of alcoholic ammonia upon an alcoholic solution of the ester at ordinary temperatures, but the product so obtained is dark vol. xxix. (n. s. xxi ) 4 50 PROCEEDINGS OF THE AMERICAN ACADEMY. colored and must be treated with bone-black in acid solution, and re- crystallized from hot water. We have also prepared mucobromamide by the action of ammonia upon mucobromylbromide dissolved in anhy- drous ether. I. 0.3496 grm. substance gave 0.2426 grm. C02 and 0.0555 grm. H20. II. 0.3640 grm. substance gave 0.5312 grm. AgBr. III. 0.2033 grm. substance gave 10.7 c.c. of moist nitrogen at 18°. 5, and under a pressure of 769 mm. in. c Calculated for C4HBrO,NH4 18.68 " i. 18.92 Found. II. H 1.20 1.77 Br 62.26 62.44 N 5.45 6.13 Mucobromamide is nearly insoluble in ether, chloroform, or benzol, and is readily soluble in hot, more sparingly in cold alcohol. It dis- solves very readily in hot water, and separates almost completely as the solution cools in colorless dendritic needles, which melt with de- composition at about 170°. Slight impurities apparently lower very greatly this point of melting aud decomposition. Alkaline hydrates added in excess carry it completely into solution in the cold without the formation of any color, but, if an excess be not added, the solution rapidly turns brown or black. From the colorless solution the imme- diate addition of acid reprecipitates the amide unchanged. It may be recrystallized from hot concentrated hydrochloric acid, or from dilute sulphuric acid, but after long heating with concentrated hydrochloric acid, or boiling for a shorter time with distillable hydrobromic acid, mucobromic acid melting at 120° is formed. The amide dissolved readily in fuming nitric acid, and on long standing this solution depos- ited well formed crystals of dibrommaleinimide melting at 218°. Dilute solutions of chromic acid with an excess of sulphuric acid effected little change even on loner boiling, but in concentrated solution oxidation took place and dibrommaleinimide was formed. This same body was also formed when the amide was heated with dry bromine at 100°. We made many attempts to make from the mucobromamide by the action of hydroxylamine, phenylhydiazine, or aniline, the correspond- ing derivatives, but in no case were we successful. Either the amide was recovered unchanged, or in a few cases in working with the free bases at higher temperatures a deeper decomposition ensued. HILL AND C0RNEL1S0N. — CROTONOLACTONES. 51 Mucochloric Acid. Toward hydroxylamine mucochloric acid behaves in essentially the same way as mucobromic acid, although we have been unable as yet to prepare the anhydride of the oxime, the body which was most readily formed from mucobromic acid. The methyl ester of muco- chloric acid when treated with ammonia gave mucochloramide without difficulty. Mucochloroxime, C4H3C12N03 . \ H20. — Mucochloric acid was dis- solved in a small amount of methyl alcohol, and one equivalent each of hydroxylamine and sodic carbonate dissolved in a little water were added. After the lapse of a few minutes the solution was acidified with hydrochloric acid and the oxime immediately separated. Analy- ses of the air-dried material (III.) or of the substance dried over sul- phuric acid (I. and II.) showed that in each case one half-molecule of water was retained. Since dichlormaleinimide was quickly formed at temperatures below 100°, no attempt was made to dehydrate the sub- stance by heat. I. 0.2220 grm. substance gave 0.2014 grm. C02 and 0.0471 grm. H20. II. 0.2008 grm. substance gave 0.2992 grm. AgCl. III. 0.1927 grm. substance gave 0.2851 grm. AgCl. IIL c Calculated for C4II3C1„N03 . A H.,0. 24.87" i. 24.79 Found. II. H 2.07 2.36 CI 36.79 36.58 36.85 Mucochloroxime crystallizes in slender felted needles, which are readily soluble in alcohol or ether, and but sparingly soluble in cold water. It dissolves readily in hot water, and if the solution is quickly cooled, it separates apparently unchanged, but on heating for a short time with water it is converted into the acid ammonium salt of dichlor- maleic acid. The identity of the acid extracted by ether from the acidified solution was established by the melting point of its anhydride (119°). The melting point of the oxime cannot be determined with precision, but it lies in the neighborhood of 90°. The melted oxime passes into the dichlormaleinimide which solidifies and again melts at 175-177°. Methylester of Mucochloroxime, C4H2C12N03CH3. — The methyl ester of mucochloroxime is formed when a cold solution of the oxime 52 PROCEEDINGS OF THE AMERICAN ACADEMY. in methyl alcohol is allowed to stand for twenty-four hours, or when mucochloric acid is dissolved in its own weight of methyl alcohol, hydroxylamine hydrochlorate dissolved in a little water added in some- what more than equivalent quantity, and the whole allowed to stand at ordinary temperatures. If the solution is boiled even for a short time the chief product is dichlormaleinimide. The ester may also be made by the action of hydroxylamine hydrochlorate upon methyl mucochlo- rate, provided the solution in methyl alcohol be allowed to stand for several days. 0.1676 grm. substance gave 0.2437 grm. AgCl. Calculated for C4H2C12N03CU3. Found. CI 35.86 35.94 The methyl ester of the mucochloroxime crystallizes in flattened needles, the melting point of which varies with the mode of heating. When introduced into a heated hath the substance melts promptly when the temperature is held at 135°. At higher temperatures the melted ester again solidifies, with the formation of dichlormaleinimide, which in its turn melts at 173-174°, The ester is readily soluble in alcohol, ether, or chloroform, and but sparingly soluble in benzol or carbonic disulphide. It is also sparingly soluble in cold water, but dissolves more readily in hot water, and, if the solution is quickly cooled, it separates in long felted needles. On longer heating with water it is converted into dichlormaleinimide. The ester may also be converted into the imide by dissolving it in a solution of sodic carbonate, and acidifying with hydrochloric acid, provided the solution is cold, and moderately concentrated, and the acid is immediately added. Mucochloramide, C^ICIoOoNFL. — Mucochloramide is readily formed when ammonia gas is passed into a solution of methyl muco- chlorate in dry ether. The hard dark-colored mass which separates is ground up in a mortar with small quantities of cold water. The greater part of the coloring matter is thus removed, and the residue may then be recrystallized from boiling water. 0.2115 grm. substance gave 0.3606 grm. AgCl. Calculated for C4HC1202NH, Found. CI 42.27 42.15 Mucochloramide is readily soluble in alcohol, sparingly soluble in ether, and almost insoluble in chloroform. It dissolves readily in hot water, and as the solution cools the greater part is deposited in the HILL AND CORNELISON. — CROTONOLACTONES. 53 form of small oblique prisms, which melt at about 16G°. Like muco- bromamide it dissolves readily iu solutions of the alkaline hydrates, and is reprecipitated unchanged if the alkaline solution is immediately acidified. On boiling for some time with concentrated hydrochloric acid, it is converted into mucochloric acid, as was shown by the melt- ing point, 124°. MUCOPHENOXYBUOMIC AdD. Mucophenoxybromoxime, C4H8(OC0H5)BrNO3. — We have found that hydroxylamine readily reacts upon mucophenoxybromic acid, and that the corresponding oxime is formed both in alkaline and acid solution. It may conveniently be prepared by adding to a solution of the acid in 20 times its weight of 50 per cent methyl alcohol the equivalent amount of hydroxylamine hydrochlorate. After standing for a short time at ordinary temperatures the oxime begins to sepa- rate in beautiful clusters of long needles, and in a few hours the reaction is completed. 0.2717 grm. substance gave 0.1791 grm. AgBr. Calculated for C4H3(0CeII5)BrN03. Found. Br 27.97 28.04 Mucophenoxybromoxime is readily soluble in alcohol or ether, and insoluble in chloroform or benzol. It is almost insoluble in cold water, dissolves but sparingly in boiling, and separates apparently unchanged if the solution is quickly cooled. On long boiling it is decomposed, and the solution then contains ammonia and phenoxy- brommaleic acid. The solution was acidified with hydrochloric acid, extracted with ether, and the acid removed from the ethereal solution with sodic carbonate. The alkaline solution then gave on acidification small clustered needles, which melted at 104-105° when rapidly heated. On sublimation a crystalline body was obtained which melted at 91°. According to Hill and Stevens, phenoxybrommaleic acid, when quickly heated, melts at 104-105°, and we found that then- acid when sublimed yielded the anhydride melting at 91°. The oxime shows great variation in melting point (120-135°), according to the mode of heating, and we have not yet been able to show that it is converted into the phenoxybrommaleinimide by heat. The melted body did not solidify on further heating, nor did it crystallize on cooling. 5-4 PROCEEDINGS OF THE AMERICAN ACADEMY. MUCOPHENOXYCHLORIC AdD. 3£ucophenoxychloroxime, C4H3(0C6H5)C1N03. — This body is formed by the action of either bydroxylaniine, or its hydrochlorate, at ordinary temperatures upon mucoplienoxychloric acid. As in pre- vious cases we used diluted methyl alcohol as a solvent. The oxime began to crystallize almost immediately, and the separation was in a short time complete. The material used in Analysis I. was pre- pared by the action of the free base ; that in Analysis II. with the hydrochlorate. I. 0.2025 grm. substance gave 0.1202 grm. AgCl. II. 0.2137 grm. substance gave 0.1294 grm. AgCl. Calculated for Found. C4H3(0C6H5)C1N03. i. II. 14.70 14.68 14.97 CI Mucophenoxychloroxime crystallizes in clustered needles, which melt at 112-125°, according to the mode of heating. It dissolves readily in alcohol or ether, but is insoluble in chloroform or benzol. It is almost insoluble in cold water, dissolves somewhat more freely on heating, and crystallizes out unchanged if the solution is quickly cooled. On longer boiling it is decomposed, and the solution then contains the acid ammonium salt of an acid which we have not yet further examined. RICHARDS. — ATOMIC WEIGHT OP BARIUM. 55 II. CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF HARVARD COLLEGE. A REVISION OF THE ATOMIC WEIGHT OF BARIUM. SECOND PAPER: THE ANALYSIS OF BARIC CHLORIDE. By Theodore William Richards. Presented May 10, 1893. Table of Contents. PAGE Introduction 55 Properties of Baric Chloride ... 56 Preparation of Materials 61 Ratio of Baric Chloride to Baric Sul- phate 67 Solubility of Argentic Chloride . . 69 PAGE Ratio of the Chlorides of Silver and Barium 74 Ratio of Metallic Silver to Baric Chlo- ride 80 The Atomic Weight of Barium ... 84 Introduction. During the time occupied by the recent investigation of baric bro- mide,* similar analyses of baric chloride were being made. The problem involved was far more complicated than the one already pub- lished, because of the solubility of argentic chloride in water ; hence the present paper, which might well have come first, was necessarily postponed. The results of the continued investigation confirmed the high value of the atomic weight of barium, 137.43, obtained from the results of the investigation of baric bromide. The balance and weights, and all the precautions relative to the weighings, have been already described in detail. f The remarks made before about the constancy of the weights, and the various comparisons with the Washington standard, naturally apply with equal force to the present work. At the close of the two years' investigation the weights were again compared, and were found to have remained surprisingly constant in relative value. * These Proceedings, XXVIII. 1. Zeitschr. Anorg. Chem., III. 441 ; IV. 160. t Ibid., XXVI. 242 ; also XXVIII. 5. 56 PROCEEDINGS OF THE AMERICAN ACADEMY. The following atomic weights of various elements entering into the present research will be assumed throughout the paper: — Oxygen =16.000 Hydrogen = 1.0075 Bromine = 79.955 Silver = 107.930 Chlorine = 35.456 Sulphur = 32.065 Properties of Baric Chloride. The properties of baric chloride are so familiar as to need little further description.* The colorless flat tables of the rhombic system contain, as is very well known, two molecules of water together with the usual slight excess. Half of this water is given off at 60-65°, f and the rest at somewhat higher temperatures. All the water is re- gained at ordinary temperatures in ordinary moist air. According to Filholf the crystals possess a specific gravity of 2.66, while according to Joule and Playfair,|| Schiff,§ and Schroeder If they pos- sess a specific gravity of about 3.05. This great discrepancy is too serious to neglect, hence a redetermination was made. The apparatus for the determination of the specific gravity was simi- lar to that already described,** and toluol was again used as the liquid to be displaced. 5.701 grams (in air) of powdered crystallized baric chloride were found to displace 1.5864 grams of toluol at 24°, or 1.8408 grams of water at 4°. Hence the specific gravity sought is 3.097. The reason for at least a part of the discrepancy is undoubtedly to be found in the enclosure of varying amounts of mother liquor iu the crystals. The specific gravity of anhydrous baric chloride has been much more definitely determined. According to Quincke,ft the value of this constant is 3.851 at 0°; according to Favre and Volson,$$ it is 3.844 at 17°; and according to Schroeder,§§ it is 3.879 at 4°. Although without doubt the mean of these values would have served the present purpose, a new determination was made, as much to check the previ- * See Graham Otto (Micliaelis), III. 658; also Watt's Dictionary (Muir), I. 441, etc. t Lescceur, Compt. Rend., CIV. 1511. | Graham Otto (Michael is), loc. cit. || Chem. Soc. Mem., II. 401 ; III. 57, 199. § Annalen,CVIII. 21. "IT Jahresbericht 1879, p. 31. ** These Proceedings, XXVIII. 14. tt Graham Otto (Micliaelis), III. 659. # Ibid. §§ Berichte d. deutsch. c. Gesell. 1878, 2131. RICHARDS. — ATOMIC WEIGHT OP BARIUM. 57 ous determinations in the present research as to serve for the basis of reduction to vacuum. 4.8620 grams (in air) of baric chloride which had been dried to constant weight at 220° were found to occupy the same volume at 24° as 1.2608 grams of water reduced to 4°. Hence the specific gravity of baric chloride under these conditions is 3.856, a value which agrees essentially with those given above. Here, as in the case of baric bromide, the question whether the water of crystallization is absolutely expelled at a red heat is one of the greatest importance. All early experiments have assumed that this is the case, without attempting to disprove the contrary assumption. The first stumbling block in the way of the careful investigation of the problem is the decomposition of baric chloride in the air at high tem- peratures. Marignac's statement that the salt may be ignited at a dull red heat without the loss of a trace of chlorine * has not been con- firmed by the present experience. As a matter of fact, baric chloride loses almost as much halogen as the bromide, upon gentle ignition in the air.f Hence in order to determine the true weight of baric chlo- ride present in all those cases where the salt was ignited in this way, it was necessary to determine the amount of baric hydroxide and baric carbonate existing in the solution of the ignited salt. Pure boiled water which had been prepared solely in platinum was used for mak- ing this solution, and phenol phthalein and methyl orange were used as indicators. Of course, the loss of 0.00355 gram of chlorine would correspond to a correction of -f 0.00055 gram to the weight of the baric chloride if the alkaline earth were in the form of carbonate, +0.00185 gram if it were in the form of hydroxide, or +0.00275 gram if it were in the form of oxide. Since it is somewhat uncertain whether the portion which is determined by phenol phthalein really exists when dry in the form of hydroxide, and not in the form of oxide, the mean of the last two figures, 0.0023 gram, was adopted in the work which follows. In the case of baric bromide, this correction was calculated for the hydroxide. The corrections are in any case so very small that the possible error introduced by either method of procedure is quite unimportant. Another possible error of about the same magnitude is to be found in the fact that pure baric carbonate is not wholly without influence upon phenol phthalein. In order to observe the behavior of baric chloride under gradually * Liebig's Annalen, LXVIII. 215; CVI. 165. t Schultze, Journ. f. prakt. Chem., [2], XXI. 407. 58 PROCEEDINGS OF THE AMERICAN ACADEMY. increasing heat, specimens of very pure material were heated at many different temperatures by means of a Berzelius spirit lamp, and finally raised to redness. The residues were then dissolved in water and the alkalimetric corrections were determined and applied in the manner which has just been described. Below are given the experimental data. (1.) 3.17+ grams of pure baric chloride which had been dried to con- stant weight (for twelve hours) at 260°+ lost 0.27 milligram on heating to somewhat over 300°, 0.18 milligram more on heating to about 450°, and yet 0.30 milligram upon ignition at a medium red heat. To neutralize the baric hydroxide found upon dissolving the salt, about 0.4 cubic centi- meter of the standard solution of hydrochloric acid was required. Since a cubic centimeter of this solution corresponded to a milligram of silver, it is evident that the correction to be added to the weight of baric chloride amounted to 0.09 milligram. (2.) About 3 grams of baric chloride dried at 400° lost no weight on heating to redness. The alkalimetric correction was +0.12 milligram. (3.) About 5.0 grams of baric chloride dried at 400° lost 0.17 milli- gram on heating to dull redness. The alkalimetric correction to be added to the weight of the baric chloride amounted to 0.12 milligram. (4.) 5.46 grams of baric chloride dried at 250° lost 0.66 milligram on heating to dull redness, 0.22 milligram of this loss being accounted for by the alkalimetric correction. (5.) A repetition of Experiment 4 showed a loss of 2.0 milligrams between 250° and dull redness, only 0.10 milligram of this amount cor- responding to the alkaline earth found. (6.) About 4.0 grams lost 1.20 milligrams between 300° and dull red- ness, the alkalimetric correction being only 0.06 milligram. (7.) 2.61 grams of baric chloride dried at 300° lost 0.40 milligram upon being ignited to a bright red heat in a stream of nitrogen. 0.15 milligram of this loss was due to substitution of oxygen for chlorine. (8.) 6.37 grams of baric chloride which had been dried at 200° lost 1.75 milligrams of water upon being heated to 300°±, and 1.30 milli- grams more upon being partly fused at a bright red heat in a stream of nitrogen. The alkalimetric correction was +0.74 milligram.* (9.) 3.6 grams of baric chloride dried at 300° lost 0.32 milligram upon being heated to dull redness, only 0.04 milligram being due to loss of chlorine. (10.) 3 11 grams of baric chloride which had been prepared by drying large clear crystals at 160° lost 0 70 milligram on heating to dull redness. The alkalimetric correction was 0.29 milligram. (11.) 2.65 grams heated to dull redness lost 0.17 milligram upon * See page 77 of this paper. RICHARDS. — ATOMIC WEIGHT OF BARIUM. 59 being further heated to bright redness. The alkalirnetric correction was +0.25 milligram. (12.) 1.5 grams heated at 250° lost 1.0 milligram on fusion in a stream of dry pure hydrochloric acid gas. No trace of alkali or acid was found in the concentrated solution of the residue, but the addition of 0.10 cubic centimeter of hundredth normal hydrochloric acid made the neutral solu- tion perceptibly acid to methyl orange. (13.) 5.36 grams dried at dull redness lost no weight on fusion in hydrochloric acid at bright redness. The residue was wholly neutral. (14.) 3.92 grams dried at dull redness lost 0.13 milligram on fusion in hydrochloric acid. The heat was so great that the crucible was attacked. (15.) 4.28 grams of baric chloride dried at dull redness gained 0.23 milligram on partial fusion in hydrochloric acid. This specimen, as well as the previous one, was absolutely neutral. Action of Heat on Baric Chloride. No. of Exp. About 200°. About 29(P. About 300 \ About 400 J. Dull Redness. Bright Kedness. Fused in HCI. 1 Per cent. Per cent. +0.021 Per cent. +0.012 Per cent. +0.006 Per cent. -0.003 Per cent. Per cent 2 -0.004 -0.004 3 +0.001 -0.002 4 +0.009 -0.004 5 +0.038 -0.002 6 +0.029 -0.002 7 +0.014 -0.006 8 +0.036 +0.009 -0.011 9 +0.008 -0.001 10 +0.013 -0.009 11 -0.003 -0.010 12 +0.066 0.000 13 0.000 0.000 14 • +0.003 0.000 15 -0.005 0.000 Aver. +0.038 +0.023 +0.014 +0.001 -0.003 -0.009 0.000 60 PROCEEDINGS OF THE AMERICAN ACADEMY. The foregoing table contains a summary of all the experiments, reduced to a common basis. Thus the column headed 'k Dull Red- ness " shows that on the average baric chloride ignited at about 500° or 600° loses three parts in one hundred thousand of its weight. The true weight of the baric chloride was computed in each case by add- ing the alkalimetric correction to the last weight taken ; and then the percentage of excess or deficiency in weight was computed for each temperature and tabulated in its appropriate place. The variations noticeable in the amounts of water retained at the lower temperatures are probably in great part due to the varying hygroscopic condition of the atmosphere at the time of drying, as well as to inaccuracies in the temperatures recorded. There are necessarily three standards for determining the true weights of baric chloride forming the basis of the tahle above : one involves the weight of the salt ignited at dull redness plus the alkali- metric correction (Experiments 1, 2, 3, 4, 5, 6, 9, and 10) ; another involves the weight of the salt ignited at bright redness plus the same correction (Experiments 7, 8, and 11); and the third involves the weight of the salt fused in hydrochloric acid, without any correction (Experiments 12, 13, 14, and 15). A comparison of the results of Experiments 11, 13, and 15 with all the earlier experiments is suffi- cient to show that the three standards are very nearly, if not quite, identical ; in other words, that baric chloride which has been dried at dull redness, or even at 400°, does not lose an essential amount of water upon being heated nearly to 900°. Nevertheless, below 400° — which is about the point at which the salt begins to show signs of partial decomposition — very noticeable amounts of water appear to be held by the so called anhydrous salt. The conclusion regarding the expulsion of water from the baric chloride is so important that it seemed desirable to prove it in another way. Hence two determinations of the water of crystallization of perfectly homogeneous powdered pure baric chloride were made by two distinct methods : — (16.) In the first case 3.10784 grams (in vacuum) of the powdered crystals were dried at a medium red heat by means of a spirit lamp to the constant weight of 2.64851 grams (in vacuum) and cooled in a vacuum over phosphoric anhydride. The solution of this salt required 1.00 c.c. of standard hydrochloric acid (of which the cubic centimeter corresponded to a milligram of silver) to render it neutral to phenol phthalein, and 0.85 c.c. more to reach the neutral point with regard to methyl orange. RICHARDS. — ATOMIC WETGHT OF BARIUM. 61 Hence the alkalimetric correction is found to be 0.00025 gram, and after adding this weight to the observed weight of the dried salt, the true per cent of water present in the original crystals is found to be 14.7717. (17.) Again, 5.02356 grams (in vacuum) of the same sample were dried in a platinum crucible at about 400° by means of the spirit lamp, and then fused in a stream of pure dry hydrochloric acid gas in a double crucible. The constant outflow of hydrochloric acid was sufficient to prevent the absorption of sulphuric acid from the products of the com- bustion of the illuminating gas, hence it was possible to use this source of heat. Indeed, the highest temperature obtainable by means of a powerful Bunsen burner is required to fuse the salt. After fusion, the salt was found to weigh 4.28148 grams (in vacuum), having lost 14.7720 per cent. These results (14.7717 and 14.7720) agree within the limit of the probable experimental error ; therefore very satisfactory confirmation is siven to the conclusion arrived at before. Hence baric chloride is a compound of sufficient definiteness to form a very satisfactory basis for the determination of the atomic weight of barium. The other properties of the salt — its solubility in water and insol- ubility in alcohol, the specific gravity of its solution, and so forth — have been studied by Gay Lussac, Gerlach, Karsten, Kopp, Mulder, and many others, and further remark upon them is unnecessary here. Preparation of Materials. Baric Chloride. — The ready crystallization of baric chloride upon the addition of alcohol to a strong aqueous solution of the salt fur- nishes an admirable method for the preparation of baric chloride in a pure state. This method of crystallization was freely resorted to in preparing the pure salt used in the preliminary experiments ; but the details of the preparation need not be dwelt upon. The spectro- scopic test of the purity of the salt has already been described at length.* The purest specimens of baric chloride used in the final experiments were prepared by three methods. According to the first of these methods, after the salt had been crystallized two or three times from water, it was precipitated five times in succession by the addition of strong pure hydrochloric acid to its concentrated aqueous solution. Each time the mother liquor was wholly discarded, and the crystals were carefully drained. The last precipitate was then dis- solved in water, and reprecipitated twice successively by means of the * These Proceedings, XXVIII. 7. 62 PROCEEDINGS OF THE AMERICAN ACADEMY. purest alcohol. These and the subsequent operations were conducted in platinum vessels. A portion of the salt thus formed (I. a) was used for Analyses 22, 25, 31, and 32. The remainder of the crystalline powder was washed four times with pure alcohol on the filter pump, dried, ignited, and fused in a large platinum crucible. The trans- lucent cake of anhydrous baric chloride was dissolved in the purest water, and the solution was allowed to stand until it had become neutral. After having been filtered from the baric carbonate, which was of a faintly brownish tinge because of the presence of impurity taken from the platinum during the fusion, the baric chloride was precipitated twice more by means of alcohol. The last crystals were dissolved in the purest boiling water, recrystallized by cooling, and finally washed once with cold water. The very pure salt thus formed was used for Analyses 20, 21, 33, and 35 (Sample I.c). The last mother liquor was freed from most of its salt by the addition of pure alcohol ; and the very small amount of salt remaining in solution was recovered by the evaporation of the large volume of alcoholic solu- tion and analyzed (Sample Lb, Analysis 34). The second chief specimen of baric chloride was treated in a different way. A large quantity of the commercially pure salt (" purissimum ") was dissolved in water, and was then evaporated nearly to dryness with excess of pure hydrochloric acid on a steam bath, to decompose the thiosulphuric acid which is usually present. The hot dilute solution of the residue was treated with an excess of hydrogen sulphide, and the whole allowed to stand in the dark for a long time in a tightly corked stout flask. The clear supernatant liquid was siphoned off from the trace of precipitate and made alka- line with baric hydroxide which had been recrystallized six times.* After a suitable time the clear liquid was again decanted and filtered from the sulphides insoluble in the alkaline solution, and all the sulphur from the filtrate was expelled by boiling with hydrochloric acid. The resulting solution was of course free from the heavy metals, but it contained marked traces of calcium, strontium, sodium, potassium, and possibly magnesium. After two successive crystalli- zations from water the salt was still found to contain very small amounts of these metals, but after two precipitations by alcohol the spectroscope was unable to show any impurity in it except a trace of the ever present sodium, upon very careful testing. The two hundred grams of pure salt thus obtained were then ignited in small portions * See these Proceedings, XXVIII. 19. RICHARDS. — ATOMIC WEIGHT OF BARIUM. 63 at a dull red heat. The residue was dissolved in a platinum dish and crystallized, and the crystals were dissolved in pure water. After filtration the solution was acidified by hydrochloric acid and again brought to crystallization. A portion of the salt thus formed (Sample Il.a) was used for Analyses 19, 26, 30, and 36, while the remainder was a^ain ignited in platinum, and was recrystallized twice more after filtration and very faint acidification with hydrochloric acid. The last mother liquor was evaporated to dryness in platinum, and served for Analyses 23, 27, and 37 (Il.b). The purest crystals were partly dried over a steam bath, powdered in an agate mortar, and exposed to a moist atmosphere until constant in weight. These crystals (Sample IT.c) served for Analyses 28, 29, 38, 39, 40, and 41. An essentially similar preparation, containing somewhat more occluded water, was used for Analyses 42 and 43. A third sample of baric chloride was prepared from a part of the carbonate used for the preparation of baric bromide, and hydrochloric acid which had been distilled twice in platinum. This was purified in the usual fashion, and served for Analysis 24. These preparations each gave every evidence of being pure. After most careful fractional separation of the greater part of the barium no trace of calcium or strontium could be found, even in large samples, by means of the spectroscope. The trace of sodium visible undoubt- edly crept in during the fractional treatment used to prepare the specimen for spectroscopic examination, for the salt itself before the fractionation showed no more sodium than the surrounding air. The effort was made wholly to free the salt from silica by means of re- peated ignition and solution, and the use of platinum vessels ; but no proof can be given that it did not contain one or two parts in a hundred thousand of this impurity. The quantitative comparison of these preparations which is made further on, furnished conclusive proof of their similarity, at any rate. If the results are reduced to the basis of 100.000 parts of argentic chloride, and arranged accord- ing to the sample of baric chloride used, the following averages are obtained.* * From these determinations all those made according to Stas's first and second methods have been omitted, since these methods proved to be unsatis- factory. Hence Preparation I.b is not represented in the table. The high result from Sample Il.a (Experiments 26 and 36 are parts of the same analysis) was due to a slight mechanical error. 64 PROCEEDINGS OF THE AMERICAN ACADEMY. Parts of Baric Chloride corresponding to 100.000 parts of Argentic Chloride, Preparation I.a, Experiments 22,25, 72.G49 tt I.c, u 20, 21, 35, 72.651 a Il.a, << 26, 36, 72.668 H Il.b, it 23, 27, 37, 72.650 it II.c, l( 28, 29, 38, 39, 40, 41, 42, 43, 72.657 « III. it 24, 72.650 Total average, 72.654 It may be remarked that the purest specimen (Preparation II.c) gave an average result which is very close to the average of all the specimens. Silver. — A number of different preparations of this metal, which served as one of the fundamental standards of reference, were used during the course of the work. For the final experiments pure argentic chloride remaining from earlier work was reduced by means of pure sodic hydrate and invert sugar, the invert sugar having been prepared in the first place by heating a strong solution of the purest obtainable cane sugar with a little hydrochloric acid at 100°. After a very thorough washing the metal resulting from this reduction was fused upon hard-wood charcoal in the flame of an ordinary blast lamp, and divided into two parts. One half of this silver was purified at once by electrolysis accord- ing to the manner described in a former paper,* and was used in Analyses 30, 31, 32, 35, 36, and 37. The other half was dissolved in the purest nitric acid, and precipitated again with hydrochloric acid; then the argentic chloride was digested with aqua regia, very thor- oughly washed with water, and again reduced to the metal. The last product was fused into large buttons upon sugar charcoal, and puri- fied by electrolysis as before. The current obtained from seven or eight gravity cells is amply sufficient for the purpose. This doubly refined silver was used in Experiments 33, 34, 38, 39, 40, 41, and 43. The silver used in Experiment 42 was prepared in the first place by Mr. H. F. Brown, according to the cuprous amnionic sulphite method of Stas.f Although the substance thus prepared gave every outward evidence of perfect purity, it was once again purified, this time by means of the electrolytic method used in the case of the previous specimens. * These Proceedings, XXVIII. 22. t Untersuchungen, etc. (Aronstein), pp. 34 and 113. RICHARDS. — ATOMIC WEIGHT OF BARIUM. 65 The three samples of beautiful crystals of silver thus obtained by- electrolysis naturally contained a minute amount of imprisoned mother liquor, which is best expelled by fusion. Many precautions were taken to prevent the absorption of an impurity during this process. In the first trials pure hard-wood charcoal was used to support the metal during its fusion in the flame of the ordinary blast lamp (Ex- periments 30, 31, and 32). Afterwards a cupel of sugar charcoal was substituted for the wood charcoal. Such cupels are readily made by mixino- finely powdered pure sugar charcoal with about a third of its weight of pure powdered sugar, and igniting in closed moulds. The silver used in Analyses 35, 36, 38, 42, and 43 was fused in this way. A possible absorption of sulphur from illuminating gas being feared, a number of buttons of silver were fused upon the sugar charcoal cupels by the use of a blast lamp in which only pure alcohol, vapor- ized in a steam bath, was burnt. Such silver was used in Experi- ments 33, 34, 37, 39, and 40. All of these specimens of silver had been cooled with care in the centre of a strongly reducing flame, to prevent the absorption of oxygen ; and any button which showed a trace of such absorption was of course discarded. Proof has already been given * that no oxygen is held after such treatment ; but in order to '■ make assurance doubly sure " three grams and a half of the purest crystals were placed upon a boat made of sugar charcoal, enclosed in a stout porcelain tube, and fused in a vacuum (Experi- ment 41). The results obtained from these different varieties of silver may well be recorded here, although the details must be reserved until nearly the end of the present paper. The first specimen of silver fused upon wood charcoal was used in experiments which were vitiated by other serious errors, so that the results from it need not be given. The figures given below represent the amount of silver required exactly to precipitate 100.000 parts of baric chloride.f (1.) First specimen, sugar charcoal, illuminating gas, 103.600 (2.) " u alcohol, 103.604 (3.) Second " (( illuminating gas, 103.600 (4.) << alcohol, 103.599 (5.) « «( fused in vacuum, 103.599 (6.) Third " « illuminating gas, 103.600 Average, 103.600 * These Proceedings, XXV. 198. t The results obtained from Stas's second method, described later, are omitted from this series. vol. xxix. (n. s. xxi.) 5 66 PROCEEDINGS OF THE AMERICAN ACADEMY. Of these figures (2), (5), and (fi) represent single results, and the other figures represent averages. Since a number of samples of baric chloride and a number of different methods of comparison were used in the individual cases, the averages, which to some extent eliminate the differences, are more reliable than the single results. It is evident that the different varieties of silver were quite as nearly alike as the accuracy of our present processes demands. The investigation of the silver closed with a direct determination of the amount of sulphur present in a sample which was purposely made under conditions as favorable as possible for the retention of this impurity. Twelve grams of silver, which had been prepared by the sulphite method and fused three times successively for a long period in the flame of an ordinary blast lamp, were dissolved and precipi- tated by electrolysis in a solution of silver nitrate prepared from the same silver. The residual electrolyte, which must have contained all the sulphur, was much diluted, freed from silver by the addition of hydrochloric acid, and evaporated on the steam bath to free it from nitric acid. This solution yielded 1.4 milligrams of baric sulphate, corresponding to 0.2 milligram of sulphur. Hence under these con- ditions the silver held only g-Q.V 07 °f i*8 weight of sulphur ; and it is fair to conclude that according to the usual method of subjecting the silver only to a single brief fusion not more than 200,000 °f lts weight would be absorbed. This result is in entire accordance with the quantitative and qualitative results already described ; it shows that silver fused once in illuminating gas may be used without appreciable error for all ordinary work. The preparation of pure water, alcohol, sulphuric and nitric acids, and all other necessary materials, has been discussed in previous papers.* Standard Solutions. — Several standard solutions were so con- stantly used during the research that they may well be described at the beginning. In the first place, exactly a gram of silver was dis- solved in a slight excess of nitric acid, with the usual precautions,! and diluted exactly to a litre. A solution of pure hydrochloric acid was then made of equivalent strength by comparison with pure sodic carbonate, and verified by comparison with the silver solution. This last comparison is not easily made, for the argentic chloride separates but slowly from a solution so dilute. It was found most convenient * These Proceedings, XXVI. 245-249 ; XXVIII. 23. t Ibid., XXV. 198; XXVIII. 24. RICHARDS. — ATOMIC WEIGHT OF BARIUM. 67 to add a very slight excess of silver, to warm, agitate, and filter the solution, and to determine the excess in the filtrate by means of an equivalent solution of ammonic sulpho-cyanate, after the method of Volhard. Sanger* has pointed out that the solubility of argentic chloride seriously affects the accuracy of this method unless the argentic chloride is filtered off before the addition of the sulpho- cyanate ; and the experience of this Laboratory had independently led to the same conclusion. It is to be hoped that Clemens Winkler observed this precaution in his interesting analyses of cobaltous and nickellous chloride ; f if not, the atomic weights obtained are probably too high. The three equivalent solutions of argentic nitrate, hydrochloric acid, and ammonic sulphocyanate each corresponded exactly to a milligram of silver for each cubic centimeter. These solutions are referred to whenever a standard solution is mentioned below. It is needless to say that the measuring apparatus was all calibrated with great care ; moreover, for the final experiments the solutions were weighed in small flasks provided with tightly fitting graduated pipettes, as well as measured. This last precaution was hardly neces- sary, however. Of course, the ordinary methods of volumetric analysis are not applicable to the determination of atomic weights ; but when the greater part of a reagent has been weighed out, the last few milli- grams may be added by the measurement of a very dilute solution with an accuracy quite equal to that obtainable upon a balance. $ Ratio of Baric Chloride to Baric Sulphate. It is well known that such a ratio as that of baric chloride to baric sulphate could form at best but a very poor basis for determining the atomic weight of barium. § An experimental error in the process is necessarily magnified many times when referred to the constant sought. Nevertheless, since five distinguished chemists have tried this method, it was thought expedient to attempt it, in order to show if possible the cause of the great discrepancies in their results. The investigation of the ratio occupied several months, but the re- port of it will be greatly abbreviated. The principal cause of error was * These Proceedings, XXVI. 34. t Zeitschr, Anorg. Chem., IV. 10. t This point, among others, seems to have been misunderstood by G. Hin- richs, of the College of Pharmacy, St. Louis. (Chem. News. LXVTII. 171.) § Ostvvald, Allgemeine Chemie, I. 23. These Proceedings, XXVIII. 11. 68 PROCEEDINGS OF THE AMERICAN ACADEMY. found in the very decided occlusion of baric chloride by baric sulphate, a fact which has already been discussed in a previous paper.* This cause of error was never wholly avoided, although by repeated treat- ment of the baric sulphate with sulphuric acid, all but a few tenths of a milligram of chlorine were found to be expelled. Whether or not the baric chloride is occluded in the form of a so called " solid solu- tion," as E. A. Schneider suggests in a recent paper with regard to the occlusion of ferric sulphate by baric sulphate, f the fact remains that an appreciable amount of baric chloride is very firmly held. Possibly nothing short of the true solution of all the baric sulphate in sulphuric acid could drive off all the chlorine ; and such treatment would be likely to introduce errors as large as the one which it would obviate. Besides this inclusion of baric chloride, which tends to raise the apparent atomic weight of barium, another circumstance, the solu- bility of baric sulphate in water, may work in the same direction. This second cause of error was quite eliminated in the present work. A third cause of error with an opposite tendency exists; namely, the obstinacy with which baric sulphate retains the last traces of the ex- cess of sulphuric acid and water used to precipitate it. It must be borne in mind that the retention of 0.2 milligram of either substance counterbalances numerically the retention of nearly two milligrams of baric chloride. Two series of determinations were made, one by the direct addi- tion of pure sulphuric acid to a concentrated solution of pure baric chloride in a platinum crucible, the other by precipitation in dilute solution after the usual fashion. In the latter case the baric sulphate in the filtrates, usually amounting to about a milligram and a half, was determined by evaporation in large platinum dishes. In each series every specimen of baric sulphate was of course heated with successive drops of pure sulphuric acid at a dull red heat until the weight became constant. Eight experiments which were not compli- cated with mechanical errors are recorded in the note-book. The two series gave the same result, a hundred parts of baric chloride yielding 112.073 parts of baric sulphate. The highest result was 112.087, and the lowest 112.060. Berzelius found 112.17.3, Turner 112.19, Thomson about 112.15, Struve 112.094, and Marignac 112.011. * These Proceedings, XXVI. 258. t Zeitschr. fur Physikal. Chem. (1892), X. 425. RICHARDS. — ATOMIC WEIGHT OF BARIUM. 69 Owing to the chemical errors which have just been pointed out, the results given above are of no real value. For this reason the indi- vidual data are omitted. The uncertainty of the result 112.073 is probably as much as a unit in the second decimal place ; therefore the atomic weight of barium computed from it, 137.43, cannot be re- lied upon to within 0.2. The different errors tend to eliminate one another, so that the result is by chance near the true one. Solubility of Argentic Chloride. Since it is evident that no satisfactory results are to be obtained from the ratio just discussed, the careful investigation of the ratios of silver and argentic chloride to baric chloride assumes even greater importance than before. The solubility of argentic chloride is the most serious difficulty which is to be overcome in this investigation. Gay Lussac knew that the filtered liquid which remains after mixing equivalent amounts of solutions of argentic nitrate and a chloride invariably gives a precipi- tate with an excess of either reagent, showing that some argentic chloride must remain in solution. Mulder, in 1857,* recognized very clearly the same fact, and pointed out, with the help of a great num- ber of detailed experiments, just what influence this fact has upon the titration of silver by Gay Lussac's method. Mulder wrongly ascribed the solubility of argentic chloride to the presence of the alkaline nitrate resulting from the decomposition, for he believed that the curdy precipitate was wholly insoluble in pure water, or in water acidified with a reasonable amount of nitric acid. Six or eight years later Stas,f without knowing of Mulder's work, clearly recognized the difficulty, but fell into serious blunders in in- terpreting it. Before 1872, however, when his noted researches on the " chemical statics " of argentic chloride and bromide began to appear,J he had become familiar with Mulder's work, and had adopted many of his points of view. In these last reseaches Stas found that freshly precipitated volumi- * Mulder, " Essayeer-Methode van bet Zilver," Scheikundige Verhande- lingen en Onderzoekingen, 1 Deel, 1 Stuk, 1857. Translated by Grimm, " Die Silberprobirmethode," Leipzig, 1859. t " Untersuchungen," etc., translated by Aronstein (Leipzig, 1867), pp. 46, 56, 59, and 295 | Annates de Cliim. et de Pliys., 4th series, XXV. 22 ; and 5th series, III. 145 and 289. 70 PROCEEDINGS OP THE AMERICAN ACADEMY. nous argentic chloride was soluble to a very large extent (over ten milligrams to a litre) in the purest water, and that nitric acid or the alkaline nitrates had no appreciable effect upon the solubility. The pulverulent precipitate obtained by continual shaking he found to be much less soluble, ouly 0.7 milligram being dissolved by a litre of pure water. Nitric acid increases to a sensible degree this last quan- tity, the increase being approximately proportional to the amount of acid present. Mulder knew well that it was possible wholly to precipitate either silver or chlorine by the addition of a sufficient amount of the other reagent, but wrongly assumed that the amount of the reagent required was a direct measure of the solubility of argentic chloride. This mis- conception probably arose from a theoretical discussion of very doubt- ful value; its effect was to make Mulder's estimate of the solubility rather excessive. Stas recognized the fact for the first time, that much more silver is required to precipitate the chlorine in solution as silver chloride than is enough directly to combine with it. A very tedious series of experiments led him to the conclusion that for every mole- cule of argentic chloride dissolved just three molecules of argentic nitrate or of a chloride were required to complete the precipitation. In 1881 Cooke called attention to some of the same facts.* Much time has been spent during the present research upon the subject, for its importance with regard to the analysis of baric chlo- ride was very evident. The experiments were largely a repetition of Stas's, with occasional variations suggested by the case in hand. Mulder and Stas having wholly done away with the old idea of the insignificance of the action of light on the haloid compounds of silver, the greatest care was taken with regard to the protection of the pre- cipitates from diffused light. Work with them was performed wholly in a dark room lighted by means of illuminating gas shining through thick orange glass ; and even here the flasks were always covered by two or three thicknesses of black cloth when left to themselves. In determining the solubility of argentic chloride a number of ex- periments were made, first by celorimetiic and " opalimetric " methods. The coloration produced by hydric sulphide and the opalescence pro- duced by hydrochloric acid were compared with similar phenomena produced by known amounts of silver. Very much experience is needed to obtain satisfactory results with either of these methods, so that a more direct method of determination was sought. * These Proceedings, XVII. 7. Cooke also studied the action of hot water upon argentic chloride. RICHARDS. — ATOMIC WEIGHT OF BARIUM. 71 An opalescent solution produced by the addition of hydrochloric acid or argentic nitrate to a solution of argentic chloride often requires many weeks to deposit its suspended salt, even when warmed and shaken, if the Mask is kept wholly in the dark. The effort to evapo- rate the solution by boiling in vessels partially closed to exclude the dust, proved equally unsatisfactory. Hence in order to weigh very small amounts of dissolved argentic chloride it was necessary to evap- orate the solutions upon the steam bath in suitably protected glass dishes. In order to determine the magnitude of any possible errors which might arise from this method of treatment a number of check experiments were made. In the first place dilute solutions of argentic nitrate were evaporated to small bulk with the addition of nitric acid. The evaporated solution was filtered through a small paper, and the ash of the washed filter was dissolved in nitric acid and titrated with a weak standard solution of amnionic sulphocyanate,* after the method of Volhard. Three experiments showed that an amount of silver equal to about 0.02, 0.01, and 0.00 milligram respectively had been retained by the filter. These experiments proved that the precautions taken to prevent dust or foreign 'materials from reaching the surface of the solutions during the process of evaporation had been sufficient for their purpose ; also, that no very important amount of silver was held by the paper. Having determined these facts, the next step was to determine whether all of a very small known amount of argentic chloride could be recovered. A few preliminary experiments showed that in order to reduce and recover all of the silver by the ignition of the filter paper it was necessary to keep the precipitate in a finely divided state. This end is easily accomplished by using a solution dilute enough to prevent the clotting of the opalescent chloride. Thus one cubic centi- meter of the weak standard hydrochloric acid precipitated enough argentic chloride from a slight excess of the nitrate to require after reduction by the burning of the filter 0.97 and 0.99 cubic centimeter of the sulphocyanate solution. Although the evaporation was con- ducted in dull diffused light, the argentic chloride was not discolored. f These experiments showed that a small amount of argentic chloride could be nearly all recovered from a small volume of solution. The next experiments were made with large volumes of solutions con- taining about 0.2 gram of silver, a very little baric nitrate, and a small * See page 66 of this paper. t Compare Pohl, Jahresbericht 1851, p. 369. Quoted by Mulder, Silber- probirmethode, p. 19. 72 PROCEEDINGS OF THE AMERICAN ACADEMY. amount of nitric acid, besides the measured amount of standard hydrochloric acid. Two experiments, after filtration, ignition, and titration as before, indicated that 1.01 and 1.02 milligrams of silver were precipitated by one cubic centimeter of hydrochloric acid, instead of just one milligram. Again, with all the other conditions just the same, 0.50 c.c. of hydrochloric acid precipitated in two cases 0.62 and 0.59 milligram of silver. In order to determine whether the pres- ence of large amounts of baric nitrate might make any essential differ- ence, these experiments were repeated with the addition of five grams of the purest preparation of this salt,* quite free from chlorine. In this way 0.50 c.c. of hydrochloric acid yielded 0.57 milligram of silver and 0.20 c.c. precipitated 0.27 milligram of silver. The last six experiments showed that the presence of a considerable excess of silver in solution causes the precipitation of a slight excess of silver with the chloride. Two experiments with neutral solutions treated in the same way showed much more marked gains ; one exposed to the air during evaporation for eight hours gained 0.13 milligram of silver ; and the other, evaporated for twenty-four hours, the water being replaced from time to time, gained 0.32 milligram. These gains might have been due to the action of organic matter in the absence of nitric acid, or the action of alkali dissolved from the glass ; in any case, the necessity of the presence of free acid was shown. In the work which follows it was several times necessary to evapo- rate strongly acid solutions of argentic chloride containing only a very slight excess of silver, hence parallel determinations under these con- ditions were very essential. In this way three successive experiments yielded respectively 0.38, 0.38, and 0.37 milligram of silver, instead of 0.50 milligram, which should have been obtained ; one experiment gave 0.52 instead of 0.60, and two more gave respectively 0.90 and 0.93, instead of 1.00. Hence, when much nitric acid and very little silver are present, the amount of metal found is too small instead of too great. These determinations all together show that the evaporation of solu- tions of argentic chloride upon the water bath is a process carrying with it small errors, some plus and some minus, which are reasonably constant under constant conditions. They show also that the presence of baric nitrate does not in the least alter the accuracy of the deter- * These Proceedings, XXVIII. 16, 19. RICHARDS. — ATOMIC WEIGHT OF BARIUM. 73 initiation, which depends chiefly upon the relative amounts of argentic nitrate and nitric acid present in each case. They moreover furnish a means of correcting results obtained by the method in hand. A solution containing two tenths of a gram of argentic nitrate, three- grams of nitric acid, and a very small known amount of argentic chlo- ride, evaporated for four or five hours, yields about 0.00006 gram too much silver, and a similar solution containing only a few milligrams of argentic nitrate yields about 0.0001 gram too little. The bearing of these facts upon the subject will presently be made more evident. It was now possible to determine the solubility of argentic chloride less crudely. Half a litre of the purest water, after being shaken occasionally with the purest argentic chloride for two hours, yielded 0.50 milligram of silver, and a similar preparation, which had been allowed to stand for twenty-four hours with more frequent shaking, yielded 0.G0 milligram. The solubility of this argentic chloride is thus found to be about 1.5 milligrams per litre. The result is not far from the mean of the colorimetric tests.* To determine the influence of nitric acid upon the solubility, four grams of this substance were added to over a litre of pure water, and the mixture was shaken with the same argentic chloride. 300 cubic centimeters of the solution yielded 0.47 milligram of silver, and two portions of 400 cubic centimeters each yield 0.56 and 0.57 milli- gram of silver respectively. After adding the appropriate correction, and taking the mean of these determinations, the solubility is found to be about 2.3 milligrams per litre. In a number of analyses exactly equivalent amounts of baric chlo- ride and argentic nitrate in dilute solution were shaken together, and both chlorine and silver were determined in the filtrate. These analy- ses were primarily to determine the atomic weight of barium, and will be described later; but since the chlorine was found to be present in the filtrate in amounts almost exactly equivalent to the silver, they may be used also to measure the solubility of argentic chloride in solutions containing about 2.5 cubic centimeters of nitric acid, and 3 to 8 grams of baric nitrate. In one case, where the solution had been shaken with the argentic chloride for ten days, the solubility was found to be about 1.6 milligrams per litre ; in another, violently shaken, but allowed to stand for a shorter time, it was found to be about 1.7 * It is noteworthy that Kohlrausch and Rose have found the solubility of argentic chloride to be 1.52 milligrams per litre at 18°. This determination was based upon the electrical conductivity of the solution. (Zeit. Physikal. Chem.,XII. 234.) 74 PEOCEEDINGS OF THE AMERICAN ACADEMY. milligrams per litre ; and in a third, much less violently shaken, but allowed to stand for a week, it was as much as 2.7 milligrams per litre. In each of these cases the temperature was about 20°. These results are, as one would expect, between the limits set by Stas ; for the precipitate was in the state usually observed, having lost its first flocculence, and not having been wholly converted into powder. A careful series of experiments was made to confirm Stas's state- ment that the chloride is absolutely insoluble in water containing an excess of argentic nitrate. It is not needful further to describe these experiments, except to state that the confirmation was wholly satis- factory. In the course of this work I was able to detect repeatedly one part of chlorine in thirty million parts of water, by careful com- parison in strong sunlight after treatment with excess of argentic nitrate. The very faint opalescence of course very soon disappears, owing to the blackening of argentic chloride by the powerful light. Ratio op the Chlorides of Silver and Barium. The properties of the two chlorides having been studied as far as they are concerned in the present work, it was now possible to com- pare their molecular weights. Very definite advice with regard to the precautions necessary in such a comparison has been given by Stas,* and most of his suggestions have been found of use. One recommendation, however, although most useful for ordinary work, is of questionable propriety when applied to the most precise work. In order to prevent argentic chloride from being dissolved by the water used for washing it, Stas advises the exclusive use of a very dilute solution of argentic nitrate for the purpose.! I have found it very difficult wholly to free argentic chloride' from traces of the nitrate, even by long continued shaking with pure water, and of course this difficulty is increased by the presence of additional impurity. The small amount of argentic nitrate which is certain to be left behind, according to the method of Stas, decomposes and blackens a little of the chloride upon fusion, even in the dark ; and no suitable correc- tion can be applied to counteract the error. Accordingly, pure water was always used for the final washings during the present work. * Mem. de l'Acad. Belg., Vol. XLIII. Part II. p. 66. I Professor Cooke also suggested this method of procedure. (These Pro- ceedings, XVII. 7.) RICHARDS. ATOMIC WEIGHT OF BARIUM. (5 These wash waters were all evaporated, and the small amount of argentic chloride present was determined in the manner which has already been described. It is thus tolerably certain that the neces- sary errors of the analytical process did not exceed the tenth of a milligram. In two preliminary experiments very pure baric chloride was ig- nited at a dull red heat to constant weight, dissolved in water, and corrected for the amount of alkali found. It was then precipitated, when perfectly cool, by pouring it into a slight excess of silver dis- solved in nitric acid, and diluted with about forty times its weight of cold water. The precipitate was very thoroughly shaken for several days, and washed by decantation until the wash water was no longer acid. It was then transferred with great care to a Gooch perforated crucible, washed for some time longer, dried at 180°, and weighed. Both of the precipitates were slightly colored, and each lost about to oo o °f ^s weight upon fusion in a covered porcelain crucible. In calculating the results it was necessary to subtract the amount of argentic chloride equivalent to the small excess of hydrochloric acid usually added in determining the amount of alkali formed by the ignition, as well as to add the amount of argentic chloride found in the wash water. The corrected results are given below, the details having been omitted; but the particulars will be given with regard to all the important determinations. PRELIMINARY DETERMINATIONS. 2 AgCl : BaCl,. No. of Exp. Corrected Weight of Baric Chloride. Corrected Weight of Fused Argentic Chloride. 100,000 parts of Argentic Chloride correspond to 11 parts Baric Chloride. Atomic Weight of Barium. 18 19 Grams in Vacuum. 6.7295 5.4597 Grams iu Vacuum. 9.2637 7.5161 n = 72.644 72.640 137.411 137.400 Average, .... .... 137.406 Since the experiments were carried on in the dark, the purplish color of the fused argentic chloride indicated that argentic nitrate had been occluded, in spite of the considerable care used in washing the precipitate. In order to determine whether this occlusion had been 76 PROCEEDINGS OF THE AMERICAN ACADEMY. due to the concentration of the silver solution, an experiment was made by pouring a solution of a definite amount of pure baric chloride into a very strong solution of argentic nitrate. 2.6088 grams (cor- rected) of baric chloride yielded 3.5929 grams (corrected) of carefully washed argentic chloride. The ratio deduced from these figures is 2 AgCl : BaCl2 = 100 : 72.009, a result almost 2^0 lower than tne last; and the consequent value of the atomic weight of barium is only 137.31. The argentic chloride was of a very deep purple color. The outcome of these experiments showed the necessity of Stas's usual practice of having the silver solution very dilute ; in the subse- quent work the dilution was usually 1 : 100, and sometimes even greater.* Moreover, the nitrate was always in future poured gradu- ally into the baric chloride, so that until the very last an excess of chlorine should be present in the solution. In the first series of the final group of experiments the baric chlo- ride was ignited at a dull red heat in air or nitrogen ; in the second series the salt was ignited and fused at a bright red heat in a stream of pure hydrochloric acid ; and in the third it was not ignited at all. The simplest manner of explaining the method of procedure adopted in the three series of analyses which follow is to give the detail of one of the experiments. Experiment 20. Very pure baric chloride, which had been pre- pared by many successive precipitations from aqueous solution by hydrochloric acid and alcohol (Sample I.c), was coarsely powdered and gradually heated to bright redness by means of a Berzelius spirit lamp in a double platinum crucible. In order to prevent as much as possible the decomposition of the salt, the long continued ignition was conducted in a stream of pure nitrogen. A large part of the salt was fused during the process. After having been cooled for eight hours over phosphoric oxide in a desiccator, the tightly covered inner cru- cible was rapidly weighed, and the weight of baric chloride was found to be 6.36793 grams in the air. Since the atmospheric pressure and temperature at the time of weighing were respectively 76 cm. and 20°, the salt would have weighed 6.36900 grams in a vacuum. The solution of the baric chloride in pure boiled water in a large platinum dish was perfectly clear, and the crucible itself was found to have lost only 0.02 milligram in weight. The clear liquid upon being titrated * In some cases Stas used only fifteen to thirty cubic centimeters of water for every gram of silver, in spite of his good advice to the contrary. Mem. Belg., etc., XLIII. Part II. p. 11. RICHARDS. ATOMIC WEIGHT OF BARTUM. 77 at once with standard hydrochloric acid* was found to require 3.40 cubic centimeters of the solution in order to remove the color of the trace of phenol phthalein added, and 0.64 cubic centimeter more to reach the point at which methyl orange began to show an excess of acid. An excess of yet 0.26 cubic centimeter was added in order to be certain of the end-point. According to page 57, the correction to be added to the weight of the baric chloride corresponding to the first addition of acid was 0.71 milligram, while the correction correspond- ing to the second addition was 0.03 milligram. Hence, the true weight of the baric chloride in vacuum, if none had been decomposed by heat, would have been 6.36974. The solution was transferred with the greatest care to a glass stop- pered Erlenmeyer flask, and diluted to about one hundred and fifty cubic centimeters with cold water. To it was then gradually added in the dark room the cold solution of about 6.72 grams of silver (about 3^ excess). The metal had been dissolved in thirteen cubic centimeters of the purest nitric acid, the lower oxides of nitrogen had been expelled, and the whole had been diluted to over six hundred cubic centimeters. The precipitate was shaken occasionally with the liquid for several days in perfect darkness ; and the filtrate, which could contain no trace of chlorine, was separated by means of the indispensable Gooch crucible. After having been parsed several times through the cruci- ble, in order to avoid the loss of any shreds of asbestos, the filtrate was set aside. The precipitate was violently agitated with renewed portions of cold water until the wash water was absolutely neutral, and was then transferred to the crucible by means of a jet of water. The last particles of argentic chloride were removed by shaking the stoppered flask violently with a little water. Subsequent testing with ammonia and nitric acid showed that no argentic chloride remained in the flask. The thirteen hundred cubic centimeters of wash water were evaporated, and the small amount of resulting argentic chloride was collected upon a small washed filter of the best quality. After the rapid ignition of the filter paper in a porcelain dish, the ash was found to contain 0.70 milligram of silver, according to Volhard's method. Hence, 0.93 milligram of argentic chloride had been dissolved. The main mass of the precipitate was dried at 180° ; and after having been carefully weighed it was transferred to a porcelain crucible, weighed again, fused in an air bath, and weighed yet again. It is needless to * See page 66 of tins paper. 78 PROCEEDINGS OP THE AMERICAN ACADEMY. say that organic matter was carefully excluded during all the opera- tions, but nevertheless the argentic chloride was not absolutely color- less after fusion. Perhaps the trace of phenol phthalein and methyl orange may have caused a slight decomposition. It will be seen how the introduction of these indicators was avoided before the close of the investigation. The final corrected weight of the argentic chloride was found as follows : — Grams. Weight in air of AgCl in Gooch crucible = 8.7G657 Correction to vacuum (T° = 22°, H = 76 cm.) = +0.00062 Weight in vacuum of AgCl in Gooch crucible = 8.76719 Loss on fusion in porcelain crucible = — 0.00050 Weight of main mass of fused AgCl = 8.76669 Weight of AgCl found in wash water = +0.00093 Total argentic chloride = 8.76762 Corresponding to 0.26 c.c. standard HO added in titration = —0.00035 Corrected weight of argentic chloride in vacuum = 8.76727 Since it seemed hardly probable that either of the chlorides was in this case weighed more accurately than within 0.1 milligram, the fifth place of decimals is omitted below. This practice was always adhered to unless the fifth place seemed to have some significance. Experiment 20 then leads us to conclude that the ratio of argentic chloride to baric chloride is 8.7673 : 6.3697 = 100.000 : 72.653. Hence, the atomic weight of barium is easily found to be 137.435, a result which agrees almost exactly with that obtained from baric bromide. The other analyses were conducted in a similar manner. In some cases, where it was suspected that the wash water carried with it a trace of finely divided asbestos, the whole quantity was filtered through a washed filter, and the weight found, if appreciable at all, was added to the weight of the Gooch crucible. The table given below is proba- bly comprehensible without further remark, except perhaps the state- ment that the two figures given in the third column represent the corrections for the baric hydroxide and baric carbonate respectively. The weights in the following cases are reduced to the vacuum standard. RICHARDS. ATOMIC WEIGHT OF BARIUM. 79 RATIO OF ARGENTIC AND BARIC CHLORIDES. First Series. No. of Exp. 20 21 22 23 24 Baric Chloride weighed. A lk. cor. See p. 57.* BaCl2 cor- rected for Alkali found. Argentic Chloride found (Fused). AgCl corresp. to extra HCl. AgCl in Wash Water. Total Argentic Chloride. Ratio 2 AgCl : BaCl2 = 100 : x. Atomic Weight of Barium. grains. 6.3600 3.7764 3.5846 1.5084 3.2160 c c. 13.4 10 6 J0.3 )0.2 i 0.0 (0.7 0.3 \ 0.3 1 0.9 grams. 6.3697 3.7765 3.5846 1.5085 3.2163 grams. 8.7667 5.1973 4.9332 2.0759 4.4254 m. gr. 0.35 0.20 0.13 0.24 0.64 m. gr. 0.93 0.80 1.10 0.86 2.32 grams. 8.7673 5.1979 4.9342 2.0765 4.4271 72.653 72.654 72.648 72.646 72.650 137.435 137.440 137.423 137.417 137.429 Average 72.649 137.428 The second series of determinations is distinguished from the first by the fact that the baric chloride was fused in an atmosphere of pure dry hydrochloric acid, beiug contained in a platinum crucible provided with a perforated cover. No phenol phthalein was added to the salt after its solution, and only a trace of methyl orange, in order to prove its absolute neutrality. Hence, the danger of a slight reduction of silver by compounds of carbon was much lessened. In all other respects the analysis was conducted in a manner similar to that employed in the last series, except that even greater care was used. RATIO OF ARGENTIC AND BARIC CHLORIDES. Second Series. No. of Exp. Weight of Fused Baric Chloride. Weight of Fused Argentic Chloride. AgCl cor- resp to extra HCl. AeCl in Wash Water. Total Argentic Chloride. Ratio 2 AgCl : BaCL = 100 : x. Atomic Weight of Barium. 25 26 27 grams. 1.52384 5.36010 3.92244 grams. 2.09730 7.37590 5.39771 m.gr. 0.00 0.10 0.12 m.gr. 0.20 0.30 1.47 grams. 2.09750 7.37610 5.39906 72.650 72.669 72.650 137.429 137.481 137.429 Average 72.6563 137.446 * Compare also these Proceedings, XXVIII. 12, also 26. 80 PROCEEDINGS OF THE AMERICAN ACADEMY. The third series differed from the two others in that the haric chlo- ride was not ignited at all. In the two analyses thus made, the true weight of the anhydrous baric chloride taken in each case was calcu- lated from the weight of the homogeneous powdered crystallized salt actually weighed. The determinations of the water of crystallization given on page 61, which were made at this time, upon this identical sample, showed with great certainty that the percentage of anhydrous salt present must have been 85.2282. Of course, no indicator of any kind was added to the solution before the precipitation. The argentic chloride after having; been fused and cooled was of a most beautiful pearly lustre, without a trace of grayness ; and in every respect the analyses appeared to be satisfactory. RATIO OF ARGENTIC AND BARIC CHLORIDES. Third Series. No. of Exp. Weight of Crystallized Baric Chloride. Calculated Weight of Anhydrous BaCl2. Weight of Fused AgCl. AgCl in Wash Water.* Total Argentic Chloride. Ratio 2 AgCl: BaCL, = 100 : x.' Atomic Weight of Barium. 28 29 grams. 7.00617 3.85330 grams. 5.'J7123 3.28410 grams. 8.21488 4.51770 milligrams. 4.0 2.2 grams. 8.2189 4.5199 72.6524 72.6587 137.435 137.453 Average 72.6555 137.444 The discussion of all these results will be found under a later head. Ratio of Metallic Silver to Baric Chloride. In determining this ratio the solubility of argentic chloride plays a much more damaging part than in the determination of the previous one. The uncertainty may easily become as great as thirty or forty milligrams in the weight of the silver, — an amount so enormous as to render accurate determination apparently impossible. Aside from this uncertainty, due to the existence of the two end-points, the ob- * Since the excess of silver was very slight, and a portion of the argentic chloride in the nitrate was determined by the evaporation of the strongly acid solution of baric nitrate, it was necessary to apply the correction of +0.0001 gram, described on pages 72 and 73, to the amount of argentic chloride found in the wash water. The values given above have been thus corrected. RICHARDS. — ATOMIC WEIGHT OF BARIUM. 81 serration of the point at which a drop of one solution or the other absolutely ceases to produce a precipitate is a very trying matter. The physical condition of the observer, and his practice in observation of this sort, have much to do with his determination of the presence or absence of a faint cloud at the top of the vessel under observation. The phenomenon is one which gradually fades away as more and more silver or hydrochloric acid is added, the last opalescence ap- pearing only after some time. Often when no cloud is visible at the end of an hour, a very evident one is to be seen in two hours. The amount of light at hand is also an important factor in the determina- tion. Temperature produces a very important effect on the solubility of argentic chloride ; and this effect is tripled when it is referred to the solution which is being added. The results are thus far less satis- factory than similar results obtained with the bromide. For preliminary experiments the method used by Stas in his early analyses * was adopted. The method consisted in the addition of an excess of silver to the solution of baric chloride, a standard solution of hydrochloric acid being used to determine this excess. At first the atomic weight thus obtained corresponded very closely to that ob- tained by the last ratio, but as the work proceeded the apparent atomic weight of barium steadily rose. The reason for this anomaly was not clear for a long time ; but in the end it was evident that the apparent accuracy of the first results had been due to the mutual elimination of two opposite errors. The occlusion of argentic nitrate by argentic chloride tended to require too much silver, and the method of titration required almost equally too much hydrochloric acid. When more care was taken with regard to the dilution of the argentic nitrate, and as the observer became more and more practised in the delicate observation of the end point, the first error steadily decreased, and the second steadily increased. Hence, they no longer counterbalanced one another. A long time was spent in finding that this first method, used by Stas in his early analyses of lithium, sodium, potassium, and ammonium chlorides, was quite false. Even relative results by this method, such as Stas's comparison of the different varieties of silver, can have no important value, because it is wellnigh impossible to make all the conditions absolutely constant. It is almost needless to say that these remarks apply only to this early method, and not to Stas's later work. Naturally he made errors at first. It is much to be regretted, however, that he did not point out mrre clearly after- * " Untersuchungen," etc. Leipzig, 1867. vol. xxix. (n. s. xxi.) 6 82 PROCEEDINGS OF THE AMERICAN ACADEMY wards the serious nature of these errors, which he clearly appreciated. Most of the investigations of Stas remain models of precision; and no one who has not attempted practically to follow in his footsteps can form an adequate idea of the greatness of his success. Stas's study of the " Statics of Argentic Chloride " has been already referred to. As a result of this long and valuable series of experi- ments, he concluded that the true end-point must be half way between the two end-points obtained by titrating in opposite directions. When this true end-point had been reached, he found that the supernatant liquid was rendered equally opalescent by equivalent amounts of hydrochloric acid and argentic nitrate. These conclusions suggested two possible methods for the solution of the problem before us ; and it is worthy of remark that at least one of these methods had been suggested by Mulder fifteen years before Stas's publication.* The unsatisfactory results obtained by the old method inclined one to subject the newer methods also to a careful testing before going further ; accordingly a long series of experiments, not unlike those described by Stas, were instituted to prove them. In the first place, pure thoroughly washed argentic chloride which had been made and kept in darkness was shaken with a dilute solution of pure baric nitrate and nitric acid ; and then the two end-points were determined by adding first hydrochloric acid, and afterward argentic nitrate, un- til no more cloudiness could be observed in either case. The results were not very satisfactory, for a reason which appeared later. Never- theless, many experiments were made in this way, showing that the observation of Stas was at least approximately true. The observation with regard to the equality of opalescence could not be confirmed at once, because the method requires much practice; but in the end the confirmation of it was much more satisfactory than the previous one. It was clear, however, that both methods depended upon a question of judgment rather than upon definite weighing or measuring. In default of a more satisfactory method a series of experiments upon the atomic weight of barium was made by determining both end-points and taking the mean as the true one, according to Stas's second method. During this series a curious fact became manifest. As the alternate addition of hydrochloric acid and silver nitrate pro- ceeded, the necessary amount of each solution increased to an enor- mous degree. For instance, in Experiment 33 below, the first difference between the end-points was 16 milligrams of silver, while * Silberprobirmethode, p. 94. RICHARDS. ATOMIC WEIGHT OF BARIUM. 83 the third was as much as 34 milligrams. This change was probably due to the fact that the argentic chloride produced by the successive additions of the solutions was of the more soluble flaky variety, while the greater part of the precipitate had been so thoroughly shaken be- fore as to become much less soluble. A larger amount of the salt in solution naturally required more of each reagent to precipitate it. This phenomenon was the chief cause of uncertainty in the attempt to confirm Stas's results. Probably the reason for his overlooking it lay in the fact that he rarely titrated backward and forward many times with the same solution. In the experiments tabulated below, the baric chloride was ignited in the air, and the small amount of chlorine lost was allowed for as usual. The calculated and carefully weighed amount of pure silver was dissolved in nearly three times its weight of the purest nitric acid in a Bohemian flask provided with bulb tubes for arresting the spray. The bulbs were washed out into the flask, the nitrous fumes were expelled by heating upon the steam bath for a long time, the con- tents of the flask were diluted and cooled, and the argentic nitrate was added to the baric chloride in a glass-stoppered flask. In every case the first readings of the end-points are recorded in the table as being the most trustworthy, although in some cases as much as three weeks were spent upon a single determination in titrat- ing backward and forward. The weights are, as usual, reduced to the vacuum standard. RATIO OF SILVER TO BARIC CHLORIDE. First Series: Stas's Second Method. No. of Experi- ment. Corrected Weight of Anhydrous Baric Chloride. Mean Amount of Silver required. Difference be- tweeu Extreme Amounts of Ag. Ratio 2 Ag : Bad, = 100 :x.~ Atomic Weight of Barium. 30 31 32 33 34 grams. 5.9717 5.4597 3.4728 9.0726 0.6950 grams. 6.1872 5.6580 3.5988 9.4010 0.7199 grams. 0.0032 0.0140 0.0050 0.0160 0.0029 S6.517 96.495 96.499 96.507 96.541 137.431 137.383 137.389 137.408 137.482 96.511 137.419 84 PROCEEDINGS OP THE AMERICAN ACADEMY. The objections to this method have been stated already ; accuracy with it is a question of judgment and keenness of sight upon the part of the observer, and absolutely constant conditions of temperature, illumination, and agitation on the part of the substance in the flask. Although great pains were taken to have these conditions as con- stant as possible, it cannot be claimed that they were absolutely so. Moreover, the method is so hopelessly tedious that it was finally abandoned. In the search for a new method, the following scheme was devised. A slight excess of one or two decigrams of silver over and above the amount required to precipitate a weighed quantity of baric chloride was weighed out with the greatest care. This was dissolved and added to the baric chloride as before. The mother liquor, contain- ing argentic nitrate, baric nitrate, and nitric acid, but no trace of chlo- rine, was filtered off with the greatest care through asbestos, and the precipitate was thoroughly shaken and washed with pure water. The wash waters were collected separately and evaporated on the steam bath in a suitably covered dish to small bulk. The trace of argentic chloride thus precipitated from them was filtered off and washed upon a very small filter, and the few milligrams of argentic nitrate in the filtrate were added to the original mother liquor. From this mixture of liquids all the silver was precipitated and weighed as argentic bromide.* The excess of silver originally weighed out is thus very simply calculated from the weight of the argentic bromide without involving a personal equation of any sort. Three analyses were made according to this method. f * For the preparation of the liydrobromic acid used, see these Proceedings, XXVIII. 17. The precipitate was assumed to contain 57.445 per cent of silver. t After the present paper had gone to press, notice was received from Pro- fessor Mallet that Stas had suggested to him, in a letter dated January 27, 1887, a method very similar to the one just described. J. W. Mallet, Stas Memorial Lecture, p. 33. [November 12, 1893.] RICHARDS. ATOMIC WEIGHT OF BARIUM. 85 RATIO OF SILVER TO BARIC CHLORIDE. Second Series: First Neav Method. No. of Exp. Corrected * Weight of Baric Chloride. Silver weighed out. grams. 6.72090 5.73383 4.08060 Weight of Sil- ver found iu Filtrate. t Weight of Sil- ver correspond- ing to BaCl2. Katio 2 Ag : BaCl„ = 100 : x. Atomic Weight of Barium. 35 36 37 grams. 6.36974 5.36010 3.92244 grams. 0.12097 0.18154 0.01680 grams. 6.59993 5.55229 4.06380 96.512 *96.539 96.522 137.419 137.476 137.440 Average 96.524 137.445 A second new method of accomplishing the desired ohject consisted in the actual determination of the silver and chlorine left in a solution obtained by mixing amounts of baric chloride and argentic nitrate supposed to be equivalent. The solution and first five wash waters filtered off from such a care- fully prepared mixture, which had been shaken occasionally for about a week, were thoroughly mixed and divided exactly iu half. One half was evaporated with hydrobromic acid, and the other with ar- gentic nitrate, and the small amounts of argentic bromide and chloride were determined by reduction and the use of Volhard's method. Subsequently it appeared that there was a slight error in the method, due to the difficulty of determining accurately such small quan- tities of substance. By the evaporation of solutions known to contain nothing but pure argentic chloride, it was found that the amount of silver obtained was always a little less than enough to correspond to the chlorine. Several experiments showed that the error from this cause amounted to very nearly one tenth of a milligram, hence this amount is added below to the amount of silver found in the half of the filtrates which was evaporated with hydrobromic acid. * For the corrections, see page 79 of this paper. These three analyses are portions of Experiments 20, 26, and 27. t This weight, computed from the argentic bromide, has been corrected for a very slight excess of hydrochloric acid. The correction, which was of course added, amounted to 0.00026 gram in the first case, 0.00007 in the second, and 0.00009 in the third. Compare data on page 79. 86 PROCEEDINGS OF THE AMERICAN ACADEMY. EATIO OF SILVER TO BARIC CHLORIDE. Third Series : Second New Method. No. of Exp. 38 39 40 41 Corrected* Weight of Baric Chloride. Weight of Silver taken. Silver found in one half of Filtrate, plus correction of 0.0001 gram. Silver corre- sponding to Chlorine found in the other half of Filtrate. Corrected Silver re- quired by Baric Chloride. Ratio 2 As: Bad., = 100 : x. Atomic Weight of Barium. grams 4.2815 2.6488 5.9712 3.2841 grams. 4.43558 2.74417 6.18547 3.40245 grams. 0.00080 0.00063 0.00065 0.00060 grams. 0.00076 0.00054 0.00118 0.00052 grams. 4.4355 2.7440 0.1865 3.4023 96.528 96.531 96.520 96.526 137.453 137.459 137.436 137.449 Average 96.526 137.449 During a large part of the considerable time consumed by these analyses, experience with the comparison of opalescence was gradu- ally being acquired. The details of this method have been so fully given by Stas that further mention of them need not be made here. For the final experiments the baric chloride was dissolved in a large amount of water, and decomposed by a very slight excess of argentic nitrate in very dilute solution, the silver having been weighed and dissolved with the greatest care. After the mixture had been shaken occasionally for a number of days, it was allowed to settle, and two portions of twenty-five cubic centimeters each were taken out with a pipette. A little hydrochloric acid produced a more evident opales- cence in one of these portions than an equivalent amount of silver did in the other, showing that too much silver had been taken. Known * The last two of these determinations formed a continuation of Experi- ments 28 and 29 already given, and the calculation of the amount of anhydrous baric chloride need not be repeated. The actual weight of ignited baric chlo- ride taken in Experiment 39 was 2.64851 grams (in vacuum) ; but this was found to require a correction due to loss of chlorine amounting to 0.00025 grams. The corrected weight is given in the table. Since the salt used in Experiment 38 was ignited in hydrochioric acid gas, no correction was neces- sary. In Experiment 38 the weight of silver given in the third column is 0.00020 gram less than the amount actually weighed, because 0.20 cubic centi- meter of standard hydrochloric acid was added to determine the neutrality of the baric chloride before the precipitation. RICHARDS. — ATOMIC WEIGHT OP BARIUM. 87 amounts of hydrochloric acid were therefore added very gradually to the main mass of the liquid, until, after agitation and settling, two por- tions of the clear supernatant liquid treated respectively with argentic nitrate and hydrochloric acid showed like intensity of opalescence. This similarity of opalescence was further confirmed by measuring the heights of parallel unknown columns of the two liquids which had been adjudged equally cloudy. From the average of a great number of readings, one could feel reasonably certain of the result within two or three per cent, corresponding to an error of not over one or two tenths of a milligram of silver in the volume of solution used. For example, in the first analysis recorded below, it was found in three preliminary trials that about 0.30 cubic centimeter of the stand- ard hydrochloric acid should be added to the remaining mixture in order to reach the desired point. This amount was therefore added, and two portions were again taken out, treated as before, and compared. Solution to w 'hich Solution to which HC1 was added. AgN03 was added. 9.5 cm. in height appeared eq ual to 10.0 cm. 10.5 ft « ft .< 10.0 " 10.3 ft « « ft 10.0 " 10.1 ft « ft ft 10.0 " 9.2 u ft (C ft 10.0 " rage, 9.9 « a ft ft 10.0 " This average shows that the equality of cloudiness was now within the probable experimental error. It was borne in mind, however, that a fifth of the original solution had been used up in the preliminary tests. Hence, the true amount of hydrochloric acid which should have been added to the original solution must be £ X 0.3 = 0.375, or practically 0.4 cubic centi- meter. Hence, since this amount of hydrochloric acid is of course a direct measure of the excess of silver originally taken, 0.4 milligram should be subtracted from the original weight. It is needless to say that all possible care was used in these final determinations. The reiteration of the countless details would be wearisome, and hence these details are left to be inferred from what has been already written. The method is the simplest and easiest of the four ; and when one has acquired experience in the comparison of opalescence, it is perhaps also the most accurate. 88 PROCEEDINGS OF THE AMERICAN ACADEMY. RATIO OF SILVER TO BARIC CHLORIDE. Fourth Series: Stas's Third Method. No. of Exp. Weight of Crystallized Baric Chloride. Weight of Anhydrous Baric Chloride. Weight of Silver taken. Excess of Silver found. Total WTeight of Silver. Ratio 2Ag: BaCl, = 100 : x. Atomic Weight of Barium. 42 43 grains. 7.63356 12.01793 grams. 6.50022* [10.23365] grams 6.7346 10.6067 milligrams. 0.4 4.4 grams. 6.7342 10.6023 96.525 96.523 137.448 137.442 Average 96.524 137.445 The Atomic Weight of Barium. The data based upon the two more satisfactory ratios detailed in this paper have been grouped in seven series of determinations. In the first three series, including ten individual experiments upon the first ratio, the individual results are grouped together according to the condition of the baric chloride used as the starting point. This method of grouping was adopted because the analytical conditions were neces- sarily quite uniform, and the initial state of the salt was the most important variable. This problem having been solved satisfactorily, the accurate decision of the true end-point in the method of Gay Lussac remained the most important incidental question to be settled during the determination of the second ratio. Hence, the last four series were grouped according to the method of titration. From the Ratio of Baric Chloride to Argentic Chloride. Atomic Weight of Barium if 0 = 16.000. I. Five experiments, BaCl2 ignited in air = 137.428 II. Three experiments, BaCl2 ignited in HC1 = 137.446 III. Two experiments, BaCl2 not ignited = 137.444 137.439 * This specimen of baric chloride was dehydrated by ignition in a stream of hydrochloric acid. The specimen whose weight is recorded in brackets below it was not ignited ; but the weight of the anhydrous salt was computed by means of the proportion 7.63356 : 6.50022 = 12.01793 : x. RICHARDS. — ATOMIC WEIGHT OF BARIUM. 89 From the Ratio of Baric Chloride to Silver. I. Five experiments, Stas's second method = 137.419 II. Three experiments, first new method = 137.445 III. Four experiments, second new method = 137.449 IV. Two experiments, Stas's third method = 137.445 137.440 Before discussing these averages it may be well to recall the results obtained from baric bromide. These last results may be divided into two groups of two series each, the essential difference between the two series depending upon their relative accuracy. From the Ratio of Baric Bromide to Argentic Bromide. I. Experiments 3 to 12 ; Ba = 137.424 II. Experiments 13 to 19 ; Ba = 137.439 137.431 From the Ratio of Baric Bromide to Silver. I. Experiments 3 to 12 ; Ba = 137.422 II. Experiments 13 to 19 ; Ba = 137.435 137.428 Total average of four groups 137.434 The four groups, including eleven series and forty-nine individual results, are thus essentially unanimous in the verdict that the atomic weight of barium is between 137.42 and 137.45. Upon closer ex- amination of details, it is apparent that some of the series are more trustworthy than others. In general, the first series of each group is considerably less reliable than the following ones, for obvious reasons. The first series of the fkst and second groups have the additional dis- advantage of having been the victims of rather unsatisfactory methods. Since the most probable errors tend in any case to lower the observed atomic weight, the striking out of these lower results appears all the more allowable. The more trustworthy results which remain may be averaged as follows : — 90 PROCEEDINGS OF THE AMERICAN ACADEMY. From the Ratio BaCl2 : 2 AgCl ; Ba=: 137.445 u « BaCl2 : Ag2 ; Ba = 137.446 a a BaBr2 : 2 AgBr ; Ba = 137.439 u a BaBr2 : Ag2 ; Ba = 137.435 Average 137.441 This average is after all but little different from the total mean of all the determinations, 137.434; and either average may be taken without serious error for the atomic weight of barium. Indeed, in the present state of chemistry, the third decimal place is of little or no importance in a large atomic weight, and the round number 137.44 is sufficient for the most exacting work. An excellent proof of the purity of the materials and the accuracy of the work is to be found in the comparison of the weights of silver and argentic chloride. Omitting in each case the less accurate first series, we may repeat the following figures : [II.] 2 AgCl : BaCl2 = 100 : x = 72.6563 [III.] 2 AgCl : BaCl2 = 100 : x = 72.6555 Av srage 72.655S [II.] 2Ag : BaCl2 — 100 X = 96.524 [III.] 2 Ag : BaCl2 = 100 X = 96.526 [IV.] 2Ag : BaCl2 — 100 X = 96.524 Average 96.5247 Hence Ag : AgCl = 72.6559 : 96.5247 = 107.930 : 143.387 Therefore the atomic weight of chlorine is found to be 35.457, a quantity exactly identical with that found by Stas in a different way. A similar comparison of the weights of the silver and argentic bromide has been given already in the fifth column in the table on page 29 of Volume XXVIII. of these Proceedings. It remains only to compare these figures with the older ones already described. Evidently all results for barium based upon the conversion of baric chloride into baric sulphate, as well as all results including water of crystallization, are worthy of no confidence. The incomplete knowledge regarding the method of Gay Lussac mani- fested by Marignac and Dumas in 1858 is quite enough to explain their unsatisfactory and discrepant results. We should expect to find RICHARDS. — ATOMIC WEIGHT OF BARIUM. 91 the ratio of baric chloride to argentic chloride the most accurately determined of all the old ratios; and in fact Turner in 1829 obtained in this way a very good result (137.45). Upon the other hand, Marignac's single result by this method (137.14) was much too low. The description of his work is so superficial that a guess as to the reason for this discrepancy is out of place ; nor is such a guess neces- sary, for a single experiment can carry with it no important weight. The new result for the atomic weight of barium, reduced to the various standards in present use, is given below. As before, the mathematical statement of the " probable error " is omitted, because this error is so small as to be meaningless. If Oxygen = 16.00, Barium = 137.44 = 15.9G, " = 137.10 " = 15.88, " = 136.41 92 PROCEEDINGS OF THE AMERICAN ACADEMY. III. CONTRIBUTIONS FROM THE C-RYPTOGAMIC LABORATORY OF HARVARD UNIVERSITY. XXL— NEW GENERA AND SPECIES OF LABOULBENIA- CE^, WITH A SYNOPSIS OF THE KNOWN SPECIES. By Roland Thaxter. Presented November 8, 1893. In attempting some years since to obtain materials for a monograph of* the Laboulbeniaceas, the writer did not anticipate that a fifth pre- liminary paper would be reached before he felt in a position to publish a general account of the family ; yet, however much the description of so large a number of new forms without proper figures is to be regretted, the delay in this respect seems fully justified by the essential data which have been obtained during the past two years. It cannot be doubted that the number of existing forms greatly exceeds the total already known, but it seems safe to assume that the basis of knowledge now available is sufficient to illustrate, at least in a general way, the more important characteristics of the group from a morpho- logical as well as a systematic point of view. The promised mono- graph will therefore be published as soon as the necessarily numerous plates can be completed; and since a sixth preliminary paper will, if possible, be avoided, a summary of the known genera and species is appended for convenience of reference, together with a very brief note concerning certain matters relating to the general morphology and development of the group. The writer's observations, based upon an examination of several thousand specimens illustrating more than a hundred species and more than twenty genera, appear to warrant the following conclusions. The LaboulbeniaceEe, while showing no signs of any non-sexual mode of reproduction, are characterized by a well marked sexual type, closely resembling that of the simpler Floridese. They are ascomy- cetous fungi, producing usually four, sometimes eight spores in asei THAXTER. — LABOULBENIACE/E. 93 which arise by a peculiar process of budding from ascogenic cells, of which there may be from one to four, usually distinct and eventually free within the cavity of the perithecium. The ascogenic cells are developed from a carpogonium consisting of a single axile cell which is fertilized by non-motile male bodies (antherozoids) through the agency of a more or less highly developed trichogyne from which it is separated by a second axile cell. After fertilization the carpogenic cell divides by transverse septa (in ascer- tained cases) into three cells, the middle cell becoming either directly an ascogenic cell or dividing into from two to four such cells, while the other two supporting cells eventually disappear. The trichogyne varies from a simple vesicular receptive promi- nence, or short filament, to a copiously branched and highly developed organ, the numerous free receptive tips of which may be coiled in close and regular spirals. The trichogyne, the insertion of which is usually terminal, disappears immediately after fertilization is accom- plished, however highly it may be developed, its insertion becoming lateral by the further development of the perithecium. The antherozoids appear to originate in two genera exogenously from special branches. In all other genera they are produced endogenously in antheridia, the form and position of which vary in the different genera. The antheridia are either single specialized cells, which may be more or less irregularly disposed or characteristically grouped, or may consist of more highly developed multicellular bodies. In either case the antherozoids are discharged through a terminal pore in the form of roddike or nearly spherical masses of naked protoplasm. The sexes are commonly both present on the same individual, usually so placed that self-fertilization is readily accomplished, or may be completely separated on specialized individuals. In the latter case of the members of any given spore pair formed in an ascus and discharged simultaneously from the perithecium one produces a male, while the other produces a female individual, thus insuring the juxtaposition of the sexes at the new point of infection. The spores which may be formed in pairs within the asci, or less definitely disposed, are of one general type, fusiform in shape and divided by a more or less well defined septum into two parts, the rela- tive position of which in formation is reversed in development, the segment which is basal in relation to its position in the ascus becom- ing distal in its relation to the growing plant resulting from its germination. A well defined gelatinous envelope surrounds the spore when mature, and develops with the growing plant to form a sac-like 94 PROCEEDINGS OF THE AMERICAN ACADEMY. covering continuous over all its parts with the exception of the pore of the perithecium and certain portions of the sexual organs, its formation being independent of the cell divisions which take place inside it. The new forms described below include a portion of the novelties collected during the past summer, and are distributed among the genera as follows : Ceratomyces, two species ; Teratomyces, two species ; Cantharomyces, one species ; Peyritschiella, one species ; Dichomyces, two species ; Heimatomyces, one species ; Dimorphomyces, one species ; while the four new genera described, Sphaleromyces, Moschomyces, Camptomyces, and Compsomyces, include each a single species. As in previous" instances, the writer is indebted to the kindness of Mr. Henshaw for the determinations of hosts. Ceratomyces humilis, nov. sp. Hyaline becoming faintly tinged with brownish. Perithecium rather narrow, without any appendage, the apex blunt or nearly truncate, its cell rows composed of not more than seven cells. Re- ceptacle composed of from two to five superposed squarish cells. Appendage consisting of six to twelve superposed cells, the series tapering distally and producing irregularly from its inner face branches which may in turn be several times branched and may reach a length twice that of the perithecium. Spores 22 X 3 fi. Perithecia 100 X 25 p. Total length to tip of perithecium 150-185 ju. Longest branches of appendage 180 p. On legs and at the edges of the elytra of Berosus striatus Say. Kittery Point, Maine. A somewhat insignificant species allied to C. contortus, from which it is easily distinguished by its small size and by the absence of any appendage near the tip of the perithecium. It occurs more fre- quently between the terminal claws of the middle pair of legs, but is rarely found on the elytra. In two specimens the perithecia have become distinctly tinged with brown, but as a rule the whole plant is hyaline. Ceratomyces terrestris, nov. sp. Nearly hyaline with black or dark brown suffusions. Perithecia large,' slightly inflated, tapering to a bluntly rounded or truncate apex from which the sharply pointed lips project : the wall of the perithe- THAXTER. — LABOULBENIACE^E. 95 cium consisting of four series of about twelve cells each, its base formed from three small cells, below which a single similar small cell connects it with the receptacle. Receptacle consisting of three small superposed squarish cells, the upper of which gives rise to the perithecium and the appendage. The appendage, consisting of six or more superposed flattened cells becoming externally suffused with blackish brown or black (the suffusion sometimes involving the whole series as well as the entire receptacle with the exception of its basal cell), bearing on its inner side numerous hyaline branches which may in turn be once or twice branched, the lower arising from a series of small cells which may extend across the base of the perithecium on one side. Spores 15 X 2.5-3 p. Perithecium 75-90 X 22-29 p. Receptacle 25 fx long. Total length to tip of perithecium 100-140 n, to tip of main appendage 45-65 p. Longest branches of appendage 75 ft. On Lathrobium punctulatum Lee. Kittery Point, Maine. This minute and curious species is chiefly interesting from the fact that it is a terrestrial form in a typically aquatic genus, of which, how- ever, it seems to possess all the essential characters. Unlike most, of its congeners, it is among the smallest of all the Laboulbeniacea?, and is very readily overlooked. SPHALEROMYCES, nov. gen. Receptacle consisting of two superposed cells, the distal bearing the appendage laterally, and the stalk cell of the perithecium terminally. Perithecium asymmetrical, the apex somewhat pointed ; separated from its short stalk cell by three basal cells. Appendage clearly distin- guished from the receptacle, composed of a basal cell bearing a series of superposed cells, each giving rise from its inner upper angle to a single short septate branch which may bear flask-shaped antheridia. Spores once septate involved in mucus. Asci arising in a double row from a single large ascogenic cell. Sphaleromyces Lathrobii, nov. sp. Wholly hyaline or faintly yellowish. Perithecium rather slender, slightly inflated towards the base, tapering to the somewhat pointed apex which is bent inwards ; the dividing lines between the wall cells indicated by successive ridges, the distal more prominent : the short stalk cell, separated from the perithecium by three additional cells. Receptacle consisting of two obliquely superposed cells, the upper 96 PROCEEDINGS OF THE AMERICAN ACADEMY. bearing the stalk cell of the peritheciuin distally and the basal cell of the appendage laterally. Appendage borne on a large basal cell connected on its inner side with the distal cell of the receptacle and a portion of the stalk cell of the perithecium, its exterior wall usually much thickened so as to form a more or less distinct rounded promi- nence at the base of the appendage proper, which consists of a series of four or five obliquely superposed cells diminishing in size towards its apex, from each of wbich arises on the inner side a single branch ; the branches directed obliquely upwards and forming a single series, septate, cylindrical, simple or bearing near their tips single short branches or flask-shaped cells. Spores 35 X 3-3.5 ti. Perithecia 100 X 22 /x — 140 X 36 fx. Total length to tip of perithecium 160-240 jx. Total length to tip of distal branch of appendage 110-166 fi. Receptacle 38-45 /x long. On Lathrobium nitidulum Lee. and L. punctulatum Lee. Kittery Point, Maine. The writer is unable to refer this perplexing yet distinct form to any of the described genera. In general appearance it resembles Stigmatomyces, to which it might be referred if the successive cells of the appendage gave rise directly to antheridia. The resemblance, however, is superficial, since the character of the appendage is essen- tially different. Among the remaining genera it might perhaps be compared with the more simple type of Ceratomyces, from which the clear differentiation of its appendage and the structure of its perithe- cium as well as the character of its antheridia would seem to distin- guish it. Specimens from L. nitidulum are distinctly smaller than those on the larger host. The species seems to be a rare one, and inhabits the legs and abdomen of its host, where, owing to its small size and pale color, it is detected with difficulty. COMPSOMYCES, nov. gen. Receptacle consisting of two superposed cells, the distal bearing from its extremity a cluster of appendages and one or more stalked perithecia. Appendages sterile or fertile, simple or branched, septate, the fertile ones bearing one or more single one-celled antheridia sepa- rated by oblique partitions from the extremities of successive cells composing the main axis of the appendage. Perithecia symmetrical, conical, borne on two superposed stalk cells and three small basal cells, the basal stalk cell producing from its distal end a simple sterile ap- pendage. Asci 8-spored. Spores once septate. THAXTER. — LABOULBENIACE^E. 97 COMPSOMTCES VERTICILLATUS. Cantharomyces verticillatus Thaxter. This species was found not uncommonly on Sunius longiusculus at Kittery Point, Me., during the past summer, and an examination of new material shows conclusively that it is generically distinct from Cantharomyces as emended in a previous paper. Its antheridia are not compound as in that genus, but simple, more or less flask-shaped and produced in a fashion more closely resembling that of certain species of Laboulbenia. The trichogyne is remarkably developed, copiously and regularly several times branched, the free receptive tips being coiled in close and regular spirals. MOSCHOMYCES, nov. gen. Receptacle composed of a sucker-like compacted mass of parenchy- matous cells penetrating the softer chitin of the host and giving rise above to numerous free cells from the distal ends of which are pro- duced solitary stalked perithecia and appendages. Perithecium very large, subconical, pointed, the apex symmetrical, borne on two simple superposed stalk cells followed by three small basal cells ; the basal stalk cell bearing from its distal end a single simple sterile appendage. Appendages septate, sparingly branched or simple, the fertile ones stouter, bearing one-celled antheridia laterally. Asci subcylindrical, eight-spored, arising in great numbers and in many rows from a single ascogenic cell or centre. Spores minute, acicular, once septate. MOSCHOMYCES INSIGNIS, nOV. sp. Perithecia pale straw-colored, becoming tinged with brown, the lower portion slightly inflated and abruptly contracted at the base, the distal portion subconical, sometimes slightly bent to one side, the apex narrow, truncate, symmetrical, the surface marked by two series of ridges extending around the perithecium, each series composed of four distinct and prominent ridges placed somewhat irregularly and in- dicating the lines of separation between the middle and the upper and lower series of cells which form the walls of the main body of the perithecium : basal cells of the perithecium small, three in number, not distinguished from it but somewhat abruptly distinguished from the distal stalk cell, which is long, subcylindrical, sometimes inflated and curved ; the basal stalk cell usually shorter and smaller, bearing distally a single slender simple rather closely septate tapering append- voi,. xxix. (n s xxi ) 7 98 PROCEEDINGS OF THE AMERICAN ACADEMY. age, usually about as long as the distal stalk cell. The appendages, which together with the single stalked perithecia spring in groups of three or four from the distal ends of large cylindrical cells project- ing from the sucker-like receptacle (more rarely arising from the latter directly), are simple or once branched, either sterile or producing the supposed antheridia on short branches near their extremities. Spores very minute, acicular, septate near the middle, 12 X 3 /x. Asci subcylindrical, 40-45 X 7.5 li, eight-spored, the spores sub-dis- tichous. Perithecia 225-290 X 55-75 /*, the stalk cells (longest) 425 p, average 375 x 25 p. Appendages 175-375 /* long. Breadth of sucker-like receptacle 75 p. On Sunius prolixus Er. Waverly, Mass. This form, which is among the most remarkable thus far discoveied, differs from all other genera in the structure of its attachment to the host. It inhabits only the softer chitinous membranes, beneath the elytra and at the bases of the legs or between the segments, which it perforates by the intrusion of its sucker-like base. It is more nearly allied to Compsomyces than to any other known genus. The minute spores are formed in enormous numbers and discharged in masses. The asci arise from a single centre, apparently a single cell, in numer- ous rows, and are distinctly eight-spored. TERATOMYCES Thaxter. Three additional species of this genus, two of which are described below, represent a well marked type adhering closely to that previ- ously described. The discharge of antherozoids has been observed in fresh material and the antheridia prove to be the beakdike cells characteristic of all the species. The trichogyne has been observed in two species in which it differs widely ; in the one case appearing as a branched organ very similar to the ordinary terminal branches of the appendages, while in the other it is highly developed with peculiarly modified receptive tips. Teratomyces Actobii, nov. sp. Perithecia one to several, reddish brown distinctly inflated towards the base, the distal portion symmetrically conical, tapering to a blunt apex ; borne on a single short stalk cell not exceeding the appendages in length followed by three unusually large basal cells disposed as in T. mirificus. Receptacle short nearly symmetrical tiuged with THAXTEK. — LABOULBENIACEiE. 99 brownish or nearly hyaline, consisting of three superposed cells, the basal small and narrow, the sub-basal squarish, the distal large, rounded and followed by the circular series of small cells from which arise the numerous appendages which in general resemble those of T. mirificus though 2jroP0,'tionately stouter and more intricately branched. Spores 26 X 3.7 p. Perithecia 120-137 X 37 p. Stalk cells 75-100 p. Longest appendages 150 p. Three basal cells of receptacle 37 X 22 p. On Actobius nanus Horn. Kittery Point, Maine. This species occurred with the next on the legs of its host. It differs from T. mirificus in the form of its perithecia, which are subconical, as well as by its hyaline or nearly hyaline symmetrical receptacle. The short stalk cell and large basal cells of its perithecia serve also to distinguish it. a Teratomyces brevicaulis, nov. sp. Perithecia several, purplish brown, long, slender, straight or slightly curved, cylindrical or slightly inflated near the middle, taper- ing abruptly to the almost truncate apex, much longer than the stalk and basal cells together, the latter concolorous with the perithecium, the stalk cell nearly hyaline. Receptacle nearly symmetrical, black and quite opaque, except the partly translucent basal cell ; above the opaque portion expanding abruptly to form the broad distal portion from the numerous small cells of which arise, around the edge, the circle of crowded appendages which surround the perithecia. Larger appendages faintly tinged with brownish purple, consisting of a rather short basal cell bearing a short series of superposed external branches, the uppermost consisting of a large, long basal cell curved and some- what inflated distally, where it bears externally a series of two to five secondary branches similarly shaped which may in turn be similarly branched, the ultimate branchlets bluntly pointed with numerous slightly oblique septa or terminating in long beak-like cells (autheridia). Spores 33 x 4 M. Perithecia 110-120 x 23 p. Stalk cells 50 X 15 p. Receptacle 85 X 50 p. Longest appendages 100 p. On Actobius nanus Horn. Kittery Point, Maine. This curious species is quite distinct from that with which it was associated on the same host, and differs in the form of its perithecia and appendages as well as the peculiar and abrupt distal expansion of its receptacle. Both of the types, which occurred at the tip of the abdomen of the host, have three perithecia. 100 PROCEEDINGS OP THE AMERICAN ACADEMY. Cantharomtces pusilltjs, dov. sp. Perithecium becoming reddish brown, inflated just above the base, the distal portion conical tapering to a blunt symmetrical apex, borne on a rather short narrow stalk cell bent towards the appendage and separated from the perithecium by three small subtriangular basal cells. Receptacle consisting of a very small basal and a much larger rouuded sub-basal cell, more or less suffused with brown, which gives rise to the stalk cell of the perithecium and the appendage. Anthe- ridial appendage consisting of a large squarish basal cell followed by the antheridium proper, which is primarily a large squarish cell, its outer half or more becoming divided by anastomosing septa into numerous small cells, the inner portion also showing a division into two or three larger cells ; the whole bearing terminally a series of usually three superposed flattened cells strongly constricted at the septa and giving rise distally to from one to three simple cylindrical nearly hyaline sparingly septate branches, usually exceeding the peri- thecium in length. Spores 18 X 2 /x. Perithecia 22-26 X 30-55 ft. Total length to tip of perithecia 80-85 ft, to tip of appendages 90-120 ft. On Trogophlceus sp. York, Maine ; Waverly, Mass. This species is perhaps the smallest of the known forms of Laboul- beniaceas, and is somewhat difficult to discover and remove from the legs or elytra of its host, where, however, it is not rarely found. Owing to its minute size the detailed structure of the antheridium was not plainly made out, neither was any discharge of antherozoids noticed. It corresponds so closely, however, to the structure characteristic of the genus as emended, that there can be little doubt of the correctness of its generic reference. CAMPTOMYCES, nov. gen. Receptacle consisting of two superposed cells, the upper bearing the short-stalked perithecium laterally and the antheridial appendage terminally. Perithecium narrow, with coarse-lipped asymmetrical apex. Appendage consisting of a single large basal cell bearing the antheridium terminally. Antheridium multicellular, subcorneal, with a prominent terminal pore for the discharge of the numerous roundish antherozoids. Trichogyne developed as a small vesicular prominence above a permanent ear-like appendage which arises laterally from the young perithecium. Ascogenic cells two in number. THAXTER. — LABOULBENIACE.E. 101 Camptomyces melanopus, nov. sp. Perithecium tinged with brownish, slightly inflated towards the base, its distal half narrow, tapering gradually to the rounded apex below which on one side is a rounded projection ; borne on a large subtriangular stalk cell surmounted by three smaller basal cells. Receptacle narrowly funnel-shaped, tapering to a pointed base and consisting of a large basal cell, slightly translucent near its lower extremity, but otherwise becoming wholly opaque, followed by a flat- tened sub-basal cell from which the mature perithecium with its stalk projects nearly at right angles to the long axis of the receptacle, while distally it bears the appendage. Appendage bearing terminally the sub- conical slightly asymmetrical antheridium. Spores about 25 X 3.5 p. Perithecium 130-150 X 30-33 /x. Total length to tip of antheridium 110-125 ft. Greatest width 25 p. Antheridium 25 X 16 p. On Sunius prolixus Er. Waverly, Mass., and York, Maine. This curious form affords an additional example of a highly de- veloped type of antheridium which has neither the peculiar honey- comb-like appearance of Cantharomyces and Haplomyces nor the more simple type of Peyritscbiella and its allies, Dichomyces and Heimatomyces. It is distinguished from the two genera first named by having a strictly terminal pore without appendages of any kind. It shows a clearly defined cavity within which the spermatia are formed which is surrounded on three sides (wholly near its base) by several rows of cells not symmetrically disjtosed. The species occurs rather rarely on the abdomen of its host, the perithecia being usually directed forward. Peyritschiella geminata, nov. sp. Hyaline. Receptacle asymmetrical, consisting of a single basal cell followed by three successive more or less definite transverse rows of cells. The lowest of these rows is the most variable and irregular, consisting of from two to four cells, one of which, larger than the rest, is an axile cell continuing the basal cell directly, while the remaining one to three cells are cut off from it on one side, each successive cell smaller and placed higher in the series, the outermost and uppermost bearing one of the sterile appendages characteristic of the genus. The second row also consists of a larger axile cell, which is free for a short distance on one side, and on both sides of which are cut off, as in the first row, from two to four cells, the smaller uppermost ones on both sides giving rise to from one to three appendages according to the 102 PROCEEDINGS OF THE AMERICAN ACADEMY. number of the cells. The upper row is either symmetrical or asym- metrical according as it gives rise to two or to only one perithecium. In the first instance it consists of an axile cell, above which are pro- duced two sterile appendages, and a variable number of cells cut off laterally as in the lower rows, but nearly symmetrical in size and shape, the outermost bearing sterile appendages. If one perithecium only is produced the row is asymmetrical and a greater number of appendages appear on one side of the perithecium than on the other. Perithecia very slightly inflated near the base, tapering abruptly but slightly to the spreading apex, which is four-lobed, the lobes rounded, large, and prominent. Spores about 37 X 3.7 /x. Perithecia 75-80 X 18-22 p. Total length to tip of perithecia 220-260 p (150 p in specimens from the smaller host). On Pterostichus luctuosus Dej. and P.patruelis Dej. Kittery Point, Maine. Unlike the other species, this form not infrequently produces two perithecia, the arrangement of its distal cells in such cases closely resembling that of Dichomyces. The main body of the receptacle is however asymmetrical, and the minute antheridium as far as can be ascertained from the material obtained occurs only on one side. It forms nevertheless an additional point of connection between the two genera, which may ultimately have to be united, despite the extreme differences between the type species in either case. Dichomyces infectus, nov. sp. Receptacle consisting of a short basal cell succeeded by an axile cell placed vertically, on either side of which a series of three obliquely superposed cells forms a blackened border. The remainder of the receptacle consisting of two successive transverse symmetrical rows of cells, the lower row made up of three central and several smaller ex- ternal cells terminating on either side in a short blunt projection below the prominent antheridia, beside which arise single sterile ap- pendages. The distal row is composed of seven cells, the external cells on either side not extending beyond the bases of the perithecia and destitute of appendages ; two appendages arise between the peri- thecia, one on either side. Perithecia two, closely approximated, arising from single broad flattened cells, short and stout, tapering slightly towards the subtruncate apex, which is destitute of papilla? or appendages. Perithecia 66 X 22 p. Receptacle 60 X 40 p. On Xantholinns obsidarius Melsh. Waverly, Massachusetts. Two specimens of this small form were fonnd at the tip of the abdomen of its host. THAXTER. — LABOULEENIACE.E. 103 DlCHOMYCES IN^QUALIS, nOV. Sp. Receptacle as in D. furciferus, its fork-like projections prominent and indistinctly septate, the distal row of cells bearing a single peri- theciuni but symmetrical except that the submedian cell, above which a second perithecium arises in D. furciferus, is much reduced in size. Appendages ten to twelve, one at the base of each antheridium, two above the median cell of the distal row, and three to four borne one from each of the three to four cells of the distal row external to the submedian cells, all arising as in D. furciferus. Perithecia large, slightly inflated towards the base or subcylindrical, tapering abruptly at the extremity to a subtruncate apex destitute of appendages. Spores 26 X 3.5 fx. Perithecia 100 X 25 p. Receptacle, length to base of perithecium 92 p ; length to tips of lateral forks 110-130 p; greatest breadth 50-60 p. Total length to tip of perithecium 180- 190 p. On Philonthus debilis Grav. Kittery Point, Maine, and TVaverly, Massachusetts. This species occurs sometimes in company with D. furciferus on the abdomen, more rarely on the legs and thorax of its host. It is at once distinguished by its solitary perithecium, which is destitute of the ear-like appendages peculiar to the last named species. The presence of a single perithecium necessitates a modification of the generic diag- nosis in this respect, but despite the absence of the usual pair, which seems to be invariable, the bilateral symmetry of the plant is other- wise maintained. More abundant material of D. furciferus shows the presence of an appendage placed beside each antheridium, a character also found in both the new species just described. Heima.tomyces aurantiacus, nov. sp. Pale straw-colored, the cell contents including numerous rather bright orange granules or oil globules. Perithecium, exceeding the tip of the receptacle by from one fifth to one quarter of its length, small, slender, the tip usually curved outwards, the lips rather prom- inent. Receptacle slender, the basal cell suffused with brown below, the sub-basal cell small and flat, the two succeeding cells elongate, the outer shorter and continued above by an unusually large basal cell of the perithecium : distal portion of the receptacle as in H. bm-ealis com- posed of three cells, the two lower very long and narrow, subtriangu- lar, obliquely superposed, their lower extremities nearly touching the sub-basal cell. Perithecium 50 X 14-15 p. Total length to tip of 101 PROCEEDINGS OF THE AMERICAN ACADEMY. receptacle 85-90 tt ; to tip of perithecium 100-110 it. Greatest breadth 22 /x. On Desmopachria convexa Aube. Kittery Point, Maine. A rare species occurring on the right elytron near the middle of its distal half. Distinguished from H. orientalis, which occurs also very rarely on the same host, and If. Bidessarius, by its slender form and orange color as well as the details of its structure. It is a very deli- cate species, seldom found in good condition. DlMORPHOMYCES MTJTICUS, nOV. sp. Male individual as in D. denticulatus, the basal cell more or less suffused with blackish. Female individual. Receptacle consisting of three superposed cells, the basal cell becoming subtriangular through the successive separa- tion, from its upper angles on either side, of a transverse series of cells, each of which gives rise to a single perithecium or a single sterile appendage in regular succession. Appendages simple, single, septate, seldom equalling the perithecia in length. Perithecia one to six, becoming slightly brownish and curved, subclavate, notched on one side below the truncate or bluntly rounded asymmetrical apex and destitute of any tooth-like outgrowth. Spores 22-25 X 3 it. Peri- thecia 75-90 X 15 it. Total length to tip of perithecia 90-130 it. On Falagria dissecta Er. Maine and Massachusetts. This species corresponds essentially in structure with D. denticu- latus, from which it is readily distinguished by its perithecia, which are larger and without the peculiar tooth-like appendage of the last named species. The male individuals can hardly be distinguished ex- cept for the suffusion of the basal cell. In the following synopsis the genus Hesperomyces Thaxter has been dropped as synonymous with Stigmatomyces, and all published species have been included, without regard to certain probable cases of synonymy. The arrangement suggested is entirely provisional, yet indicates in a general way the natural sequence of the genera. THAXTER. — LABOULBENIACE^E. 105 I. ANTHEROZOIDS ENDOGENOUS. A. Antheridia composed of several Cells. § Dioecious Dimorphomyces 1. On Falagria dissecta Er denticulatus. 2. On Falagria dissecta Er muticus, §§ Monoecious. * Antheridium borne on an appendage free from the receptacle. a. Antheridium lateral below a terminal branch of the appendage Cantharomyces 1. On Bledius assimilis Bledii. 2. On Bledius armatus Er occidentalis. 3. On Trogophlceus sp pusillus. b. Antheridium terminal tipped by a spine-like process. Haplomyces 1. On Bledius ornatus Lee calif ornicus. 2. On Bledius rubiginosus Er texanus. 3. On Bledius emarginatus Say virginianus. c. Antheridium terminal with a prominent apical pore. Camptomyces 1. On Sum us prolixus Er melanopas. ** Antheridium united to the body of the receptacle from which its pointed apex projects. a. Peritheoia terminal, free from the asymmetrical recep- tacle. Terrestrial Peyritschiella 1. On Platynus cincticollis (Say) curvata. 2. On Platynus cincticollis (Say) minima. 3. On Plerostichus luctuosus Dej. and P. patruelis Dej geminata. 4. On Philonthus debilis Grav nigrescens. b. Perithecia terminal, free from the symmetrical recep- tacle. Terrestrial Dichomyces 1. On Philonthus debilis Grav furciferus. 2. On Philonthus debilis Grav incequalis. 3. On Xantholinus obsidarius Melsh infectus. c. Perithecium wholly or partly united to the asymmet- rical receptacle on one side. Aquatic Heimatomyces 1. On HaMplus riificollis DeG. and Cnemidotus muticus Lee Halipli. 106 PROCEEDINGS OF THE AMERICAN ACADEMY. 2. On Laccophilas maculosus Germ., L. hyalinus Dej., and L. minutus Sturm paradoxus. 3. On Laccophilus maculosus Germ. . . . appendiculatus. 4. On Laccophilus hyalinus Dej. and L. minutus Sturm melanurus. 5. On Laccophilus maculosus Germ. Hydro- porus spurius Lee. and sp. indet. . . . marginatum. 6. On Laccophilus maculosus Germ, and Hydro- porus spurius Lee rhyncostoma. 7. On Laccophilus maculosus Germ. . . . lichanophorus. 8. On Laccophilus maculosus Germ, and Hydro- porus spurius Lee uncinatus. 9. On Laccophilus maculosus Germ hyalinus. 10. On Laccophilus maculosus Germ, and Hydro- porus sp affinis. 11. On Laccophilus maadosus Germ., Hydroporus spurius Lee., and gen. indet simplex. 12. On Bidessus granarius Aube Bidessarius. 13. On Desmopachria convexa Aube horealis. 14. On Desmopachria convexa Aube .... aurantiacus. B. Antheridia composed of single Cells. Dioecious Amorphomyces 1, On Falagria dissecta Er Falagrice. 2. On Bledius basalis Lee jloridanus. § Monoecious. * Antheridia borne in definite series on the appendages. 7 Antheridia springing directly from successive cells of the appendage. a. Appendage solitary, bearing the antheridia in sev- eral vertical series Helminthophana 1. On Nycteribia Dufourii Nycteribice. b. Appendage solitary, bearing the antheridia in a sin- gle vertical series Stigmatomtces 1. On Drosophila nigricornis Loew entomophila. 2. On Musca domestica L Baeri. 3. On Chilocorus bivulnerus Muls virescens. c. Appendages numerous, arising directly from the receptacle, bearing the antheridia in a single vertical series Idiomyces 1. On Deleaster dichrous Grav Peyritschii. THAXTER. — LABOULBENIACE^E. 107 yy Antheridia borne on branches of the appendages. d. Appendages several, the antheridia borne on lateral branches in a single vertical series . . Cokethromyces 1. On Cryptobium pallipes Grav. and C. bicolor Grav. Cryptobii. 2. On Lathrobium nitidulum Lee setigerus. 3. On Lathrobium jacobinum Lee. and L. collare Er. jacobinus. e. Appendage single, with terminal sterile branches ; an- theridia borne below its successive septa as lat- eral branches, often branched or irregular Rhadinomyces 1. On Lathrobium nitidulum Lee. and L. punctidatitm Lee cristatus. 2. On Lathrobium fulvipenne Grav., L. punctidatum Lee, and L. angular e Lee pallidus. ** Antheridia not borne in any definite series on the appendage. a. Receptacle of two superposed cells, the upper bearing several appendages and one or more stalked perithecia Compsomyces 1. On Sunius longiusculus Mann verticillatus. b. Receptacle closely multicellular, bearing numerous ver- tical cells from which arise terminally several appendages and solitary stalked perithecia Moschomyces 1. On Sunius prolixus Er insignis. c. Receptacle typically nine-celled ; the appendages two or more, terminal, the inner fertile . . . Laboulbenia 1. On Anchomenus viduus Pz., A. albipes, and Platy- nus extensicollis Say anceps. 2. On Harpalus pennsylvanicus DeG arcuata. 3. On Antennophorus caput-carabus armillaris. 4. On Acrogenys hirsuta Maclean australiensis. 5. On Brachinus mexicanus Dej. and spp. indet. . Brachini 6. On Patrobus longicornis Say and P. tenuis Lee. . brachiata. 7. On Casnonia pennsylvanica Dej Casnonice, 8. On Catoscopus guatemalensis Bates Catoscopi. 9. On Clivina dentifemorata Putz Clivince. 10. On Bembidium spp. indet. compacta. 11. On Anisodactylus baltimorensis Say compressa. 12. On Harpalus pennsylvanicus DeG conjerta. 13. On Platynus extensicollis Say contorta. 14. On Coptodera Championi Bates Coptoderce. 108 PROCEEDINGS OF THE AMERICAN ACADEMY. 15. On Pcederus Uttorarius Grav., P. obliteratus Lee, P. rujicollis Fabr., and sp. indet cristata. 16. On Bembidium bimaculatum Kirby curtipes. 17. On Harpalus pennsylvanicus DeG elegans. 18. On Platyaus cincticollis (Say) elongata. 19. On Chlcenius ce?ieocephahts Dej., C. Chrysoceph- alus Rossi, Callistus lunatus Fabr., and Ap- tinus mutilatus Fabr europcea. 20. On Chlcenius vestitus F fasciculata. 21. On Anisodactylus Harrisii Lee, A. nigerrimus Dej., and A. interpunctatus Kirby .... Jilifera. 22. On Bembidium lunatum Duft., Anchomenus al- bipes F., and A. marginatus L jlagellata. 23. On Platynus cincticollis (Say) fumosa. 24. On Galerita janus Fabr., G. mexicana Dej., G. atripes Lee., and sp. indet Galeritce. 25. On Platynus extensicollis Say gibber osa. 26. On Gyretes sericeus Lab., G. compressus Lee., and G. sinuatus Lee Guerinii. 27. On Gyrinus fratemus Coup., G. affinis Aube, G. aralis Say, G. conjinis Lee., G. consobrinus Lee., G. plicifer Lee., G. ventralis Kirby, G. urinator I\Y\g., and spp. indet Gyrinidarum. 28. On Harpalus pennsylvanicus DeG Harpali. 29. On Bradycellus rupestris Say injlata. 30. On Galerita leptodera Chaud longicollis. 31. On Bembidium varium Oliv. and spp. indet. . luxurians. 32. On Galerita mexicana Chaud., G. nigra Che v., and G. cequinoctialis Chaud mexicana. 33. On Calleida pallidipennis Chaud minima. 34. On Morio georgice Pal Morionis. 35. On Nebria pallipes Say, N. Sahlbergi Fisch, N. Gregaria Fisch, N. Brunnea Duft., and N. Villce Dej Nebrice. 36. On Pachytelis mexicanus Chaud Pachytelis. 37. On Panagceus crucigerus Say, and P. fasciatus Say Panagcei. 38. On Platynus extensicollis Say, P. ceruginosus Dej., and sp. indet parvula 39. On Platynus melanarius Dej., P. ruficornis Lee., and P. extensicollis Say paupercula. THAXTER. — LABOULBENIACE.E. 109 40. On Bembidium spp. indet pedicellata. 41. On Pheropsophus cequinoctialis Linn., P. margi- natus Dej., and spp. indet Pheropsophi. 42. On Philonthus debilis Grav., P. cunctans Horn, P. micans Grav., P. eequalis Horn, P. cali- fornicus Mann., and spp. indet Philonthi. 43. On Olisthopus parmatus Say, Stenolophus lim- balis Lee, S. fuliginosus Dej., Badister maculatus Lee, Harpalus pleuriticus Kirby, Agonoderus pallipes Fabr., and gen. indet. . polyphaga. 44. On Eudema tropicum Hope, Chlcenius auricollis Gory., and Dolichus ? sp proliferans. 45. On Pterostichus adoxus Say, P. luctuosus Dej., P. mancus Lee, and sp. indet Pterostichi. 46. On Quedius vernilis Lee Quedii. 47. On Platynus extensicollis Say ....... recta. 48. On Brachinus crepitans L., B. explodens Duft., B. scolopeta F., and (?) Platynus cincticollis Say Rougetii. 49. On Platynus extensicollis Say scelophila. 50. On Schizogenius Uneolatus Say, and S. ferrugi- neus Putz Schizogenii. 51. On Anophthalmia Menetriesii Motsch., A. angus- tatus Lee, and A. MotschidsJcyi Schm. . subterranea. 52. On Bembidium spp. indet. . truncata. 53. On Stenolophus ochropezus Say „ umbonata. 54. On Anomoglossus pusillus Say, Chlcenius cestivus Say, 0. cumatilis Lee, C. cursor Chev., 0. leucoscelis Chaud., C. pZoridanus Horn, C. pennsylv aniens Say, C. ruficaudis Chaud., C. sparsus Lee, G. texanus Horn, C. tricolor Dej., C. viridicollis Reiche, Omophron ameri- canum Dej., 0. m'mbatum F., and spp. indet., Patrobus longicornis Say, Platynus extensicol- lis Say, Pterostichus adoxus Say, P. luctuosus Dej., P. corvinus Dej., P. oaudicollis Say, and Nebria pallipes Say variabilis. 55. On Bembidium littorale Pz., B. fasciolatum Duft., B. punctulatum Drap. , B. lunatum Duft., B. obsoletitm, Dej., and spp. indet vulgaris. 5G. On Crepidogaster bimaculata Boh zanzibarina. 110 PROCEEDINGS OP THE AMERICAN ACADEMY. d. Receptacle two-celled, appendage single, bearing a series of branches superposed in a single row. Sphaleromyces. 1. On Lathrobium nitidulum Lee. and L. punctula- turn Lee Lathrobii. e. Receptacle of numerous cells superposed in a single series giving rise directly on one side to fertile appendages, one or two peritbecia, and sterile appendages in the order named .... Ch^etomtces 1. On Pinophilus latipes Er Pinophili. f. Receptacle consisting of a primary axis of several to many superposed cells and a secondary series of smaller cells irregularly placed and bearing numerous bristle-like appendages Acanthomyces 1. On Atranus pubescens Dej lasiophora. 2. On Lathrobium longiusculum Grav. and sp. indet. Lathrobii. 3. On Lathrobium fulvipenne Grav brevipes. 4. On Othius fulvipennis Fab furcatus. 5. On Anophthalmias Bilimeki Sturm hypogceus. 6. On Colpodes evanescens Bates longissimus. g. Receptacle multicellular, bearing distally two ap- pendages on either side at the base of a stalked perithecium. . . Thaxteria ( Giard nee Sacc.) 1. On Mormolyce phyllodes Hagenb Kunkelii. h. Receptacle of three cells terminated by a horizon- tal series of numerous cells bearing a circle of appendages and one or more stalked perithecia surrounded by them Teratomyces 1. On Acylophorus pronus Er mirifiens. 2. On Actobitts nanus Horn Actobii. 3. On Actobius nanus Horn brevicaulis. II. ANTHEROZOIDS EXOGENOUS. Typically aquatic. a. Receptacle of few of many superposed cells, running into the branch bearing appendage on one side and the wall of the perithecium on the other, the wall cells of the latter superposed in four many-celled rows Ceratomyces 1. On Tropistemus glaber lib. and T. nimbatus Say mirabilis. THAXTER. — LABOULBENIACE^I. Ill 2. On Tropisternus glaber Hb. . . . . . . camptospoms. 3. On Hydro comb us fimbriatus Melsh. and Philhy- drus ductus Say rostratus. 4. On Tropisternus glaber Hb. and T. nimbatus Say JiUformis. 5. On Tropisternus nimbatus Say minisculus. G. On Lathrobium punctulatum Lee terrestris. 7. On Berosus striatus Say contortus 8. On Berosus striatus Say furcatus. 9. On Berosus striatus Say humilis. Receptacle parenchymatously multicellular, numer- ous perithecia and appendages arising from its cup-shaped extremity Zodiomyces. 1. On Hydrocombus lacustris Lee, H. jimbriatus Melsh., and gen. indet vorticellarius. 112 PROCEEDINGS OF THE AMERICAN ACADEMY. IV. ON THE FORMATION OF VOLATILE COMPOUNDS OF ARSENIC FROM ARSENICAL WALL PAPERS. By Charles Robert Sanger. Received November 9, 1893. In cases of chronic poisoning from arsenical wall papers, the form in which the arsenic leaves the paper has been a long disputed ques- tion. Certain well defined symptoms were easily attributable to the local irritant action of arsenical dust in the form of arsenious oxide, the copper greens, etc., and the extension of these symptoms to the digestive organs could be referred to the same cause. Yet, on the other hand, under conditions unfavorable to the separation of arseni- cal dust, as, for instance, where an arsenical paper underlay one that was perfectly free, cases of poisoning have been very frequent, and to explain them the formation of a gaseous or volatile arsenical compound was assumed. Experiments to decide this point have been numerous since chronic arsenical poisoning by wall papers was recognized, but none of them until recently have been decisive, as they were with few exceptions carried on under unfavorable conditions and by methods that in point of exactness left much to be desired. Each side has had its adherents, the one asserting that chronic poisoning must be due entirely to dust, the other declaring that the cause lay in the absorption of an arsenical gas, or in the combined action of gas and dust. Many, not believing that the absorption of dust could alone lead to the numerous extreme or obscure cases that have been observed, have used the lack of proof of gas formation as an argument against the possibility of chronic poisoning from arsenical surroundings, although they admitted the local action of dust in certain cases. Further, those who denied the possibility of arsenical poisoning from wall paper under any circum- stances considered their position strengthened by the failure to establish definitely the formation of a gaseous compound. It was in this doubtful state that the question lay at the time when this investigation was begun, some seven years ago. My work was SANGER. — VOLATILE COMPOUNDS OF ARSENIC. 113 frequently interrupted and delayed, but about two years ago, I arrived rather reluctantly at results which seemed to show the improbability of the formation of an arsenical gas, though the possibility was as clear as ever. But the extension of bacteriological methods to the investigation has thrown a new light on the matter, and during the past year I have been able to satisfy myself of the correctness of the investigations which have shown that a volatile arsenical compound can be formed by action of certain moulds on organic matter contain- ing arsenic. The nature of the compound is still unknown, and will require extended investigation before it is determined; but it certainly does not seem to be arseniuretted hydrogen, which was the compound commonly supposed to be formed. Historical. The danger from arsenical colors in living rooms seems to have been to some extent foreseen before cases of chronic poisoning actu- ally arose, for an order of the Prussian government of January 12, 1815,* directed that the color should be rubbed from green walls when wet. and not when dry. Gmelin,f in 1839, was the first to call attention to the danger from arsenical wall papers, and considered that the arsenic might be volatil- ized as some product of the fermentation of the organic matter with which it was mixed. Riedehf in 1844, says that "air can take up arsenic acid and hydrogen can form arseniuretted hydrogen." Von Basedow, § in 1846, considered that the arsenic might be liber- ated in the form of kakodyl. He gives some cases of poisoning in rooms painted with an arsenical green on a lime ground, and lays stress on the peculiar garlic odor which he notices. This he com- pares to the odor obtained by boiling arsenious oxide with acetic acid, and says it is similar to that wbich arises from stuffed birds. The occupants of the room noticed no such odor. Von Basedow notes that the arsenic decreases in the paper by decay, but no analyses are given in support of this assertion. Krahmer,|| in 1852, was the first to institute experiments to deter- * Krahmer, Deutsche Klinik, 1852, XLIII 481. t Ref. by Eulenberg, Die Lekre v. d. sr!,ddl. v. gift. Gasen, 1865, p. 413, to Knrlsruher Zeitung, November, 1839. J Kef. by Krahmer, lac. cit., to Ann. der Staatsarzneikunde, X. 407. § Preuss Med. Zeitung, 1846, X. 43. || Loc. cit. VOL. XXIX. (n. S. XXI.) 8 114 PROCEEDINGS OP THE AMERICAN ACADEMY. mine whether a volatile arsenical compound was liberated. He ridiculed the ideas of Riedel and v. Basedow, and considered that the arsenic could not even leave the walls as dust. His experiments were as follows. Four grams Neuwiecler green were mixed with paste and lime taken from a damp part of a ground floor wall, and placed in a two-necked WoulfF bottle. After 19 days he observed no odor of garlic, nor indeed could he observe any in the five years that elapsed before his results were published, during which time the mixture remained in the bottle. Five hundred litres of air were then drawn through the bottle for 21 days, the air first passing through alcoholic potassic hydroxide. The solution was then neutralized by hydrochloric acid, and, "on treatment with arsenic free zinc, gave no arsenic." We have here no idea of the delicacy or accuracy of the test used. Next, a solution of argentic nitrate was substituted for the potassic hydroxide and air drawn through for fifteen days. No change was observed in the silver solution, but it was not tested for arsenic. He then passed the air from the bottle for five days through a hard glass tube heated in two places and obtained no mirror of arsenic. He next thinks that the arsenic compound, if formed, should be mixed with hydrogen before absorption, and so places a hydrogen generator in front of the bottle. The hydrogen generated from 133 grams of zinc, after filtration by cotton wool, was passed through the bottle, then through a solution of alcoholic potassic hydroxide, then silver solution, and finally through the hot tube. He found no trace of arsenic in the solutions or in the tube. The test is not given. During the whole time there was no evidence of mould in the bottle. Krahmer states that he had lived for eight years in a dry and airy room, the walls of which had received during that time three coats of Neuwieder green, probably five pounds at a time. He never felt any trouble from its presence. He repeatedly examined the dust for color, but never found any. To decide the question as to arsenic in the dust, he drew fifty litres of the air of the room through nitric acid, and added to it the dust that had fallen on a piece of paper during that time. The acid was then neutralized with sodic carbonate, evapo- rated, and the sodic nitrate melted. The residue was dissolved in water, filtered, and the filtrate tested with zinc and hydrochloric acid. He found no arsenic, which is hardly to be wondered at in the presence of so much nitric and nitrous acids. He could not find copper in the filtrate, but this was probably lost by ignition and filtration. If Krahmer's experiments had been otherwise properly conducted, SANGER. — VOLATILE COMPOUNDS OP ARSENIC. 115 the method of examination for arsenic would alone have deprived them of value. Not finding, by his unskilful test, any arsenic in the dust, he concludes at once that arsenic cannot be liberated at all, and uses the result as an argument against the possibility of arsenical poisoning from wall papers. Arnd,* in 1855, offers the following remarkable explanation : " Ar- senates are decomposed by carbon dioxide, setting free arsenious oxide which is volatilized. By evaporation of the water from the paste, arsenical particles are carried off. Sulphide of arsenic on a lime ground is decomposed with evolution of arseniuretted hydrogen, as the lime takes away the sulphur, forming calcic sulphide." Eulenbergf mentions the following experiment made in 1857 by Halley and Williams, but I can find no account of it or reference to it elsewhere. Several sheets of filter paper were soaked in ammonia- cal argentic nitrate and hung up in a closed room of which the paper contained . Schweinfurth green. Gas lights were kept burning ten hours, and after they were put out the room was kept closed ten hours longer. The papers, which were blackened, were digested with hydrochloric acid for thirty minutes, and a piece of copper was laid in the solution. A gray deposit was formed on the copper, and the latter, after being washed and dried, was heated in a bulb tube. A white ring was obtained and said to be arsenic, though no further proof was given. Yet, if it were arsenic, which is not decided, it mi^ht have had its source in the dust of the room, as well as from a volatile compound. Halley, in 1858, in a letter to the London Times, $ tests the air of a room containing an arsenical wall paper by merely hanging up sheets of paper soaked in ammoniacal argentic nitrate, in which he observes after some time the formation of numerous "well defined crystals of arsenious oxide, visible under a low power microscope." I am inclined to think that this experiment is merely the foregoing, subjected to a newspaper condensation, which has omitted essential details. Campbell, § in 1858, made the following experiment. Strips of arsen- ical paper, about one square foot in all, were placed in a bottle contain- in"' a thermometer and fitted with a double bored cork. Through one * Verhandl. d. Verein f. Staatsarzneiwissensch. in Berlin, 1855, 1. 47. t Loc. cit., p. 416. J Pliarmac. Journ. and Transactions, 1858, p. 428. § Ibid., p. 520. 116 PROCEEDINGS OF THE AMERICAN ACADEMY. boring passed a right-angled hard glass tube, terminating in a spiral, which could be heated by a broad burner. The other boring held an exit tube passing to a solution of potassic hydroxide. Gas jets were kept burning in the room, and air, at a temperature from 16 to 60°, was drawn from the bottle through the solution. Several trials were made, each lasting one hour, and in no case was arsenic detected. How the arsenic was tested for is not stated, but it is not surprising that none was found. The absurdity of the experiment was shown by Taylor* a week or so later. Abel,f in the same year, made the following experiments at the instance of the British government. A room was selected having a green unglazed paper containing one fifth of a grain of arsenious oxide per square inch (20 grams per square meter). 1. The room was closed for 36 hours, and then the air was led between four and five hours through a solution of argentic nitrate, and then through a tube containing asbestos soaked in ammoniacal argen- tic nitrate. No arsenic was found in solution or tube. 2. Five gas flames were burned in the room for several hours, and at the end the air was led through the same reagents with the same negative result. The method used in testing is not given. 3. A glass tube, 3.5 ft. long and 2 in. in diameter, was filled with small pieces of arsenical paper, warmed to 32°, and connected with the reagents as above. The air of the room was led through for a week, and from time to time gas burners were lighted. 4. The air was passed through a strong solution of sulphurous acid and then led for 24 hours through the apparatus. 5. The products of combustion of a gas burner were concentrated by a funnel and led through the apparatus for three days, and again (6) for four days. 7. The paper was roughened by friction and hot air led through for nine days. 8. Strips of arsenical paper with decomposing paste were placed in the same tube, and hot air led over for nine days, with products of gas combustion also. In none of Experiments 3 to 8 was arsenic found in either silver solution or tube. Yet as the method of test- ing for arsenic is not given, and we do not know its accuracy, the experiments are thus deprived of considerable value. * Pharmac. Journ. and Transactions, 1858, p. 553. t Ibid., p. 556 ; also British Review, October, 1858, and Schmidt's Jahrbuch, 1859, XLV. 20. SANGER. — VOLATILE COMPOUNDS OF ARSENIC. 117 9. 600 grains (39 grams) of emerald-green were distributed in a lot of cotton wool and placed in a jar, into which a tube was plunged containing cotton wool and connected with the silver solution. Air was drawn through at 32° for some time, but no arsenic was detected in the solution. Phillips,* in 1858, repeated the experiment of Halley. Two clos- ets were used, one containing 48 sip ft. (4.46 sq. m.) of a paper containing 11.8 grains arsenious oxide, as Schweinfurth green, to the square foot (8.3 grams per square meter), and another containing 53 sq. ft. (4.92 sq. m.) of the same paper. In each closet were placed two dishes, one containing a solution of potassic hydroxide and the other ammoniacal argentic nitrate, besides a sheet of paper saturated with the latter. The closets were closed 72 hours and = 12, in which 12 is a known skew symmetric matrix, and <£ denotes the transverse of . Cayley's solution of this equation fails for those matrices of which —1 is a latent root. If 12 is real, and is both skew symmetric and orthogonal (i. e. if O-1 = 12 = — £2), the product of two of Cayley's expressions gives every real solution of this equation. Thus, let be any real solution of the equation <£ 12 = 12, of which —1 is a latent root. A polynomial = /(<£) in integer powers of <£ with real coefficients can be formed, containing every factor in the identical equation to , except <£ + 1, and such that 4>2 = <&, $ 12 = 12 $.* * The latent roots of being ±1 of multiplicity m and n respectively, 9r> 9~l> each of multiplicity pr, for r = 1,2, ... i, the identical equation to

= 0 (1 - 2 $) J then — 1 is not a latent root of , and 0 is a solution of the equation Q (f> = Q. The matrix <£0 — (1 — 2 4>) is also a solution of this equation; and we have o = 1. Therefore 0 = if/ we then have ^ = — if/. Since <£0 = O-1 if/, and since <£02 = 1, we have fi-^fi-1^ = 1. Therefore, if/ Q, \f/ = — if/ O if/ = \f/U~lif/ = fi. Consequently if/ is also a solution of the equation. The matrix has no latent root equal to — 1. The matrix , and therefore 0 , are real ; if/ is then a real skew symmetric matrix, and consequently has no latent root equal to — 1. Therefore hoth (f> and if/ may be represented by Cayley's expression. Therefore the most general real solution of the equation Cl cf> = 12 , in which 12 is a real skew symmetric matrix such that £22 = —1, is 0 = fi-1 (Q - Y)-1 (Q + Y) (O — Y')"1 (« + Y'), in which Y and Y' are arbitrary real symmetric matrices. The ex- pression (fi — Y)-1 (O + Y) (O — Y')"1 (fl + Y>) is equally general. Clark University, Worcester. # = [{* + !)» -(1 + 1)"]™ Ft(0) F2(0) F,(tf>) (- 1)» (2)— ^i(-l) F2(-l) ' • Fd-1)* Since 0 is real, its imaginary latent roots occur in pairs which are conjugate imaginary, and have the same multiplicity. From which it follows that * is real. 180 PROCEEDINGS OF THE AMERICAN ACADEMY. VII. CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY, UNDER THE DIRECTION OF E. L. MARK, XXXIX. ON SOME LAWS OF CLEAVAGE IN LIMAX. A PRELIMINARY NOTICE. By C. A. Kofoid. Communicated by E. L. Mark, January 10, 1894. The following is a statement of the results obtained from the study of cleavage in Limax, and of the literature of cell lineage in other invertebrates. It is desirable to confirm my results by a study of cleavage in other forms before the publication of my final paper, and it has therefore seemed best not to defer the presentation of the conclusions to which I have arrived. A few words in regard to the usage of terms will be necessary. The egg is regarded as having the animal pole uppermost, and the terms right and left, upper and lower, are used as resident in the egg itself. Or, to express it in another way, a miniature observer is imagined as placed in the principal (vertical) axis of the egg, with his head at the animal pole, facing the part or parts of the egg under discussion, and the terms right and left, upper and lower, are used as determined by this observer. By " a gen- eration of cells " is meant all those cells which are removed from the ovum by the same number of cell divisions, regardless of the time of appearance or position of such cells, i. e. the word is used in its literal sense. This is not the usage of Fol (75) or Blochmann ('81), who employ the term in its literal sense with reference to the blastomeres through the four-cell stage, but thereafter use it to designate successive sets of four micromeres, naming them in the order of their appearance in time. As is well known, cells cleave in sets of fours throughout the spiral period of cleavage. The cleavage of the individual cells of the set may be synchronous or successive, and the cleavage of KOFOID. — LAWS OF CLEAVAGE. 181 any given set may or may not coincide with that of the other sets of the same generation ; but whatever the modifications, each cell in its origin bears a close relation to three other cells, and these sets of four related cells of co-ordinate origin will be called quartets. During spiral cleavage the egg is made up of a num- ber of superposed quartets, and it may be compared to a house of as many stories, each story representing a single quartet of four cells. The regions of the cleaving egg occupied by the four blasto- meres of the four-cell stage and their derivatives during the spiral period will be called quadrants, and the four primitive blastomeres and their respective derivatives will be designated by the letters a, h, c, d, taken in the order in which the hands of a clock move. In Nereis and Umbrella, these designate the left anterior, right anterior, right posterior, and left posterior quadrants respectively. The term spiral will be used to indicate the divergence imme- diately after cleavage of the centre of the nucleus of the upper one of two daughter cells from the vertical plane passing through the corresponding portion of the lower cell and the vertical axis of the egg. The spiral will be a right spiral when the divergence is toward the right (as defined above), and a left one when the divergence is toward the left. Or, using Blochmann's ('81) com- parison to the hands of a clock, when the egg is observed from the animal pole the spiral is a " right " one if the divergence of the upper cell is in the direction of the motion of the hands of the clock, a "left" spiral if the divergence is opposite the motion of the hands of the clock. By this method of nomenclature all those spirals whose spindles stand in similar positions with reference to the vertical axis are given the same name. It should be noted in this connection, that this divergence, or apparent shifting of cells, in Limax at least, is the result of the obliquity of the plane of division, and is predetermined by the position of the spindle. This fact, whatever may be the cause of the particular position of the spindle, is the immediate basis of the phenomenon termed the "spiral." The position of the spindle primarily determines the position of the daughter cells, though mechanical environment may secondarily modify that position. The nomenclature of spirals as followed by Blochmann ('81), Lang ('84), Wilson ('92), Heymons ('93), and others, presents no 182 PROCEEDINGS OF THE AMERICAN ACADEMY. constant basis of reference. Not only have these authors named homologous spirals differently, as Lang ('84, p. 325) and Heymons ('93, p. 256), but no one of those named except possibly Lang has used the same method of naming all the spirals discussed. In some cases the lower cell of a pair of daughter cells is regarded as the fixed one, in other cases the upper cell is so regarded ; or, to express it differently, in some cases peripheral cells are regarded as fixed, in other cases axial ones. In general it seems to have been the custom to consider the larger of the daughter cells as fixed, and the budding smaller cell as the movable one. Two reasons may be cited for the employment of the relative size of the daughter cells as a basis for the nomenclature of the spiral. (1) The larger cell occupies more nearly the position of the mother cell, and it is therefore natural to regard the smaller cell as the movable one. (2) In the first spiral the larger cells (macromeres) are basal, and the micromeres upon them are there- fore regarded as the movable cells, and this basis adopted in the first spiral is suggested for other spirals. Though this refer- ence of the spiral to the relative size of the cell may furnish a logical basis for nomenclature of spirals where cleavage is un- equal, it cannot furnish one for those spirals, or eggs, in which cleavage is equal. Nor has this basis when once adopted been consistently followed in every case, — as, for example, in Wilson's paper on Nereis ('92). On page 391 he says: " A careful study of the embryo through these changes shows that all of the cell divisions conform to the spiral type. . . . It is also easily seen in the divisions of the secondary micromeres (a2, b2, c2, X). Each of them divides somewhat obliquely (cf. Figs. 25, 26, 33) so that one of the cells lies somewhat lower than the other, and in most cases the lower cell is obviously smaller than the upper. The difference in size is very great in the case of X and x', but is much less in the case of the others (a21, a22, Fig. 33). (In the speci- men shown in Figs. 25, 26, on the other hand, there is no appre- ciable difference in size, but I have never seen a case in which the upper cell is the smaller.) If this group of cells be followed around the embryo from right to left (against the hands of a watch), the upper (larger) cell always comes first ; i. e. the first division of the second group of micromeres takes place in a left-handed spiral, like the second division of the first set of micromeres." There is no escape from the conclusion that in this case the KOFOID. — LAWS OF CLEAVAGE. 183 cell of reference is the larger one, as it was in the case of the " third and fourth cleavages " (cf. Wilson, pp. 387 and 388). That it is the " larger " rather than the " upper " cell, will be seen when in Wilson's Fig. 21 (reproduced in outline in my Plate I. Fig. 5) we apply to the spirals foreshadowed in the spindles of the cells dl, X (— d'2), and D, the method employed by him in naming this spiral (a2, b2, c2, X). Following around the embryo from right to left the "upper " cell (indicated by the upper end of the spindle) comes first in all three cases, but the spirals are not all given the same name. The "larger" cell comes first in d1 and X, and the spirals are called left-handed spirals. The smaller cell comes first in D, and the spiral is called a right-handed spiral (p. 391). — Let me call attention, in passing, to the fact that, in the system of nomenclature I have proposed, the three spirals above referred to would be given the same name. They would all be called right spirals, and in this similarity of name would be recognized the similarity of the position of spindles, and the fact that in passing from right to left the upper cell always comes first. The basis on which my system rests is not the varying size of the cells, but the more fundamental factor of position. — Up to this point in cleav- age Wilson has consistently used his system of nomenclature, but upon the next page (p. 392), in discussing the third division of the primary micromeres a1, b1, c1, d1, resulting in the formation of the cells a1, b1, c1, d\ and the rosette cells a13, blz, c13, d13, he abandons the larger cell as the basis of reference, as will be seen in the following quotation : " The four primary micromeres (a1, b1, c1, d1) bud forth four small cells at their inner angles (at the upper pole) which arrange themselves in a very regular apical rosette, the cells of which alternate with the central micromeres (Figs. 27, 28, etc.). The position of the spindles is the same as in the first division of a1, b1, c1, d1 ; i. e. the division follows a right-handed spiral, but the character of the division is very different since the smaller cells are formed at the central instead of the peripheral angles of the cells (i. e. towards instead of away from the vertical axis of the embryo)." In this case the fact that " the position of the spindles is the same " is cited as a basis on which the spiral is named. In my Fig. 6, Plate I., is reproduced in outline Fig. 25, Plate XV., of Wilson's paper. If we apply to the cells c1 and c13 the test mentioned by Wilson on page 391, and name the spiral according to the size of the cell that " comes first," we must call it, not a right-hand, but a left- 184 PROCEEDINGS OF THE AMERICAN ACADEMY. hand spiral. This is the only case in which the similarity of the position of the spindles is recognized as a factor in the nomenclature of spirals. By this change in the basis of nomen- clature he has recognized the interesting fact that there exists in the three successive divisions of the primary micromeres (a1, b1, c1, d1) an alternation in the direction of the spirals. This alternation, as described and named by him, is inde- pendent of that which exists in the first three divisions of the macromeres. It belongs to an entirely separate system, and its relation to successive generations of cells is neither suggested nor discussed. Not only does the system of nomenclature based on the size of the cells fail to furnish a logical basis for cases of equal cleavage, but it also fails to furnish such a basis for the comparison of cleavage in different forms, for the cells in which the yolk is lodged in the progress of cleavage are by no means homologous cells in different species (cf. Nereis, Neritina, and Umbrella). It therefore seems to me very much more logical to base the nomenclature of spirals upon constant spatial relations than upon the relative sizes of the daughter cells, which are inconstant, or upon the apparent greater shifting of one of the daughter cells referred to no constant plane. The student of cell lineage finds his task much complicated by the various systems of nomenclature already employed, and it seems a pity to introduce still another to add to the confusion. But as all systems heretofore employed fail to recognize the fundamental importance of sucessive generations of cells, and as these are the basis of my treatment of the subject, it has been impossible to adopt any of the existing systems. At the same time, it is believed that the system proposed, resting as it does upon generations of cells, is adapted to all forms of spiral and radial cleavage, and will furnish a satisfactory and convenient means of comparison in these cases. Its applicability to spiral cleavage will now be discussed more fully. Each individual cell of an egg in spiral cleavage can be traced back to one of the four blastomeres a, b, c, d, i. e. it belongs to a definite quadrant. It also belongs to a definite quartet or " story " of the egg; it is likewise removed from the ovum by a definite number of cell divisions, i. e. belongs to a definite generation. Any system of nomenclature involving these three factors will both localize the cell and outline its ancestry. Each cell of the KOFOID. LAWS OF CLEAVAGE. 185 spiral period of cleavage may therefore be designated by three characters: (1) a letter, as a, b, c, d, indicating the quadrant ; (2) a first exponent indicating the generation, as a*', a5-, etc. ; (3) a second exponent indicating the quartet or story, as a4-1, a4,2, a7,10, etc. Generations are numbered starting with the ovum as the first generation. The number of cells doubles with each succeed- ing generation, and after the third generation, i. e. after the four-cell stage or the first quartet, the number of quartets is also doubled. The quartets or stories are numbered from the vegeta- tive toward the animal pole. Thus in the eight-cell stage the lower quartet is designated by the exponent 1, and the upper quartet by the exponent 2. This principle is followed in the nomenclature of all quartets. During the period of spiral cleav- age, the two daughter cells (or quartets) resulting from the cleavage of any given cell (or quartet) never lie in the same plane, and the lower cell (or quartet) of the two is always designated by an odd exponent and the upper by an even one. Thus, when ft5.3 ^ — ^5.3) divicieSj the resulting cells (quartets) are a.6-5 ( — d6-5) and a6,6 ( — «2T> must all be shifted one place to the right ; i. e. in the direction of the hands of the watch, as indicated by the long arrows outside the limits of the figure in Figure 10. In my interpretation the quartet a3-d3 is the quartet designated as a2-d2, and vice versa. The change is made upon the following grounds. (1.) The spindle in the cell b, for example, Figures 8, 9, indi- cates that the cell b3 (Fig. 8, == my rf6-2) will lie above and to the right of the cell b (my rf6-1), and perhaps higher (Figs. 9, 10) than the upper derivative (my fZ6-4) of the cell b2 (my d5-2), and that it will lie in contact with that derivative (deA) and with the lower derivative (r<6-3) of the cell a2 (my a5-2). This is my interpreta- tion of the position of the cell bs, as foreshadowed by the spindles of Figures 8, 9. However, in Figure 10 Blochmann lias placed the cell b3 on the left side of the cell b, and in contact with a2l and b2, a position directly contradictory to that indicated by the position of the spindle of the cell b in Figures 8, 9. KOFOID. — LAWS OF CLEAVAGE. 193 (2.) The spindle of the cell b2 in Figures 8, 9, indicates that the products of its division will lie upon the upper left side of the cell b, the upper derivative being in contact with the cell b3 and with the lower derivative. Blochmann's interpretation, on the other hand, throws the upper derivative completely out of the quadrant of the cell b, and over upon the cell c of an adjoining quadrant. The position of the cells az-dz and a2l-d2l, as figured by Blochmann in Figure 10 can be explained only on the assumption of a rotation of both sets of spindles (or of the upper derivatives of the spin- dles) as shown in Figure 9 about ninety degrees to the left. The fact that he has not observed such a rotation renders its existence all the more improbable. There remains, however, important evidence in favor of Bloch- mann's view, namely the presence in the cells a2 and c2 of " eine Anhiiufung von kleinen stark lichtbrechenden Kornchen." After the cell divisions indicated in Figures 46, 47 (Plate II. Figs. 8, 9), all of these granules are found in the cells a2 and c2, hence the idea of their derivation from the cells a2 and c2. Such a view must presuppose (1) that in the equal division of the cells a2 and c2 all of these granules go to on$ of the supposed daughter cells, i. e. to a2 and c2 ; (2) that granules similar to those of the cells a2 and c2 cannot arise in other cells derived, like a2 and c2, from the macromeres. If, on the other hand, the view advanced here as to the interpretation of Figure 48 (Plate II. Fig. 10) is correct, we may suppose that in the equal division of the cells a2and c2the granules accumulated in them were shared by the two daughter cells o„ and a5, c2 and c3 (using Blochmann's designations, but not his interpretation), and thus the granules were divided and be- came less conspicuous. At the same time the cells a2l and c2 of the third set of micromeres, originating, as I believe they do, from the macromeres b and d, show granules similar to those of the cells a2 and c2 of the second set of micromeres. This may afford an escape from the dilemma as to the relation of these groups of cells in which Blochmann is placed by the contradic- tory evidence afforded on the one hand by the position of the spindles, on the other by the presence of the granules in the cells. With regard to the origin of the cells of the quartet a2n-d2n of Figure 50 (Plate II. Fig. 11), I wish to present a view which differs from that of the author. As the nomenclature indicates, he derives this group from his quartet a2l-d2 of Figure 48 (Plate vol. xxix. (n. s. xxi.) 13 194 PROCEEDINGS OF THE AMERICAN ACADEMY II. Fig. 10). There are indications, however, that they were re- ally derived from the apical quartet a^-d-^ ; for (1) their nuclei are nearer those of the apical quartet ; (2) the cells of the apical quartet are much smaller after the cells a2u-d211 appear than be- fore ; (3) a2-d2 have just arisen by a recent division, whereas some time has elapsed since the first division of the apical quartet, Figure 45 (Plate II. Fig. 7). If this interpretation holds, the cells «! and d211 result from the division of ax, while br and a2u come from b, etc. The inner part of the "cross" would thus arise from the first group of micromeres and the outer part from the third group. These suggested changes reduce the cleavage of Neritina to complete agreement with that of Nereis, Umbrella, and Limax. Wilson ('92, p. 439) compares the cleavage of Nereis with that of the poly clad Discoccelis, and also with that of the gaster- opods — taking Neritina and Crepidula as types — as follows : " Up to a late stage in the spiral period (twenty-eight cells) ev- ery individual blastomere and every cell division [in Nereis] is represented by a corresponding blastomere and a corresponding cell division in the embryo of the polyclad and in that of the gas- teropod." This statement must imply, it seems to me, some other interpretation of the cleavage of Neritina than that given by Blochmann himself, although Wilson makes no statement to that effect, but on the contrary says (p. 442), " It is impossible to explain the differences between the annelidan and molluscan cross by assuming inaccuracy of observation on Blochmann's part, since the pole cells of the lateral arms show a peculiar granulation that may be seen in the parent cells (a2-2, c2-2) from which they rise." The reduction of Neritina to agreement with Nereis and Umbrella in the manner I have suggested brings it also into har- mony with the law of alternation of spirals, and this affords an- other presumption in favor of the correctness of the revision here proposed. That the spirals do alternate in Neritina can be seen on an examination of the short arrows indicating my interpretation of the genetic relation of the cells of Neritina (Plate II.). The arrows show the relation of the cells, the head of the arrow in every case lying in the derivative which, according to my inter- pretation, is the upper one. The following table will assist in the comparison of the two interpretations and the determination of the spirals. KOFOID. — LAWS OF CLEAVAGE. 195 Revised. Blochmann. Spiral. Nomenclature of Cells. Generation. Cell Stage. Nomenclature of Cells. Left . . . a3, b3, c3, d3 3 4 a, b, c, d Right . . f«4'2\ 3 X 4 8 1 j> a Left . . . Left . . . [ «5-4 \, a4.2 \ ] a5-3 ^ a } I a5-1 J> a S 5 16 12 i S>ai Right . . Not figured. Right . . Right . . {:::;>«-} {%>*} {£>**} 6 28 24 20 a2 ^ * Left . . . Left . . . {£>-} {£>*■} 7 36 36 d m -v. 2 ^>^7! Only two spiral cleavages of the seventh generation are figured. It will be noted that in Plate II. Fig. 10, the order in position of the quartets of the sixth generation taken from the vegetative toward the animal pole is as follows: a61-rf61, ae-3-dGS, a6A-d6A, and aG-2-di)2, instead of a61-d6\ a62-d62, a63-dS3, and a6A-d6A, as in Limax, Plate I. Pig. 2. This change in the succession of quartets is due to the fact that in Neritina the quartet a52- d52 by virtue of the spiral occupies a position in the furrows between the macromeres a6i-dsl. Therefore, when the cleavages resulting in the sixth generation take place, the upper derivatives (a6-2- d62) of the quartet a51-d51 come to occupy a position nearer the animal pole than the upper derivatives (a6A-d6A) of the quartet a52-d5-2. The application of the system of nomenclature proposed in this paper is not thereby involved in difficulty, for the designation of the daughter cells is in all cases predetermined by that of the mother cell, according to the basis set forth on page 186, the upper derivative in every case receiving the even exponent and the lower one the odd. Thus, notwithstanding the secondary 196 PROCEEDINGS OF THE AMERICAN ACADEMY. changes induced in the position of the quartets by the accumu- lation of yolk in the basal quartet, the cells in ISTeritina retain the same nomenclature that is given to the homologous cells of a form such as Limax, presenting the typical undisturbed super- position of quartets. In all the literature examined there occurs but a single case where the position of an indicated spindle, contradicts the princi- ple of alternation of spirals. This is found in Lang ('84, Taf. 34, Fig. 20) in the division of the cells aey-de^ forming aex-dex and ae6- de6. Little can be said with regard to this except that we do not know the full nuclear history of this cleavage, and that it occurs at the close of the spiral period, immediately over the group of " Scheitelzellen " that has just sunk below the level of the remaining ectoderm. This may have produced secondary modifications in the direction of the spiral. It remains for me to speak of the influence of yolk upon cleav- age. Reference to Plate I. Fig. 2, shows that in Limax the cell b53 belonging to the quartet of this generation, which is the last to divide, is a member of the smallest quartet of the fifth generation. It will further be seen that the order of the division of the quar- tets, viz. first, second, fourth, third, is also the order of size from largest to smallest, the quartet of largest cells, i. e. those presum- ably with the greatest amount of yolk being the first to divide in this generation, as it was also in the preceding generation. This phenomenon is by no means confined to Limax. It is found in a great variety of forms, and the greater the amount of the yolk the greater seems to be the tendency of the cells of the yolk-laden quartet to divide before those of the smaller quartets. This can be illustrated by a comparison of Figure 1 (Plate I.), which shows the egg of Limax approaching the sixteen-cell stage, and Figure 7 (Plate II.), which shows ISTeritina approaching the same stage. In Limax there is little yolk, and the cleavage of the two quartets is almost coincident. In Neritina, however, where the yolk is more abundant, the lower quartet has divided, and its nuclei have assumed a " resting condition " before the cleavage of the upper quartet has fairly begun. Thus we have a well marked twelve- cell stage. The same tendency of the yolk-laden cells to cleave before the others is seen in the later stages of ISTeritina, and in- deed in many forms of Molluscan cleavage. It is also found in Nereis and Discoccelis. It is not, however, confined to the spiral KOFOID. — LAWS OF CLEAVAGE. 197 type of cleavage, for in Echinocyamus (Theel '92), where the cleavage is of the radial type, we find the meridional cleavages, which divide the quartets into "octets," starting with the fourth cleavage (eight- to sixteen-cell stage) at the vegetative pole, and reaching the apical quartet of the animal pole in the sixty-cell stage. Not only does the quartet having the largest cells divide first, but where there is an inequality in the size of the cells of a quartet, the largest cell or cells may divide first, as is the case in Discoccelis (Lang '84) and Unio (Lillie '93). In view of these facts I wish to raise the question as to the applicability to the individual cells of cleaving eggs of the law first formulated by Balfour (To) and concisely expressed in his " Comparative Embryology " ('80, Vol. I. p. 95), as follows : " Where the yolk spherules are fewest, the active protoplasm is necessarily most concentrated, and we can lay down as a general law that the ve- locity of the segmentation in any part of the ovum is, roughly speaking, proportional to the concentration of the protoplasm there ; and that the size of the segments is inversely propor- tional to the concentration of the protoplasm. Thus the seg- ments produced from that part of an egg where the yolk spherules are most bulky, and where therefore the protoplasm is least con- centrated, are larger than the remaining segments, and their formation proceeds more slowly." It is true that, of two eggs otherwise similar, the one with the larger amount of yolk in general cleaves more slowly. For ex- ample, Umbrella, which has a large amount of yolk, requires four days to reach the stage attained by Limax in one day. But of two cells of Limax or Umbrella during the cleavage period, that one is the first to divide which is the larger and presumably has the greater amount of yolk, and which in Balfour's terms has its protoplasm less concentrated. Not only this, but in the cases cited the greater the amount of the yolk the greater is the ten- dency of the division of the yolk-laden cell to precede that of the cell with less yolk. There are, to be sure, many cases where Balfour's law actually applies, as in the frog's egg ; but do not the cases cited above, belonging as they do to numerous and widely distributed classes of animals, form an important exception to the law as he has formulated it ? A paradox is thus presented. Yolk appears to delay cleavage in the cells of the frog's egg, to hasten it in the cells of the snail's egg. Yolk also appears to delay the development of an organism as a whole (cf. Limax and 198 PROCEEDINGS OF THE AMERICAN ACADEMY. Umbrella) while it may (in Limax and Umbrella) at the same time apparently hasten the cleavage of those cells of the organ- ism in which it is more abundant. This apparent conflict of statements has its foundation in the hypothesis that the amount of yolk alone is the decisive factor in the determination of the rapidity of cleavage. But there are other factors to be consid- ered, especially the quality of both yolk and protoplasm, and in these there may ultimately be found some solution of the diffi- culty. In the case of the presumably undifferentiated blasto- meres of the cleavage stages of Limax (and Umbrella) the differ- ence in the rapidity of cleavage is apparently correlated with the greater or less absolute amount of protoplasm in the individual cells. The amount of protoplasm, in turn, is dependent on both the quality of the yolk and the activity of the protoplasm. The yolk, by contributing to the amount of protoplasm in the larger cells, may thus indirectly hasten their division. In this, how- ever, the appropriation of the yolk by the protoplasm is the im- portant factor, for in case the protoplasm fails to appropriate the yolk with sufficient rapidity, the division of the yolk-laden cell may be delayed as in the frog's egg. This delay may depend on either one or both of the two factors, — quality of protoplasm and quality of yolk. These same factors probably determine the differences in the rates of cleavage of different eggs. While this does not afford a solution of the difficulties encountered in at- tempting to harmonize the facts here presented with Balfour's law of cleavage as influenced by yolk, future inquiries in the direction suggested may lead to a better understanding of the factors determining the nature of cleavage.* Cambmdge, Mass., December 22, 1893. * The substance of this paper was presented at the annual meeting of the American Morphological Society, in New Haven, Conn., December 29, 1893. KOFOID. — LAWS OP CLEAVAGE. 199 LITERATURE CITED. Balfour, F. M. '75. A Comparison of the Early Stages in the Development in Ver- tebrates. Quart. Jour. Micr. Sci., Vol. XVI. pp. 207-226, PL X. July, 1875. '80. A Treatise on Comparative Embryology, Vol. I., xi -4- 492 -f- xxii pp. London, 1880. Blochmann, F. '81. Ueber die Entwicklung der Neritina fluviatilis, Mull. Zeitschr. f. wiss. ZooL, Bd. XXXVI. pp. 125-174, Taf. VL-VIII. 19. August, 1881. Bobretsky, N. '76. Studien iiber die Embryonale Entwicklung der Gasteropoden. Arch. f. mikr. Anat., Bd. XIII. pp. 95-169, Taf. VIII.-XIII. 1876. Conklia, E. G. '91. Preliminary Note on the Embryology of Crepidula fornicata and of Urosalpinx cinerea. Johns Hopkins Univ. Circ, Vol. X. No. 88, pp. 89, 90. May, 1891. '92. The Cleavage of the Ovum in Crepidula fornicata. Zool. Anzeiger, Jahrg. XV. No. 391, pp. 185-188. 1892. Fischer, P. '80-87. Manuel de Conchyliologie et de Paleontologie conchylio- logique ou Histoire naturelle des Mollusques vivants et fossiles. xxiv -f 1569 pp., 23 PI. Paris, 1880-87. Fol, H. '75. Etudes sur le developpement des Mollusques. ler Memoire : Sur le developpement des Pteropodes. Arch. Zool. exp. et gen., Tom. IV. pp. 1-214, PL I.-X. 1875. '76. Etudes sur le developpement des Mollusques. 2me Memoire: Sur le developpement des Heteropodes. Arch. Zool. exp. et gen., Tom. V. pp. 105-158, PI. I.-IV. 1876. Haddon, A. C. '82. Notes on the Development of Mollusca. Quart. Jour. Micr. Sci., Vol. XXII. pp. 367-370, PL XXXI. 1882. Heymons. R. '93. Zur Entwicklungsgeschichte von Umbrella mediterranea, Lam. Zeitschr. f. wiss. ZooL, Bd. LVI. pp. 245-298, Taf. XIV-XVI. 1893. Kowalevsky, A. '83. Embryogenie du Chiton Polii (Phillippii), avec quelques Re- marques sur le Developpement des autres Chitons. Ann. Musee 200 PROCEEDINGS OP THE AMERICAN ACADEMY. Hist. Nat. Marseille, Zool., Tom. I., Mem. No. 7, 55 pp., 8 PI. 1883. Lang, A. '84. Die Polycladen (Seeplanarien) des Golfes von Neapel und der angrenzenden Meeresabschnitte. Eine Monographie. Fauna u. Flora d. Golfes v. Neapel, XL, ix -|- 688 pp., 54 Holzschn., Atlas 39 Taf. Leipzig, 1884. Lillie, F. R. '93. Preliminary Account of the Embryology of Unio complanata. Jour. Morph., Vol. VIIL pp. 569-578, PI. XXVIII. Metcalf, M. M. '93. Contributions to the Embryology of Chiton. Studies Biol. Lab. Johns Hopkins Univ., Vol. V. No. 4, pp. 249-267, PI. XV., XVI. Rabl, C. '79. Ueber der Entwicklung der Tellerschnecke. Morph. Jahrbuch, Bd. V. Heft 4, pp. 562-655, Taf. XXXII.-XXXVIIL, 7 Holzsch. 1879. The'el, H. '92. On the Development of Echinocyamus pusillus (O. F. Miil- ler). Nova Acta Reg. Soc. Sc. Upsala, Ser. III. Vol. VI., 57 pp., PI. 9. 1892. Wilson, E. B. '92. The Cell Lineage of Nereis. A Contribution to the Cytogeny of the Annelid Body. Jour. Morph., Vol VI. No. 3, pp. 361-480, PI. XIII.-XX. '93. Amphioxus and the Mosaic Theory of Development. Jour. Morph., Vol. VIIL No. 3, pp. 579-638, PI. XXIX.-XXXVIII. EXPLANATION OF PLATE I. Figures 1-4, Limax, drawn with camera lucida. Figure 1. Eight-cell stage from anterior end, X 375; p.g., polar globules; s.c, segmentation cavity. " 2. Twenty-four-cell stage from right side, X 375. " 3. Twenty-five-cell stage from the animal pole, X 375. " 4. Same egg from vegetative pole, X 375. Figures 5, 6, Nereis after Wilson ('92). " 5 After Plate XIV. Fig. 21, " Rear view of twenty-two-cell stage. Division of A', spindles of cs, d3." " 6. After Plate XV. Fig. 25, " Thirty-two- (four-) cell stage, right side- view. Third spiral cleavage of a1, bl, c1, dl, in progress." KOFOID. — LAWS OF CLEAVAGE. 201 PLATE I. 3. 4. 202 PROCEEDINGS OP THE AMERICAN ACADEMT. EXPLANATION OF PLATE II. Neritina after Blochmann ('81). The arrows, indicating the derivation of cells, and the nomenclature placed within the limits of the figures are mine ; the rest of the nomenclature is repro- duced from Blochmann's figures. It is to be observed that his lettering, a, b, c, d, is from right to left, whereas my own is that generally accepted, — from left to right. The explanation of these figures is taken from Blochmann, and the term '' generation " is used in the sense in which he has employed it. Figure 7. Taf. VII. Fig. 45. " Formation of the cells of the sixth generation, "l1 V ci! ^l1- Rhl, polar globules." " 8. Taf. VII. Fig. 46. "Formation of the cells of the seventh and eighth generations, a3 b3 c3 d3 and a2' b2l c2l d.2l." " 9. Taf. VII. Fig. 47. "The same stage in profile." " 10. Taf. VII. Fig. 48. " Completed twenty-four-cell stage." The long arrows, indicating a change of interpretation, are mine. " 11. Taf. VII. Fig. 50. " Twenty-eight-cell stage ; a2u b2u c„» J2", cells of the ninth generation." " 12. Taf. VII. Fig. 51. " Thirty-six-cell stage ; vz derived from a.2u, vzx from c2i [my 67-3], d2lu from d2l, azl from a3, etc. KOFOID. — LAWS OF CLEAVAGE. 203 PLATE II. 204 PROCEEDINGS OF THE AMERICAN ACADEMY. VIII. FURTHER OBSERVATIONS UPON THE OCCURRENCE OF DIAMONDS IN METEORITES. By Oliver Whipple Huntington, Ph. D. Presented October 11, 1893. Ever since Haidinger, in 1846, described a cubic form of graphite in the Arva iron, and Rose suggested that the crystals were pseudo- morph after diamond,* the possibility of finding the gem itself in meteorites has been conceived. In 1886, H. Carvill Lewis, after his study of material from the greatest depths of the South African diamond mines, predicted the discovery of diamonds in meteorites.! Hence it was no great surprise when graphite having the hardness and form of the diamond was found by Jerofeieff and Latschinoff in the Novo Urei meteorite. J No colorless material, however, seems to have been found till E. Wein- schenk described, in 1889, a minute quantity of transparent grains from the Arva iron, hard enough to scratch the ruby, and burning in oxygen to form carbonic acid gas.§ A fresh interest was added to the study of the subject two years ago when Dr. A. E. Foote brought numerous pieces of the Canon Diablo iron from Arizona, and Professor Koenig of Philadelphia announced that a piece of it contained diamonds, his statement being based upon the appearance of a black vitreous substance lining a cavity, and having a hardness beyond the sapphire. Dr. Foote says in his paper on the subject, " The most interesting feature is the discovery for the first time of diamonds in meteoric iron." || He then refers to the cubic * Pogg. Ann., Bd. LXVII. pp. 437-439, 1846. t British Association, 1886, p. 667. J Verhandl. der kais. russ. Mineral. Gesellschaft, 2d Series, Vol. XXIV. p. 263, 1888. § Annalen des k.-k. naturhistorischen Hofmuseums, Vol IV. p. 93, 1889. || Proceedings of the American Association for the Advancement of Science, Vol. XL. p. 4. ? o 1 o o o 3 O 3 ~ "J re I I-1 IO lO r+ 5' — = =a ° re 00 o c 53 > o 3 "0 I- H m HUNTINGTON. — DIAMONDS IN METEORITES. 205 form of graphite in the Arva iron, described so long ago by Haidinger, and the cliftonite found by Fletcher in the Youndegin iron, but makes no mention of Weinschenk's description of colorless transparent diamonds in the Arva iron. The Mineral Cabinet of Harvard University was presented, through the liberality of Francis Bartlett^ Esq., with one of the original masses of the Canon Diablo meteorite, and it was at once examined for diamonds. A piece weighing one hundred grams was dissolved in acid, and, after the usual process of getting rid of the carbon, silica, etc., a small quantity of fine powder was separated by its higher specific gravity and examined under the microscope. It appeared to be made up of dark and light grains, the latter showing unmistakably the lustre of the diamond. The colorless particles were separated mechanically upon the stage of the microscope, and looked to the eye like white beach sand. One of these grains was mounted in the point of a metallic lead pencil, and it wa9 found that, though so small, it would readily cut glass with the characteristic singing noise of a glass-cutter's diamond, and it would also readily mark upon a polished sapphire. There was not enough of the clear material obtained at the time for a chemical test, and on accouut of the association of the diamond grains with amorphous carbon, such a test would not have been conclusive without a perfect mechanical separation. This seemed at the time sufficient proof that the material was diamond, and a portion of it was exhibited before the meeting of the American Academy of Arts and Sciences on May 11, 1892, and at that time the author showed a large nodule of graphite from the Sevier County iron and said : — " As is well known, graphite separates from melted cast iron when it slowly cools, and the connection of these masses with the iron meteorites indicates that they also were formed by the slow cooling of masses of melted iron, very possibly thrown up by volcanic action from the interior of some planet. The high specific gravity of the earth suggests the conjecture that its interior is a mass of iron, and the metallic iron which has been found in the deep-seated eruptive rocks, as in Greenland and the South of France, gives support to the hypothesis that these rocks may be the matrix of the diamond, and that the diamond crystals may have separated during slow cooling from the melted metal forced up from below. « Thus : — " Graphite separates out from meteoric iron. It also separates during the cooling of cast iron. 206 PROCEEDINGS OF THE AMERICAN ACADEMY. " Graphite is further found associated with the iron of the deep- seated rocks. " Graphite, crystallizing iu the form of diamond, is found in meteoric iron. Diamonds are also found iu meteoric iron. " The greatest deposits of native diamonds appear to have come from great depths. " Does it not appear plain that the meteoric diamonds, which are obviously in their original matrix, may explain the origin of the terrestrial diamonds ? " My object has been merely to point out a possible relation, and not to advocate a theory, but in this relation we have a clue which may possibly lead to a solution of one of the most obscure problems of mineralogy." It was further suggested that pressure might be a determining factor in the crystallization of carbon. At the time this seemed a somewhat large assumption to make from so small data, but since then M. Moissan has followed up the same idea by most interesting experiments, iu which he finds that carbon can be made to crystallize out of melted wrought iron in the form of diamond if the iron is allowed to cool under pressure ; as is the case when the melted metal expands in setting against the crust formed by quenching the exterior of the mass with water.* May it not be that, if we could reach thus a pressure commensurate with that produced by the crust of the earth, the resulting diamonds would attain the size and perfection of those now found in the famous diamond fields of South Africa? In December, 1892, M. C. Friedel,f apparently not having seen the papers published by the author seven months before, made a thorough examination of some of the Canon Diablo iron. The diamonds that he isolated were of the carbonado variety, and in order finally to prove that the black grains were diamond he burnt a portion in oxygen. Thus he proved chemically that the substance was carbon, and since the hardness was beyond the ruby he considered it suffi- ciently determined that it belonged to the variety diamond. The proof of the hardness of the diamond, however, does not rest with the fact that diamonds will scratch the sapphire or the ruby, since this may be done by other substances, notably " carborundum," but diamond alone will cut diamond, and this test had not been applied to the material * Comptes Rendus, Tome CXVI. No. 6, Feb. 6, 1893, p. 218. t Bulletin de la Socie'te Francaise de Mineralogie, Tome XV. No. 9, Decem- ber, 1892, p. 258. HUNTINGTON. — DIAMONDS IN METEORITES. 207 from the Arva, Novo Urei, or Canon Diablo meteorites. Especially since the work of Daubree and others has shown how numerous the allotropic forms of carbon may be, varying in hardness, specific gravity, aud other physicial characters, it seemed of value to get a more accurate determination of the hardness of this new material, since no distinct crystals had been observed. At the suggestion of Mr. George F. Kunz, the author undertook to dissolve many pounds of the Canon Diablo iron, in order to obtain enough diamond dust to use at the Columbian Exhibition for cutting and polishing rough diamonds, thus demonstrating in a practical way the fact that we have true diamonds in meteorites. A method of dissolving iron slightly different from that described by the author in " Science"* was adopted, in order to find if possible the paragenesis of the diamond material. About two huudred pounds of the iron was examined, and the most promising pieces were dissolved. The object of working on so large a scale was twofold. In the first place, to obtain sufficient material for the practical test of cutting and polishing rough diamonds, and in the second place to discover how the substance occurred, and whether it crystallized in the usual form. Of course there was also a possibility, though not a probability, of happen- ing upon a crystal of considerable size. The specimens of iron selected were successively suspended by a platinum wire in a platinum bowl, filled with hydrochloric acid nearly saturated with chloride of iron and then slightly diluted with water. The iron was made the positive pole of a battery consisting of twelve gravity cells, the bowl forming the other electrode. When the conductivity of the solution was properly adjusted by means of the chloride of iron it was found that only the pure iron of the meteorite would dissolve, thus setting free the plates of schreiberseit and taenite, as well as the graphite and other impuri- ties, bringing out the crystalline structure of the meteorite in wonder- ful perfection. A partially dissolved piece is shown a little less than the natural size in Figure 1. Unfortunately, the sketch does not bring out the plates as prominently as they appear in the specimen, and the differences of tint, which are very striking in the original, cannot be reproduced in a black and white print. If the acid was too strong, or the temper- ature of the solution slightly increased, all the distinction of parts at once disappeared, and it dissolved en masse, leaving a great quantity of slime in the bottom of the dish. * Vol. XX. No. 492, July 8, 1892. 208 PROCEEDINGS OF THE AMERICAN ACADEMY. The iron was dissolved in two- or three-ounce pieces, and it at once appeared that there was a wide variation in its composition. Most of it contained no diamonds whatever. One piece, however, in the pro- cess of dissolving showed an irregular vein running through it consist- ing of a white vitreous substance varying in width from a fine line to nearly four millimeters. On trying the hardness of the vein-stuff, it was found to be beyond the ruby. On further examination, it ap- peared to be a mixture of iron, a sulphide of iron, silica, amorphous carbon, and diamond, so that it finally had to be crushed in order to get rid of all the iron. The hardest grains were isolated by the method described in the previous papers, and when examined under the microscope one minute but perfect octahedron of diamond was found, transparent and col- orless. It was separately mounted on a microscope slide, but soon disappeared, and in its place were found only some very minute angular fragments. A second crystal was afterwards isolated, but dis- appeared in like manner, suggesting that they had been formed under pressure, and when exposed in a warm room had exploded. This called to mind the fact that broken diamonds are so often found in the South African mines, though the cases may not be analogous, since in the latter instance the fragments have never been found associated. About half a carat of diamond powder was finally obtained, being separated by its specific gravity from a very large quantity of amor- phous carbon. The particles varied from colorless through yellow and blue to black. Many of them appeared to be angular fragments, though some of them looked not unlike hyalite except for their more brilliant lustre. Several perfect little octahedrons were found that did HUNTINGTON. — DIAMONDS IN METEORITES. 209 not break up, and Figure 2 is an exact reproduction of a drawing of one of these crystals, traced through a camera lucida attached to a microscope. The original specimen measured only a little over a hundredth of an inch in diameter, but viewed through a two-thirds objective it showed distinctly the hexakis octahedral planes, the curved edges, striatums, etc. exactly as seen in the drawing. Unfortunately, the print cannot show the adamantine lustre and clear water of the crystal. A portion of the powder was sent to Mr. Kunz for the diamond- cutting experiment, but it was tried at first hastily and without success. The author, however, was so sure of the nature of the material that he repeated the experiment in company with Mr. Kunz, on Monday, Fig. 2. September 11, 1893, in the Tiffany Pavilion of the Mining Building of the World's Columbian Exhibition. A newly planed wheel had been provided, through the courtesy of Messrs. Tiffany & Co., but in order to make the experiment doubly sure a diamond set for cutting was placed in position on the wheel, and left there for a period of five minutes, while the wheel was making twenty-five hundred revolutions per minute. On removing the diamond it was found that there had been no appreciable shattering of the edges by friction, and that it would take a long time for its own powder to be produced in sufficient quantity to have any effect. The wheel was then charged with the residue from the Canon Diablo iron, mixed with oil in the usual way. The diamond immedi- ately gave out a sharp hissing sound, making it apparent at once to an expert that the material was cutting, and in a few minutes it was found that a face had been ground down and polished. Two other vol. xxix. (n. s. xxi.) 14 210 PROCEEDINGS OF THE AMERICAN ACADEMY. diamonds were cut and polished in the same way. The last specimen was a complex twin, and when a projecting angle was applied to the wheel the cutting was naturally slow, but none the less effective, and in all respects the residue from the Canon Diablo iron was found to act exactly like any other diamond powder. Thus the combined evidence of the author's work with that of Friedel * and Moissan f establishes the fact that the Canon Diablo iron contains true diamonds, and not any new allotropic form of carbon. It is surprising that crystals are so rare, but, as this paper shows, they do occur. It has been suggested, since the Canon Diablo iron has been found in such quantities, that it may have a terrestrial origin, but such an idea is at once dispelled by an examination of the exterior surface of the largest mass thus far obtained, apart from other considerations. The ma^s referred to was purchased by Professor J. P. Cooke from Mr. W. W. Howell of Washington, D. C. It weighs 1087 pounds, and is at present in the Harvard Museum. It is seldom that irons which have not been seen to fall retain, at the time they are found, any of the original features of their crust, but this specimen is a marked exception. It is a roughly spherical mass somewhat flattened in one plane. One of these flattened surfaces shows signs of fusion, but is further characterized by deep pittings looking almost like bullet holes, except that it frequently happens that they are larger on the interior than at the orifice. Occasionally they are winding, but all preserve the same general direction. They vary in size from one sixteenth of an inch in diameter to two inches, and reach in some cases a depth of from three to four inches. These cannot be explained by unequal heating or by the erosive action of the air, owing to the fact just mentioned, that they are larger in the interior than at the surface. This character is shown in a reproduction of a small photograph, (Plate I. Fig. 1). The other side of the mass (Plate I. Fig. 2) shows large concave surfaces, as if pieces of six or seven inches in diameter had flaked off, or the hollows had been scooped out by the action of pneumatic drills. For some time these two utterly different surfaces were a puzzle, but a closer examination disclosed troilite in the depths of some of the small cavities (Plate I. Fig. 1) while on the side of the larger pittings (Plate I. Fig. 2) the same mineral could be seen exposed and * Bulletin 39.07 38.76 14.52 14.40 2.96 43.78 3891 14.46 2.96 * Probably some undeoomposed oxichloride. t This sample was undoubtedly heated too high when drying. BENEDICT. — DOUBLE HALOIDS. 219 An examination of the above two sets of averages establishes beyond a doubt the value of the ignition method of analysis in these com- pounds. With these data the formula nearest approaching their values is SbCl3 . SbOCl . 2 KC1. Calculated * for SbCl3, SbOCl, 2 KC1. Found. Sb 43.88 43.78 CI 38.89 38.91 K 14.31 14.46 O 2.92 2.96 100.00 100.11 It is evident, therefore, that this salt has the formula SbCl3 . SbOCl . 2 KC1. SbCl3.2KCl [Hexagonal]. These crystals are tabular, consisting of a regular hexagonal prism terminated by basal planes. With the polarizing microscope it gives a very perfect uniaxial figure of a positive sign. In this figure the yellow ring is the inner- most, therefore p < v. It being quite permanent in the air, no difficulty was experienced in making the above examinations. Hot or cold water immediately de- composes it. It is soluble in a solution of tartaric acid, and also readily soluble in dilute hydrochloric acid, from which it can be crystallized. The antimony and chlorine were determined by precipitation, and the potassium was weighed as potassic chloride after ignition in a current of dry carbon dioxide gas. In this residue there was not a trace of any insoluble matter (absence of Sb20.,) after ignition. Here, all the SbCl3 was volatilized, giving again the data for the calculation of the percentage composition. Calculated for SbCl3, 2 KC1. Found. Sb 31.95 30.44 CI 47.20 47.21 K 20.85 22.34 100.00 99.99 The loss of 1.51 % in antimony and the gain of 1.49% in potassium were at first difficult to explain, but on taking into consideration the * Atomic weights used: — Sb 120; CI 35.456 ; K 39.14; O 16. 220 PROCEEDINGS OF THE AMERICAN ACADEMY. fact that the per cent of chlorine in SbCl3 is 47.04, in KC1 is 47.53, and in this salt is 47.20, it is evident that an excess of potassic chloride would explain the discrepancy from the calculated results, and not ma- terially interfere with the per cent of chlorine. The crystals of this salt are, as has been stated, hexagonal plates. These plates pile up in layers, leaving the prism faces very much striated. This process of formation would tend to the occlusion of mother liquor rich in potassic chloride. As a matter of fact, a micro- scopical examination of the smallest, clearest, and most perfect crystals revealed numberless cavities with air bubbles and bubbles of liquid. This was the only salt in which such cavities and bubbles were found. SbCl3 . 2 KC1 [Monoclinic]. a : b : c = 1 : 1.381 : .9974. 0 = - 68° 57'. Angles between Normals. A B A n OT0 A TTO = 56° 6' OTO A 021 = 36° 45' A C A m 0T0 A 1T0 = 56° 0' OTO A 0T1 = 56° 0' C D m m' 1T0 A 100 = 34° 0' 0T1 A Oil = 68° 0' c c C m no a no = 68° 3' 1T0 A 0T1 = 56° 0' O' O" • B 0' TTl A Til = 61° 20' TTO A TTl = 46° 15' BENEDICT. — DOUBLE HALOIDS. 221 This salt is quite stable in the air. It is the dimorphous form of the previous salt. Found. 32.13 47.20 20.70 Calculated for SbCl3 . 2 KC1. Sb 31.95 CI 47.20 K 20.85 100.U0 100.03 Here there is no great difference between the antimony and potas- sium percentages as before. Microscopic examination showed no cavities or bubbles of enclosed liquid. SbCl3 . 2 KC1 . 2 H20 [Monoclinic]. OO I a:b:c = 1 : 1.3798 : .9358. 0 = 78° 30'. Angles between Normals. A B 110 A T10 = 111° 40' B C T10 A 001 = 99° 30' A A' 110 A 1T0 = 68° 48' B' C TT0 A 001 = 99° 28' A C 110 A 001 = 80° 35' B B' T10 A TT0 = 68° 42' A' C 1T0 A 001 = 80° 28' B 0 T10 A Til 0 C Til A 001 = 34° 31' = 64° 57' 222 PROCEEDINGS OP THE AMERICAN ACADEMY. The fact that this salt effloresced so readily at first prevented the making of any accurate crystallographic measurements. However, it was found that at a temperature of about — 5° the salt retained its lustre for a sufficient length of time to afford very accurate measure- ments and crystallographic study. These crystals were of a tabular form, and upon looking through the plate with a polarizing micro- scope no interference figure could be discerned. As the formula indicates, this is a hydrate of the two previous salts. The water of crystallization is easily lost, even at a temperature of 15°. The weather at this period of the investigation was extremely cold, in consequence of which it was found that when the salt is kept in a room with the temperature at about : — 5°, it retains its water of crystallization for a period long enough to admit of weighing. The salt in the form of large isolated crystals was taken from the mother liquor, hastily dried between sheets of filter paper, crushed in a mortar, and finally rubbed between filter papers. Then it was transferred to a glass-stoppered weighing bottle, the whole operation occupying not more than 80 seconds at a temperature of — 5°. From the weighing bottle it was quickly removed as desired ; — the loss in weight of the bottle showing the amount of salt used. The determination of the water of crystallization presented several difficulties. Heating the salt in an air bath at 90°-95°, it will not come to constant weight in ten hours. If heated to 110°, it continues to lose in weight for a long time ; in fact, there is a decomposition accompanied by the liberation of chlorine. Analysis of the salt heated to 95° showed that no CI had gone off, but when heated to 110° there was a marked diminution of the percentage of chlorine, indicating a decomposition. At a tem- perature of 104° the salt will lose weight rapidly for a while, and then lose it very slowly for hours. Evidently the point where the last portion of water goes off and the point where the chlorine begins to be given off are nearly coincident. A great many determinations were made, and by estimating the chlorine every time the point could be very closely determined. Almost all the water will be given off in a desiccator over sulphuric acid, but it is a long time before the last trace is gone (thirty hours). 1. In a desiccator after 30 hours, the per cent of H20 = 8.743 2. Heated in air bath 90-95°, at end of 4 hours = 8.476 3. " " 140° for 1 hour = 9.874 4. " " 105-107° for 5 hours = 10.24 5. " " 100-105° " " = 8.765 BENEDICT. — DOUBLE HALOIDS. 223 Nos. 1 and 5. The water has all been given off, and these are the two best values for the per cent of water. No. 2. The water is not all off yet, as shown by heating another hour, when per cent = 8.62. Nos. 3 and 4 both showed by analysis that chlorine had been evolved. The dried salts from Nos. 1 and 5 were used in the ignitions, and these gave results corresponding to the formula SbCl3 . 2 KC1. Calculating the percentages for the additional water of crystallization, the following result was obtained : — Calculated for SbCl3 . 2 KC1 . 2 H20. Found. Sb 29.16 29.07 CI 43.07 43.07 K 19.02 18.53 H20 8.75 8.76 100.00 99.43 In the ignited residue there was no insoluble portion of Sb203. The evidence warrants the selection of SbCl3 . 2 KC1 . 2 H20 as the formula for this salt. The conditions under which these salts form are closely related. So much so that it has been difficult to establish the exact conditions under which each salt may be isolated. At first the work was wholly tentative, using a saturated solution of potassic chloride, and adding crystallized antimonious chloride, filtering off any insoluble portion. In this way were obtained the salts SbCl3 . SbOCl . 2 KC1, and SbCl3 . 2 KC1 (hexagonal). But in order more closely to study the relations of each to the others, a series of experiments was carried out in watch glasses. To about 5 c. c. of a saturated solutiou of potassic chloride were added 5 grams of crystallized antimonious chloride ; when this had all dissolved, as it will in the cold, more SbCl3 was added, till the solu- tion was so to speak saturated with SbCl3. In all, about 30 grams of SbCl3 were added. In each of a series of watch glasses was placed 1 c. c. of the above solution, and to the several glasses increasing amounts of a saturated solution of potassic chloride were added, — from 1 c. c. in the first to 5 c. c. in the last. In each a white pre- cipitate was formed which redissolved on heating. Then the reverse was tried, i. e. 2 c. c. SbCl3, 1 c. c. KC1, down to 5 c. c. SbCl3. 1 c. c. KC1. There was a precipitate in each watch glass, which however dissolved on heating. The watch glasses were placed on the window sill, with moderately cold weather, and ex- amined at intervals. 224 PROCEEDINGS OF THE AMERICAN ACADEMY. Designating the salts by Roman numerals : — SbCl3 . SbOCl . 2 KCl I. SbCl3. 2 KCl [hexagonal] II. SbCl3 . 2 KCl . 2 H20 III. Contents of Watch Glass. Nov. 11, A. M. Nov. 11, P. M. Nov. 13, 8.30 a. m. Nov. 13, 4 p. m. (Colder.) Nov. 14, 6 p. m. lc.c.SbCl3sol. 1 c.c. KCl " Monoclinic prisms. I. Same. I. Same. I. Same. I. lc.c.SbCl3sol. 2 c.c. KCl " I. I. General mass I., but the edges rounded off. A few II. Indications of III. A few II. Two crystals of III. 1 c.c. SbCl3 sol. 3 c.c. KCl " I. I. I. a few remain. II. predominates. III. a good crystal. Chiefly II. Several III. Remains of a few I. lcc. SbCLjSol. 4 c.c. KCl " I. I. Chiefly cubo-oeta- hedrons of po- tassic chloride. II. present. I. but few remain. Chiefly II. with amorphous KCl. lcc. SbCI3sol. 5 c.c. KCl " Clouds up. Possible in- dications of I. Precipitate at bottom a few I. floating. Chiefly cubo-octa- hedrons of KCl. No signs of I. A very few II. Chiefly cubo- octahedrons with a few 11. 2cc.SbCl3sol. 1 c.c. KCl " down to 5c.c.SbCl3 " 1 c.c. KCl " — — Scum of white oxichloride. Same. During these experiments a number of mixtures of the two salts were crystallized in dishes under varying conditions. Regarding salt I. (SbCl3 . SbOCl . 2 KCl), the evidence indicated that it crystallizes best from solutions by spontaneous evaporation in a cool place. Solutions should be preferably not too concentrated, and contain no large amount of free acid, i. e. no acid used in assisting the solution. The hexagonal salt II. (SbCl3 . 2 KCl) results by evaporating either by heat or a vacuum desiccator a solution of almost any proportion BENEDICT. — DOUBLE HALOIDS. 225 of SbCl3 and KC1 with or without free acid ; preferably concentrated solutions. This salt is the one most readily obtained of all the salts. The hjdrated salt III. (SbCl3 . 2 KC1 . 2 H20) crystallizes from moderately dilute solutions by exposure to severe cold, — 10°. The presence of free hydrochloric acid does not interfere in auy way, and possibly assists in the crystallization, provided it is not in too great an excess. Lastly, the monoclinic form of SbCl3 . 2 KC1. The hexagonal salt is readily obtained, and, if it is left in the mother liquor, after a few days undergoes a transformation from the hexagonal to the mono- clinic. This is apparently accomplished by the building up of the hexagonal plates into a dumb-bell form, after which the planes of the monoclinic form are assumed, accompanied by an entire reconstruction of the crystal. The mutual relations of these salts require further considera tion. If a solution is made up as directed for preparing salt I. (SbCl3. SbOCl. 2 KC1), the addition of one drop of water beyond a certain point will cause a white precipitate, which is soluble upon heat- ing, but which reapjy-'ars on cooling. If, however, a crop of I. crystals is taken from the solution, and the mother liquor be divided in two portions, water can be added to one of them in considerable quantity, till finally one drop produces a cloudiness, which disappears on heating, but reappears on cooling. If now the second half of the mother liquor be added to the dish, the precipitate will not redissolve in the cold, but on heating it goes completely into solution, and is not thrown out on cooling. If this solution was left to evaporate after removing the cryst als of I., there would soon be a deposit of II. (SbCl3.2KCl, hexagonal). But it is found that the addition of water as above described will promote the crystallization of more I. (SbCl3. SbOC1.2KCl). When a solution depositing I. is placed at a temperature of — 5° then III. will immediately be deposited. When the salt I. is dissolved in hydrochloric acid just strong enough to effect the solution, there is deposited chiefly II., together with a few I. When the salt II. is dissolved under the same conditions, there is dissociation, and the potassic chloride crystallizes out as cubes or cubo- octahedrons at first. Later, the salt will crystallize out itself as II. As has been before stated, II. hexagonal goes into II. monoclinic on long standing at a moderate temperature. Considerable difficulty was experienced in attempting these crystallizations during rainy VOL. XXIX. (n. S. XXI.) 15 226 PROCEEDINGS OF THE AMERICAN ACADEMY. weather. The excessive moisture caused a white scum of oxichloride on every receptacle of the various solutions. According to all analogy regarding the constitution of double salts, there would be three possible forms for the double chlorides of anti- mony and potassium ; i. e. SbCl3 . KC1 ; SbCl3 . 2 KC1 ; SbCl3 . 3 KC1. Singularly enough, as yet only the second one is obtained, and that is dimorphous. Solutions were made containing molecular proportions for SbCl3 . KG and SbCl3 . 3 KG. Each solution was divided into two parts, one of which was acidulated with hydrochloric acid, and the other not. The only resulting salts were: 1st solution, i. e. SbG8 . KG = I. and II. 2d " " SbG3 . 3 KG = at first a lot of cubo-octahedrons of potassic chloride ; later, II. Although so many different forms were mentioned by previous writers, none such have been definitely proven in this research. Re- garding the statement of Poggiale,* that the SbG3 . 2 KG crystallizes in cubes, it may be mentioned that in the preparation of that salt a number of crystals appeared under the microscope as perfect cubes. But on isolating them and nursing them till they were of sufficient size to admit measurement, they were proven to be the hydrated salt III. (SbG3 . 2 KG . 2 H20), with the angle not 90° but 80°. It is simply a different habit of the tabular crystal with no octahedral face. It may be possible that he observed the cubes of potassic chloride, which will crystallize from many of these solutions. The mention of a triclinic crystal by Fehling * can be accounted for by the peculiar habit of this same salt III. of developing more rapidly on one side than another, giving to the eye a decidedly triclinic appearance. The trimetric form of Rammelsberg f may be explained in two ways. First, the salt III. crystallizes often as per diagram, in which the plane of symmetry is through the dotted line, and O represents the octahedral face.t but the lower face is wanting where the crystal rests on the dish. At first sight this was taken for a trimetric crystal, but the angle of AAC is different from that of O A C, and no interference figure is seen with the polarizing microscope. Secondly, the salt (SbG3 . 2 KG, monoclinic) crystallizes in dia- * This paper, page 213. f Ibid., page 214. \ Ibid , page 221. BENEDICT. — DOUBLE HALOIDS. 227 mond-shaped tablets, but the plane of symmetry is parallel to A, and octahedral faces are all wanting.* This occurs when there is a thin layer of mother liquor in the crystallizing dish. Obviously, this would be taken at first sight for trimetric, but measure- ment of angles and stauroscopic observations in- disputably prove otherwise. No crystals were obtained crystallizing in any manner like leaves. In some cases there was obtained, when an exceedingly thin layer of liquid was allowed to evaporate, an arborescent formation as a thin coating on the bottom of the dish. This was SbCl3 . 2 KG. The salt Jacquelaine f described as in oblique rhombic prisms, with a formula SbCl3 . 2 KG, may have been the salt described here as the mouocliuic form of II., but it is rather singular that he made no men- tion of the hexagonal form which is the first to be obtained. On the other hand, it is quite possible that he meant the monoclinic salt, SbG3 . SbOG. 2 KG, which is the first to crystallize from molecular solutions of SbG3 . 2 KG. The writer is engaged in an investigation of the double salts of antimonious chloride with the bivalent metals. * This paper, page 220. t Ibid., page 212. 228 PROCEEDINGS OP THE AMERICAN ACADEMY. CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF HARVARD COLLEGE. CERTAIN BROMINE DERIVATIVES OF RESORCINE. By C. Lorixg Jacksox axd F. L. Duxlap. Presented February 14, 1894. The research described in the following paper was suggested by the observation* tbat tribromnitroresorcine diethylether, when heated with a solution of sodic ethylate, lost two atoms of bromine, which were replaced by two of hydrogen, so tbat C6Br3N02(OC2H5)2 was converted into C6BrH2N02(OC2II5)2. Considering the easy removal of these two atoms of bromine, it seemed possible that the tribrom- resorcine diethylether, which differs from it only in the absence of the nitro group, might behave in a similar way. This expectation has not been fulfilled, as our experiments have shown that tribromresorcine diethylether does not give up bromine to sodic ethylate at the boiling point of alcohol, and that the dibrom ether is not attacked below 200°, and therefore is, if anything, more stable than the symmetrical tri- brombenzol, which according to Blau t is slowly converted into di- bromanisol at temperatures of 120° to 130°. The behavior of free tribromresorcine was next studied, and it was found that bromine could be removed with great ease, even water at its boiling point decomposing this substance, while solutions of sodic carbonate, or sodic hydrate, or sodic ethylate dissolved in alcohol acted even more readily. These results show that the ethoxyl groups do not loosen the attachment of the bromine atoms to the benzol ring, whereas the hydroxyl groups have as much of this loosening effect as do the nitro groups in tribromdinitrobenzol. We have not succeeded in determining the nature of the organic compounds produced by removing bromine from tribromresorcine, as they were brown amorphous bodies, which could not be brought into * These Proceedings, XXVII. 315. t Monatsh. f. Chem., VII. 630. JACKSON, DUNLAP. — BROMINE DERIVATIVES OF RESORCINE. 229 a state of undoubted purity for analysis. From their properties, however, it can be inferred that they are substituted resorcine ethers formed by a reaction involving the hydroxyl groups of the tribroni- resorcine. Sodium malouic ester gave a similar unmanageable product with tribromresorcine, but in addition to this acetylentetracarboxylic ester was isolated. This substance could hardly be formed here other- wise than by the action of brommalonic ester on sodium malonic ester, and the brommalonic ester must have been formed by the replacement by hydrogen of one or more atoms of bromine in the tribromresorcine. It follows, therefore, that in this respect also the two hydroxyl radicals behave like the two nitro groups in tribromdinitrobenzol, in which a similar .replacement of bromine by hydrogen has been observed when this substance is treated with sodium malonic ester. The tribromresorcine diethyl ether melts at 68° -69°, and is formed by the action of an excess of bromine on the dibromresorcine diethyl- ether, melting point 99°-100°. This latter substance is prepared by the addition of bromine to a solution of resorcine diethylether in glacial acetic acid. The tribrom product was shown to have the bromine atoms in the symmetrical position to each other by treatment with nitric acid, which converted it into the tribromnitroresorcine di- ethylether melting at 101°, which is made from symmetrical tribrom- trinitrobenzol. * The action of nitric acid on dibromresorcine diethyl- ether was entirely different, as the product was a new dinitroresorcine diethylether melting at 126°, formed by the replacement of the two atoms of bromine by two nitro groups. A similar replacement of bromine by nitro groups was observed when the tribromresorcine was treated with fuming nitric acid, the product being the bromdinitro- resorcine described by Typke f and Fevre. t As this substance was obtained by the latter from dibromnitroresorcine, in this case also bromine was replaced by a nitro group, and several other cases are to be found in the chemical literature although they are far from common. The most important point established by the observations given above is that in these cases hydroxyl or ethoxyl radicals have the same effect upon bromine which they have upon hydrogen attached to the benzol ring; for just as phenols are much more susceptible to the replacement of their hydrogen by halogens or nitro groups than * These Proceedings, XXV. 183. t Ber. d. ch. G., XVI. 555. t Bull. Cliem. Soc, XXXIX. 591. 230 PROCEEDINGS OF THE AMERICAN ACADEMY. the corresponding hydrocarbons, so the bromine can be replaced by other radicals, including the nitro group, much more easily in tribrom- resorcine than in tribrombenzol ; while on the other hand the phenol ethers hold their hydrogen or bromine as firmly as the corresponding hydrocarbons. The one exception we have met with to this statement is the action of nitric acid upon dibromresorcine diethylether, in which the bromine is replaced by nitro groups, so that in this case the ether behaves like the free tribromresorcine rather than like the correspond- ing tribrom ether. This exception is the more remarkable, because we have found that toward other reagents, such as sodic ethylate, this dibrom ether shows the same great stability which characterizes the tribromresorcine diethylether. As we could not remove the bromine from the tribromresorcine diethylether with sufficient ease, and the free tribromresorcine, although it gave up its bromine easily, gave unmanageable products since the hydroxyl groups took part in the reaction, we undertook the study of the tribromresorcine diacetate, which we hoped might give products that could be purified and analyzed ; but these experiments did not lead to the desired result, as the acetate was converted into tribrom- resorcine by the reagents used, even boiling with water producing this effect to a slight extent, and consequently the products were those already mentioned as obtained from this substance. When treated with fuming nitric acid, the diacetate gives tribromnitroresorcine di- acetate which melts at 161°. The sodium salt of tribromresorcine was also studied. This was obtained in white crystals by precipitating an ethereal solution with alcoholic sodic ethylate, and seemed to contain two molecules of alcohol of crystallization. It was remarkable on account of its slight stability, as even in vacuo or in an atmosphere of hydrogen after standing for a short time it suddenly turned black, when there was a considerable evolution of heat and alcohol vapor was given off. The black residue gave up sodic bromide to water, and also a black soluble salt, while a red product was left behind. The black salt on the addition of acid gave a red precipitate similar to the product insoluble in water, but we were unable to bring either of these red amorphous substances into a condition undoubtedly fit for analysis ; and although some analyses led to a very probable formula and reaction for the decomposition we can give no weight to these results, as we have no proof that the samples analyzed were homogeneous. There is no question that these substances are very similar to the products obtained from tribrom- resorcine and sodic ethylate or carbonate. JACKSON, DUNLAP. — BROMINE DERIVATIVES OF RESORCINE. 231 Dibromresorcine Diethylether, C6H2Br2(OC2H5)2. Resorcine diethylether, prepared by Pukall's method,* was dis- solved in glacial acetic acid and bromine added to it until no more was taken up and a yellow color appeared in the solution. Crystals were soon seen to separate, and these were recrystallized from alcohol until they showed a constant melting point, when they were dried in a desiccator and analyzed with the following results: — I. 0.2262 gram substance gave, according to the method of Carius, 0.2644 gram of argentic bromide. II. 0.2272 gram substance gave 0.2628 gram of argentic bromide. Calculated for Found. C6H2Br„(0C2H5)2. I. II. Bromine 49.38 49.74 49.22 Properties. — The dibromresorcine diethylether crystallizes in white silky felted long slender prisms terminated by one plane nearly at right angles to the sides; they melt at 99°-100°, and are slightly soluble in hot water ; readily soluble in ether, benzol, and carbonic disulphide ; very soluble in acetone and chloroform ; also soluble in ligroin. Alcohol is the best solvent to use in purification. In a sealed tube at 100° sodic ethylate does not react with dibrom- resorcine diethylether, nor was any action observed until the tube was heated to 200°, when, upon distillation of the product with steam, a few oil drops passed over. This experiment shows that the dibrom- resorcine diethylether is no more reactive than symmetrical tribrom- benzol, if in fact as much so, since Blau has converted this latter substance, by heating it at 120°-130° with sodic methylate, into di- bromphenol and its methylether.f As the interest in this work lay in the study of the effect of other radicals on the atoms of bromine rather than in the nature of the compounds formed, it was not thought worth while to spend the time necessary to prepare the product in quantity sufficient for its identification. Dibromresorcine diethylether does not react with aniline in a sealed tube at temperatures ranging from 150° to 175°, and sodic phenolate at 230° also seems to have no effect upon it. It cannot therefore be called a very reactive substance. Tribromresorcine Diethylether, C6HBr3(OC.2H5)2. The crystals of dibromresorcine diethylether were treated with an excess of bromine, and then heated upon a steam bath until the excess * Ber. d. ch. G., XX. 1141. t Monatsh. f. Chcm., VII. 630. Calculated for C6HBr3(OC2H5)2. Found. 59.55 59.70 232 PROCEEDINGS OF THE AMERICAN ACADEMY. of bromine had evaporated. The reddish oil which resulted soon solidified and was recrystallized from alcohol until it showed the con- stant melting point of 68°-69°, when it was dried in a desiccator and analyzed. 0.2222 gram of the substance gave, by the method of Carius, 0.3177 gram of argentic bromide. Bromine This body crystallizes beautifully in long pearly fibres, often col- lected in bunches and sometimes branched, which melt at 68°-69° and are practically insoluble in cold water; very soluble in benzol and carbonic disulphide ; readily soluble in ligroin and chloroform ; quite soluble in acetone, but less soluble in ether. Alcohol is the best sol- vent for purification. This tribrom ether reacts but very slightly with sodic ethylate at the boiling temperature of alcohol, and no more strikingly when boiled with sodic amylate. When boiled with aniline, the tribrom ether gave an amorphous black mass from which nothing definite could he ob- tained. Inasmuch as the sodic ethylate had so little action on this body, it was deemed unnecessary to carry this line of work farther, for the reasons given under the dibrom compound. Constitution of Tribromresorcine Diethylether. Tribromresorcine diethylether dissolves readily in fuming nitric acid, giving the body C6Br3N02(OC2H5)2, as was proved by its melting point, 100°-10l°. As one of us and Warren # prepared this substance from symmetrical tribromtrinitrobenzol, this experiment proves the symmetrical position of the three bromine atoms in tribromresorcine diethylether. Dinitroresorcine Diethylether, C6H2(N02)2(OC2H5)2. When dibromresorcine diethylether is treated with fuming nitric acid, the result is different from that obtained with the corresponding tribrom compound. To study this action, some of the dibrom ether was added to an excess of fuming nitric acid, when it immediately dissolved, after which it was treated with excess of water and the reddish oil which separated allowed to solidify. It was then recrys- * These Proceedings, XXV. 183. JACKSON, DUNLAP, — BROMINE DERIVATIVES OP RESORCINE. 233 tallized from alcohol till it showed the constant melting point of 12G°, after which it was dried at 100° and analyzed. 0.1313 gram of the substance gave 13.5 c. c. of moist nitrogen at a temperature of 26° and a pressure of 746.6 mm. Calculated for C6(N02)2H2(0C2H5)2. Found. Nitrogen 10.94 11.19 Properties. — The dinitroresorcine diethylether crystallizes from alcohol in radiated clusters of long white needles, or of smaller curved needles forming woolly masses. It is of a very faint yellow color and melts when pure at 126°. It is readily soluble in chloroform and acetone ; slightly soluble in ligroin ; soluble in benzol ; tolerably soluble in carbonic disulphide ; soluble also in ether. Alcohol is the best sol- vent for purification. Tribromresorcine, C6H Br3(OH)2. Tribromresorcine was prepared by the method of Benedikt,* as this yields better results than the treatment of resorcine dissolved in water with bromine, used by Hlasiwetz and Barth.f For this purpose the calculated amount of bromine is run into a solution of resorcine in cold glacial acetic acid. After the addition of the bromine this solu- tion is somewhat warm, and upon cooling a finely crystallized product separates, which is best washed free from acetic acid by decantation. If more than the calculated amount of bromine is added, pentabrom- resorcine is formed in addition to the tribromresorcine. The yield by this method is slightly over 57 per cent of the theoretical. A great many conflicting statements occur in the chemical literature regarding the melting point of tribromresorcine. Hlasiwetz and Barth, who discovered it, give no melting point. Typke.J in some work on tribromdioxyazobenzol, obtained a body with a constant melting point of 104°. The amount of this body that he had was too small for analysis, but a careful comparison of its properties, especially its melting point, its crystalline form, and its solubility, with that of tribromresorcine he had prepared himself, convinced him of the iden- tity of the two substances. Rudolph Benedikt § by the action of tin and hydrochloric acid upon Stenhouse's pentabromresorcine, obtained tribromresorcine which * Monatsh. f. Chem., IV. 227. \ Ber. d. ch. G., X. 1578. t Ann. Chem., CXXX. 357. § Ibid., XI. 2168. 234 PROCEEDINGS OP THE AMERICAN ACADEMY. melted at 1110. He also gives analyses which agree very well indeed for the body in question. Von Pechmann,* three years later, obtained a body by the action of an excess of bromine upon benzolresorcinephthaleni, which he called tribromresorviue on the ground of the melting point of 103°. He gives no analysis of this body, notwithstanding the fact that Benedikt had previously found its melting point to be 111°. Beilstein in his "Handbuch" gives only Typke's mehiug point of 104°. All of the tribromresorcine we have made use of in this research melted at 1110, the point given by Benedikt, and repeated recrystal- lization from water, which seems to be preferable to all other solvents, failed to raise its melting point higher. According to Hlasiwetz and Barth,f tribromresorcine when recrys- tallized from water contains water of crystallization, which it loses at 100°. As the presence of water of crystallization might explain the differences in the melting points given by previous observers, we examined the tribromresorcine in this direction. Some of it was crys- tallized from hot water and dried in the air to a constant weight. This air-dried salt, which remained constant in a desiccator over cal- cic chloride, was heated to 100° for ten hours, when 0.2169 gram of the substance lost but 0.4 milligram in weight. Therefore tribromresorcine when recrystallized from water does not contain water of crystalliza- tion which it loses at 100°, as Hlasiwetz and Barth state. In the recrystallization of the tribromresorcine from hot water, especially from boiling water, it was noticed that the crystals had a light pink color, so a study of the action of boiling water upon this body was taken up. One gram of the pure tribromresorcine was boiled in a 500 c. c. flask, fitted with an inverted condenser, with 250 c. c. of water. The greater part of the solid went into solution, and after the boiling had proceeded but a very short time the liquid became pink in color, which deepened into a dark red after the action had been carried on for twenty-four hours. The portion of the tribromresor- cine which did not go into solution melted under the boiling water, its color also changing from a pink to a very dark red after the boiling had been continued for some time. At the end of twenty-four hours' boiling, we found that all of the tribromresorcine had not been changed into the red body, for on cooling crystals of a light pink color separated from the solution, which had the melting point of * Ber. d. ch. G., XIV. 1061. t Ann. Chem., CXXX. 357. JACKSON, DUNLAP. BROMINE DERIVATIVES OF UESORCINE. 235 111°. Even further boiling, it seemed, failed to complete the change. Upon the addition of argentic nitrate to this solution, a heavy pre- cipitate of argentic bromide was obtained, showing that bromine had been removed from the tribromresorcine. The red body which formed the organic product in the reaction was amorphous, and could not be recrystallized from any solvents we could find. It was thought that possibly in a sealed tube the reaction would run much more smoothly, converting all the tribromresorcine into the amorphous red body. Consequently, quite a number of experiments were tried with varying amounts of the substance, heated from twelve to twenty-five hours at different temperatures, but without success. The tribromresorcine, it seems, is incapable of complete transformation into the red body at temperatures below 200°, while at 200° it is carbonized after heating for a short time. As the solubilities of the red body are practically the same as those of tribromresorcine, all methods for purification failed, and it could not be brought into a state fit for aualysis. It is probable, however, that it contains more hydroxyl groups than the tribromresorcine, forming perhaps in the first place a substituted pyrogallol, (for in alkaline solutions the red body immediately blackens.) which is afterwards converted into more complex substances by reactions in which the hydroxyla bear a part. Aniline or sodic alcoholates also acted easily with the tribromresor- cine, but although much bromine was evidently removed in the form of hydrobromic acid or sodic bromide, in no case could any organic product be isolated in a state fit for analysis. Sodium Salt of Tribromresorcine, C6HBr3(ONa)2 . 2 C2H5OII. This salt cannot be obtained from an aqueous solution, although it seemed to be formed, since upon treating tribromresorcine with a dilute solution of sodic carbonate carbonic dioxide was evolved, but the solution at once turned black and evidently underwent a deep- seated change. The following method for obtaining it was accordingly adopted. One gram of pure tribromresorcine was dissolved in absolute ether, and to this was added less than the calculated amount of sodium dissolved in absolute alcohol (the calculated amount of sodium is 0.13 gram). A precipitate immediately separated, consisting of short flat prisms having a silky lustre. These were filtered off and thoroughly washed with absolute ether. In the preparation of this body, t lie ether must be in large excess ; if the reverse was true (that is, if the absolute alcohol were in excess) no precipitate was formed. The 236 PROCEEDINGS OF THE AMERICAN ACADEMY. crystals were transferred, after washing with absolute ether, to a desiccator and allowed to dry in a stream of hydrogen. Some of the salt thus prepared was submitted to analysis. All the operations with this substance must be carried on as quickly as possible, since it has a strong tendency to decompose, forming a black mixture, and this change takes place even in vacuo or in an atmosphere of hydrogen, although not so quickly as when in contact with the air. Even during the weighing in a closed tube, the substance often decomposed, but the following analyses were made with the unchanged substance, although in all the analyses made the samples used were light brown in color, showing that slight decomposition had taken place. I. 0.1357 gram of the salt gave 0.0396 gram sodic sulphate. II. 0.1398 gram of the salt gave 0.0416 gram sodic sulphate. Calculated for Found. C6HBr3(ONa), . 2 C2H50H I. II. Sodium 9.52 9.45 9.64 These results indicate that the salt contained two molecules of alcohol of crystallization, and that alcohol was present in the body was evident, for a great many times, in attempting to weigh it, the contents of the weighing tube would suddenly turn black, giving off heat, while a liquid collected in large drops on the sides of the tube which was recognized as alcohol. Properties. — The sodium salt of tribromresorcine crystallizes in short flat prisms of a white color and a silky lustre. Its most strik- ing property is its great instability, decomposing rapidly with the elim- ination of alcohol, the salt becoming black in color; when dissolved in water, the solution immediately turns black. A freshly prepared lot of this sodium salt was divided into two portions, one of which was put into a small tube through which a rapid stream of hydrogen was passing, the other was put into a test tube which was surrounded by a freezing mixture of salt and ice, the substance in the test tube beinsf in contact with the air. It was found that the sodium salt in the stream of hydrogen (at ordinary temperature) blackened more rapidly than did the salt which was in the freezing mixture in contact with the air. This experiment proves that the salt is more stable in the cold than at ordinary temperatures, and also that the blackening is not due to oxidation. Since then we had proved that this behavior was not due to oxidation, it seemed possible that the black mixture obtained by drying at 100° might have the same percentage com- position as the dried salt. Accordingly the amount of sodium was Calculated for CBHBr3(0Na)2. Found. 11.76 11.40 JACKSON, DUNLAP. — BROMINE DERIVATIVES OF RESORCINE. 237 determined in a specimen of the salt which had been allowed to blacken and dried to a constant weight at 100°. The black body is slightly hygroscopic, so it must be weighed in a weighing tube. 0.2052 gram of the substance gave 0.0722 gram of sodic sulphate. Sodium An attempt to determine the amount of alcohol present by heating some of the freshly prepared salt to 100° gave the following results : — 0.5944 gram of the substance lost at 100° 0.1068 gram in weight. Calculated for C6Br3H(0Na)2 . 2 U2H5OH. Found. Alcohol 19.05 17.97 This determination, although one per cent too low, is as near as we could expect it to come, inasmuch as it was impossible to obtain a sample for analysis which was perfectly white ; but there is a more serious objection to this result and the determination of the sodium in the salt dried at 100°, as we have no proof that alcohol was the only substance volatilized during the decomposition and subsequent drying. In fact the determinations of the amount of sodic bromide contained in the dried mixture indicated that a certain quantity of hydrobromic acid might have been evolved. The only analyses of the salt, therefore, which are above suspicion, are those first given of the white uudecomposed preparation. The following experiments were tried to prove the nature of this decomposition of the salt. When the black decomposition product was dissolved in water, and dilute sulphuric acid added, a flocculent reddish brown precipitate was formed, which very much resembled ferric hydrate in appearance. The clear filtrate from this gave a very heavy precipitate of argentic bromide upon addition of argentic nitrate. This shows that the decomposition was produced by the elimination of a portion of the bromine from the ring. The amount of sodic bromide formed during the reaction was determined by treating the blackened salt with water of varying temperatures and for different lengths of time, and then the solutions were acidified with dilute sul- phuric acid, the precipitate filtered off, and bromine determinations made in the clear filtrate. The results obtained were not constant, varying from 32.01 per cent to 37.90 per cent of bromine, and this variation may be explained by the supposition that a certain amount of hydrobromic acid escaped during the decomposition. 238 PROCEEDINGS OP THE AMERICAN ACADEMY. The study of the red flocculent precipitate obtained by the addition of sulphuric acid to an aqueous solution of the sodium salt of tribrom- resorcine has not led to any trustworthy results. Treatment with water and ether showed that it was not homogeneous, and the different fractions thus obtained could not be brought into a state fit for analysis, as no means of proving that they were pure could be found. Some of these analyses gave numbers agreeing fairly well with the very probable formula [C6H2P>rOH]202 ; but for the reason just given it is impossible to accept this formula as established by them. Tribromresorcine when treated with aqueous sodic carbonate dis- solves with the evolution of carbonic dioxide, the solution immediately blackening, as has been stated already ; upon acidification with dilute sulphuric acid a reddish brown flocculent precipitate was formed similar to that mentioned above, and the clear filtrate from this body gave a very heavy test for bromides. That in this case the blackening was not due to oxidation was shown by the fact that it took place in an atmosphere of hydrogen, even when the solution had been boiled to insure the absence of dissolved oxygen. Sodic or potassic hydrate has the same effect upon tribromresorcine as the carbonate. The red precipitate obtained upon acidifying has not yet been obtained in crystals, and when heated decomposed without melting, giving off hydrobromic acid. This body is doubtless the same as that obtained by the action of water on tribromresorcine, and also from the black- ened sodium salt by acidifying its aqueous solution, as has been described above. Action of Sodium Malonic Ester on Tribromresorcine. This reaction was taken up to see if another case of the peculiar substitution of bromine by hydrogen, already noticed by one of us,* could not be obtained. After a great many trials the best mode for studying the reaction was found to be as follows. Five grams of sodium were dissolved in 250-300 c. c. of absolute alcohol, and to this were added thirty-five grams of malonic ester. Twenty-five grams of tribromresorcine were then added, and this turned dark and afterward dissolved. After boiling; the mixture for three hours in a flask fitted with an inverted condenser, the alcohol was distilled off and the con- tents of the flask were acidified with dilute sulphuric acid. A dark red oil separated, which was dissolved in ether, this solution then dried over calcic chloride, and the ether distilled off. Upon standing several * These Proceedings, XXIV. 1, 256. JACKSON, DUNLAP. — BROMINE DERIVATIVES OP RESORCINE. 239 days, the residual oil turned rather viscous, and crystals were seen permeating the mass. These crystals may be obtained by absorbing the oil with bibulous paper or by distilling in vacuo. By crystalliza- tion from dilute alcohol the constant melting point 76° was obtained, and as the crystals are white, contain no bromine, and their solu- bility, mode of crystallization, and shape correspond with those of acetyleutetracarboxylic ester, there can be no doubt that they are this substance. This acetyleutetracarboxylic ester could hardly have been formed otherwise than by the action of brommalonic ester on sodiummalonic ester, and the brommalonic ester it would seem must have been produced by the interchange of a bromine atom of the tri- bromresorcine with one of the atoms of hydrogen of the malonic ester. It follows from this reasoning, therefore, that we have here another case of the curious replacement of bromine by hydrogen, so often observed in this Laboratory, and that the two hydroxyl groups of the resorcine exercise the same action on one or more of the atoms of bromine as the two nitro groups in the tribromdinitrobeuzol. The following fruitless attempt was made to isolate the other organic de- rivative formed. If the red oil absorbed by the bibulous paper, after the acetyleutetracarboxylic ester has been separated as completely as possible, be treated with benzol, it is divided into a soluble and an insoluble portion. The insoluble portion, by precipitating an alcoholic solution of it with benzol, came down in a flocculent form, and when dry was light brick-red in color. It was evidently closely related to the substance obtained by the action of water or sodic carbonate on the tribromresorcine, and like that could not be purified. Action of Nitric Acid on Tribromresorcine . Fuming nitric acid reacts easily with tribromresorcine. The product was recrystallized from benzol and upon analysis gave the following results : — 0.2219 gram of the substance gave by the method of Carius 0.1479 gram of argentic bromide. Calculated for C6HBr(N02)2(0H)2. Found. Bromine 28.67 28.37 This analysis and its melting point, 192°-193°, show it to be the body heretofore described by Typke * and Fevre.f * Ber. d. ch. G., XVI. 555. t Bull. Chem. Soc, XXXIX. 591. 240 PROCEEDINGS OF THE AMERICAN ACADEMY. We have found, in attempting to prepare the diacetate of dinitro- bromresorcine, that dinitrobromresorcine would not react with acetyl chloride when heated with it at 100° for six hours in a sealed tube. Water does not react with dinitrobromresorcine in a sealed tube at temperatures varying from 100° to 150°. At temperatures between 150° and 190° the body is entirely decomposed, leaving a residue in the tube which seemed to be carbon, as it was insoluble in all solvents. It is strange that this substance should be less affected by water than the tribromresorcine, which is slowly acted on by water at its boiling point. The dinitrobromresorcine reacts with sodic etbylate, but, as in the case of the tribromresorcine, the resulting amorphous mass could not be purified or crystallized. Behavior of Tribromresorcinediacetate. While the tribromresorcine itself reacts readily with boiiing water and sodic ethylate, but gives unmanageable products, and the di- and tribromresorcine diethyl ethers do not act with these reagents, it was thought that possibly the diacetate of tribromresorcine miglit be better to work with than either of the ethers or the free phenol, because it holds an intermediate place between the acid tribromresorcine and its neutral ethers. Instead of using Claassen's method* of boiling penta- bromresorcine with acetic anhydride, then with glacial acetic acid, and finally with water, the tribromresorcinediacetate was directly prepared from tribromresorcine by treating it in a sealed tube for four or five hours at 100° with either acetyl chloride or acetic anhydride, prefer- ably the former. It is best to use a slight excess of acetyl chloride, and after the heating is finished to decompose the excess by means of water. The diacetate is easily recrystallized from alcohol, and was identified by its melting point, 108°. After boiling tribromresorcinediacetate with water for twenty-four hours, the water was slightly pink in color, and gave a feeble test for bromides with argentic nitrate. This is due, no doubt, to the fact that water at 100° saponifies a small part of the diacetate, yieldiug tribrom- resorcine, which in its turn is acted upon by the boiling water, as before noted. In the hope of avoiding this saponification, a sealed tube containing the diacetate was heated over night at 157° with ordinary acetic acid, but the saponification was not prevented, for the result was the same as if the free tribromresorcine had been heated with water alone to that temperature. Sodic ethylate in alcoholic * Ber. d. ch. G., XI. 1439. JACKSON, DUNLAP. — BROMINE DERIVATIVES OF RESORCINE. 241 solution when warmed with tribromresorcinediacetate deposited a pre- cipitate which when filtered ofF soon turned black. This body was undoubtedly the decomposition product of the sodium salt of tribrom- resorcine. On account of the readiness, therefore, with which the diacetate was saponified, it was not worth while to carry on this line of investigation further. IVibromnitroresorctnediacetate, C6(NOo)Br3(OC2H30)2. The diacetate of tribromresorcine when added to an excess of fuming nitric acid immediately dissolved, and after standing several minutes the solution became very warm and began to boil, the body in solu- tion becoming completely oxidized; but if just before the oxidation took place, when the solution was already hot, it was poured into an excess of water, the resulting }'ellowish oil soon solidified. The sub- stance was purified by crystallization from alcohol until it showed the constant melting point of 161°, when it was dried at 100° and analyzed with the following results : — 0.1752 gram of the substance gave, by the method of Carius, 0.2086 gram of argentic bromide. Calculated for CcN02Br3(0C2H30),. Found. Bromine 50.42 50.69 The diacetate therefore behaves toward nitric acid like the tribrom- ether rather than like the tribromphenol itself, for in the latter and in the dibromether atoms of bromine are replaced by nitro groups, as already described earlier in this paper. Properties. — The tribromnitroresorcinediacetate forms good-sized prisms terminated by two planes at an obtuse angle to each other ; it is very pale yellow, almost white, and melts at 161°. It dissolves slowly in ligroin, is readily soluble in chloroform, benzol, and ether, quite soluble in acetone, and tolerably soluble in carbonic di sulphide. In order to establish if possible the position of the three bromine atoms in tribromresorcine, attempts were made to saponify the tribrom- resorcinediethylether, which has the three bromine atoms arranged symmetrically. When saponified in a sealed tube with aqueous hydro- chloric acid, however, the product was a red body similar to that formed by the action of water on tribromresorcine. Accordingly, in hopes of a better result, dry hydrochloric acid gas was passed for one hour over tribromresorcinediethylether heated in an oil bath to 200°, but no saponification took place. vol. xxix. (n. s. xxi.) 16 242 PROCEEDINGS OF THE AMERICAN ACADEMY. XL CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF HARVARD COLLEGE. ON MUCOPHENOXYCHLORIC ACID. By Harris Eastman Sawyer. Presented by H B Hill, February 14, 1894. It was found some years ago, by Hill and O. R. Jackson,* that mucobromic acid was decomposed by treatment with baric hydrate, and that the course of the reaction varied, according as the solution was made strongly or feebly alkaline. In the first case, formic, di- bromacrylic, and brompropiolic acids were formed. The product in the latter case, as has been shown in a more recent investigation,! was mucoxybromic acid. Hill and Stevens, pursuing the subject fur- ther, studied the action of potassic phenylate on mucobromic acid, and obtained a mucophenoxybromic acid whose constitution was established by its conversion into phenoxybromacrylic and phenoxybrommaleic acids. $ They made, however, no attempt to prepare the corresponding series of chlorine compounds, because of the difficulties then attendant on the preparation of mucochloric acid. As it has become possible to obtain this acid in any desired quantity, it has seemed desirable to continue this work, which is now of more interest because of the rela- tionship between the substituted mucobromic and mucochloric acids, and the recently described crotonolactones of Hill and Cornelison.§ The author would here express his deep obligations to Professor H. B. Hill, at whose suggestion the research was undertaken ; and to Mr. Howard Nash, for assistance in some of the analytical work. MUCOPHENOXYCHLORIC ACID. The following method has been found most advantageous for the preparation of this acid. A solution of 17.6 grams of potassic hydrate * These Proceedings, Vol. XVI. p. 188. t These Proceedings, Vol. XXII. p. 316. \ These Proceedings, Vol. XIX. p. 262. § These Proceedings, ante, p. 42. SAWYER. — MUCOPHENOXYCHLORIC ACID. 243 and of 25 grams of crystallized phenol in 25 c. c. of water is cooled to 10°, and 13.2 grams of powdered mucochloric acid stirred in, as rapidly as is possible without rise of temperature. After standing for half or three quarters of an hour, at ordinary temperatures, the potassium salt of the new acid separates as a finely crystalline precipitate, which, after a second strong cooling, is filtered off and washed with a little ice-cold water. From a solution of this salt, hydrochloric acid precipitates muco- phenoxychloric acid as an oil, which solidifies on standing, and may be recrystallized from benzol. From a solution of the pure potassium salt, however, the acid is precipitated at once, in the crystalline con- dition. I. 0.1750 gram substance gave 0.3412 gram C02 and 0.0517 gram H20. II. 0.2602 gram substance gave 0.1652 gram AgCl. HI. 0.3102 gram substance gave 0.1966 gram AgCl. in. Carbon Calculated for C4H2(0C6H5)C103. 52.98 i. 53.17 Found II. Hydrogen Chlorine 3.09 15.67 3.28 15.70 15.68 Mucophenoxychloric acid is sparingly soluble even in warm water, still less in cold. On the spontaneous evaporation of its aqueous solu- tion, it crystallizes out in clusters of long, thin plates, melting at 91°. It is readily soluble in alcohol, ether, chloroform, and benzol, even in the cold, and is extremely soluble in hot benzol. It is dissolved but sparingly by carbonic disulphide and ligroin. Potassic 3fucophenoxy chlorate, KC4H(OC6H5)C103 . H20. An aqueous solution of the crude potassium salt, saturated at 50°, is filtered while warm, and cooled sharply in a freezing mixture. After some minutes, the salt separates in small rhombic plates which contain one molecule of crystal water. Thoroughly dried by pressure, these suffer no material loss of weight on exposure to the air ; over sulphuric acid there is a diminution in weight corresponding to about one fifth of the molecule of water; at 95°,after a further loss approxi- mately equivalent to the remainder of the molecule, browning and slight decomposition set in, making the exact determination of the crystal water impossible. 244 PROCEEDINGS OF THE AMERICAN ACADEMY. I. 0.3778 gram air-dried salt gave 0.1166 gram K2S04. II. 0.6117 gram air-dried salt gave 0.1878 gram K2S04. III. 0.3891 gram air-dried salt gave 0.1188 gram K2S04. Calculated for Found. KC4H(OC6H5)01O3 . H20. i. II. in ium 13.84 13.89 13.79 13.71 Baric Mucophenoxychlorate, Ba[C4H(OC6H5)C103]2. It was first attempted to prepare this salt by neutralizing a cold aqueous solution of the acid with baric carbonate, and after filtration allowing it to evaporate in vacuo over sulphuric acid. This method, which gave satisfactory results with mucophenoxybromic acid,* proved fruitless here, for the solution rapidly became acid, and deposited crystals which, from their form and melting point, 91°, were evidently the free acid. The salt was finally prepared by precipitation of an aqueous solution of the pure potash salt, saturated at 50°, with a similarly saturated solution of baric chloride; on chilling in a freezing mixture, a crystalline precipitate appeared, whicli was sucked off, and washed with small quantities of ice-cold water, until the washings gave no precipitate with argentic nitrate. The salt crystallizes in small anhydrous rhombic plates, which are extremely soluble, even in the coldest water; and as its solution is so readily decomposed, no attempt was made to recrystallize it for analysis. When pressed dry between filter papers, it remained unchanged in the air, lost only a trifling amount of weight over sulphuric acid, and but a trifle more at 105°. I. 0.2982 gram substance gave 0.1161 gram BaS04. II. 0.2152 gram substance gave 0.0858 gram BaS04. Calculated for Found. BatC^OCcII^ClOa].,. I. II. Barium 23.35 23.60 23.47 No detailed study has yet been made of the decomposition mentioned above ; it has, however, been noticed that there is no formation of baric bromide, and that phenol and baric oxalate are set free. Argentic Mucophenoxychlorate, AgC4H(OC6H5)C103. This salt is best prepared by adding a solution of the potassium salt to an excess of moderately dilute argentic nitrate solution. There is an immediate precipitate of white granular crystals, which is increased * These Proceedings, Vol. XIX. p. 264. SAWYER. — MUCOPHENOXYCHLORIC ACID. 245 in quantity by sharp cooling, and scratching. The salt is sparingly soluble in water, more readily in warm than cold, and is deposited by the warm concentrated solution in short compact prisms, which con- tain no crystal water. 0.4115 gram substance gave 0.1771 gram AgCl. Calculated for AgC4H(0C6H6)C103. Found. Silver 32.40 32.40 A solution of this salt, on standing, undergoes a decomposition simi- lar to that taking place in a solution of the barium salt. Phenoxychloracrylic Acid. The reaction taking place when mucophenoxychloric acid is decom- posed by an excess of alkaline hydrate, is analogous to that occurring when mucophenoxybromic acid is treated in like manner, the products being formic and a substituted acrylic acid.* Phenoxychloracrylic acid is most advantageously prepared in the following manner. Po- tassic mucophenoxychlorate is added to an equal weight of potassic hydrate dissolved in twice its weight of water, the mixture being gently heated on the water bath. On cooling, potassic phenoxychlora- crylate separates in well formed crystals ; these are dissolved in water and the acid is precipitated from their filtered solution by the addition of hydrochloric acid. When recrystallized from hot water, and dried over sulphuric acid, it gave the following results: — I. 0.1862 gram substance gave on combustion 0.3738 gram C02, and 0.0616 gram H20. II. 0.2362 gram substance gave 0.1702 gram AgCl. III. 0.2018 gram substance gave 0.1458 gram AgCl. Calculated for Found. C3H2(OCr)H5)ClC02. I II. III. Carbon 54.41 54.75 Hydrogen 3.51 3.68 Chlorine 17.92 17.84 17.89 The acid dissolves readily in alcohol, ether, benzol, ligroin, and carbon disulphide ; is nearly insoluble in cold water, and only spar- ingly soluble in boiling water, from which it may be crystallized in long silky needles, melting at 104°-105°. * These Proceedings, Vol. XIX. p. 267. 246 PROCEEDINGS OP THE AMERICAN ACADEMY. Potassic Phenoxychloracrylate, KC3H(OC6H5)C102. When a hot aqueous solution of phenoxychloracrylic acid is neutral- ized with potassic carbonate, and partially evaporated, potassic phe- noxychloracrylate separates from the concentrated solution in well formed plates, which may be recrystallized from hot water, although quite soluble in the cold. These crystals are permanent in the air, and lose no weight at 110°. The air-dried salt was analyzed with the following results : — I. 0.3310 gram ignited with H2S04 gave 0.1206 gram K2S04. II. 0.1929 gram ignited with H2S04 gave 0.0707 gram K2S04. Calculated for Found. KC3H(0C6H6)C102. I. II. Potassium 16.49 16.34 16.43 Calcic Phenoxychloracrylate, Ca[C8H(OC6H5)C102]2 .5 H20. A hot aqueous solution of the acid is neutralized with calcic car- bonate and concentrated, when the calcium salt crystallizes out in clusters of long needles, which are permanent in the air. They lose most of their crystal water over sulphuric acid, and the whole at 110°. I. 0.7847 gram air-dried salt lost at 110° 0.1343 gram H20. II. 0.3240 gram air-dried salt, ignited with H2S04, gave 0.0847 gram CaS04. Calculated for Found. Ca[CsH(0C6H5)C102]2.5H20. I. II. Water 17.14 17.12 Calcium 7.62 7.69 0.1684 gram salt, dried at 110°, gave, on ignition with H2S04, 0.0525 gram CaS04. Calculated for Ca[C3H(OCr)H5)C102]2. Found Calcium 9.19 9.17 Baric Phenoxychloracrylate, Ba[C3H(OCcH5)C102]2 . 5 H20. The hot aqueous solution of the acid is neutralized with baric car- bonate, evaporated until a thin film appears on the surface, and then allowed to cool, when the barium salt separates in clusters of long silky needles. These crystals appear to be slightly efflorescent, losing on exposure to the air for several days less than a molecule of water. SAWYER. — MUCOPHENOXYCHLORIC ACID. 247 I. 0.3G40 gram salt dried between filter papers gave, on ignition with H.,S04, 0.1346 gram BaS04. II. 0.7799 gram salt dried between filter papers lost, at 110°, 0.1140 gram ILO. Calculated for Found. Ba[C3U(OC6H6)C102]2 . 5 H20. i. II. Water 14.47 14.60 Barium 21.52 21.64 I. 0.1382 gram salt dried at 110° gave, on ignition with H2S04, 0.6076 gram BaS04. II. 0.1535 gram salt dried at 110° gave, on ignition with H2S04, 0.6754 gram BaS04. Calculated for Found. Ba[C3H(0C,iII5)C10.2]2. I II. Barium 25.75 25.85 25.87 Argentic Phenoxychloracrylate, AgC3H(OC6H5)C102. On adding argentic nitrate to a hot aqueous solution of amnionic phenoxychloracrylate, there is precipitated crystalline argentic phenox- ychloracrylate, which may be recrystallized from hot water in beau- tiful feathery needles. Dried over sulphuric acid it gave the following results : — I. 0.1759 gram salt gave 0.0821 gram AgCl. II. 0.1165 gram salt gave 0.0547 gram AgCl. Calculated for Found. AgC3H(0C6H5)C102. I. II. Silver 35.35 35.13 35.34 Phenoxychlormaleic Acid. The structure of mucophenoxychloric acid has been further charac- terized by its oxidation to phenoxychlormaleic acid. When a solution of mucophenoxychloric acid is heated with argentic oxide, there is deposition of metallic silver, and formation of argentic phenoxychlormaleate, which crystallizes from the hot filtered solution in minute stellate clusters. If the silver be precipitated with hydro- chloric acid, and the filtered solution strongly concentrated, we obtain, on cooling, fine needles of phenoxychlormaleic acid. It has, however, been found advisable, on account of the sparing solubility of the silver salt, to prepare the acid in another way. Hill and Cornelison found that mucophenoxybromic acid might advantageously be converted to phenoxvbrommaleic acid by boiling 248 PROCEEDINGS OP THE AMERICAN ACADEMY. the aqueous solution of its oxime ; * and the same procedure has here been followed for the oxidation of mucophenoxychloric acid. Muco- phenoxychloroxime is boiled with about twenty times its weight of water, until nothing separates on cooling. This solution is acidified with hydrochloric acid and extracted with ether, and the phenoxy- chlormaleic acid is removed from its ethereal solution by sodic car- bonate, from which, in turn, it is separated by a second acidification and extraction ; the ether is allowed to evaporate, and the crystalline residue is purified by sublimation in a current of carbonic dioxide. The anhydride thus obtained is dissolved in a small quantity of water, which, on spontaneous evaporation, deposits the acid as a mass of interlacing needles. When thoroughly dried by pressure between filter papers, these crystals lose no weight in the air; but over sulphuric acid they suffer a gradual loss equivalent to one molecule of crystal water. I. 0.2948 gram air-dried substance lost over H2S04, 0.0195 gram H20. II. 0.2769 gram air-dried substance gave on combustion 0.0860 gram H,0 and 0.4660 gram C02. III. 0.1810 gram air-dried substance gave 0.0995 gram AgCl. Calculated for Found. C4H2(0C6U5)C104 . H,0. I. II. in. Water 6.91 6.62 Carbon 46.06 45.90 Hydrogen 3.46 3.45 ov Chlorine 13.62 13.60 0.1652 gram substance dried over H.,S04 gave 0.0960 gram AgCl. Calculated for . C4H2(OCliH3)C101. Found. Chlorine 14.64 14.38 The acid containing water of crystallization has no definite melting point, while that dried over sulphuric acid melts at 115°-122°, accord- ing to the mode of heating. It dissolves readily in water, alcohol, ether, chloroform, and benzol, is somewhat less soluble in ligroin, and only sparingly in carbonic disulphide. From hot benzol or ligroin it crystallizes in minute needles. The identity of acids prepared by the two methods already described has been determined, in the absence of a definite melting point, by the identity of their anhydrides. * These Proceedings, ante, p. 53. SAWYER. — MUCOPHENOXYCHLORIC ACID. 249 Phenoxychlormaleic Anhydride, C4(OC0II5)ClO3. This substance, purified by sublimation, was analyzed with the following results : — I. 0.2000 gram substance gave 0.0441 gram H20, and 0.3929 gram ( O,. II. 0.1054 gram substance gave 0.0664 gram AgCl. Calculated for Found. C^OCeHgJClO* I. II- Carbon 53.45 53.56 Hydrogen 2.23 2.45 Chlorine 15.81 15.60 It melts sharply at 97°, sublimes in thin curving plates, and dissolves readily in water, alcohol, ether, chloroform, and benzol, but only spar- ingly in ligroin and carbonic disulphide. Baric Phenoxychlormaleate, BaC4(OC6H5)C104 . 4 H20. This salt is most advantageously obtained in the following manner. Baric hydrate is added to a solution of the acid ammonium salt, pre- pared by boiling a solution of mucophenoxychloroxiine, until the re- action is faintly alkaline. Baric bromide is then added in sufficient quantity to replace all the ammonium by barium, and on the addition of alcohol the new salt is precipitated in nodular aggregations of needles. It is filtered off, pressed dry, and redissolved in water, from which it is redeposited, on spontaneous evaporation, in bundles of long silky needles. When dried for analysis by pressure between filter papers, its weight remains constant in the air ; but over sulphuric acid it suffers a loss corresponding to more than three molecules of crystal water, and at 110° there is a further loss, equivalent to the remainder of a fourth molecule. I. 0.6161 gram air-dried salt lost over H2S04, 0.0871 gram, and in addition at 110°, 0.0090 gram H20. II. 0.1830 gram air-dried salt gave 0.0960 gram BaS04. III. 0.1918 gram air-dried salt gave 0.1004 gram BaS04. Caleulated for Found BaC4(OC,;lI.-,)U104.4H20. I. II. III. Water 16.02 15.60 Barium 30.50 30.84 30.77 0.2060 gram salt, dry at 110°, gave 0.1273 gram BaS04. Calculated for BaC4(OC6H5)C104. Found. Barium 36.29 36.33 250 PROCEEDINGS OF THE AMERICAN ACADEMY. Argentic Phenoxychlormaleate, Ag2C4(OC6H5)C104. This salt was prepared by precipitating a solution of the acid with argentic ni-trate, as well as by heating a solution of mucophenoxy- chloric acid with argentic oxide. In both cases it crystallized in clusters of colorless and somewhat opaque dendritic needles, which were only sparingly soluble in hot water, and even less so in cold. 0.3590 gram air-dried salt gave 0.2252 gram AgCl. Silver Calculated for AgA(0C,.,H5)C104. Found. 47.29 47.22 Fig. 1. Largest Mass of the Smithville Iron. v<"':C;';r:. Fig. 2. Largest Section of the Smithville Iron, Etched. Shows at the right a nodule of graphite two inches in diameter. HUNTINGTON. — SMITH VILLE METEORIC IRON. 251 XII. CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF HARVARD COLLEGE. THE SMITHVILLE METEORIC IRON. By Oliver Whipple Huntington, Ph. D. Presented February 14, 1894. In the early part of last summer Mr. Herman Meyer sold three new masses of meteoric iron from Smithville, De Kalb County, Tennessee, to Professor Ward, of Rochester, New York. After the largest mass had been sawed into slices these were kindly sent by Professor Ward to the writer for examination, and with it was sent the following letter describing the find : — Smithville, De Kalb Co., Tennessee. Three siderites, weighing about seven, fifteen, and sixty -five pounds each. In November, 1892, Mr. John D. Whaley ploughed up the medium-sized meteorite. In a few days thereafter Mr. Berry Cantrell on the adjoining farm of James Beckwith ploughed up the large one at about two hundred feet distance from the first. These meteorites were carefully kept in the families of the respective finders. During December, after diligent search of some weeks, the third meteorite and the smallest was discovered and kept in the family of J. D. Whaley. On February 25, 1893, 1 obtained the medium-sized meteorite, and asked for thorough search of all their fields during the spring ploughing. This was done. Several pits were dug cov- ering nearly the entire space of three hundred feet in all directions from the spots where the siderites had been found. The ground was looked over after each rain, and I myself looked over the ground well on two occasions, and no piece was found. Only after being sure the three pieces were all that fell did I announce the meteorites to the public, which I did, after purchasing the other two, on July 12. The spot where found is three eighths of a mile south from Smithville and Lebanon Pike, two miles from Smithville, and on extreme south- west field of J. D. Whaley and adjoining field of James Beckwith. I satisfied myself that the meteorites were original and distinct, and that all was as represented. Since in my possession they have been strictly guarded. Herman Meyer, August 18, 1893. Cashier Bank of Carthage, Tennessee. 252 PROCEEDINGS OP THE AMERICAN ACADEMY. This locality is about forty miles southeast from the spot ou Cany Fork where the Carthage meteorite was found. There is no question that the irons were found as stated in the above letter, but one familiar with the well known and widely dis- tributed Cocke County iron could not but be struck with its close resemblance to the Smithville meteorites. The largest mass of the Smithville iron, as shown in the plate, Fig- ure 1, is roughly spherical, with no sigus of original crust, but marked by one deep pitting which once contained troilite now nearly weathered away, or possibly melted out during the flight of the body through the atmosphere. That the specimen had lain for a long time in the soil is evident from its being coated with a thick covering of magnetic oxide of iron, the magnetism being strong enough to attract iron nails with considerable force. This covering, however, does not fully con- ceal certain very typical features of the iron, — a marked silvery white- ness, a very striking and ready octahedral cleavage, a slightly yellowish metallic foil separating the crystalline plates of iron, — also numerous nodules of a very cleavable troilite, embedded in graphite and granular schreiberseit ; while perhaps the most striking feature of all is a nodule of fine-grained compact graphite nearly two inches in diameter. This is probably a larger mass of meteoric graphite than any on record. The only one which compares with it is that formerly described by J. Lawrence Smith in the Sevier County iron.* The weight of the Smithville nodule cannot be accurately estimated, as it was not observed till the mass had been sawed into slabs. It appeared to be nearly spherical, with a diameter as great as the longest dimension of the dumb-bell shaped nodule described by J. Lawrence Smith, making the total mass of the former considerably greater than that of the latter. Figure 2 of the plate represents an etched surface of the largest section of the sixty-four and a half pound specimen of the Smith- ville iron, and here again are very striking features. The natural size of the section is nine and a quarter inches by seven. Near the lower right-hand corner can be seen, somewhat indistinctly outlined, the large nodule of graphite just described, while the rest of the plate brings out quite markedly the peculiar features of the Widmanstattian figures. In the first place, nodules of graphite and troilite are abundantly scattered over the surface. Usually these nodules are troilite embedded in graphite, and this in its turn is sur- rounded by schreiberseit, though there is considerable variety in the * American Journal of Science, 3d ser., Vol. XI. p. 392, 1876. HUNTINGTON. — SMITHV1LLE METEORIC IRON. 253 relative arrangement of these three minerals in the individual nodules. In places the schreiberseit widens out into bright patches between the Widmanstattian plates, especially in proximity to the troilite nodules. This inequality in the distribution of the schreiberseit gives a very varied appearance to the etched surface, and areas selected from oppo- site ends of the slab shown in the plate (Fig. 2) could not possibly be identified by the Widmanstattian figures alone. In this character the iron very closely resembles those of Arva and Sarepta. Several analyses were made of the Smithville iron, though in the opinion of the writer such analyses are of very little value on account of the difficulty of sampling, since the nickel and cobalt must vary with the ttenite plates, while the phosphorus would depend upon how much schreiberseit happened to be in the mixture selected for analysis. Choosing, however, as uniform material as possible, the average analy- sis gave : — • Iron 91.57 Nickel 7.02 Cobalt 62 Copper Trace Phosphorus .18 Residue, mainly Cliftonite .... .15 99.54 Comparing this with the irons most closely resembling it we have : — Greenbrier Co.,* Jennie's Creek,T Cosby's Creek, J Sevier Co.,§ West Virginia. Wayne Co., W. Va. Cocke Co., Tenn. Tenn. Iron 91.59 91.56 87.00 93.80 94.03 TVirkpl 7 11) Cobalt '.J «8'31 «1200 *•"« "* Copper Trace Phosphorus .08 .13 Carbon — — .50 Residue , 1 2 Loss — — .50 .10 .10 99.50 100.00 100.00 98.56 98.57 * Mineralogical Magazine, Vol. VII. p. 183, 1887. t American Journal of Science, 3d ser., Vol. XXXI. p. 145, 1886. t Ibid., 1st ser,, Vol. XXXVIII. p. 250, 1840 ; Ibid., 1st ser , Vol. XLIII. p. 354, 1842. § Ibid., 2d ser., Vol. IV. p. 83, 1847. U By difference. 254 PROCEEDINGS OP THE AMERICAN ACADEMY. No analyses have been made of the meteorites of Waldron's Ridge,* Claiborne Co., and Lebanon,f Wilson Co., Tennessee, but they are generally accepted as identical with the Cocke County iron. Of the above analyses the Greenbrier County is probably the most reliable, and this it will be seen is almost identical with the Smithville. The analysis of Jennie's Creek is of very little account, since it consists of a determination of the percentage of the iron alone, the balance in weight being set down as nickel and cobalt, while the analyses of the Cocke County and Sevier County, unquestionably the same iron, differ widely from each other. In marked contrast to this variation is a collection of analyses placed in comparison by Fletcher, when describing the Greenbrier County specimen now in the British Museum. $ Greenbrier Co. Trenton. (L. Smith.) Rio Juncal. (Daniour.) Seneca River. (Shepard.) Staunton. (Santos.) Iron Nickel Cobalt Copper Phosphorus Sulphur Residue 91.59 7.11 0.60 Trace 0.08 Trace 0.12 91.03 7.20 0.53 Trace 0.14 0.45 92.03 7.00) 0.62) 0.21 92.12 7.54 0.08 0.26 91.44 ( 7.56 (0.01 0.02 0.07 0.02 0.14 99.50 99.35 99.86 100.00 99.86 On glancing at the above table one would suppose that it exhibited parallel analyses of the same iron, whereas they are actually irons so widely separated by their physical characters as well as by their geographical distribution that in our present knowledge of the subject we are forced to regard the resemblance of the chemical analyses as an accidental coincidence. Although a mere chemical analysis seemed insufficient to identify the Smithville iron with the Cocke County, or Sevier County iron, oue point in its composition seemed important. After dissolving a portion * American Journal of Science, 3d ser., Vol. XXXIV. p. 475, 1887. t Ibid , Vol XXXIII. p. 118, 1887. t Mineralogical Magazine, Vol. VII. p. 183, 1887. HUNTINGTON. — SMITHVILLE METEORIC IRON. 255 of the Sinithville iron in hydrochloric acid, assisted by a battery,* a black residue was obtained, consisting mainly of small graphitic crys- tals, with a predominence of cubo-octahedral forms, but showing also perfect little cubes without any modifications, and others with their edges truncated by the dodecahedron, and occasionally bevelled by a very obtuse tetrakis hexahedron. This was evidently the now well known form of meteoric graphite, first seen by Haidinger in the Arva iron,f named Cliftonite by Fletcher in the Youndegin iron, X but also found by him in the Cocke County iron. While examining a nodule of graphite formerly obtained by J. Lawrence Smith from the Sevier County iron, it was accidentally broken and showed in its interior what appeared to be a skeleton octahedron of graphite three eighths of an inch in diameter, and with all but one of its faces sufficiently perfect for measurement by an application goniometer. This striking feature at once suggests that this also may be a pseudomorph after diamond. With the Cliftouite from the Sinithville iron were to be seen numerous white glassy grains. On digesting the residue for a long time with hydrofluoric acid, most of the white grains disappeared, but a few remained entirely unaffected by the acid. These appeared as very brilliant transparent angular fragments, and exhibited a hard- ness sufficient to scratch the ruby. As only a few grains were obtained from the amount of iron placed at the disposal of the writer, no further experiments could be made, but in all probability the grains were diamond. In a letter from Professor Ward dated October 13, 1893, he says, " The ' Smithville ' seems to cut harder than any iron which we have yet undertaken " ; and Fletcher says of the Youndegin iron, " The large specimen was cut on the premises of the Museum by means of hack-saws, and was found to be so hard that three weeks were re- quired for the severance of a fragment of which the cut face is not two aud a half inches square." § May not this unusual hardness be accounted for by the presence of minute diamonds? In view of the constant occurrence of graphite in meteoric iron, and the frequent appearance of the same in the crystalline form of diamond, (the carbonaceous material apparently varying in hardness from one * These Proceedings, Vol. XXIX. p. 204. t Pogg. Annalen, Bd. LXVII. p. 437, 1846. } Mineralogical Magazine, Vol. VII. p. 121, 1887. § Ibid., p, 123. 256 PROCEEDINGS OF THE AMERICAN ACADEMY. and a half on the usual scale to the maximum hardness in the crystals of true diamond as found by the writer in the Canon Diablo iron,*) may it not be that all meteoric iron contains diamonds at least of the car- bonado variety, only waiting for a sufficiently careful search to reveal them, and that the peculiar toughness of these irons, so imperfectly accounted for by the network of Widmanstiittian plates, may be ex- plained by the presence of the hardest substance known in a state of excessively fine division? Glancing now for a moment at the accounts of the several masses of the Cocke County iron, we can see why this same iron has kept coining to light year after year ever since 1840. That it is given a different name each time is not to be wondered at, since the prices paid for meteorites by collectors puts a great premium on names. In the first description of the Cocke County iron, in 1840, Dr. Troost says at the opening of his paper : " During my excursions through East Tennessee I had seen small fragments of native iron, and had heard of large masses of it. It being considered a precious metal, all that was known about it, and the place where it was found, were kept a profound secret." t He then goes on to state how a few small pieces had come into his hands, and that he believed the original mass weighed two thousand pounds. He also adds that he went to Buncombe County, North Carolina, where there was said to be a great quantity of it, but found none. Two years later Professor C. U. Shepard described a mass in the museum of the East Tennessee University at Knoxville, and quoted a letter in regard to it : — " It is a portion of an irregular mass, which was given me about five years since. The mass, as you have been informed, was discovered in Cocke County. The proprietor resisted for some time all impor- tunities to discover where it was ; believing it to be some metal of great value. I assured his agent that it was native iron, and probably meteoric. After he became satisfied of its character, many individuals examined it, in place. It was entirely insulated on the ground, and weighed about seven or eight hundred pounds. Specimens were ob- tained from it and dispersed through the country. " It was my intention to have purchased and transported the entire * These Proceedings, Vol. XXIX. p. 204. t American Journal of Science, 1st ser., Vol. XXXVIII. p. 250. HUNTINGTON. — SMITHVILLE METEORIC IRON. 257 mass to Knoxville, until I learned that Dr. Troost, Geologist of the State, had obtained the refusal of it. He has conveyed it since to Nashville."* Later on, Shepard quotes two letters describing the original Cocke County mass : " The large mass of meteoric iron found some years ago in Cocke County (on a creek called Cosby's) fell into the hands of some persons who tried to break it with sledge-hammers; but not succeeding, they placed it upon what is here called a 'log-heap,' where, after roasting for some time, it developed certain natural joints, of which advantage was taken with cold chisels and spikes for its separation into fragments. These were put into a mountain wagon, and transported thirty or forty miles to a sort of forge, and there hammered into ' gun-scalps,'! and other articles of more common use. Some remnants of the mass fell into the hands of Dr. Troost. The original mass was one of rare character, and ought to have been pre- served entire. Much of it was composed of large and perfect octa- hedral crystals. Its weight was about a ton. Another mass weighing one hundred and twelve pounds was found near the locality of the larger one. This also was malleable, very white, and easily cut with a sharp instrument. It was picked up by a mountaineer, who, suppos- ing it to be silver, asked fifteen hundred dollars for it. After retain- ing it for some years, he finally sold it to a friend of mine for a small sum, who transferred it to Dr. Troost." Second letter : " The weight of the mass has been variously esti- mated; but I am certain it was never weighed prior to its being broken up. It was probably about two thousand pounds. In figure, it was an oblong square block. I saw several very regular octahedral crystals which had been detached from the exterior angles of the mass. I had formerly supposed that the whole of it had been taken to Lary's forge, in Sevier County, and the greater part of it there wrought into 'gun-scalps'; but very recently I have been informed that part of it was taken to the forge of Peter Brown, in Green County, and there forged. I understand that a man by the name of McCoy had a neat bar forged from it for making a gun-barrel, which, to use the expres- sion of Brown's son, " was as bright as silver." In conversation young Brown informed me that he thought a piece of the iron in its natural state still remained. On searching, it was found by a little girl of the * American Journal of Science, 1st ser., Vol. XLIII. p. 354, 1842. t The forged iron bar before being bored for a gun barrel is called in Ten- nessee a " gun-scalp." vol. xxix. (n. s. xxi.) 17 258 PROCEEDINGS OF THE AMERICAN ACADEMY. family. It weighs rather more than a pound, and had been preserved by the family as a nut-cracker." * About the year 1880, a fragment of iron, said to weigh eleven pounds, was found in Greenbrier County, West Virginia, in regard to which Fletcher says : " The finder and his official agent, thinking it a piece of rich iron ore, searched unsuccessfully for a vein : the specimen itself was taken to a country smith's shop, heated, and cut with a cold chisel ; the fragments were distributed as specimens of iron ore. Some time afterwards, two of them, weighing respectively 63 oz. and 31 oz., were given by the agent to Mr. Matthew A. Miller, Civil Engineer, of Richmond, Virginia ; convinced of their meteoric origin, he immediately tried to recover the pieces already distributed, but after travelling several hundred miles was forced to the conclusion that they were irrecoverably lost." f Later still Mr. George F. Kunz describes three small pieces from Wayne County, West Virginia, the largest of the three only weighing a little over half a pound, the rest of the original mass having been broken up and distributed.! The Wayne County specimens were sent to the writer by Mr. Kunz before being described, and were identified at once as more of the Cocke County iron. In con- nection with this iron a supposed fall of a meteorite in the direction of Wayne County some five years previous was quoted, but any one familiar with meteorites knows how frecpjently such accounts have to be rejected, even when the observer thinks the specimen fell at his feet, and that he picked it up while still too hot to hold. Another piece of the Cocke County iron was given to Professor N. S. Shaler in 1887, purporting to be a sample of a vein of native iron of indefinite extent, near Lebanon, Wilson County, Tennessee. It may be that this fragment was originally broken off from the Smith- ville irons and that they were the iron vein referred to, since they were found not far from the Lebanon Turnpike, though in the adjacent county. The map on the opposite page, drawn to scale, will give an idea of the distribution of the counties in which the various irons under discussion were found. At first one might be led to suppose that these numerous masses of iron resulted from a wide-spread shower, but on looking over the * American Journal of Science, 2d ser., Vol. IV. p. 84, 1847. t Mineralogical Magazine, Vol. VII. p. 183, 1887. t American Journal, 3d ser., Vol. XXXI. p. 146, 1886. HUNTINGTON. — SMITHVILLE METEORIC IRON. 259 records it is obvious that portions of the original mass from Cocke County were distributed far and wide ; and since only small portions have found their way into museums, it is not surprising that they have been turning up so frequently. Thus we have : — Reported Weight. Collected. Cocke County 2,000 lbs. A few small pieces Sevier County 112 " 112 lbs. Wayne County 27 « 2 « Greenbrier County 11 " Less than 6 " Wilson County Walclron's Ridge, ) 18 " 18 " Tazewell County x * *' \ /V Thus we see that about a ton of this iron is still missing, and, though some of it has been used up at a blacksmith's forge, other pieces are known to have been deliberately buried and dug up again to sell as samples of iron or silver mines, while others still have been carried to long distances, treasured by the owners as silver or some other metal of value, and this has been going on for more than fifty years. At this late date, therefore, it is impossible to form any conclusion in regard to the original distribution of the irons under discussion, or to determine whether only one mass originally fell, or whether there was a large shower, spreading over a considerable area, as in the case of the Canon Diablo meteorite recently studied. That in several cases specimens in our cabinets under different names came from this original Cocke County mass is more than 260 PROCEEDINGS OP THE AMERICAN ACADEMY. probable, but it may well be that in other cases the identity of characters only indicates that the several masses were parts of the same meteoric shower. The multiplication of unnecessary names is unfortunate, and for the simplicity of our catalogue it is desirable that such questions as the present should be investigated. In conclusion, the writer wishes to thank Professor H. A. Ward, of Rochester, New York, for his kindness in having the plates made with which this paper is illustrated, and for his extreme courtesy in furnishing material for analysis, as well as sending for inspection the large mass sawed into slabs. LAWS. — COEFFICIENTS OF SELF-INDUCTION. 261 XIII. CONTRIBUTIONS FROM THE PHYSICAL LABORATORY OF THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY. XLII. — ON AN APPARATUS FOR THE MEASUREMENT OF COEFFICIENTS OF SELF-INDUCTION, AND THE INVESTIGATION OF THE PHENOMENA OF ALTER- NATING CURRENTS. By Frank A. Laws. Presented by Charles R. Cross, January 10, 1894. The increasing use of alternating currents in telephony and elec- tric lighting has, during the last few years, directed attention to the measurement of coefficients of self-induction. In this paper certain apparatus devised for this purpose, together with methods for its use, will be described. Further discussion of the various methods of measurement proposed and of experimental results obtained is reserved for a later article. For investigations pertaining to telephony it is particularly desira- ble to measure inductances by means of alternating currents of suit- able known frequencies, and of the same order of magnitude as those used in practice. In all such work the reduction of the results is greatly simplified if the E. M. F. employed follows the sinusoidal law. Therefore the aim at the outset was to produce a dynamo whose E. M. F. should vary in this manner, and whose frequency of alterna- tion should be about that which is normal in telephonic work. Under these requirements there have been designed in the Rogers Laboratory two dynamos, or arrangements of dynamos. The first of these was planned during the winter 1890-91, and was used by Mr. L. Derr in connection with his thesis in 1891-92. This machine was provided with a stationary armature having Lord Kelvin's zigzag winding on the outer surface of a hollow cylinder. This was placed in the gap between a star-shaped piece of iron and a disk with a corresponding star-shaped opening, the two pieces being similarly and concentrically placed and forming the pole pieces of the machine. Owing to the 262 PROCEEDINGS OF THE AMERICAN ACADEMY. shape of the pole pieces the field between them is variable, being denser at the points of the stars, so that when the field moves past the armature an E. M. F. will be set up in the latter. One disadvantage of this design is that the E. M. F. is small, owing to the very large non-effective field ; for it is only the variation of the field that is of service in producing the E. M. F. By properly shaping the pole pieces a curve of E. M. F. was obtained, which, when plotted, was found to be nearly sinusoidal in form. The second machine was constructed by Messrs. C. L. Norton and P. H. Thomas, of the class of 1893, and forms the subject of the present paper. The special form of dynamo used was designed in accordance with a suggestion from them. The frequency is some- what low, being 400 complete alternations per second. When another machine is built, it will be so designed that the superior limit will be raised to 2,000 complete periods per second. Both the machine of 1890 and that of 1892 embody the same general features. Each has two independent armatures, one fixed, the other movable in phase relation to the first. This relation is capa- ble of measurement. At the outset it was determined that in these machines the armatures should contain no iron, and that they should remain stationary, in order to eliminate any possible microphonic action at the brushes. The machine, which is shown in place ready for use in Figure 1? will now be described. It consists, in brief, of two dynamos, with revolving fields mounted on one shaft to insure definite phase rela- tion. The armature of one is fixed, while that of the other may be given any desired angular advance or retardation in reference to the first by means of a tangent screw. This advance can be read off from the graduated head of the screw. The fields are independent of one another. In addition to the dynamos the shaft carries a con- tact arrangement for mapping wave-forms, and a stroboscopic disk for determining the speed of rotation. One of the field magnets is shown in plan in Figure 3 and in section in Figure 2. It consists of two star-shaped pieces of cast iron, keyed to the sh ft and having hubs of such a length that tho faces of the teeth are -fa" apart, thus leaving a gap in which is the armature, \" thick. The castings are recessed to receive the field coils, which are about one square inch in section and contain 250 turns of No. 16 B. & S. double cotton-covered wire. The coils were wound on forms, paraffined and taped, and were inserted when the machine was assembled. The terminals are brought out through bushed holes and Fig. 1. LAWS. — COEFFICIENTS OF SELF INDUCTION. 263 fastened to composition collector rings, which are carried and insulated by blocks of vulcanite mounted on the tips of the teeth of one of the castings. The brushes of spring copper are supported from the base plate. The eight teeth were milled from the round by a special Brown and Sharpe cutter, and have the form of a sinusoid laid off radially from the base circle. Of course, owing to the spreading of the lines of force, the particular form to be given to the teeth in order to obtain the best action can only be ascertained by trial. The armature consists of a vulcanite disk, 10£" diameter, \" thick, having a 4" hole in the middle. From this hole radiate sixteen equally spaced saw-cuts -fa" deep, ^" wide, and 2\" long, the outer extremi- ties of alternate pairs of cuts being connected by grooves concentric with the disk. In these cuts and grooves the armature wires are laid in shellac. The windings at present consist of three sections having respectively 10, 10, and 20 turns of No. 30 B. & S. copper wire. The total resistance. is about ten ohms. The manner of winding as well as the action of the dynamo, by varying the length of active wire, are evi- dent from Figure 3. The armatures are carried by built up hard-wood rings, \0\" outside diameter, S\" inside, and Ifa" deep. The ring in one dynamo is attached to a brass disk and hub, which fit the shaft and are rigidly bolted to the frame of the machine. In the other dynamo, the armature, ring, disk, and hub are attached to a worm-wheel, and the whole is movable on a sleeve supported firmly from the frame of the machine. Through this sleeve passes the shaft without contact. The worm-wheel gears into a tangent screw provided with a divided head, the gearing being such that 1° on the wave form is represented by 20° on the head. Backlash is prevented by a spiral spring. See Figures 2 and 3. Between the two dynamos is a brass disk, keyed to the shaft and provided with an insulated collector ring and contact point. By the side of this disk is a radial arm, revolving freely about the shaft, and capable of being clamped at any desired point to a fixed graduated circle. The arm carries the contact spring, made of watch spring and supported on its lower side very near its extremity to prevent the lengthening of the arc of contact by vibration. The spring can be moved out of contact with the point by a fine-pitched screw. This contact arrangement we have found to work satisfactorily at as many as 3,000 revolutions per minute. The shaft of the machine is of steel, 1" in diameter, 29" long, and runs in composition bearings, each of which is 5£" long. The frame of the machine is heavy, weighing 175 pounds, and is designed so 264 PROCEEDINGS OF THE AMERICAN ACADEMY. as to be very rigid. Driving is effected by means of a 2" belt run- ning over a 5" pulley. The motor (1 H. P.) is of the Sprague type. i i | -JJotU CHE({) i — r TiT '4. 2 O > •J o Variations of speed are obtained by the use of pulleys of different sizes on the motor shaft. Both machine and motor rest on a brick pier, the machine being LAWS. — COEFFICIENTS OF SELF-INDUCTION. 265 set in sulphur; mechanically, the arrangement leaves nothing to he desired. It has been run continuously at the rate of 3,000 revolutions per minute, without perceptible jarring. The machine was at first supplied with a mechanical counter, which could be thrown on or off by means of an electro-magnet operated by a key. We found, however, that the arrangement was uusatisfac- Fj&,3 Field Mac.net and Armature tory, as variations of speed were introduced by its friction and inertia. Moreover, such an arrangement gives only the average speed during a given time ; we therefore abandoned this counter, and adopted a stroboscopic device. A ray of light is reflected upon the revolving segmented disk from a mirror attached to an electrically driven tuning-fork of known rate. With such an arrangement the desired speed may be attained, and variations from it measured by noting the apparent motion of the disk. 266 PROCEEDINGS OF THE AMERICAN ACADEMY. The first test to which the apparatus was put was to obtain the judgment of a person having a trained ear as to the quality of the tone produced by its current in a telephone. His opinion was that the tone was pure, only the fundamental being detected. The wave form was then mapped. In planning this work it occurred to us to employ the method of projection of potential whenever it was necessary to measure a P. D. 9."*1" 4 •^-— • — ^^ J Yi \\ J \V > li \V . 11 II II II r II Jl II \\ \\ \\ .t II ji It 1 \\ \\ \\ .j li \ - II jl It \\ \\ \\ 2. j \ / f i ■ ■ ...i _ ■ 1 . \ 3.o' to' io" to' iou' no- w Fic.hAKMf.ruRe A . Tiive Sine Cume Dotted it,o tic1 or E. M. F. at any instant. The opposing P. D. was obtained by the fall of potential due to the passage of a direct current through a suit- able resistance. As the opposing P. D. was not alternating, the above mentioned contact arrangement was used in place of the ordi- nary key, and its value was determined by a suitable Weston volt- meter. This method of dealing with instantaneous E. M. F.'s was originally suggested by Professors Foster and Maxwell, their object LAWS. — COEFFICIENTS OF SELF-INDUCTION. 2G7 being to obtain a balance between a known E. M. F. and the P. D. between the terminals of a resistance through which a current of known absolute magnitude was flowing.* In order to obtain sensitiveness at the final balancing, a condenser one microfarad in capacity was inserted in shunt with the galvanometer, the discharge of the condenser serving to increase the deflection. The galvanometer was of 3,000 ohms' resistance. A telephone was "Tin ho'' to*' $0. /oo lie ihq' FigS Armatur'b B .True Sine Curve Dotted. /to* /to' used to obtain the preliminary balance. In Table I. are given the results of the determination of the wave forms of both dynamos, one section of 10 wires being used. In the case of armature A (movable) a series of checks is appended. The whole series was gone through with and then repeated in the reverse direction, the person making * Telegraphic Journal, 1874, II. 317. 268 PROCEEDINGS OF THE AMERICAN ACADEMY. TABLE I. — Determination of Wave Form. Armature A. Ampere Turns, 2,875. — Speed, 1,465. Armature B. Ampere Turns, 2,150. Revs, of Mic. Head. E. M F. in Volts. Check. Position of Contact. E. M. F. in Volts. 0 +0.49 +0.48 o 11 +1.00 * —0.91 —1.20 10 —0.26 1 —2.52 —2.53 9 —1.72 H —4.01 —4.01 8 —2.80 2 —5.21 —5.19 7 —3.86 2* —6.38 —6.36 6 —4.98 3 —7.09 —7.21 5 —5.93 8* —8.04 —8.03 4 —6.75 4 —8.61 —8.61 3 —7.34 4i —9.01 —8.96 2 —7.86 5 —9.11 —9.12 1 —8.30 *>h —9.12 ' —9.12 360 —8.50 6 —8.95 —8.94 359 —8.56 6* —8.52 —8.54 358 —8.54 7 —8.00 —7.95 357 —8.30 n —7.09 —7.08 356 —7.91 8 —6.11 —6.12 355 —7.26 8* —4.80 —4.80 354 —6.62 9 —3.53 —3.48 353 —5.86 9£ —1.94 —1.96 352 —4.93 10 —0.60 —0.60 351 —3.85 10* +1.12 — 350 —2.67 — — — 349 —1.60 — — — 348 —0.32 — — — 347 +0.74 LAWS. — COEFFICIENTS OF SELF-INDUCTION. 269 the readings not knowing the previous results. The most probable sources of deviation are backlash, variations of field current, and change of speed. The checks show sufficient precision of the appa- ratus for the method of procedure to be described presently. These results are plotted together with true sine curves in Figures 4 and 5. The curves proved to be very nearly symmetrical, so that only one half has been given. The saturation curve connecting maxi- mum E. M. F. and field current was taken with armature A. The results are given below, and are plotted in Figure 6. ^fottfc ♦ .6 } t 1 x i » ■ » ■ ■ .ric.C Saturation CvRve i m ib TABLE II. — Data of Saturation Curve. Speed, 1,465. Field Current in Amperes. E. M. F. in Volts. Field Current in Amperes. E. M. F. in Volts. 3.40 5.56 8.06 8.10 4.12 6.16 9.38 8.52 4.93 6.74 11.12 9.23 6.10 7.40 14.40 10.10 7.08 7.70 15.00 10.18 270 PROCEEDINGS OF THE AMERICAN ACADEMY. Methods of Using the Machine. The following methods are all based on the assumption that we have a sinusoidal E. M. F. of known period, acting in a circuit con- taining only true resistances and inductances. The current at any instant will be expressed by the equation i = = sin I Pi — tan~ l — - • VR2 + P'L2 \ R) The maximum value of the sine term being unity, we have 1 _ 1 E ~ V-R'2 + R2L2' I and E being the maximum current in, and maximum E. M. F. im- pressed upon, the part of the circuit under discussion. Also the lag LP between current and E. M. F. is expressed by tan 6 = -^5- . From R these two equations we have E . E L = y~ sin 6, and R = — cos 6. Our aim is, therefore, to obtain the values of the impedance, angle of lag, and frequency, and to calculate the coefficient of self-induction and the resistance for any given periodicity by the above formulas. (a) If we make up a circuit of the movable armature, a known resistance of negligible inductance, and the unknown coil, all in series, we may, by means of the measuring arrangement already mentioned, obtain E and /, the latter by the maximum P. D. around the known resistance. The angle 6 would best be found by taking measurements on both sides of the zeros of current and E. M. F., and interpolating for the zero points. The difference of readings of the micrometer so obtained divided by 20 gives the angle 6. L and R are to be calcu- lated by the above formulae. (b) If we have a carefully made plot, like Figure 6, connecting the maximum E. M. F. and field current of one dynamo, we may proceed as follows. As the wave forms are alike, both following the sinu- soidal law, we may apply Poggendorf's method of balancing E. M. F.'s directly, using a suitable telephone instead of a galvanometer. To do this make up a circuit of the stationary armature, known inductionless resistance, and unknown coil. Keep the field current of this dynamo constant. Attach the second armature through the telephone, first to the terminals of the known resistance. Adjust the phase relation for LAWS. — COEFFICIENTS OF SELF-INDUCTION. 271 a minimum of sound in the telephone; then adjust for silence by vary- ing the field current. Read the micrometer head and value of exciting current. Repeat this process at the terminals of the unknown coil. These measurements furnish the data necessary for the calculation of L and R. 6 is given by difference of micrometer readings divided by 20, as before. Owing to lack of exact similarity of E. M. F. waves it would be necessary in practice to adjust by a minimum, annulling the fundamental. £ OQ — * ==*=QTeii V"'V<*» J Fig. 7. (c) The trouble of taking the preliminary curve, as well as any uncertainty which its use may introduce, can be avoided by the con- nections shown in Figure 7. This method differs from the preceding only in the use of a derived E. M. F. for balancing, and in the method of its adjustment. For con- venience give Rt such a value that the P. D.'s around Rx and R shall be about equal. Adjust the phase of Armature A by the tangent screw, and adjust for amplitude by varying the field current, final adjustment being obtained by the slider. Now change the connection to the terminals of the unknown coil, and again adjust for silence, making the adjustment for amplitude in this case by the slider alone. If the two values of the slider resist- T 7? ance are R* and R3, then — = ^ . 6 is given by the micrometer Jb r±r% head as before. Fig. 8. (c?) "We may use a differential telephone, that is, a telephone hav- ing two opposed windings of equal influence and resistance, in the manner shown in Figure 8. 272 PROCEEDINGS OP THE AMERICAN ACADEMY. i?T is a resistance equal to that of one coil of the telephone and of negligible inductance. Adjustment is to be made until with the same value of tA the telephone is silent when in place, as shown, and E when interchanged with ET then Rx = -=. 6 is given by micrometer readings as before. (e) We may also use the differential telephone to indicate when an unknown inductance is neutralized by a condenser of known variable capacity, by placing one of its coils in series with the inductance and condenser, and the other in series with a variable inductionless resist- ance i?l5 shunted around the above condenser, inductance, and tele- phone coil, if the resistance E1 and capacity C be varied until silence is obtained. Then L = -tt— and E = i?i. It will be desirable Or- to employ an air-condenser to eliminate absorption effects, though this procedure has the disadvantage of requiring a very bulky and costly condenser when small inductances are to be measured. With this apparatus we may also determine the electrostatic ca- pacity of a condenser when not complicated by absorption effects. For instance, by the first method, we should insert in series with the condenser an inductionless resistance of such a magnitude that the angle of advance could be determined with accuracy. Then I 1 1 E~ J i tan * = CEP' 7 V C = -FPn—- — r ' and R — 1 cos V'- EP sin ^ I HE = 0, * = 90° and 0 = -^- The above are the chief uses to which the apparatus may be put. We hope in the near future to investigate the transmission of tele- phonic and telegraphic signals, and to study various small induction coils such as are used in telephony, as well as the errors introduced into results by the above methods owing to lack of fulfilment of the assumed law by the E. M. F. Rogers Laboratory of Physics, September, 1893. ROBINSON. — ALSINEiE. 273 i XIV. CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD UNIVERSITY, NEW SERIES, NO. VI. By B. L. Robinsox. Presented January 10, 1894. I. — THE NORTH AMERICAN ALSINEIE. The following provisional treatment of the North American Alsinece is based principally upon the material in the Gray Herbarium. An effort, however, has been made to see as large a representation of the group as possible, and with this end in view a number of other large and valuable collections have been visited ; namely, the herbaria of the Department of Agriculture, Columbia College, the Missouri Botanical Garden, the Philadelphia Academy of Natural Sciences, the Boston Natural History Society, and the private collections of Mr. William M. Canby of Wilmington, Del., Mr. John Donnell Smith of Baltimore, and Mr. Walter Deane and Mr. Edward L. Rand of Cambridge, Mass. Those in charge of all these collections have most kindly allowed their Caryophyllacece to be examined, and all such specimens as showed any noteworthy peculiarities to be brought to Cambridge for critical study. From the herbaria of the Canadian Geological Survey, Iowa Agricultural College, and Natural History Society of Wilmington, Del., the entire representation of Alsinece has also been most obligingly forwarded to the Gray Herbarium for examination. While the study of such copious materials has obvious advan- tages, it by no means lessens, but rather increases, the difficulties of classification. In a long series of specimens individual variation becomes apparent in a way which can scarcely be realized if study is confined to fewer plants. Specific limitations often become very obscure, and sharp definition almost or quite impossible. Several instances of such practically confluent species occur in the present group, and in adjusting specific lines there is consequently great room for difference of opinion. Free criticism and friendly aid will be very vol. xxix. (n. s. xxi.) 18 274 PROCEEDINGS OF THE AMERICAN ACADEMY. welcome, as much may thus be contributed to fuluess and accu- racy in the final treatment of this group for the " Synoptical Flora." To the botanists, both professional and amateur, who have most kindly assisted the author in his work upon the present paper, he would express sincere appreciation and cordial thanks. To the following he is especially indebted : Professors William Trelease and N. L. Britton, Mr. J. H. Redfield, Mr. F. V. Coville, Dr. J. N. Rose, Mrs. T. S. Braudegee, Miss M. E. Carter, Professor J. Macoun and Mr. J. M. Macoun, Professors T. C. Porter, L. H. Pammel, W. W. Bailey, and O. D. Allen, Messrs. William M. Canby, John Donnell Smith, Walter Deane, Edward L. Rand, Theodor Holm, and J. F. Collins. In the study of Sjjergularia an extensive series of Californian forms of that difficult genus, contributed to the Gray Herbarium by Mrs. T. S. Braudegee, has proved invaluable. In matters of synonymy and the citation of literature Dr. Watson's Bibliographical Index has again been a most useful guide. CARYOPHYLLACE^E, Tribe I. SILENE^E. Including gen- era 1-7. (See Proc. Am. Acad, xxviii. 124.) Tribe II. ALSINEiE. Sepals free or slightly united at the very base. Petals more or less contracted but not uuguiculate below. Corona absent. Flowers mostly small. Styles distinct to the base. * Stipules none. h- Capsule cylindric, more or less elongated, often curved, dehiscent by twice as many teeth as there are carpels. 8. Holosteum. Sepals 5. Petals 5, white, subentire or denticu- late toward the apex. Stamens 3-5, very rarely 10. Styles 3 (occa- sionally 4 or 5), longitudinally stigmatic. Pod unicellular. Seeds numerous, dorsally flattened, i. e. parallel with the incumbent cotyle- dons ; the radicle prominent upon the ventral surface. Inflorescence umbelliform. 9. Cerastium. Sepals in our species 5. Petals as many, retuse or bifid, very rarely subentire, white. Stamens 10, or sometimes fewer. Styles 5 (4 or 3). Capsule usually exceeding the calyx, often curved. Seeds numerous, more or less laterally compressed. -*- ■>- Capsule ovoid or oblong relatively short, dehiscent by as many or twice as many teeth as there are carpels. -w. Styles usually fewer than the sepals, when of the same number opposite them. 10. Stellaria. Sepals 5 (-4). Petals 5 (-4, rarely abortive or absent), always more or less deeply bifid, often divided almost to the base, white. Stamens 3-10. Styles 3-4, rarely 5. ROBINSON. — ALSINE^E. 275 11. Arenaria. Sepals 5. Petals as many, white or nearly so, entire or emarginate (very rarely minute or wanting). Stamens 10, or often fewer by abortion. Styles 3 or 4. Seeds many. ++ -M- Styles as many as the sepals and alternate with them. 12. Sagina. Sepals 5, (rarely 4). Petals as many, entire or emarginate, white, rarely obsolete. Stamens usually 5, less frequently 3-10. Valves of the capsule as many as the sepals, and opposite them. Seeds several to many. * * Stipules present, scarious : petals undivided. 13. Spergularia. Sepals 5. Petals 5 (rarely fewer or none), reddish or white. Stamens commonly 10. Styles 3 (very rarely 5) ; ovary 1-celled ; valves of the capsule as many as the styles, when 5 in number alternate with the sepals. Seeds often margined. Leaves linear or filiform. 14. Spergula. Sepals 5. Petals 5. Stamens 10 (rarely 5). Styles 5 ; ovary unilocular, many-ovuled. Valves of the capsule 5, opposite the sepals. Seeds acutely margined or narrowly winged. Leaves narrow, linear, verticillate and fascicled in the axils. Tribe III. POLYCARPE^E. Including genera 15-18. (See Proc. Am. Acad, xxviii. 126.) 8. HOLOSTEUM, L. (6'Aos, all, and 6ar iov, bone ; used iron- ically, since the plants are soft and weak.) — A small genus of Old World annuals and biennials much resembling Cerastium except in inflorescence and seeds. The commonest species is adventive in America. — Gen. no. 928; Reichb. Icon. Fl. Germ. v. t. 221; Gay, Ann. Sci. Nat. ser. 3, iv. 23 ; Benth. & Hook. Gen. i. 148. H. umbellatdm, L. Finely glandular-pubescent, somewhat glau- cous : stems 3-18 inches high : leaves sessile, ovate-oblong : umbels 3-12-flowered, terminal upon long naked peduncles; pedicels 8-12 lines long, some of them reflexed : filaments shorter than the calyx. — Spec. 88 ; Eng. Bot. i. 27. Locally naturalized in Pennsylvania, New Jersey, and Delaware, Porter, Austin, Canby, Small. (Adv. from Eur.) 9. CERASTIUM, L. Mouse-ear Chickweed. (iclpas, a horn, from the elongated curved capsules.) — Annuals or perennials, usually pubescent and often viscid. Leaves usually flat. Flowers white, borne in more or less expanded leafy or naked cymes. A genus distinguished from Stellaria and Arenaria somewhat by habit, but chiefly, although not always satisfactorily, by the form and dehiscence 276 PROCEEDINGS OF THE AMERICAN ACADEMY. of the capsule. — Gen. no. 376; Seringe in DC. Prodr. i. 414; Grenier, Flora, 1840, pt. 1, 266 ; Reichb. Icon. Fl. Germ, v.-vi. t. 228-236 ; Benth. & Hook. Gen. i. 148 ; Pax in Engl. & Prantl, Nat. Pflauzenf. iii. 1 b, 80. § 1. Strephodon, Seringe, I. c. Styles 3-5 : teeth of the cap- sule finally circinate-revolute from the tip. — Our species have pubes- cent leaves. C. Tesanum, Britton. Annual, viscid : stems several, slender, nearly erect, leafy below, nearly naked and dichotomous above : leaves oblanceolate or spatulate, 6 lines to 2 inches in length, very pubescent or subcinereous on both surfaces : flowers rather small : petals bifid : styles 3-4 (-5 ?) : capsule H-2 times the length of the calyx. — Bull. Torr. Club, xv. 97. — Hills, Blanco, Texas, Wright; New Mexico on the Mongollons, Greene ; Arizona, Santa Cataline Mts., Lemmon. (Mex., Palmer ; Lower Calif., Brandegee.) C. maximum, L. Stoloniferous perennial with stems simple or nearly so, erect or decumbent, becoming a foot or more in height : leaves linear or lanceolate, attenuate : flowers very large for the genus, 1 inch in diameter, borne on erect pedicels in simple or branched cymes : sepals oblong or narrowly ovate, obtuse, 3-4 lines long : petals obovate, much exceeding the calyx, deeply notched at the apex : capsule symmetrical, much exserted at maturity. — Spec. 439 ; Ledeb. Icon. Fl. Ross. t. 242 ; Fenzl in Ledeb. Fl. Ross. i. 399 ; Seem. Bot. Herald, 51. C. grande, Greene, Pitt. ii. 229. — Alaska. (Siberia.) Asiatic specimens of this species, identified at the St. Petersburg Gardens, show that the capsule becomes cylindric and much longer than figured by Ledebour. § 2. Orthodox, Seringe. Styles normally 5 : teeth of capsule erect or spreading; the edges sometimes slightly reflexed. — DC. Prodr. i. 415. — Our species have pubescent leaves. * Flowers comparatively small : petals 1-1 \ times as long as the sepals. ■t- Pods 1-lf times as long as the calyx : introduced or doubtfully indigenous weeds. C. viscosum, L. (Mouse-ear Chickweed.) Annual, viscid pu- bescent, 3 inches to a span high : leaves oval or elliptic-oblong, very obtuse ; the lowest narrowed below to a short margined petiole : flowers small, at first densely clustered at the ends of the branches, becoming laxer in fruit, but even the longest pedicels not exceeding the acute sepals (lf-2 lines in length): bracts herbaceous: petals scarcely equalling the calyx: stamens frequently 5. — Spec. 437; Hook. f. Arc. PI. 288 ; Wats. Bibl. Index, 101 ; Wats. & Coulter in ROBINSON. — ALSINEJ3. 277 Gray, Man. ed. 6, 88. G. vukjatum, Linn, in herb. ; Torr. & Gray, Fl. i. 187 ; Gray, Man. eds. 1-5 ; and others. G hirsutum, Muhl. Cat. 46. G. ylomeratum, Thuill. as used by Hooker, f. and others. G. connatum, Beck, JBot. 55. Depauperate forms with few flowers and short capsule have been regarded as indigenous, being the G. viscosum, var. tenellum, Grenier, 1. c. 266, and the G. semidecan- drum^ auct. (not of Linn.). — Widely distributed in the United States and Canada, but much less common than the following; probably introduced from Europe. Delicate specimens apparently to be referred to this species, but with minute apetalous flowers, have beeu collected at San Diego, Calif., Orcutt. C. vulgatum, L. (Common Mouse-ear Chick weed.) Peren- nial, viscid-pubescent, a little taller and more spreading than the last : leaves oblong, obtusely pointed : flowers larger : the lower pedicels in fruit considerably exceeding the calyx : bracts herbaceous : sepals 2-3 Jines long, obtuse, often purple-tipped, appearing acute through the infolding of the scarious margins : petals as long as the calyx. — Spec. erl. 2, 627 ; Regel, Ost-Sib. i. 432 ; Wats. Bibl. Index, 101 ; Wats. & Coulter in Gray, Man. ed. 6, 88. G. viscosum, Linn, in herb. ; Torr. & Gray, Fl. i. 187 ; Gray, Man. eds. 1-5, etc. G fulvum, Raf. Prec. Decouv. 36. G. triviale, Link, Enum. Hort. Berol. i. 433. — Very common in fields, etc., but also often remote from habitations and cultivated ground, thus perhaps native. Flowering through the summer. C. semidecandrum, L. Near the two preceding, but smaller and with shorter leaves : the bracts, at least the upper ones, conspicuously scarious-margined : pedicels in fruit longer than the calyx. — Spec. 438; G vulgatum, var. ? semidecandrum, Gray, Man. ed. 5, 94. — New Jersey, Britlon, Peters, to Norfolk, Va., Britton, Small. (Ad- ventive from Europe.) -t- -t- Pods 2-3 times as long as the calyx : indigenous species. C. brachypodum. Pale green, finely pubescent and sometimes very viscid : leaves linear-oblong to oblanceolate, obtusish, seldom more than an inch in length : flowers in more or less open dichotomous cymes ; pedicels, even the lower ones, only equalling or little exceeding the capsules, erect or deflexed, straight or gently curved, not hooked. — G. nutans, Raf., var. brachypodum, Engelm. in herb. — St. Louis, Mo., EiKjelmann, April, 1842, April-May, 1845; westward and southward to Nevada, Anderson, 238, Watson, 156; Arizona, Palmer; New Mexico, Fendler, 61; Texas, Wright. (Mex. Schaffncr, Palmer.) G. tenellum, Fenzl, mentioned in Watson's Index (but never pub- 278 PROCEEDINGS OF THE AMERICAN ACADEMY. lished ?), represented by Druinmond's no. 30 of his 3d Texan Coll., appears to be only a more slender form of the above. Exactly the same thing, however, has been found at Milledgeville, Ga., by Dr. Boy kin (Short Herbarium), thus considerably extending the range of the species. In raising Dr. Eugelmanu's variety to specific rank I follow the suggestion of Dr. N. L. Britton. Certainly, the more restricted geographic range and the absence of connecting forms indi- cate its distinctness from G. nutans. A very leafy and velvety tomen- tose form from Willow Spring, Arizona, Palmer, is worthy of mention. Var. compactum. Inflorescence capitate-umbellate : pods very slender. — C nutans, Raf., var. compactum, Eugelm. in herb. — A marked variety or form from the Bad Lands of Nebraska, Hayden ; Belknap, N. Texas, Hayes ; Ealse Washita, Ind. Terr., Palmer, 18G8. C. nutans, Raf. A pubescent and viscid annual, 8-18 inches high : stems branched : leaves oblong-lanceolate, acute ; the lowest narrowed toward the base : flowers numerous in an open dichotomous cyme : calyx about 2 lines in length : petals somewhat exserted, oblanceolate, bifid : pedicels elongated, ascending or spreading, tending to be hooked or nodding at the summit : capsule 4-6 lines long, nod- ding but curved upward. — Prec. Decouv. 36, & Desv. Journ. Bot. iv. 269 (1814) ; Gray, Gen. ii. t. 114. C. longepedunculatum, Muhl. Cat. 46. C. glutinosum, Nutt. Gen. i. 291. C. apricum, Schlecht. Linnaea, xii. 208. C. oblongifolium, Anderson, Cat. PI. Nev. 118. — Common and widely distributed from New England to the Pacific and from Hud- son Bay to New Mexico. Like the last, paler green than the other common species. Apetalous specimens have been found at Wawa, Penn., Brinton. Arizona forms of this species also differ slightly in habit, but lack technical characters for satisfactory distinction. C. sericeum, Wats. Annual : stems one or many, 1-2 feet high, stout for the genus, sericeous, very leafy below : leaves oblong- lanceolate, sessile, 1-2 inches long, 3-6 lines broad ; the lower cinere- ous with dense flocculent wool ; the upper green : flowers numerous in spreading cymes : calyx 2k lines long, scarcely exceeded by the corolla. — Proc. Am. Acad. xx. 354. — S. Arizona in the Huachuca Mts., Lemmon; Santa Rita Mts., Pringle. * * Flowers large ; petals usually twiee as long as the calyx : indigenous species. C. arvense, L. Perennial: stems several, weak, usually almost naked above : leaves linear to narrowly lance-oblong: petals obcor- ROBINSON. — ALSINE.E. 279 date : pod in the typical form scarcely longer than the calyx. — Spec. 438; Hook. Fl. Bor.-Am. i. 104; Torr. & Gray, Fl. i. 188; Ilollick & Britton, Bull. Torr. Club, xiv. 45. ?G. hybridum, Muhl. Trans. Am. Phil. Soc. ser. 1, iii. 170. G Pennsylvanicuui, Hornem. Hart. Hafn. 435. — Rocky soil, common; May-July. (Eu., Asia, & S. Amer.) Very variable in size, pubescence, relative length of its capsules, etc. Yar. angustifoliuji, Fenzl, I.e. i. 413, with cauline leaves narrowly oblono" to linear, attenuate at the base, much fascicled and 9-15 lines in length, and var. latifolium, Fenzl, 1. c. i. 412, with shorter oblong leaves, 6-8 lines long, broad at the base, are forms strikingly different in their extremes, but rather freely intergrading and often difficult to distin o-uish. The latter is perhaps a little more common in the Rocky Mountains but extends eastward to Labrador. Better marked are the following. Var. oblongifolium, Hollick & Britton. Leaves oblong or lance-oblong, obtuse or obtusish : capsule longer, H-2| times as long as the calyx. — Bull. Torr. Club, xiv. 47, t. lxiii. G. oblon- gifolium, Torr. Fl. U. S. 460; Torr. & Gray, Fl. i. 188. ? G. di- chotomum, Muhl. Cat. 46. ? G. bracteatum, Raf. Prec. Decouv. 36. — Nova Scotia to Virginia, and Montana, Scribner, to New Mex- ico, Vasey. This variety has been widely drawn by its authors to in- clude narrow-leaved forms as well as the original rather broad-leaved C. oblongifolium, extended series of specimens showing complete tran- sitions. Var. maximum, Hollick & Britton, 1. c. xiv. 47. Taller, 1-2 feet high : leaves elongated, lanceolate, acutish, 2-3 lines bread : inflorescence very spreading : capsule equalling or half exceeding the calyx. — G. oblongifolium, Torr. Pacif. R. Rep. iv. 70. G. pilosum, Brew. & Wats., Bot. Calif, i. 67, not of Ledeb. — Dixon, 111., Vasey, and in California, Punta de los Reyes, Bigelow, and elsewhere. Scarcely more than a rank growing form, but serving to connect the next species through var. Fischerianum. Var. villosum, Hollick & Britton, 1. c. xiv. 49. Densely villous: leaves rather broadly lanceolate. — G. velutinum, Raf. Med. Rep. (hex. 2), v. 359. G. villosim, Muhl. Cat. 46. Darlingt. Fl. Cest. ed. 2. 279. ? C. hirsutum, Darlingt. Florula Cest. 54. G. oblongifolium, Torr. & Gray, Fl. i. 188 in part; Wats. Bibl. Index, 101. — Lancaster Co., Penn., Porte?'. Var. Fuegianum, Hook. f. Depauperate, 2-3 inches high, with short thickish imbricated leaves and sub-solitary terminal flowers. — U. S. Expl. Exped. 119. — Specimens collected by Coulter in the 280 PROCEEDINGS OF THE AMERICAN ACADEMY. Yellowstone Park have been confidently referred to this variety by Hollick and Britton, 1. c, and Parry's no. 41 from Northwestern Wyoming is doubtless the same. C. alpinum, L. Densely silky : stems weak, matted : leaves elliptic-ovate, in the typical form only 4—5 lines long : petals notched at the apex, 1^-2 times the length of the sepals. — Spec. 438 ; Torr. & Gray, Fl. i. 188 ; Regel, Ost-Sib. i. 433, 434. G. lanatum, Lam. Encycl. i. 680. C. latifolium, Greville, Mem. Soc. Wern. iii. 429. G. vulgatum, Hook. f. Arc. PI. 288 in part. ? C. latifolium, Hart., Trimen's Journ. of Bot. ix. 205. — Arctic America from Greenland to Alaska, also in Labrador, the Hudson Bay region, and upon the Rocky Mountains of British America. (Europe and Asia.) The following varieties extend farther southward. Var. Beeringianum, Regel. Hirsute and less silky-villous, somewhat viscid above : leaves smaller, oblong. — Ost-Sib. i. 435. G. Beeringianum, Cham. & Schlecht., Linna^a, i. 62. G. vulgatum, var. Beeringianum, Fenzl in Ledeb. Fl. Ross. i. 409. — Alaska to the Rocky Mountains of Colorado and Arizona. Var. Fischerianum, Tour. & Gray. Hirsute, taller, 8-10 inches or even more than a foot in height : leaves rather thick, elliptic- lanceolate or oval-lanceolate, acute or acutish, an inch or more in length: capsule H-2 (or rarely 3) times the length of the calyx. — Fl. i. 188; Regel, 1. c. i. 438. G rigidum, Ledeb. Mem. Petr. v. 538. G. Fischerianum, Seriuge in DC. Prodr. i. 419. G. vulga- tum, vars. grandiflorum and macrocarpum, Fenzl in Ledeb. Fl. Ross, i. 409, 410. To judge from the figure in the Caiques des Dessins G. stellar ioides, M09. should be referred here also, having been placed by Seringe probably through error in § Strephodon. — A stout variety passing to G. arvense, var. maximum, but with broader more elliptic- ovate leaves and longer capsules. Alaska to Humboldt Co., Calif., Rattan. (Siberia, Japan.) The leaves are thicker and the sepals more pubescent and acute than in G. pilosum, Ledeb., to which it is also nearly related. Var. glabratum, Hook. Leaves and calyx nearly smooth. — Parry's 2d Voy. 390; Fl. Bor.-Am. i. 104. — Arctic America with the pubescent forms. (N. Eur.) §3. Dichodon, Bartl. Styles normally 3: teeth of the capsule erect or slightly spreading, not circinate-revolute. — Endl. Gen. 970. — Our species with symmetrical capsule and short glabrous leaves. C. trigynum, Vill. Perennial, with stems weak, spreading, some- what matted, smooth or glandular-pubescent, loosely 2-3 flowered: ROBINSON. — ALSINE.E. 281 leaves oblong, 3-5 (-8) lines in length ; the uppermost ovate : sepals lance-ovate or oblong, obtuse, 2-3 lines long: petals 1^-2 times the length of the calyx, obcordate, bifid nearly half way to the base : capsule oblong-conic, twice the length of the calyx ; the teeth finally spreading. — Dauph. iii. 645, t. 46 ; Fenzl in Ledeb. Fl. Ross. i. 396. Stellaria cerastoides, L. Spec. 422 • Torr. & Gray, Fl. i. 184 ; Hook. f. Arc. PI. 288. — Table-topped Mountain, Gaspe, Lower Canada, Allen; Cape Chudleigh, Hudson Strait, Bell; Labrador, Greenland, Holm. (Europe and Siberia.) A species now generally appended to Ceras- tium, but forming a transition to Stellaria. Mcsnchia quaternella, Ehrh. (Sagina erecta, L., Cerastium quaternellum, Fenzl.) An erect glaucous annual with subsimple stems, 2-3 inches high, bearing 1 or 2 erect 4-parted flowers, was found in the thirties near Baltimore by B. D. Greene. It is said to have been recently rediscovered there, but nothing more definite has been learned concerning its American occurrence. (Europe.) 10. STELLARIA, L. Chick weed, Starwort. (Stella, a star, in reference to the form of the flower.) — Low spreading herbs, sometimes a little succulent, mostly preferring a moist shaded habitat. Leaves flat or very rarely acerose. A genus conveniently but some- what artificially separated from Arenaria by the more or less deeply cleft petals. — Spec. 421, & Gen. ed. 5, no. 504; Seringe in DC Prodr. i. 396; Fenzl in Endl. Gen. 969 ; Reichb. Icon. Fl. Germ. v. t. 222-226; Benth. & Hook. Gen. i. 149; Gray, Gen. ii. t. 113; Pax in Engl. & Prantl, Nat. Pflanzenf. iii. 1 b, 79. Spergulastrum, Michx. Fl. i. 275. Micropetalon, Pers. Syn. i. 509. Larbrea, St. Hil. Mem. Mus. Par. ii. 287. § 1. Myosoton, Monch (as genus). Styles 5, alternate with the sepals : leaves ovate acute. — Method. 225. Malachia, Fries, Fl. Hall. 77. S. aquatica, Scop. Perennial, with stem strongly angled and somewhat pubescent : leaves large, ovate, or ovate-lanceolate, acute ; the upper sessile, cordate ; the lower petiolate : pedicels glandular- viscid, deflexed in fruit: petals 1^ — 2 times as long as the campanulate glandular-pubescent calyx : seeds numerous, dark-colored, tuberculately roughened. — Fl. Carn. ed. 2, i. 319. Malachia aquatica, Fries, Fl. Hall. 77. Larbrea aquatica, Seringe in DC Prodr. i. 395 (excl. syuon.). — Becoming frequent upon waste land and public grounds in the Eastern States, and more or less established along roadsides in British America, Stratford, Ont., Burgess; Nanaimo, B. C, Macoun. (Ad- ventive from Europe.) 282 PROCEEDINGS OF THE AMERICAN ACADEMY. § 2. Eustellaria, Fenzl, 1. c. 969. Styles 3-4. * Petals, except in some flowers of S. pubera, very deeply 2-parted (sometimes minute or wanting) : segments narrow. -)- Lower leaves ovate, rather abruptly contracted into slender petioles. S. media, Smith. (Common Chickweed.) A low annual : stem pubescent in lines : leaves acute ; the upper narrower sessile, the lower on pubescent narrowly margined petioles : calyx glandular- pubescent, equalled or slightly exceeded by the capsule : petals shorter than the sepals : stamens 3, 5, or 10. — Eug. Bot. viii. t. 537. Alsine media, L. Spec. 272; Walt. Car. 117. Holosteum succulenlum, L. Amoen. iii. 21 ; Nutt. Gen. i. 89 ; Torr. Fl. U. S. 159. — One of the commonest weeds in dooryards and cultivated grounds, especially in moist soil, flowering from early spring to late autumn. S. prostrata, Baldw. Annual : stems weak, elongated, prostrate, pubescent : leaves ovate, acute or shortly acuminate, the lower sub- cordate on slender ciliated petioles ; the upper cauline short-petioled or sub-sessile ; the floral reduced and bractlike : pedicels filiform : flowers smaller than in the preceding : sepals in anthesis but a line long : petals nearly twice as long : mature capsule much exceeding the calyx ; valves distinctly circinate-revolute. — Baldw. in Ell. Sk. i. 518; Torr. & Gray, Fl. i. 183; Gray, PI. Lindh. ii. 152, & PI. Wright, ii. 17; Chapm. Fl. 50. — Moist and shaded places, rocky woods ; Georgia and Florida to Texas. (Adj. Mex.) Leaves very variable in size, from 2 lines to an inch in length. The flowers in this species are distinctly smaller than in the nearly related Mexican S. cuspidata, Willd., and S. ovata, Willd. S. nitens, Nutt. Annual, slender, erect, shining: stems filiform, forked several times, leafy and slightly pubescent near the base, almost naked and quite glabrous above : leaves of two forms, the lowest (1-3 pairs) ovate, acute, only 2 lines long, on slender petioles of somewhat greater length, not always persisting; the other leaves lance-linear acute, 3-5 lines long : sepals very acute, scarious-margined, 1-3 nerved : petals half as long as the sepals, sometimes absent: capsule oblong., about equalling the calyx. — Torr. & Gray, Fl. i. 185; Torr. Pacif. R. Rep. iv. G9, & Bot. Mex. Bound. 37 ; Gray, Proc. Am. Acad. viii. 378. S. mamchioides, Fenzl, ace. to Torr. & Gray, Fl. i. 675. S. stricta, Hook. Fl. Bor.-Am. i. 96 in part. — S. California to Brit. Columbia, Macoun ; eastward to Utah, Jones ; April, May. S. graminea, L., with seldom persistent but sometimes slightly petiolate lower leaves, may possibly be sought here. ROBINSON. — ALS1NE.2B. 283 -t- h- Leaves acerose. S. Kingii, "Wats. Cespitose, minutely glandular-pubescent above: root woody : stems several, 2-7 incbes high: leaves crowded below, rigid, pungent, light green : sepals spreading or somewhat reflexed during anthesis, scarious-margined, exceeded by the narrow petals. — Bot. King Exp. 30, t. vi. f. 1-3 ; Brew. & Wats. Bot. Calif, i. 68. — Humboldt Mts., N.Nevada, Watson; S. Utah, Parry, Palmer ; July and August. •*- 4- h- Leaves all sessile or subsessile, sometimes narrow, but not acerose. ++ Bracts small, scarious. = Flowers small : petals minute or none. S. umbellata, Turcz. Smooth: stems weak, ascending from a decumbent rooting base : leaves varying from lanceolate and acute to elliptic-oblong, 3-S lines in length : pedicels filiform, sub-umbellately grouped at the ends of the branches, often deflexed : sepals small, 1—14, lines in length, glabrous, scarious-margined: capsule twice as long; the valves deeply 2-toothed ; teeth obtuse. — Cat. Baic. 5, & Fl. Baic-Dahur. i. 236 ; Fenzl in Ledeb. Fl. Ross. i. 394 ; Kegel, Ost-Sib. i. 383, 399 ; Wats. Bot. King Exp. 38 ; Porter & Coulter, Fl. Col. 13; Brew. & Wats. Bot. Calif, i. 67. S. borealis, var. Hook. Fl. Bor.-Am. i. 94. — Mountains of Colorado and Arizona to Union Co., Oregon, Cusick. (Asia.) = = Flowers of medium size : petals equalling or exceeding the calyx (except sometimes in -S. uliginosa). a. Seeds essentially smooth. S. longifolia, Mdhl. Stems sharply 4-angled, commonly 8 inches or more in height : leaves linear or linear-oblong, somewhat narrowed at each end, thickish, often ciliate toward the base ; the larger ones 1]-H inches long: flowers rather numerous in a lateral long-pedun- cled open cyme ; pedicels spreading, horizontal or deflexed : petals and capsule exceeding the sepals : seeds smooth. — Cat. 45 ; Willd. Enum. 479; Fenzl, 1. c. i. 392; Gray, Gen. ii. 38, t. 113, f. 1-5. S. graminea, Bigel. Fl. Bost. 110. Spergulastrum gramineum, Michx. Fl. i. 276. Micropetalon gramineum, Pers. Syn. i. 509. M. longifo- lia, Eat. & Wright, N. A. Bot. ed. 8, 319. — Canada to Maryland and westward to the Rocky Mts. ; June and July. (Europe and Asia.) S. longipes, Goldie. Smooth and shining or more or less glaucous, spreading at the base: branches erect, 3-12 inches high: leaves linear or lance-linear, gradually narrowed from the base to the acute apex, 1-nerved, 8-12 lines in length, spreading: flowers irregu- 284 PROCEEDINGS OF THE AMERICAN ACADEMY. larly cymose: peduncles terminal or rarely and tardily somewhat lat- eral; pedicels elongated, unequal, erect; the lowest often more or less distinctly axillary : sepals oblong-lanceolate : capsule exceeding the calyx, acutish, dark and shining ; seeds very smooth. — Edinb. Phil. Journ. vi. 327 ; Hook. Fl. Bor.-Arn. i. 95 ; Torr. & Gray, vars. a, )3, and y, Fl. i. 184; Fenzl, 1. c. i. 386. S. palustris, Richardson, Frankl. Journ. 738. S. stricta, Richardson, Frankl. 2d Journ. 15. S. Iceta, Torr. Ann. Lye. N. Y. ii. 1G9. S. glauca, Meyer, PI. Lab. 93. S. crassifolia, Wats. Bot. King Exp. 38. S. longifo- lia, Rothr. Enum. PI. Cent. Col. 35. ? Micropetalon gramineum, James, Cat. 181. — A variable species marked by its long dark- colored acutish capsule and very smooth seeds. It is widely distributed from Maine to Arctic America, and from Alaska (Siberia) southward along the Rocky Mountains to Colorado and on the Pacific slope to San Bernardino, Parish. The commoner form has acute sepals and leaves varying imperceptibly from flaccid and spreading to erect and somewhat pungent. (Var. fi of Torr. & Gray, Fl. i. 185. S. stricta, Richardson, etc.) The typical form with spreading leaves and " very obtuse " sepals is comparatively rare. The following, although the best marked varieties, are connected by innumerable puzzling inter- mediate forms. Var. lseta, Wats. Low, smooth or somewhat pubescent, 1-4 inches in height, usually very glaucous, densely leafy at the base : leaves carinate, lanceolate-subulate to linear, rather rigid, erect, 2-6 lines long, shorter than in the type, narrower than in the following. — Bibl. Index, 112. S. Iceta, Richardson, Frankl. Journ. 738 ; Hook, in Parry's 2d Voyage, 300, & Fl. Bor.-Arn. i. 96. S. stricta, var. y> Hook. 1. c. i. 96. S. longipes, var. 8, Torr. & Gray, Fl. i. 185.— Arctic America to the Rocky Mountains of Montana and Wyoming to Gaspe, Allen. (Siberia.) A very similar form has been found on the coast of New Brunswick, Fowler. The variety peduncularis of Fenzl is a boreal form somewhat intermediate between this variety and the next, and indefinitely characterized by still more elongated pedicels. Var. Edwardsii, Wats. Low, smooth or pubescent : leaves lanceolate to ovatedanceolate or even ovate, shorter than in the type : stems usually but 2-3-flowered ; the lower peduncles axillary, much longer than the others. — Bibl. Index, 113. S. Edwardsii, R. Br. in Parry's 1st Voy. 271, 308; Cham. & Schlecht. Linmea, i. 48; Hook. Fl. Bor.-Am. i. 96, t. 31 ; Hornem. Fl. Dan. xiii. t. 2290. S. nitida, Hook, in Scoresb. Greenl. 411; Cham. & Schlecht. 1. c. ROBINSON. — ALSINE.E. 285 i. 47. — British America from Labrador to British Columbia, north- ward to the Arctic regions ; Alaska. (Siberia.) b. Seeds distinctly rugose-roughened under a lens. S. GRAMINEA, L. Stems ascending, smooth and shining, 1— 2i feet high, sharply 4-augled (rhombic in cross-section) ; internodes usually elongated : leaves lanceolate or lance-linear, thickish, attenuate, fur- rowed above and with midrib prominent beneath : inflorescence a broad terminal pedunculate cyme, often accompanied by one or two smaller cymes springing at its base ; pedicels elongated, spreading, or deflexed: capsule exceeding the calyx. — Spec. 422; Eng. Bot. xii. t. 803 ; Fenzl in Ledeb. Fl. Ross. i. 391 ; Wats. & Coulter in Gray, Man. ed. 6, 87. — Introduced in moist grassy places, Nova Scotia to New Jersey, and in Northern States across the continent ; common. (Europe.) S. uliginosa, Murr. Low, weak, diffuse: stems numerous, leafy : leaves lanceolate or elliptic-lanceolate, 6-8 lines long, acute at each end : inflorescences few-flowered, pedunculate or sub-sessile, much smaller than in the last, becoming decidedly lateral, 1^ inches or less in length : flowers smaller and petals relatively shorter than in the related species: sepals very acute, If lines in length. — Comm. Gott. 1778, 55; Fenzl, 1. c. i. 393; Eaton & Wright, N. A. Bot. 442; Warming, Bot. Foren. Festkr. 1890, 216, f. 10. S. aqunfica. Poll. Palat. i. 429; Torr. & Gray, Fl. i. 186. S. alsine, Hoffm. Fl. Germ. i. 153 ; Muhl. Cat. 45 ; S. borealis, Darlingt. Fl. Cest. 274. Larbrea ullginosa, Hook. Fl. Bor.-Am. i. 93. — Swamps, Atlantic slope, Halifax (Macoun) to Maryland ; rare. (Europe.) •w- ++ Bracts foliaceous (except in S. borealis, var. corollina.) = Leaves narrowly elliptical to lanceolate or linear. S. longipes, Goldie, may be sought here, as weak specimens with solitary terminal long-peduncled flowers do not always show the scarious bracts which are developed in more vigorous plants. S. borealis, Bigel. Suberect, 6-10 inches in height, smooth or nearly so: leaves lanceolate, attenuate, 6-18 lines long, with one prominent nerve: pedicels scattered, 8-14 lines in length, often deflexed : sepals ovate-lanceolate, scarious-margined, acute or often narrowed to an obtusish apex : petals much shorter than the calyx or none: capsule narrowly ovoid, acutish, 11-2 times as long as the sepals; seeds smooth. — Fl. Bost. ed. 2, 182; Torr. & Gray, Fl. i. 185; Fenzl, 1. c. i. 381 ; Vahl, Fl. Dan. xiv. t. 2355. S. calycantha, Bong. Sitch. 127. S. crassifolia, Boland. Cat. 6. Spergulastnon 286 PROCEEDINGS OF THE AMERICAN ACADEMY. lanceolatum, Michx. Fl. i. 275. Micropetalon lanceolatum, Pers. Syn. i. 509. Arenaria calycantha, Lecleb. Mem. Acad. Petr. v. 534. A. lateriflora, Darliugt. Florula Cest. 54. — New England to New Jersey; Mendocino Co., Calif., and northward; flowering in mid summer; frequent. Var. corollina, Fenzl, 1. c. i. 382. Taller : inflorescence spread- ing and more definitely terminal : bracts reduced, the uppermost more or less scarious : petals usually present : seeds slightly roughened. — S. brachypetala, Bong. Sitch. 126 ; Torr. & Gray, Fl. i. 186. S. alpes- tris, Fries, Mant. i. 10, excl. var. S. Fenzlii, Regel, Ost-Sib. i. 399„ S. borealis, var. alpestris, Gray, Man. ed. 5, 93. — Lake Superior, Bobbins, to Oregon, Howell, and northward. (Europe and Asia.) S. crassifolia, Ehrh. Low, smooth: stems many, weak, ascending or suberect, internodes short : leaves small, numerous, thick- ish, oblong-lanceolate, acutish, 3-6 lines in length : sepals ovate- lanceolate, acuminate, somewhat exceeded by the petals and capsule : seeds distinctly roughened under a lens. — Beitr. iii. 60 ; Fenzl, 1. c. i. 383 ; Chapm. Fl. ed. 2, 608 ; Wats. & Coulter in Gray, Man. ed. 6, 87, excl. Kentucky plant. S. gracilis, Richardson, Frankl. Journ. 738; Hook. Fl. Bor.-Am. i. 97; Torr. & Gray, Fl. i. 184. S. borealis, var. (3, Hook. 1. c. i. 95. — Labrador, Martin, Allen, to Illinois, Vasey, Hill; Montana, Canby, and northward. S. fontinalis. Glabrous, stems regularly and dichotomously branched, 6-12 inches long: branches spreading: leaves spatulate- linear, obtusish, spreading, 5-10 lines long: internodes elongated, 1-2 inches in length : peduncles solitary in the forks of the branches, 1-1^ inches long, ascending : sepals 4-5, oblong, obtuse, 3-nerved : petals none: stamens 4-8: styles 3-4, very short: capsule obtuse, not exceeding the calyx. — S. crassifolia, Wats. Bibl. Index, 111 in part ; Wats. & Coulter in Gray, Man. ed. 6, 87 in part. Sagina fontinalis, Short & Peter, Transylv. Journ. Med. 28, 600 ; Torr. & Gray, Fl. i. 177. Spergula fontinalis, Dietr. Syn. PI. ii. 1597. — Cliffs of Kentucky River and Elkhorn Creek, Kentucky, Short fy Peter; Nashville, Tenn., Gattinger ; April, May. This too little known plant, of distinctly aquatic habit, merits further study with more abundant material. It is certainly distinct from S. crassifolia, Ehrh. S. humifusa, Rottb. Low, densely matted, smooth : stems prostrate or ascending, angulate, shining : leaves elliptic-ovate or oblong, acutish, 2-5 lines long, marcescent : peduncles axillary. 4-7 lines in length : sepals ovate-oblong, acute, narrowly margined : petals ROBINSON. — ALSINEiE. 287 somewhat exceeding the calyx: seeds smooth. — Act. Hafn. x. 447, t. 4, f. 14 ; Torr. & Gray, Fl. i. 184 ; Fenzl in Ledeh. Fl. Ross. i. 384. S. marginata, Cham. & Schlecht. Linuaea, i. 50. Arenaria tJnjmi- folia, Pursh, Fl. 317 ; Eaton & Wright, N. A. Bot. 132. A. Purshi- ana, Seringe in DC. Prodr. i. 414 ; Hook. Fl. Bor.-Am. i. 102. — Salt marshes and boggy slopes, Little Cranberry Island, Maine, Red- field, and coast of Oregon, Howell Bros., northward to Alaska and Arctic America. (Greenland ; N. Asia.) The commoner form, var. OVALIFOLIA, Fenzl, 1. c, has leaves ovate or sub-orbicular, crowded, but 1-3 lines in length ; var. oblongifolia, Fenzl, 1. c, has more elongated internodes and oblong leaves, 4-5 lines long. S. obtusa, Engelm. Smooth: stems prostrate, 2-3 inches long: leaves thin, ovate, acute, about 4 lines long, half as broad : flowers solitary, appearing axillary : peduncles 3-4 lines long : sepals ovate, obtuse, hardly at all scarious on the margins : petals none : capsule \\-l ^ times as long as the calyx, obtuse ; seeds brown, uuder a com- pound microscope covered with lighter colored oblong tubercles with fringed edges. — Bot. Gaz. vii. 5 ; Macoun, Cat. Canad. PL 76. S. humifusa, Macoun, Phsenog. & Cryptog. PI. of Canad. 9. — Anthracite Creek, Colorado, 9,000-10,000 feet, Brandegee ; near Macleod's Lake, Brit. Columbia, Macoun; Kootanie Pass, B. C., Dawson; June, July. S. crispa, Cham. & Schlecht. Smooth, or more rarely pu- bescent: stems numerous, weak, decumbent: leaves thin, ovate, acumi- nate, commonly crisped on the edges : pedicels solitary, 3-6 lines long : sepals lanceolate, acute, margined, 3-nerved, considerably exceeded by the acutish capsule: petals minute or none. — Linnrea, i. 51 ; Hook. Fl. Bor.-Am. i. 97 ; Torr. & Gray, Fl. i. 186, 675 ; Gray, Proc. Am. Acad. viii. 378. S. borealis, var. crispa, Fenzl ex Torr. & Gray, Fl. i. 675 ; Torr. Bot. Wilkes Exp. 245. S. borealis, var. apetala, Regel, Ost-Sib. i. 396 in part. — Mountainous regions of N. Calif, to Alaska. S. ruscifolia, Willd. Glabrous: leaves coriaceous, ovate, sub- cordate, acuminate, somewhat rigid with pungent tip : flowers rather large, terminal, pedunculate : sepals acute. — Willd. in herb. ace. to Schlecht. Berl. Gesell. Nat. Fr., Mag. 1816, 194; Cham. & Schlecht. 1. c. i. 50; Regel, Ost-Sib. i. 419. (Siberia, Kamtschatka.) "Var. arctica, Regel, 1. c. i. 420. " Low stems, scarcely an inch in length, sepals obtuse. — On the Melville Islands." = = Leaves broad, an inch or more in length. S. littoralis, Torr. Pubescent; stems decumbent, dichoto- mously branched, 8 inches in height : leaves ovate, rounded at the 288 PROCEEDINGS OP THE AMERICAN ACADEMY. base, acute or acuminate, about au inch in length, with definite intra- marofinal veins : flowers rather numerous in the forks of the branches ; peduncles becoming horizontal or reflexed : sepals 2^ lines long, acute: petals of nearly equal length, cleft almost to the base : capsule some- what shorter. — Pacif. R. Rep. iv. 69; Brew. & Wats. Bot. Calif. i. 68. — Coast of Marin Co., Calif. ; Punta de los Reyes, Bigelow ; Dillon's Beach, Congdon. In habit much resembling the Old World S. dichotoma, L., of which it may well prove a form. It differs, bow- ever, in its much more deeply cleft petals. S. pubera, Michx. (Great Chicioveed.) Perennial, decum- bent, stout for the genus : stems pubescent in lines : leaves elliptic- oblong, finely ciliate, acute or obtusish, |— 1 \ inches long, or on the late tall and mostly sterile shoots 3 inches in length : calyx nearly or quite smooth ; sepals 3-4^ lines in length : stamens 10 : capsule globose, not exceeding the calyx. — Fl. i. 273 ; Darlingt. Fl. Cest. 274 ; Torr. & Gray, Fl. i. 183. — Rocky woods, Pennsylvania to Georgia, westward to Tennessee, Kentucky, and Indiana. Professor Britton calls atten- tion to the fact that the teeth of the capsule become circinate-revolute, as in Cerastium § Strephodon. Miss E. F. Andrews states that the petals are sometimes cleft half their length, in other cases nearly to the base, which is confirmed by specimens. * * Petals retuse or shortly bifid, divided but \-\ the way to the base, com- monly considerably exceeding the calyx : species approaching Armaria. . -i- Glabrous or nearly so. S. unifiora, Walt. Weak and slender : stems decumbent or suberect, a foot in length : leaves linear, acute, or the lower lanceolate, gradually narrowed below, mucronate, 8-12 lines in length ; the floral much reduced : flowers few, solitary, on elongated slender peduncles : calyx soft in texture; sepals scarcely veined. — Car. 141 ; Torr. & Gray, Fl. i. 184; Chapm. Fl. 50. Arenaria glabra, Ell. Sk. i. 520, not Michx. ; Wood, Bot. & Fl. 56. Alsine Walteri, Gray, Gen. ii. 34. — Moist meadows, North Carolina to Florida and Alabama, Winchell ; March to May. S. macropetala, Torr. & Gray, appears to be only a form of Arenaria patida, Michx., a species in which the petals vary from entire to retuse or even bifid. S. Holostea, L. Stem sharply 4-angled, 6-18 inches high : leaves narrowly lanceolate, spreading, long-attenuate from near the rounded sessile base, often ciliated, 1^-3 inches in length, 1-nerved: sepals lance-oblong, thin, nerveless, 4 lines in length, exceeded by the large white petals : styles 3 : valves of the capsule at length circinate- ROBINSON. — ALSINEiE. 289 revolute. — Spec. i. 422 ; Reichb. Icon. Fl. Germ. v. t. 223. — Found more or less established at Train's Meadow Road, Long Island, Ruger ; Poland, Maine, Miss Furbish. (Adv. from Europe.) +- +- Glandular-pubescent. S. dichotoma, L. Stems terete, profusely and dicbotomously branched : leaves ovate to ovate-lanceolate, acute or acutish, cordate, spreading, 6-12 lines in length: peduncles 1-flowered, springing from the forks of the branches, considerably exceeding the leaves, com- monly detlexed in fruit: sepals lanceolate, acute, usually about equal- ling the petals. — Spec. 603 ; Fenzl in Ledeb. Fl. Ross. i. 378. — An Asiatic species of great variability. Var. Americana, Porter, in litt. Leaves oval, obtusish : sepals oblong, obtuse, but 1^ lines long, considerably exceeded by the rather narrow white petals. — Collected near Virginia City, Montana, 1871, by W. B. Piatt on the Hayden Survey, and sent to the Gray Herbarium by Prof. T. C. Porter. A portion of the same material has been kindly examined by Messrs. Batalin, Korshinsky, and Lipsky, who pronounce it a variety or form of S. dichotoma, near var. cordifolia, Bunge, but with more obtuse leaves and sepals. S. Jamesii, Torr. Viscid above: stem strongly angled: leaves elongated, lanceolate, attenuate, smooth, 2-4 inches in length, 3-8 lines broad near the closely sessile base : flowers in a leafy terminal panicle: sepals oblong, herbaceous, 2 lines in length. — Ann. Lye. N. Y. ii. 169 (as S. Jamesiana) ; Pacif. R. Rep. iv. 69; Torr. & Gray, Fl. i. 183 ; Wats. Bot. King Exp. 38. ?S. graminea, James, Cat. 181. — Woodlands and " creek bottoms," Rocky Mts. of Colorado, New Mexico, and Arizona to Northern California and Washington, Brandegee. S. Nuttallii, Torr. & Gray. Annual, a span high : leaves linear-oblong, obtusish ; the upper much reduced but not scarious : flowers in dichotomous racemes; pedicels horizontally spreading, 9 lines in length: corolla 6-8 lines broad. — Fl. i. 183; Fielding, Sert. PI. t. 18. Alsine Drummondii, Fenzl ex Torr. & Gray, 1. c. i. 675. Alsine Nuttallii, Gray, Gen. ii. 34. — Arkansas, Nut t all ; Louisiana, Hale; Central Texas, Wright, Lindheimer, Drummond, Hall, etc. 11. ARENARIA, L. Sandwort. {Arena, sand, a sandy place, from the habitat of several species.) — A composite genus, and, when taken as here in its more comprehensive sense, the largest of the Alsinem. Plants of wide distribution both as regards latitude and altitude, and possessing in consequence much variability in aspect; being either slender annuals or herbaceous perennials of the habit of vol. xxix. (n*. s. xxi.) 19 290 PROCEEDINGS OF THE AMERICAN ACADEMY. Stellaria, or often more densely tufted and occasionally distinctly woody at the base. Leaves sessile or nearly so, either flat and with a well developed blade or more frequently awl-shaped or acerose. Flowers of Stellaria, but with petals entire or barely retuse (sometimes more deeply cleft in A. patula, Michx.). — Gen. no. 374 in part ; Torr. & Gray, Fl. i. 178; Reichb. Icon. Fl. Germ. v. t. 216-219; Benth. & Hook. Gen. i. 150; Wats. Bibl. Index, 94; Hook. & Jackson, Index Kew. i. 178. § 1. MaiHRiNGiA, L., as amended by Fenzl. Seeds, at least when young, provided at the hilum with a light-colored spongy appendage (strophiole). Habit of Stellaria. — Fenzl in Endl. Gen. 968; DC. Prodr. i. 390; Gray, Gen. ii. t. 112. A. lateriflora, L. Stems terete, weak, often decumbent, puberu- lent : leaves elliptic-oblong or oval, obtuse or rounded at the apex, thin, puberulent, 5-10 lines long ; the veins and edges beneath covered with a fine spreading pubescence : cymes pedunculate and somewhat umbellately few-(l-6-)flowered : sepals ovate, obtuse or scarcely acute, 1^ lines long, but one third to one half the length of the obovate petals: filaments pubescent. — Spec. 423; Hook. Fl. Bor.-Am. i. 102, t. 36; Torr. & Gray, Fl. i. 182, 675. A. Pennsyl- vanica, Muhl. Ind. Fl. Lane. 169. A. buxifolia, Poir. in Lam. Encycl. vi. 362; Torr. & Gray, Fl. i. 182. A. Haenheana, Bartl. in Presl, Rel. Haenk. ii. 15. Stellaria bijlora, Pursh, Fl. 317. Mcehringia lateriflora, Fenzl, Verbr. Alsiu. 18, 38, & Ledeb. Fl. Ross. i. 371; Gray, Gen. ii. 36, t. 112. — New England to New Jersey, Colorado, Oregon, and northward to the Arctic Ocean. A. macrophylla, Hook. Stems decumbent, angled, pulverulent- pubescent : leaves lanceolate, acutish to acuminate at both ends, (less commonly elliptic, obtusish,) 1-3 inches in length, glabrous, more or less punctate: peduncles slender, terminal or becoming axillary, 1-5- flowered : sepals ovate-lanceolate, very acuminate, exceeding the petals. — Fl. Bor.-Am. i. 102, t. 37; Torr. & Gray, Fl. i. 182 ; Torr. Pacif. R. Rep. iv. 69 ; Gray, Proc. Am. Acad. viii. 378 ; Greene, Fl. Francis. 125. Mcehringia macrophylla, Torr. Bot. Wilkes Exp. 246. — Extending from San Diego (Orcutt) northward through Cali- fornia, Oregon, and Washington into British America, and eastward to Isle St. Ignace, Lake Superior, Wheeler; April to August. While the essential floral characters remain the same, there is considerable variation in the size, texture, and shape of the leaves. § 2. Ammadenia. Styles 3-5 : disk conspicuous, 10-lobed and glanduliferous : capsule globose, somewhat baccate ; seeds not strophio- ROBINSON. — ALS1NE.E. 291 late. — Ginelin, Fl. Sib. iv. 160 (proposed as genus). Arenaria § Am- modenia, Benth. & Hook. Gen. i. 151. Honkenya, Ehrh. Beitr. ii. ISO. Hulianthus, Fries, Fl. Hall. 75. Adenarium, Raf. Journ. Phys. 1818, 259. — A single stout fleshy species of maritime habitat, with axillary flowers. A. peploides, L. Perennial, glabrous: stems a span in height, stout, angled : leaves thick, ovate or obovate, 1 -nerved, shortly pointed, clasping at the broad base : sepals ovate-lanceolate, acuminate, 3£ lines in length, about equalling the petals. — Spec. 423; Pursh, Fl. 317. Alsine peploides, Crantz, Inst. ii. 406. Honkenya peploides, Ehrh. 1. c. ii. 181 ; Torr. & Gray, Fl. i. 176; Fenzl in Ledeb. Fl. Ross. i. 358 ; Gray, Gen. ii. 32, t. 110. Adenarium peploides, Raf. 1. c. 259. - Seashores, from New Jersey and Washington State, Henderson, northward. (Northern Europe and Asia.) On the Northwest Coast the commoner form is "Var. major, Hook. Taller: leaves longer, often 15 lines iu length, oblong or oblanceolate, more pointed and decidedly narrowed to the base. — Fl. Bor.-Am. i. 102. A. Sitchensis, Dietr. Syn. PI. ii. 1565. Honkenya oblongifolia, Tovr. & Gray, Fl. i. 176. Honkenya peploides, var. oblongifolia, Fenzl in Ledeb. Fl. Ross. i. 358. § 3. Merckia. Styles 3-5 ; ovary 3-5-celled : capsule large, de- pressed-globose, somewhat inflated, many-seeded ; seeds not strophio- late. — Fisch. ex Cham. & Schlecht, Linnosa, i. 59 (as genus). Fenzl, 1. c. i. 359 ; Pax iu Engler & Prantl, Nat. Pflanzenf. iii. lb, 84. Arenaria § JSIerckia, Benth. & Hook. Gen. i. 151. — A single glandular and slightly fleshy species of the Northwest. A. physodes, Fisch. Perennial, cespitose; stems weak, decum- bent, 3-6 inches in length: leaves ovate, cuspidately pointed, 4-6 lines long: flowers solitary at the summit of the stem or becoming lateral : sepals lance-oblong, acute, 3 lines in length, equalling or slightly ex- ceeding the petals: capsule 4 lines (said to hecome half an inch) in diameter. — Fisch. in DC. Prodr. i. 413; Wats. Bibl. Index, 97. Merckia physodes, Fisch. 1. c. i. 59 ; Hook. 1. c. i. 103 ; Torr. & Gray, Fl. i. 176. Stellaria ovalifolia, Hook. 1. c. i. 97 ; Hook. & Arn. Bot. Beech. 122. — Brit. Columbia to N. Alaska; July, August. § 4. Euarenaria. Styles normally 3 : capsule ovoid, dehiscent by three two-toothed or cleft valves ; seeds not strophiolate. — Arenaria of many authors, as Fenzl in Ledeb. Fl. Ross. i. 360; Regel, Radde's Relsen in Ost-Sib. i. 334; Pax in Engl. & Prantl, Nat. Pflanzenf iii. 1 b, 84, etc. 292 PROCEEDINGS OF THE AMERICAN ACADEMY. * Leaves ovate, elliptic or linear, not acerose. -*- Annuals. A. SERPYLLIFOLIA, L. (THYME-LEAVED SANDWORT.) AuUUal, finely pubescent, much branched : leaves very short, 2-3^ lines in length, ovate, acute or acuminate, rather distinctly 3-5-nerved, rounded at the base; the lowest only being narrowed to a short petiole : flowers numerous in open leafy cymes; pedicels 1-3 times the length of the ovate-lanceolate acuminate hispidulous sepals: petals small, about two thirds the length of the sepals : capsule flask-shaped. — Spec. 423 ; Michx. Fl. i. 274; Ell. Sk. i. 518; Torr. & Gray, Fl. i. 182.— Sandy soil, Lower Canada and New England to Florida and to Oregon, Washington, and British Columbia. (Nat. from Europe.) Var. tenuior, Koch. More delicate: leaves reduced: flowers smaller in a nearly naked racemose panicle : capsule more oblong. — Syuop. 117. A. leptoclados, Guss. Fl. Sic. Syn. iii. 824. — Portland, Oregon, Brandegee. Suksdorf's specimens from Mc Cloud Lake are somewhat similar. (Europe.) A. Benthamii, Fenzl in herb. A. slender annual, branched from the base; branches finely pubescent in lines: leaves short, 3-4 lines in length, elliptic-lanceolate, acute and apiculate, often punctate, narrowed to a sessile base, or the lowest to short ciliated petioles; floral leaves much reduced : pedicels filiform, many times exceeding the ovate acuminate glabrous often punctate sepals : seeds dark brown, minutely tuberculate. — Torr. & Gray, Fl. i. 675 ; Gray, PI. Wright, ii. 18; Torr. Bot. Mex. Bound. 36. A. monticolu, Buckley, Proc. Acad. Philad. 1861, 449. — Rocky ground, Texas, Drummond, Wright, Thurber, Hall, Reverchon ; New Mexico, Thurber. •»- -t- Perennials. A. ciliata, L. Minutely glandular-puberulent: stems numerous, slender, terete, leafy, densely matted, or in less exposed situations spreading and ascending, 1-5 inches long, terminally 1-3-flowered : leaves small, ovate-oblong or lance-oblong, scarcely acute, 1-3 lines in length, distinctly ciliate near the cuneate base: peduncles erect, 2-5 lines long: sepals ovate-oblong, obtuse, nerved, 1|-H lines long: petals of similar shape and equal length : stamens 8-10 : valves of the capsule rather deeply bifid exceeding the calyx. — Spec. 425 ; DC. Prodr. i. 411 ; Fenzl in Ledeb. Fl. Ross. i. 370. (High moun- tains and arctic regions of Europe.) Var. (?) humifusa, Hornem. Leaves without ciliation : sepals nerveless. — Hornem. ex Lange, PI. Groan. 132. A. Norregica, Gunn, Fl. Norv. ii. 145, t. ix. f. 7-9. A. humifusa, Wahlb. Fl. Lap. ROBINSON. — ALSINE.E. 293 129. — Rich soil, British America, Mt. Albert, Gaspe, Allen, Porter; Lake Mistassini, /. M. Macoun, and what is with scarcely a doubt the same thing at Kicking Horse Lake in the Rocky Mts., J. Macoun. (Greenland, N. Europe.) A. alsinoid.es, Willd. in herb. Minutely pubescent with slightly hooked hairs or smoothish: stems long, procumbent, moderately branched : leaves narrowly elliptic, acute, narrowed below, commonly pseudoverticillate, 8-10 lines long, punctate: flowers axillary, soli- tary at the nodes : pedicels filiform, elongated, spreading or horizontal, nearly or quite an inch in length : sepals ovate, acute, tuberculate- puuctate, 1J lines long: petals commonly smaller or wanting: seeds smooth and shining. — Schlecht. Berl. Mag. Naturf. Freunde, vii. (1816), 201; Wats. Proc. Am. Acad. xvii. 327. A. diffusa, Ell. Sk. i. 519. A. nemorosa, IIBK. Nov. Gen. et Sp. vi. 35. A. lanugi- nosa, Rohrb. in Fl. Bras, xiv.2 274, t. 63. Spergulastrum lanugino- sum, Michx. Fl. i. 275. Micropetalon lanuginosum, Pers. Syn. i. 509. Stellaria elongata, Nutt. Gen. i. 289. S. lanuginosa, Torr. & Gray, Fl. i. 187, 675. — Moist shaded ground. North Carolina to Florida and Texas, Drummond, Hall. (Mexico and S. Amer.) A more western form, represented from New Mexico by Fendler's 58 and 62 and Wright's 864, has slightly firmer stems, more numerous sub-paniculate flowers, and leaves less narrowed at the base. In all these respects it shows a transition to the following. A. saxosa, Gray. Finely puberulent but green : stems many, spreading from a rather stout root, decumbent or creeping at the base, 2 inches to a foot long : leaves numerous, opposite, not fascicled or pseudoverticillate (sometimes crowded), slightly fleshy, lance-oblong, acute, mucronate, 2-9 lines long, sessile by a scarcely narrowed base : flowers terminal and subsolitary on short simple peduncles or in stouter individuals numerous and more or less paniculate : petals almost or quite equalling the ovate-lanceolate sharply acuminate slightly fleshy sepals. — PI. Wright, ii. 18; Walp. Ann. iv. 258. Mozhringia urn- brosa, Gray PI. Fendl. 13, & PI. Wright, ii. 18, not Fenzl. — Colo- rado, Brandegee, Hooker fy Gray; Guadalupe Mts., Texas, Havard ; New Mexico, Fendler, Wright; Arizona, Rothrock, Lemmon, Jones, Rusby. (Lower California, Orcutt.) A species of wide range, occur- ring alike in rocky subrdpine regions and much lower upon sandy banks, accordingly varying much in height and diffuseness of branch- ing. The type is a condensed few-flowered form. Var. cinerascens. Somewhat more rigid, and grayish through- out with a fine pubescence: leaves pungent. — Iluachuca Mts., Ari- zona, J, G. Lemmon. (Herb. Columbia Coll.) 294 PROCEEDINGS OF THE AMERICAN ACADEMY. # * Leaves very narrowly linear, commonly acerose and often rigid and pun- gent ; western species. i- Sepals broadly ovate, obtusish, sometimes apiculate : flowers not densely aggregated. A. capillaris, Pom. Leaves chiefly grouped at the base iu fasci- cles upon a multicipital caudex, h-2^ inches long, somewhat pungent, little spreading ; the cauline few pairs, much reduced : stems 4-8 inches iu height : petals obovate, considerably exceeding the short obtuse sepals. — Poir. in Lam. Encycl. vi. 380; Regel, Ost-Sib. i. 3G6. Al- sine nardifolia, Anderson, Cat. 118. The typical glabrous form with straight leaves is comparatively rare in America but occasionally occurs with var. nardifolia, Regel, 1. c. i. 372, which is glabrous with curved leaves (A. nardifolia, Ledeb. Fl. Alt. ii. 166, & Icon. Fl. Ross. 6; Hook. Fl. Bor.-Am. i. 98, t. 32), and the more common form var. FORMOSA, Regel, I. c. i. 371, which has the stem and inflorescence glandular. (A. formosa, Fischer in DC. Prodr. i. 402 ; Hook. f. Arc. PI. 287, 322 ; Torr. Bot. Wilkes Exp. 243.) — Central California to Utah, Montana, and British Columbia. (Asia.) A. ursina. Root stout, ligneous : caudex very densely multicipi- tal and bearing very close fascicles or rosettes of very short filiform- linear thickish glaucous glandular-ciliolate white-punctate apiculate leaves (but 2-3 lines in length) : stems few, erect, branched, slender, finely glandular-pubescent, 1^—3 inches high, bearing one or two pairs of short distant spreading and rather rigid leaves : sepals nerveless : petals white, oblong, slightly emargiuate : alternate filaments glandu- liferous at the base : mature capsule considerably exceeding the calyx, about 6-seeded. — Dry hills, Bear Valley, San Bernardino Mts., S. B. S)- W. F. Parish, August, 1882. This species has much the habit of A. compacta, Coville, but much more obtuse sepals ; it is most nearly related to A. capillaris, Poir., but differs in its condensed habit and in a waxy punctation of the leaves. A. aculeata, Wats. Leaves grouped chiefly iu fascicles at the summits of a multicipital caudex, decidedly glaucous, rigid and pun- gent and with age strongly spreading, 6-12 lines in length; cauline leaves few, shorter: stems simple up to the few-flowered cymes, 4-6 inches high: petals rather narrow, elliptic-oblanceolate, obtuse, 1^—2 times as long as the sepals. — Bot. King Exp. 40, & Bibl. Index, 94. — Chiefly in mountainous districts from Oregon (Nevius, Cusick, How- ell) to N. Nevada (Watson), S. Utah, and Arizona (?). The sepals are variable and become more acute in the southern form, which is more difficult to distinguish from those varieties of A. congesta which have an open inflorescence. ROBINSON. — ALSINEjE. 205 4- *- Sepals ovate, or ovate-lanceolate, acuminate, shorter than the petals (ex- cept in A. macradenia, var. Pariskiorum). A. compacta, Coville. Root thick, ligneous: caudex much branched and bearing very closely tufted rosulately spreading subulate glaucous leaves ; the latter not exceeding 2 lines in length, minutely glandular, ciliate : stems slender, an inch or more in height, simple or sparingly branched, almost naked, the cauline leaves being few and much reduced: flowers terminal on the branches: sepals 1^ lines long, scarious-margined, thickened in the middle, attenuate. — Proc. Biol. Soc. Wash. vii. 67 ; Contrib. U. S. Nat. Herb. iv. 70, t. v. — Moun- tains of Tulare Co., Calif., Coville, August, 1891. A. congesta, Nutt. Smooth, or more rarely glandular-pubes- cent : stems slender, simple, 5-14 inches high, numerous, springing from a matted non-ligneous caudex : basal leaves gramineous-setaceous, |— 3 inches long, ciliolate-serrate near the base: cauline leaves rather dis- tant, gradually reduced: flowers sessile or shortly and unequally pedicelled in 1-3 dense heads (subtended by 1-several pairs of sca- rious-margined bracts) : sepals carinate, obscurely 3-nerved, membra- nous-margined, 2 lines long, considerably exceeded by the narrowly oblong petals : stigmas not strictly capitate. — Nutt. in Torr. & Gray, FI. i. 178 ; Torr. Frem. Rep. 87 ; Wats. Bot. King Exp. 39 ; Porter & Coulter, Fl. Col. 13 ; Brew. & Wats. Bot. Calif, i. 09 ; Greene, Fl. Francis. 123 (excl. syn.) ; Mrs. Brandegee, Zoe, ii. 161. — Rocky Mts. of Colorado and Wyoming to the Yosemite (ace. Mrs. Brandegee) and northward to Washington, Suksdorf. Var. suffrutescens. Caudex sometimes, perhaps always, very ligneous ; its branches becoming 2-3 lines in diameter, bearing fasci- cled sub-equal leaves (an inch or less in length) : flowers somewhat smaller (sepals 1^ lines long) in spherical capitate umbels : pedicels slender, 2-3 times as long as the calyx : stigmas capitate. — Brewerina suffrutescens, Gray, Proc. Am. Acad. viii. 620, under A. congesta ; Brew. & Wats. Bot. Calif, i. 69; Wats. Bibl. Index, 95; Greene, Fl. Francis. 123. — California, Cisco, Bolander Sf Kellogg; Emigrant Gap, Jones; Tulare Co., Coville $ Funston, 1504. A form too well marked in its foliage, ligneous caudex, and allium-like inflorescence to be united with the typical A. congesta, yet appearing to intergrade with it. One of the transitional forms has been collected in Sierra Valley by Lemmon. Var. subcongesta, Wats. Caudex more or less ligneous, stems smooth, glandular or pulverulent-pubescent, often knotted with enlarged nodes : flowers as in the type, but borne in more or less expanded di- 296 PROCEEDINGS OP THE AMERICAN ACADEMY. cbotomous cymes: leaves varying greatly in length and texture. — Bot Calif, i. 69 ; Bibl. Index, 454. A. Fendleri, var. subcongesta, Wats Bot. King Exp. 40, & PI. Wheeler, 6 ; Porter & Coulter, Fl. Col. 13 Rothr. Enum. PI. Col. 35. A. Fendleri, var. glabrescens, Wats. Bot King Exp. 40, & Bibl. Index, 95, differs only in its still looser inflo rescence, and should doubtless be referred hither. — Rocky Mts. of Colorado and Arizona, Newberry, to British America, Cypress Hills, Macoun, westward to Oregon, Howell, and California, Sierra Co., Lemmon, Donner Pass, Torrey. These puzzling and inconstant forms are intermediate between this species and A. capillaris, A. Fendleri, and A. macradenia. From the last they scarcely differ save in their smaller flowers and slightly in the form of the petals. A. macradenia, Wats, (revised). Glabrous or nearly so: rootstock more or less ligneous, extensively and irregularly branched : stems stout for the genus, 6-15 inches high, knotted with the eularged nodes : leaves chiefly cauliue, glaucous, rigid, pungent, J— 2 inches long : flowers larger than in the related species, in an open cyme : sepals fleshy subcarinate, 2| — 2| lines long, with membranous margins : petals considerably exserted, obovate or oblong with an obtusish sometimes auricled base : stamiueal glands moderately developed : stigmas sub- capitate. — Proc. Am. Acad. xvii. 367 in part. — S. California, Mo- have River, Palmer, 41 ; Antelope Valley, Oliver ; S. Utah, Parry, 20, Palmer ; Arizona, Palmer ; Mt. Agassiz, Lemmon. These plants with their exserted petals and ligneous caudices best represent Dr. Watson's species as described, Palmer's no. 41 being the first men- tioned type. Var. (?) Parishiorum. Smooth or minutely glandular-pubescent : caudex scarcely ligneous, densely multicipital : stems slender; nodes not conspicuously enlarged : leaves chiefly basal : petals narrowed at the base, shorter than or barely equalling the sepals, the latter fully 3 lines in length : stamineal glands very large. — A. macradenia, Wats. 1. c. xvii. 367 in part. — Common on mountains bordering on the Mojave Desert, S. B. $ W. F. Parish, no. 1330. +_ «_ ^ Sepals lanceolate to lance-linear, attenuate, equalling or exceeding the petals. ++ Flowers cymose, not densely aggregated. A. Fendleri, Gray. Rather pale and glaucous, finely glandular- pubescent above: stems numerous, erect, leafly, 4-15 inches high, closely aggregated upon the summit of a thick root: basal leaves setaceous, gramineous, ciliolate or quite smooth, 2-4 inches in length, somewhat pungent ; the cauline gradually shorter, connate and sheath- ROBINSON. — ALSINE.E. 297 ing at the base : the interuodes an inch or two long : inflorescence dichotonious, few to many flowered : sepals attenuate, glandular, nearly equalling the obovate petals (2^-3 lines in length) : capsule commonly a fourth shorter. — PI. Fendl. 13 ; Torr. Pacif. R. Rep. iv. G9 ; Wats. Bot. King Exp. 40, exclusive of var. glabrescens ; Porter & Coulter, Fl. Col. 13. — Chiefly in the Rocky Mts., but sometimes among the sage-brush of the plains. Nebraska, Engehnann; Colorado to New Mexico, G. R. Vasey ; San Francisco Mts., Arizona, Lemmon ; Los Angeles, Cal., Nevin. The var. diffusa of Prof. Porter's Fl. Col. is a greener form from the Rocky Mts. of Colorado and Wyoming with a more lax and spreading inflorescence and often although not always larger flowers. It intergrades with the type, so that in herbarium specimens at least its separation is often unsatisfactory. A form col- lected by Prof. Porter in the Garden of the Gods, and possessing very small flowers (sepals 1£-1^ lines in length) upon curved and spread- ing branches, is perhaps equally worthy of varietal distinction. *+ *+ Flowers densely fascicled at the summit of the stem. A. Franklinii, Dougl. Caudex of numerous procumbent more or less elongated branches, covered with somewhat persistent dried leaves : stems quite smooth, erect, simple, 3-5 inches high, somewhat rigid but fragile, bearing 3-6 pairs of narrowly subulate pungent spreading smooth or ciliolate and minutely scabrous leaves (5-9 lines long) : cymes dense, sub-involucrate : sepals elongated, attenuate, pun- gent with slightly spreading tips, 1-nerved, 4-6 lines long, distinctly exceeding the petals. — Hook. Fl. Bor.-Am. i. 101, t. 35; Torr. & Gray, Fl. i. 178 ; Torr. Bot. Wilkes Exp. 244. — Sandy soil, Oregon, Douglas, Lyall, Howell, Nevius, Cusick ; Washington, Suksdorf. Specimens collected by Douglas at source of the Missouri may well have been the next species. A. Hookeri, Nutt. Caudex densely multicipital : stems 1-4 inches high, pubescent: leaves shorter than in the last: flowers smaller and petals about equalling or slightly exceeding the sepals. — Nutt. in Torr. & Gray, Fl. i. 178. A. Franklinii, var. minor, Hook. & Arn. Bot. Beech. 326; Wats. Bibl. Index, 95; Coulter, Man. Rocky Mt. Bot. 35. A. Franklinii, Engelm. Trans. Am. Phil. Soc. n. ser. xii. 186; Coulter, 1. c. 35 in great part; Hook. & Jackson, Index Kew. 179. — Rocky Mts., lat. 40°, Nultall ; Colorado, Vasey ; Wyoming, Hayden, Parry, Porter, Greene, Sheldon ; Plains of Green River, Gray; Nebraska, Rydberg, Webber; Montana, Tweedy. This species with much the habit of the preceding differs in its much denser caudex and constantly pubescent stem, as well as in the differences 298 PROCEEDINGS OF THE AMERICAN ACADEMY. indicated. The stem is terete even in a dried state, while the stems of A. Franklinii in drying become furrowed and angulate, as though slightly fleshy. § 5. Alsine, Wahlenberg (as genus, not Linn.). Capsule ovoid, 3- valved ; valves entire ; seeds not strophiolate : matted perennials or delicate annuals, usually with narrow linear subulate or acerose leaves. — Fl. Lapp. 127; Fenzl in Ledeb. Fl. Ross. i. 341 ; Regel, Radde's Reisen in Ost-Sib. i. 337 ; Pax in Engl. & Prantl, Nat. Pflanzenf. iii. 1 b, 82. # Palustrine perennial with weak elongated stems, narrow linear or lance- linear leaves and axillary long-peduncled flowers. A. paludicola. Glabrous, flaccid : stems several, subsimple, procumbent, rooting at the lower joints, sulcate, shining, leafy throughout: leaves uniform, flat, 1-nerved, acute, spreading, f-1^ inches long, 1-3 lines in breadth, often punctate, somewhat connate, slightly scabrous upon the margins: peduncles solitary in the axils, 1-2 inches long, spreading or somewhat deflexed : sepals nerveless, not at all indurated, acutish about half the length of the obovate petals. — A. palustris, Wats, (not Gay). Bot. Calif, i. 70, & Bibl. Index, 97 ; Greene, Fl. Francis. 124; Mrs. Brandegee, Zoe, ii. 341. Alsine palustris, Kellogg, Proc. Calif. Acad. iii. Gl. — Abundant in swamps about Fort Point near San Francisco, Bolander, Kellogg $• Harford ; also in swamps near San Bernardino, Parish Bros. ; May to August. * * Terrestrial annuals of the Atlantic Slope and Alleghany Mountains, rarely extending to the interior in the Southern States, essentially glabrous: sepals obtuse, soft in texture, scarcely or not at all nerved. A. G-rcenlandica, Spreng. Somewhat fleshy : root at first simple, later of many delicate fibres : stems few to many, decumbent or erect, subsimple, 2-8 inches long, bearing 1-5 flowers : leaves linear, obtuse, l£-7 lines long, at first in a dense, more or less rosulate cluster at the base ; the cauline 2-4 pairs : sepals broadly ovate, 1^—2 lines in length: petals obovate, about twice as long, entire or notched; capsule subglobose to oblong, more or less contracted to a point. — Syst. ii. 402 ; Torr. & Gray, Fl. i. 180 ; Torr. Fl. N. Y. i. 95, t. 15. A. glabra, Torr. Fl. U. S. 455, not Michx. ; Bigel. Fl. Bost. ed. 2, 180. Alsine GrosJikmdica, Gray, Man. ed. 2, 58. Stellaria Grcenlandica, Retz. Fl. Scand. ed. 2, 107; Vahl, Fl. Dan. vii. t. 1210. ? S. Labra- dorica, Schrank, Pfl. Lab. 24 ; Meyer, PI. Lab. 93. — Rocky soil, chiefly but not always at higher altitudes, Greenland to the coast of Maine, Bath, Gambel ; Bar Harbor, Rand ; also at Middletown, Conn., Osbom, Wright ; locally abundant in the White, Green, Adirondack, ROBINSON. — ALSINE.E. 299 Catskill, and Shawangunk Mts. ; reported from Kittatinny Mts. of N. W. New Jersey (Britton, Bull. Ton-. Club, xi. 128), and recently found in the mountains of Pennsylvania (according to Prof. Porter), of S. Virginia (Small S? Heller), and in N. Carolina (Small), where it had passed as a form of A. glabra, Michx., having been previously collected on Roan Mountain by Gray fy Carey, Smith, aud Scribner* A. glabra, Michx. Glabrous, loosely matted, many-stemmed : stems weak, slender, suberect, very leafy, G— 12 inches high: leaves narrowly linear, spreading, thin, nerveless, equalling or exceeding the internodes : peduncles filiform, elongated, spreading, 1-flowered : corolla rather broad, considerably exceeding the calyx : sepals ovate- oblong, obtuse, nerveless, lh lines in length, somewhat exceeded by the ovoid capsule. — Fl. i. 274; Torr. & Gray, Fl. i. 180 in part. Alsine glabra, Gray, Man. ed. 2, 58 ; Chapm. Fl. 49. — On rocks in mountains of N. Carolina, Michaux ; Table Mountain, Gray ; Table Rock, S. Carolina, Vasey ; Stone Mt., Ga., Gray ; De Kalb Co., Ga., Small; also in the Arroyo of Lamben, near the Mexican boundary, Parry. A. brevifolia, Nutt. Glabrous: stems erect, filiform, 2-5 inches high, with spreading branches: leaves linear or lance-linear obtuse, nerveless, slightly fleshy, 1-4 lines long, commonly much shorter than internodes : sepals ovate-oblong, obtuse, only a line in length, with a distinct thin white margin: petals rather conspicuous, obovate, 2J— 3^ lines in length, widely spreading : capsule ovoid, acuminate, a third longer than the calyx ; valves ovate, acuminately narrowed almost to the tip. — Nutt. in Torr. & Gray, Fl. i. 180. Alsine brevifolia, Chapm. Fl. 49. — On rocks, Georgia, Tatnal County, Nuttall ; Stone Mountain, Canby, Gray, Small; April, May. Apparently the most rare and local eastern species. * * * Terrestrial annuals of the Pacific Slope: sepals neither indurated nor very strongly nerved, -t- Seeds much flattened, and margined. A. Douglasii, Torr. & Gray. Thinly glandular-pubescent and somewhat viscid, or nearly glabrous: stems much branched, 2-15 inches high: leaves attenuate to filiform points: peduncles filiform: flowers numerous, larger than in the related species, 4-5 lines in diameter : sepals ovate, thin-margined, obscurely or more or less dis- tinctly ribbed : petals obovate, conspicuous • capsule subglobose ; the valves rounded at the apex ; seeds large, smooth, or with fine radiate * For further discussion of this species see page 328. 300 PROCEEDINGS OF THE AMERICAN ACADEMY. striation, reuiform, broadly margined. — Fl. i. 674 ; Durand, PI. Pratt. 83; Brew. & Wats. Bot. Calif, i. 69 ; Greene, Fl. Francis. 124. A. verna, /?, Hook. & Arn. Bot. Beech. 325. Greniera Douglasii, Gay, Ann. Sci. Nat. ser. 3, iv. 27. Alsine tenella, Torr. Bot. Mex. Bound. 36 (ex char, et hab.). — Barren hillsides and grassy slopes, S. Arizona, Palmer, and S. California to Oregon, Howell, Henderson ; some smaller flowered specimens with seeds of A. Douglasii have been collected by Tkurber in San Diego, Calif. -i- -t- Seeds not flattened nor thin-margined. A. Howellii, Wats. Finely glandular-pubescent: stem terete, purple, diffusely branched, more than a foot high : leaves rather thick, obtuse, 4-7 lines in length ; the floral much reduced : flowers 2|— 3 lines in diameter : petals oblong, little exceeding the ovate glandular nerveless sepals: capsule ovoid, pointed; valves narrowed to an acutish apex ; seeds dark, slightly tuberculate-crested. — - Proc. Am. Acad. xx. 354. A. Californica, Brewer. Smooth : the delicate filiform stems branching from the base, erect, 2-4 inches in height : leaves very short, slightly fleshy, 1-2 lines in length, obtuse : flowers 4 lines in diameter : petals oblong, about twice the length of the ovate-oblong nerveless or inconspicuously ribbed sepals : seeds small, finely rough- ened.— Brewer in Boland. Cat. 6; Brew. & Wats. Bot. Calif, i. 69 ; Greene, Fl. Francis. 124. A. brevi folia, var. (?) Californica, Gray, Proc. Calif. Acad. iii. 101. — Dry hills, sandy soil. Central Calif, to Grant's Pass, Oregon, Howell. A. pusilla, Wats. Smooth, very diminutive, l£-2 inches high: stems purplish, filiform, branched from the base : leaves obtusish, but 1-2 lines in length: sepals not so strongly nerved as in the preceding, 1-1 1 lines in length : petals minute or wanting : seeds minute, smooth. — Proc. Am. Acad. xvii. 367. A. Californica of Bot. Calif, (in part). — Plains about Yreka, Calif. (Greene), to the Dalles of the Columbia (Howell Bros.), and White Salmon, Wash. (Suksdorf). * * * * Annuals or slender-stemmed loosely matted perennials, 5-15 inches in height : sepals lanceolate, acuminate or attenuate, strongly 3-5-nerved. •*- Puberulent. A. tenella, Nutt. Finely glandular-pubescent: stems very slen- der, dichotomously branched almost from the base, 3-8 inches in height: leaves attenuate from a connate prominently ribbed base to a filiform often curved apex, 3-5 lines long ; the uppermost considera- bly reduced : pedicels filiform, several times as long as the strongly ROBINSON. — ALSINE^E. 301 3-ribbed sepals ; the latter equalled or more or less exceeded by the oblong petals : valves of ovoid capsule exceeding the sepals ; seeds small, margined with a fine muriculate crest (under a strong lens). — Nutt. in Torr. & Gray, Fl. i. 179; Eaton & Wright, N. A. Bot. 133 (excl. Arkansas spec.) ; Macoun, Bot. Gaz. xvi. 286. A. tenuifolia, var. Americana, Fenzl ex Torr. & Gray, Fl. i. 074. Greniera tenella, Gay, Ann. Sci. Nat., ser. 3, iv. 27. Alsine tenella, Torr. Bot. Wilkes Exp. 243. — Rocky places, flowering in midsummer ; Oregon, Nuttall, Tolmie, Hall, Howell, to British Columbia, Kamloops ace. to Macoun ; Nanaimo, Miss Cooley. Like A. Grcenlandica of the Eastern States, this species seems to occur either in mats or in a segregated state. In the former condition it considerably resembles A. stricta, Michx., but is to be distinguished by its small flowers and puberulent inflo- rescence. It differs from A. patula in its broadly ellipticpetals not at all truncate or retuse. A. patula, Michx. Stem diffusely branched, 2 inches to a foot in height, often almost filiform: leaves spreading, slightly fleshy: inflorescence dichotomous ; pedicels filiform, spreading: sepals lan- ceolate, attenuate, with 3-5 prominent converging nerves, slightly indurated, a little over 2 lines in length, usually minutely glandular: petals twice as long, entire or retuse, obcordate : the obtuse valves of the capsule about equalling the calyx; seeds black, minutely rough- ened. — Fl. i. 273 ; Torr. & Gray, Fl. i. 180 ; Gray, Man. ed. 5, «J1 ; Hill, Bun. Torr. Club, xvii. 172; Macmillan, Bot. Gaz. xv. 332. A. Pitcheri, Nutt., and (?) A. tenella, Nutt. in Torr. & Gray, 1. c, so far as Arkansas plants are concerned. Alsine microsperma, Fenzl ex Torr. & Gray, Fl. i. 674. A. patula, Gray, Man. ed. 2, 58; Chapm. Fl. 49. A. Pitcheri, Wood, Class-Book, 260; Chapm. Fl. (Suppl.) 608. Stellaria macropetala, Torr. & Gray, Fl. i. 184 (Alsine macropetala, Gray, Gen. ii. 34), differing only in the slightly more deeply divided petals, which are themselves more or less variable, must be referred hither, where its identity of habit and calyx clearly indicates its affinity to be. — Kentucky to Alabama (Peters, Mohr) and Texas (Drummnnd, Meyer, HucMey), northward to Chicago (Babcock, Hill) and Cass Co., Minnesota (ace. to Macmillan). The leaves of this species are varia- ble, commonly narrowly linear or Aliform, 4-7 lines in length, but they occasionally become l£ inches long and a line wide. -»- *- Glabrous. A. stricta, Michx. Smooth, loosely matted : stems very numer- ous, slender, ascending, 3-15 inches high, leafy nearly to the middle : leaves subulate-setaceous, conspicuously fascicled in the axils: iuflo- 302 PROCEEDINGS OF THE AMERICAN ACADEMY. rescence a loosely forked cymose panicle: petals narrowly obovate, nearly twice the length of the somewhat rigid acuminate prominently 3-ribbed sepals : capsule about equalling or exceeding the calyx. — Fl. i. 274; Ell. Sk. i. 521 ; Hook, Fl. Bor.-Am. i. 99, t. 33 (including both var. a, a weak boreal few-flowered form with erect leaves, and var. /3, the common form with spreading leaves) ; Torr. & Gray, Fl. i. 179, var. ft ; Britton, Mem. Torr. Club, ii. 37. ? A. setacea, Muhl. Ind. Fl. Lane. 1G9. A. Michauxii, Hook. f. Arc. PI. 287, 322. Alsine Michauxii, Fenzl, Verbr. Alsin. 18; Regel, Ost-Sib. i. 351, t. 8, f. 1-5. — Rocky and gravelly soil, Vermont to S. Carolina, westward to Minnesota. Var. Texana. More rigid, stems fewer, 3-7 inches high, strongly enlarged at the nodes : leaves very short, conspicuously connate ; the fascicled ones but 1-2 lines long : flowers in a small rather dense cyme : sepals almost cartilaginous, very strongly 3-nerved, appear- ing attenuate through the infolding of their margins. — ? A. stricta, var. a, Torr. & Gray, Fl. i. 179. — Rocky Hills, Texas; along the Canadian River, Gordon, April, 1848; Bigelow, August, 1853; Dallas, Hall, July, 1872: Comanche Peak, Reverrhon. 1881; Arkan- sas, Leavenworth; False Washita, Ind. Ter., Palmer ; Kansas, Norton, Smith. Owing to its definite and limited geographic range this variety may, as Dr. Britton suggests, prove worthy of specific rank. The characters, however, are not very definite, and a specimen from Potosi, Mo. (F. Peck) exactly connects it with the type. While the descrip- tion of Torrey and Gray's var. a, cited above, ill accords with the present plant and rather suggests that these authors intended their var. a to be equivalent to the var. a of Hooker, as their var. (3 was of his var. f3, yet in the Columbia Herbarium there are specimens of the southwestern plant labelled in Torrey's hand as var. a. If the Texan plant was the var. a of Torrey and Gray, it need scarcely be said that it could not have been the typical form of Michaux's species, as it was doubtless intended to be. ***** Perennials, closely matted or tufted, 1-6 inches in height: sepals acuminate, but not strongly nerved, except in A. venta. A. verna, L. Rather closely tufted : stems numerous, slender, ascending or erect, smooth, 1-5 inches high, 1 to 3 (or more) flowered ; the upper internodes commonly considerably exceeding the leaves : leaves linear-subulate, flat, rather strongly 3-nerved, usually erect and never squarrose : peduncles filiform: sepals ovate-oblong, acutish to acuminate, strongly 3-nerved, 1|— If lines long, exceeding the obo- vate or oblanceolate obtusish petals : capsule somewhat surpassing the ROBrNSON. — ALSINE^E. 303 calyx. — Mant. i. 72 ; Seringe in DC. Prodr. i. 405 ; Hook. Fl. Bor.- Am. i. 99 ; Torr. & Gray, Fl. i. 181. A. j uniperina, Pursh, Fl. 318; Hook. Fl. Bor.-Am. i. 98; Torr. & Gray, Fl. i. 179, G74. Alsine verna, Bartl. Beitr. ii. 63. — A widely distributed species with numerous but ill defined varieties. The smooth typical form appears to be common in the Rocky Mts. of British America (Macoun), and has been found on Mt. Albert, Lower Can. {Allen). A far more frequent form is Var. hirta, Wats. Finely glandular-puberulent upon stems, peduncles, and calyx : leaves nearly or quite smooth. — Bot. King Exp. 41 ; Porter & Coulter, Fl. Col. 14 ; Rothr. Enum. PL Col. 35. A. hirta, Wormsk. Fl. Dan. x. t. 1G46. A. propinqua, Richardson in Frankl. Journ. 738. Alsine verna, var. hirta, Fenzl in Ledeb. Fl. Ross. i. 349. A. rubella, var. hirta, Lange, PI. Groenl. 132. A. pro- pinqua, Lange, Fl. Dan. xvii. t. 2903. A. hirta, Warming, Bot. Foren. Festskr. 1890, 229. — From Greenland to Alaska, southward to Smugglers' Notch, Vt. (Pringle), and along the Rocky Mts. to Arizona (Lemmon) ; also in the San Bernardino Mts. ( Wright). Var. rubella, Hook. f. Depauperate, minutely glandular-puber- ulent or very rarely smooth : peduncles and sepals purplish tinged, the latter less strongly nerved. — Journ. Linn. Soc. v. 82. A. Giesekii, Hornem. Fl. Dan. ix. t. 1518. A. hirta, var. glabrata, Cham. & Schlecht. Linnaea, i. 56. Alsine rubella, Schrenk in Fenzl, 1. c. i. 349. A. verna, var. glacialis, Fenzl fide Wats. Bibl. Index, 99. — Occur- ring with and often scarcely to be distinguished from the preceding. A. Rossii, Richardson. Dwarf and closely tufted, glabrous : leaves crowded, narrowly linear, 3-edged, obtusish, slightly fleshy : stems many, |— 1^ inches long, filiform, usually ending in a solitary peduncle, more rarely branched and several-flowered : sepals attenuate, slightly fleshy, not at all rigid and scarcely or not at all ribbed, 1— 1£ lines long: petals oblong, nearly equalling the calyx, often minute or none: capsule shorter than the calyx. — Richardson in Frankl. Journ. 738 ; R. Br. in Parry's 1st Voy. App. 272 ; Hook. Fl. Bor.-Am. i. 100 ; Torr. & Gray, Fl. i. 181 ; Porter & Coulter, Fl. Col. 14. A. elegans, Cham. & Schlecht. Linnaea, i. 57. A. stricta, Wats. Bibl. Index, 98, in part, not of Michaux nor Wahlenberg's Alsine stricta of the Old World, which is surely distinct. Alsine Rossii, Fenzl, Verbr. Alsin. 18. — Mountains of Colorado, Hall $• Harbour, Coulter, Wolf; Wyo- ming, Parry ; British America, Bourgeau, to Arctic America. A. Nuttalli, Pax. Glandular-puberulent or tomentulose through- out : root single, vertical, rather stout : stems many, loosely matted and much branched near the base; branches ascending or erect, leafy : 304 PROCEEDINGS OP THE AMERICAN ACADEMY. leaves subulate-acerose, rigid, pungent, tending to be squarrosely spreading, connate, 3-4 lines long : flowers usually numerous in spreading cymes, rarely subsolitary : sepals attenuate, acuminate, often purplish, not strongly nerved, 2-2i- lines long, exceeding the more or less pointed petals and ovoid capsule. — Pax in Engler, Jahresb. xviii. 30. A. purtgens, Nutt. in Torr. & Gray, Fl. i. 179 (not of Clem.) ; Wats. Bot. King Exp. 40. — Rocky Mts. from Wyo- ming to S. British America, westward in mountainous regions to Washington and 8. California. "Var. gracilis. Sepals narrow, elongated and still more attenuate, 2|-3 lines long : leaves less rigid, scarcely spreading or pungent. — A. punyens, var. gracilis, Gray in herb. — California Mts. above Big Tree Grove, Bolander, 4976; Long Meadow, Tulare Co., Palmer, Coville fy Fanston. Intergrading with the typical form. ****** Densely cespitose perennials with acicular or awl-shaped leaves: sepals oblong, or linear-oblong, very obtuse. +- Alpine, boreal or arctic species. *+ Petals oblong or narrowly obovate. A. Sajanensis, Willd. Cespitose : stems finely but rather densely glandular-hirsute, decumbent, very leafy below and with age sheathed at the base with the dried persistent leaves ; the upper more or less erect portion of the stems £-2L inches in length, bearing two or three pairs of short and rather distaut more or less puberulent leaves, and terminating in 1 to 3 flowers ; lower leaves linear, obtusish, rather rigid, erect, 2— 3J lines long, quite glabrous or ciliolate, less commonly glandular-pubescent, straight : segments of the calyx linear- oblong, 1-3-ribbed, glandular-pubescent, 2 lines in length : petals spatulate, equalling or half exceeding the sepals, rarely nearly twice as long (but narrower than in A. arcticd) : valves of the capsule linear-oblong, obtuse, often considerably exceeding the calyx. — Willd. in Schlecht. Berl. Mag. Natf. 1816, 200; DC. Prodr. i. 408. A. thymifolia, James, Cat. 181. A. obtusa, Torr. Ann. N. Y. Lye. ii. 170. A. bi/lora, Wats. Bibl. Index, 94, not of Linn. A. arctica, and vars. of various authors, not of Stev. Stellar ia bijlora, L. Spec. 422. Alsine bijlora, Wahlenb. Fl. Lappon. 128; Fenzl in Ledeb. Fl. Ross, i. 355. — Mt. Albert, Lower Can., Allen, Macoun, to Labrador and Behring Strait, southward to Oregon, Cusick, and along the Rocky Mountains to New Mexico, Parry, and Arizona, Letnmon. (Green- land and Siberia.) A common species widely distributed in alpine and arctic regions of the Old and New World. Its confluent varieties and forms seem largely due to individual environment. The follow- ROBINSON. — ALSINE^. 305 ing only need be mentioned : var. rigidula (Alsine bijlora, var. rigi- dula, Fenzl, I. c), with leaves erect, firm in texture and rather closely imbricated ; and var. caknosula (Alsine bijlora, var. carnosula, Fenzl, 1. c), more flaccid, with leaves spreading and slightly flesliy. ? A. laricifolia, L. Slightly woody and much branched at the base : stems clothed with linear acicular secund ciliolate-denticulate leaves : fertile branches erect, simple, 4-7 inches in height, 2-5-flow- ered : sepals 3^ lines in length, linear-oblong, 3-nerved : petals oblong or narrowly obovate, entire, twice as long as the calyx. — Spec. 424. — An alpine European species at various times reported from Alaska but still somewhat doubtful. Plants recently collected upon the Por- cupine River by J. H. Turner certainly possess much resemblance to the European plant, but differ in their shorter sepals and less leafy stems. It is not unlikely that they may prove merely a tall and long- petalled form of the preceding polymorphous species. ** -M- Petals broadly obovate, much exceeding the calyx : Alaskan. A. arctica, Stev. Stems 1-3 iuches long, glandular-pubescent : lower leaves narrow, linear, obtuse, slightly fleshy, crowded upon the bases of the stems, nearly or quite glabrous, sometimes slightly ciliated near the base, half a line in breadth ; the upper leaves a little broader, and more or less distant : flowers solitary, terminal upon slender glandular-pubescent peduncles, 5-7 lines in diameter: capsule 3i— 4 lines long, considerably exceeding the sepals : seeds minutely rough- ened and slightly crested. — Stev. in DC. Prodr. i. 404 ; Cham. & Schlecht. Linnoea, i. 54; Hook. Fl. Bor.-Am. i. 100 (excl. vars.) ; Torr. & Gray, Fl. i. 181 in part. Alsine arctica, Fenzl, Verbreit. Alsin. 18, & Ledeb. Fl. Ross. i. 355 ; Regel, Ost-Sib. i. 338, 346 (excl. var. breviscapa). — Western and Northern Alaska and adjacent islands. A. macrocarpa, Pursh. Stems 2-4 inches long, covered except near the ends with the densely imbricated lance-linear obtuse conspic- uously ciliated leaves ; the latter \ of a line broad : flowers solitary, terminal, often exceeding \ inch in diameter : valves of the mature capsule fully 6 lines in length; seeds slightly margined. — Fl. 318; Cham. & Schlecht. Linnsea, i. 55; Hook. Fl. Bor.-Am. i 101 ; Torr. & Gray, Fl. i. 182, 675. A. arctica, var. /3, grandijlora, Hook. Fl. Bor.-Am. i. 100, t. 34, f. B. Alsine macrocarpa, Fenzl, Verbreit. Alsin. 18 ; Regel, Ost-Sib. i. 354, t. 8, f. 6-9 ; A. arctica, var. breviscapa, Regel, 1. c. i. 347. Regel's elaborate subdivision of the Siberian forms of this species is not warranted in America in the absence of abundant fruiting material. — Western Alaska near the coast. (Siberia.) vol. xxix. (n. s. xxi.) 20 306 PROCEEDINGS OF THE AMERICAN ACADEMY. -t- -t- Species of the Atlantic and Gulf States, neither arctic nor alpine. A. Caroliniana, Walt. Stems several to many, glandular- pubescent and viscid above. 3-8 inches in height, densely leafy near the base: leaves linear-subulate, rigidulous, pungent, triangular in section, channelled above ; the lower imbricated and more or less squarrosely spreading ; the upper reduced, distant : cymes few- flowered ; pedicels slender, ascending : sepals oval, 1^ lines in length : petals broad, rounded at the apex. — Car. 141 ; Wats. & Coulter in Gray, Man. ed. 6, 85. A. squamosa, Michx. Fl. i. 273 ; Ell. Sk. i. 520 ; Ton-. Fl. N. Y. i. 95. A. imbricata, Raf. in Desv. Journ. Bot. (1808), i. 229. A. Rafinesquiana, Seringe in DC. Prodr. i. 409. Alsine squarrosa, Fenzl ex Gray, Man. ed. 2, 57 ; Gray, Gen. ii. 34, t. Ill ; Chapm. Fl. 49. — Pine barrens, S. New York to Florida. 12. SAG-IN A, L. Pearl wort. (Name from the Latin sagi- nare, to fatten ; the plants though small and delicate sometimes grow abundantly in otherwise barren regions and are grazed by cattle.) Low slender herbs commonly cespitose with filiform stems and subu- late or filiform leaves ; about a dozen species inhabiting the temperate and frigid parts of the northern hemisphere ; one being also widely distributed in the southern hemisphere. — Gen. n. 33 G ; DC. Prodr. i. 389 ; Gray, Gen. ii. t. 109. * Very slender, 2-5 inches high : the almost capillary stems several to many, subsimple from near the base, usually several-flowered : the lowest flowers distinctly axillary : leaves nearly filiform but flattened above, not pro- liferous in the upper axils nor forming sterile lateral rosettes ; the basal rosette seldom persisting : flowers very small, 4- or 5-parted. S. apetala, L. Very slender and commonly glandular-pubescent: stems not numerous, procumbent or nearly erect : leaves very slen- der, H-3 or 4 lines in length, scarcely flat: pedicels straight; flow- ers normally 4-parted ; petals obovate. — Mant. ed. 2, 559 ; Fenzl in Ledeb. Fl. Ross. i. 338; Torr. & Gray, Fl. i. 177. S. procumbens, var., Benth. Brit. Fl. 120. — Labrador, Allen, to Pennsylvania, infrequent but locally abundant in dry situations. A form with elon- gated capillary stems is abundant in grassy situations near Hewitt's, Bergen Co., N. J., Britton. Introduced at Auburn, California, Mrs. Ames. Alsinella ciliata, Greene, which is ambiguously character- ized in the Flora Francis. 126, as a very slender and diffuse plant of compact habit, does not differ in its described characters from this species. S. decumbens, Torr. & Gray. Annual, quite smooth or the younger parts slightly glandular : stems several, decumbent or sub- ROBINSON. — ALSINEjE. 307 erect, 2-5 inches high, subsiinple : the filiform straight peduncles exceeding the narrowly linear very acute leaves : flowers normally 5-parted : calyx appressed even in fruit, obtusish but not rounded at the base, § the length of the capsule valves : petals scarcely equalling the sepals: stamens 10. — Fl. i. 177. S. procumbens, Pursh. Fl. 119. S. Elliottii, Fenzl ex Gray, Man. ed. 2, 61. aS. subulata, Torr. & Gray, Fl. i. 178, not of Wimm. ? S/jergula nodosa, Walt. Car. 142. S. saginoides, Michx. Fl. i. 276, not of Linn. S. decumbeus, Ell. Sk. i. 523. S. subidata, Hook. Fl. Bor.-Am. i. 93. — New England to Great Plains of British America (3Iacoun), southward to Florida and Texas ; on dry, sandy ground. Var. Smithii, Wats. More slender : flowers apetalous, at least all the later ones. — Bibl. Index, 105; Wats. & Coulter- in Gray, Man. ed. 6, 89. S. subulata, var. Smithii, Gray, Man. ed. 5, 95. Unsatisfactorily distinguished from S. apetala by its 5-parted flowers. S. occidentalis, Wats. Annual, glabrous, with habit and foliage of the preceding species but with longer pedicels (usually 7-10 or 12 lines) and larger also 5-parted flowers: capsule 1£ lines in length: calyx rounded at the base. — Proc. Am. Acad. x. 344. S. procum- bens, Boland. Cat. 6 ; Torr. Bot. Wilkes Exp. 242 ? & Linncei, Gray, Proc. Am. Acad. viii. 378. Alsinella occidentalis, Greene, Fl. Francis. 125. — Vancouver's Island to Southern California, low grounds and salt marshes of the coast. The western equivalent of S. decumbens, possibly intergrading with that species. S. procumbens, L. Matted : the numerous procumbent leafy stems l|-4 inches in length: leaves smooth or ciliate, narrowly linear, obtusish aud mucronate: pedicels filiform elongated, nodding at the summit during anthesis : flowers normally 4-parted: petals considera- bly shorter than the sepals ; the latter spreading in fruit. — Spec. 128 ; Torr. & Gray, Fl. i. 177. — Moist rocks, also in paths, etc., Nova Scotia to Pennsylvania, also rarely inland as far as Michigan, Hill; flowering through the summer. (Eu., Asia, S. Amer.) Apetalous forms have been noted in the Old World. * * Stems very short, \- 2 inches long; flowers rather small, 5-parted, terminal: leaves thickish, narrowly linear to subulate, not proliferous in the upper axils but almost always forming lateral sterile rosettes about the base. S. Linnaei, Prf.sl. Matted, 1-3 inches high : stems slender, de- cumbent, rooting and often producing lateral rosettes : radical leaves narrowly linear, acutely mucronate, 3-7 lines long, forming dense and mostly persistent rosettes ; cauline leaves short, few : pedicels long, filiform, commonly recurved at the summit: flowers moderately large 308 PROCEEDINGS OF THE AMERICAN ACADEMY. for the genus : petals not quite equalling the calyx : capsule ovate, conic, even before dehiscence considerably exceeding the sepals ; the dry valves fully twice their length : stamens 5-10. — Rel. Haenk. ii. 14; Fenzl in Ledeb. Fl. Ross. i. 339; Wats. Bot. King Exp. 41. S. saxatilis, Wimm. in Lange, PI. Groeul. 133. Spergula saginoides, L. Spec. 441. Alsinella saginoides, Greene, Fl. Francis. 125. — Labrador, Allen, to Greenland, Alaska southward along the Rocky Mts. to New Mexico; also in Mariposa Co., Calif., Congdon. S. nivalis, Lindbl. Very condensed, ^-1 inch high : leaves subu- late, or linear subulate, 2-3 (rarely 5) lines long, forming one or more dense rosettes; cauliue leaves few and short: pedicels spread- ing, 5 lines in length, straight or curved but scarcely ever hooked at the summit : petals equalling the purple edged sepals, about a line in length. — Bot. Not. 1845, 66; Hook. f. Arc. PI. 287, 322; Babington, Seem. Journ. Bot. ii. 340; Wats. Bot. King Exp. 42. S. intermedia, Fenzl, 1. c. i. 339. Arenaria ccespitosu, Vahl, Fl. Danica, xiii. t. 2289. — A rare plant, first collected in America by Dr. Watson in the Uinta Mts. in 1869 (Herb. Dep't of Agric.) ; since found in Alaska, without the exact locality, Dad ; Kyska Harbor, Harrington ; and also in the Rocky Mts. of Colorado near Gray's Peak, Patterson. The species has been regarded by some authors, and perhaps rightly, as a boreal or high alpine form of the preceding. * * * Distinctly fleshy : stems not filiform, more or less branched, several- flowered : flowers 5-parted : species of the Pacific coast. S. crassicaulis, Wats. Smooth : stems several or many, branching, 1^—5 inches long: leaves linear, pungent, thickish, 2|— 7 (rarely 12) lines long ; the basal forming a rosette which may persist or not; the cauline connate by broad searious membranes: pedicels numerous, straight : petals and sepals subequal, \\ lines in length: capsule \~ \ longer. — Proc. Am. Acad, xviii. 191. Alsinella crassicaulis, Greene, Fl. Francis. 125. — Beaches of Pacific coast, Marin Co., Calif., Congdon; Monterey Co., Michener fy Bioletti ; II- waco, Washington, Henderson. Distinguished from the Japanese S. maxima, Gray, by its glabrous peduncles and calyx. * * * * Stems simple, 2-6 inches in length ; upper leaves short, proliferous, i. e. bearing fascicles of minute leaves in their axils: flowers 5-parted: petals exceeding the calyx : species of the Atlantic Slope and Great Lakes. S. nodosa, Fenzl. Perennial : stems several to many, decum- bent, rooting at the base, often 5-6 inches in length: lower leaves filiform ; the upper subulate, only a line in length, bearing a tuft of undeveloped leaves in the axils, thus giving a nodose appearance to ROBINSON. — ALSINE.E. 309 the slender stems : flowers terminal, large for the genus, 4 lines in diameter when expanded. — Verbr. Alsin. t. 18, & Ledeb. Fl. Ross, i. 340. Spergula nodosa, L. Spec. 440. — Moist sandy soil. Coast of Maine south to Cape Ann, J. Robinson ; also on both shores of Lake Superior and northward to Hudson Bay. 13. SPEE/G-UL.ARIA, Pers. (Name, a derivative of Sper- gula.~) Annuals or perennials, usually of maritime and saline habitat, with narrowly linear, often fleshy leaves. A genus of moderate size but difficult, through the natural variability of certain species, the incon- stancy of characters (such as the form of the seeds) which elsewhere are most trustworthy, and finally through an unfortunate complication in the synonymy, arising both from the most diverse views as to the number and proper limitation of the species and from the differences in the choice of the generic name. Space does not here permit any complete discussion of the last point. It may, however, be said that the name Spergularta is the only one which has ever attained a wide use, having been employed in such excellent popular manuals as Hooker & Bentham's Handb. Brit. Fl., Hooker's Fl. Brit. Isl., Garcke's Fl. Deutsch., Gray's Man., eds. 1-5, Cosson & Germain's Synop. Analyt., etc. Beside this matter of popular usage of the past, there is at present the best authority for selecting this name, since it is the one retained by the Kew botanists, and is advoeated in the Berlin recom- mendations. On the other hand, the more radical reformers of nomen- clature have attained no unanimity in regard to the proper name for the genus ; Professors Baillon, Greene, and Britton adopting l^issa. Dr. Kuntze Buda, and Mr. N. E. Brown Corion, while Prof. Pax, who in his Monograph of the Caryophyllaceoe in Engler & Prantl's Nat. Pflanzenf. used Ti'ssa, in a later publication has returned to Spergu- laria. — Syn. i. 504; Gray, Gen. ii. 27, t. 108; Benth. & Hook. Gen. i. 152. Arenaria, L. Gen. n. 374 in pirt. Corion, Mitchell, Act. Phys. Med. Acad. Nat. Cur. viii. App. 218, fide Britton in Brit- ten's Journ. of Bot. xxix. 303 ; N. E. Brown, Suppl. Eng. Bot. Tissa, Adans. Fain, des PI. ii. 507; Baillon, Hist, des PI. ix. 116; Britton, Bull. Torr. Club, xvi. 125; Greene, Fl. Francis. 126, & Man. Bay Reg. 35. Buda, Adans. 1. c. i. 507 ; Dumort. Fl. Belg. 110; Wats. & Coulter in Gray, Man. ed. 6, S9. Lepigonum, Fries, Fl. Hall. 76; Kindberg, Monogr. ; Wats. Bibl. Ind. 103. # Procumbent, or decumbent, slender, scarcely at all flesby, growing near or even on the seacoast, hut not truly saline : flowers small or of medium size: petals magenta : stipules lanceolate, elongated and conspicuous. S. rubra. Presl. Usually annual, smoothish below, finely glan- dular-pubescent above : stems spreading, wiry : leaves fiat above, nar- 310 PROCEEDINGS OF THE AMERICAN ACADEMY. rowly linear, cuspidate, 4-6 lines loug, \—^ line broad : stipules white, attenuate, 2-3 lines long : inflorescence racemiform ; pedicels truly filiform, exceeding the bracts and about twice as long as the oblong- lanceolate scarious-margined acutish glandular-pubescent sepals : cor- olla 1^ lines in diameter, scarcely equalling the calyx: capsule of the same length ; seeds " semiobovate " and minutely crested but not winged. — Rel. Haenk. ii. 9 ; Gray, Gen. ii. 28, t. 108, & Man. ed. 1, 04 excl. var. S. rubra, var. campestris, Gray, Man. ed. 5, 95. S. campestris, Aschers. ex Rohrb. in Mart. Fl. Bras, xiv.2 2G7. Arenaria rubra, L. Spec. 423 excl. var.; Bigel. Fl. Bost. 108 ; Hook. Fl. Bor.- Am. i. 98. Buda rubra, Dumort. Fl. Belg. 110. Spergula rubra, Torr. & Gray, Fl. i. 175 excl. vars. Tissa rubra, Britt. Bull. Torr. Club, xvi. 127. — An attractive species, growing about paths in dry sandy soil. Maine to Virginia, and Washington State, Suksdorf ; to San Francisco, Mrs. Brandeyee. (Europe.) S. Cleveland.].. Perennial, viscid-glandular : leaves ascending, often secund, conspicuously fascicled in the axils, almost terete and filiform, very acute and attenuate, 5-10 lines in length: flowers much as in the last, but often somewhat larger : seeds winged. — Tissa vil- /osa, Britton, 1. c. xvi. 129. T. Clevelandi, Greene, Fl. Francis. 127. T. rubra, K. Brandegee, Zoe, iv. 84. — Sandy soil, California, San Diego, Cleveland, Mrs. Brandegee / San Jose, Mrs. Bush ; and at the Presidio, San Francisco, according to Prof. Greene. A doubtful species with close affinities, as Mrs. Brandegee suggests, to S. rubra, but not readily identifiable with any form of that species, as it seems to differ in its longer, more terete and acute leaves, and winged seeds. From the South American S. villosa it differs in its lower growth, distinctly smaller flowers, shorter pedicels, and somewhat firmer and less flaccid leaves. * * Slender spreading or erect annuals of the West and Southwest, scarcely fleshy, and with short deltoid stipules. •<- Corolla pink (or white ?) more than half as long as the sepals. S. diandra, Boiss. Viscid-pubescent to nearly glabrous : leaves not fascicled, linear-filiform : pedicels slender, about 2 lines long, spreading or deflexed : sepals in fruit 1^ lines long, but little exceeded by the capsules : stamens usually but 2-3. — Fl. Orient, i. 733. Are- naria. diandra, Guss. Prodr. Sic. i. 515. Tissa diandra, Britton, Bull. Torr. Club, xvi. 128. — Texas, Druinmond, Lindheimer, Goshen, Calif., Mrs. Brandegee, Oregon, Henderson, and on sandy banks of the Columbia River in Klickitat Co., Washington, Snksdorf. In its wiry procumbent stems this species also suggests S. rubra, from which, however, it differs in its short deltoid stipules. ROBINSON. — ALSINE.E. 311 S. gracilis. Low, glabrous, 2-6 inches in height, diffusely branched : leaves ^-1 inch in length : stipules deltoid : flowers small, subglobose, 1-1 £ lines in diameter, not closely aggregated; pedicels 2-3 lines long : sepals elliptic ovate, a line or less in length, thick in the middle but scarious-margined : valves of the capsule a third to a half longer than the sepals ; seeds angled, somewhat triangular in out- line, finely but distinctly roughened. — Lepigonum gracile, Wats. Proc. Am. Acad. xvii. 367. 7issa gracilis, Brittou, Bull. Torr. Club, xvi. 128. — Sandy ground, dried ponds, etc. Dallas, Texas, Revere/ton, to S. California, Parry, Nevius, Orcutt. S. tenuis. Dichotomously much-branched, becoming 8-10 inches in height, somewhat glandular-puberulent or pubescent above : leaves 6-10 lines long : the very numerous flowers somewhat larger than in the last and inclined to be closely grouped; capsule twice the length of the ovate-oblong sepals. — Lepigonum tenue, Greene, Pittonia, i. 63. Tissa tenuis, Greene ex Britton, Bull. Torr. Club, xvi. 128. T. diandra?, K. Brandegee, Zoe, iv. 84. — California near Alameda, Greene, Mt. Eden and Tulare, Mrs. Brandegee. The affinities of this species are almost equally with S. salina, Presl, and S. diandra, Boiss., and in a large series of specimens it may not be possible to distinguish it sharply from either. The flowers, however, are more numerous and smaller and the corolla less developed than in the former ; the habit also is much more branching. From S. diandra it differs in the usually much shorter pedicels of the closely aggregated and at first subcurymboseiy arranged flowers. * * * Annuals or biennials, more decidedly fleshy, usually of maritime or saline habitat : flowers small or of medium size: corolla more or less con- spicuous, white or pink, less frequently pink-purple : stipules ovate or deltoid. S. salina, Presl. Commonly although not always pubescent: leaves often fascicled in the axils : sepals ovate to oblong-lanceolate, narrowed upward although obtuse at the summit, 2-2 i lines long : petals pink : capsule equalling or a third to a half longer than the calyx : seeds turgid, obovate, usually roughened, less frequently nearly or quite smooth (var. leiospernnim, N. E. Brown, Eng. Bot. Suppl.). — Fl. Cechica, 95; Gray, Man. ed. 5, 95; Warming, Bot. Foren. Festkr. 1890, 238, fig. 20. S. Canadensis, Don, Mill. Diet. i. 426. S. rubra, var. marina, Gray, Man. ed. 1, 64. S. media, and var. macrocarpa, Gray, Man. ed. 5, 95. Arenaria rubra, var. marina, L. Spec. 423. A. marina, Bigel. Fl. Bost. 109; Roth in Ilornem. Fl. Dan. xiii. t. 2231. Buda marina, Dumort. Fl. Belg. 110. LepigO' 312 PROCEEDINGS OP THE AMERICAN ACADEMY. num salinum, Fries, Mant. iii. 34. L. marinum, Kindberg, 1. c. 18. L. medium, Wats. Bibl. Index, 103 in part. Tissa marina, Britton, 1. c. xvi. 126. — Common on both the Atlantic and Pacific coasts, also occurring upon the Gulf coast, and not infrequent about salt lakes and in alkaline regions of the interior, especially westward. "Var. ? minor. Smaller, 2-3 inches high : flowers smaller and very numerous, on short pedicels (J-2 lines in length) and conse- quently rather densely aggregated. — Buda marina, var.? minor, Wats. in Gray, Man. ed. G, 90. — Coast of New Hampshire and Massachu- setts. An ambiguous form suggesting the western S. tenuis but smaller and with a better developed corolla. S. borealis. More slender and in well developed specimens more diffusely branched than the preceding, 2-5 inches high, glabrous : leaves opposite, seldom fascicled ; stipules ovate, broader than long, obtuse or obtusish : sepals ovate, 1-1^ Hues long, very obtuse : petals shorter, white or pink : capsule ovate-oblong, usually nearly or quite twice as long as the calyx ; seeds usually wingless and nearly or quite smooth. — Armaria rubra, (3, Michx. Fl. i. 274. (Dr. Britton, who has examined the type of Michaux's variety, pronounces it identical with this species.) A. Canadensis, Pers. (Syn. i. 504), the oldest specific name, but not to be selected for use under Spergularia, since S. Canadensis has been used by Don (Mill. Diet. i. 426) for a " pilose " and " rather hispid " plant, extending from " Canada to Carolina " and being doubtless S. salina, Presl ; Leplgonum medium, Wats. Bibl. Index, 103 in part. Tissa salina, Britt. 1. c. xvi. 127. Buda borealis, Wats. & Coulter, in Gray, Man. ed. 6, 90. — Sea-beaches and tidal marshes, Maine to Labrador. * * * * Stout and fleshy perennials of the Pacific slope : flowers large. S. macrotheca. Smooth to densely glandular-tomentose : stems spreading, ascending, 8-15 inches in height; leaves linear, acute, mucronate, 8 lines to 2 inches in length, about a line in breadth; internodes more or less developed, usually |— 1 inch long : floral bracts resembling the leaves : inflorescence inclined to be racemiform ; pedi- cels 4-12 lines in length, spreading or more or less deflexed : sepals lanceolate, acutish or somewhat attenuate to an obtuse point, thick in the middle, nearly smooth or viscid-glandular, conspicuously mem- branous-margined : petals roseate, shorter than the sepals : capsule oblong-ovoid, acutish, nearly equalling the sepals. — Arenaria macro- theca, Hornem. ex Cham. & Schlecht. Linnaaa, i. 53. Lepigonum macrothecnm, Fisch & Mey. Ind. Sem. Petr. iii. 14 ; Kindberg, Monog. 16, t. 1, f. 1 ; Wats. Bibl. Index, 103. L. Chllense, Fisch & ROBINSON. — ALSINE^. 313 Mey. 1. c. Spergularia rubra, Torr. Pacif. R. Rep. iv. 70. — Cali- fornia, chiefly on or near the coast. A polymorphous species, the varieties of which, although diverse in aspect, appear in a large series of specimens to be thoroughly connected by intermediates. "Var. leucantha. Erect or nearly so, with long internodes : leaves somewhat narrower and more erect than in the type : floral bracts reduced and inflorescence more distinctly cymose ; pedicels elongated, rather rigidly spreading or deflexed : corolla white, nearly 6 lines in diameter. — Tissa leucantha, Greene, Fl. Francis. 127. — A variety of alkaline regions of the interior. A form from Vauden (Mrs. Brandegee) , identical in habit and foliage, has magenta flowers. "Var. scariosa. Low, pale, smoothish near the base, and often very glandular-viscid above, densely leafy ; the internodes scarcely or not at all developed : leaves J— \ inch long, acute ; stipules con- spicuous, ovate-lanceolate, acuminate, 4-5 lines long : inflorescence racemiform ; pedicels not greatly exceeding the calyx. — Tissa macro- theca, var. scariosa, Britton, Bull. Torr. Club, xvi. 129. Tissa pallida, Greene ex Britton, 1. c. xvi. 129; Fl. Francis, i. 127. — Coast at Monterey, California, Torrey ; Hooker fy Gray, 1877; Gray, 1885; and at Fort Point, Brandegee. Tissa valida, Greene, Erythea, i. 107, appears to be a firmer and more erect form of the same thing, also pale and very viscid, but with more elongated internodes and dis- tinctly dichotomous cymose inflorescence. — Island of Santa Cruz, according to Professor Greene. SPERGfULiA, L. Spurry. (Name from the Latin spargere, to strew, in reference to the scattering of the numerous seeds.) — Annuals with narrowly linear slightly fleshy apparently whorled leaves ; one species common in America, having probably been introduced with grain from the Old World. — Gen. n. 875 ; Reichb. Iconogr. vi. t. 511-513. S. arvensis, L. A foot or two high : leaves numerous in rather remote whorls : inflorescence a terminal naked spreading cymose panicle ; pedicels often deflexed in fruit : petals white, equalling or slightly exceeding the sepals, 2-2i lines long : capsule ovate-globose ; seeds black minutely roughened with light-colored papillae, acutely edged but scarcely winged. — Spec. 440; Walt. Car. 142; Eng. Bot. xxii. t. 1535; Pursh, Fl. 320; Hook. Fl. Bor.-Am. i. 92; Torr. & Gray, Fl. i. 174 ; Rothr. PI. Alask. 444. S. ramosissima, Dougl. ms. according to Torr. & Gray, 1. c. — Grain fields and cultivated grounds, common, United States and Canada, northward to Alaska. (Intro- duced from the Old World.) 314 PROCEEDINGS OF THE AMERICAN ACADEMY. II. — DESCRIPTIONS OF NEW AND HITHERTO IM- PERFECTLY KNOWN PLANTS COLLECTED IN MEXICO BY C. G. PRINGLE IN 1892 AND 1893. Eriodendron tomentosa. Branchlets and petioles fuscous-to- mentose ; leaves 7-fuliate ; leaflets narrowly lanceolate, acuminate in both directions, sharply and finely serrate, sparingly hirsute upon the veins above, fuscous-tomentose beneath especially along the midrib, attenuate at the base to short scarcely distinct very tomentose petiolules : common petiole 2-3 inches long ; stipules subulate, tomen- tose, persistent : peduncles short, thick : calyx turbinate-campanulate, 1J-1J inches long, finely tomentose on the outer surface, densely woolly within ; teeth rounded, a fourth as long as the tube : petals red, linear-oblong, spreading, rounded at the apex, over 5 inches in length, 4-5 lines in breadth, silky-pubescent upon the outer surface, less pubescent or glabrate within : staminal tube 8 lines long, hirsute ; free portion of the filaments spreading, exserted from the corolla ; anthers anfractuose. ■ — ■ Collected on a barranca near Guadalajara, June, 1892 (no. 5300). This species is distinguished from Erioden- dron cBsculifolia, HBK., by its tomentose branchlets and foliage. The latter species is apparently represented by Mr. Pringle's no. 5324, from hills bordering on Lake Cuitzea, Michoacan, August, 1892, which accords in every regard with Kunth's description save that the leaflets range from 5 to 7 instead of 7 to 8. It differs, however, from Mocino and Sesse's Plate 94 in the Caiques des Dessins, in having a glabrous calyx, and stamens really shorter than the petals ; leaflets a little broader and distinctly petiolulate. As the plate shows the calyx to be pubescent and foliage as in E. tomentosa, it appears probable that it was not drawn from E. cescidifolia, as hitherto supposed, but really represents a distinct species, probably the E. tomentosa just described. Ayenia glabra, Wats. The following additional characters may be given : fruit subglobose, shallowly 3-lobed, nearly 4 lines in diameter, light green, muriculately roughened with darker colored points; seeds dull brown, strongly rugose. — Collected in the barranca of Tequila, 2 and 17 October, 1893 (no. 4576). Bursera Pringlei. Wats. (Proc. Am. Acad. xxv. 145). This spe- cies was described from fruiting specimens. Mr. Pringle's no. 4381, collected on lava beds near Zapotlan, 27 May, 1893, shows the follow- ing additional characters : staminate flowers yellowish green : sepals 5, ROBINSON. — MEXICAN PLANTS. 315 ovate, acute, | line in length : petals 2 lines long, with involute margins : stamens 10 ; anthers linear-oblong, twice the length of the filaments : disk deeply crenate. Nissolia confertiflora, Wats., var. laxior. Agreeing with the type in foliage and technical characters, but having flowers nearly twice as large (3|-4 lines in length) upon more elongated pedicels. — Col- lected in the barranca of Beltran, 5 June, 1893 (no 4379). JEschynojiene amorphoides, Rose in litt. Brya (?) amor- phoides, Wat*. (Proc. Am. Acad. xxii. 406). Excellent flowering and fruiting specimens of this plant, collected by Mr. Pringle in the barranca of Tequila, 1 October, 1893 (no. 4613), fully con firm the view expressed in a letter from Dr. J. N. Rose, that it, should be referred to the genus jEschynomene. The stamineal tube early splits on the ventral or lower side, and the pods are in all respects those of an JEschynomene. Clitoria (?) skricea, Wats. (Proc. Am. Acad. xxii. 407). Ad- ditional specimens collected on rocky hills near Guadalajara, June and July, 1893 (no. 4449), show the flowers of this species not hitherto described. Calyx short, campanulate, silky-villous, with narrow acu- minate spreading teeth equalling the tube: corolla light blue, villous upon the outer surface, 6-7 lines long; petals suhequal in length; the standard erect, open, orbicular, slightly thickened at the insertion of the claw ; ahe oblong, abruptly contracted to a short claw : style bearded upon the upper surface : stamens all united. — An anomalous plant with the calyx rather of Centrosema but the bearded style and other characters of Clitoria. Calogania pulchella, HBK., var. racemosa. Leaflets broadly ovate, acutish : all or nearly all of the flowers borne in leafless axillary racemes (4-6 inches in length). — Collected by Dr. Edward Palmer at Tequila, Jalisco, August to September, 1886 (no. 379), and by Mr. Pringle near Guadalajara, 15 September, 1893 (no. 5164). Galactia multiflora. Stems slender, woody, grayish brown, much branched, ascending, not twining: twigs, petioles, and peduncles cinereous-tomentose : leaves pinnately trifoliate ; leaflets oval or some- what obovate, rounded or retuse at the apex, dull green and glabrate above, ciuereous-tomentose beneath; veins slightly prominent beneath ; the lateral veins 7 to 9 pairs, parallel ; margin slightly thickened and revolute; terminal leaflet 15 lines in length, two thirds as broad; the lateral slightly smaller: racemes axillary, 2-3 inches in length: buds acutely lanceolate in outline: calyx purplish, silky-villous, deeply 4- cleft, 4 lines in length ; the upper segment oblong, acuminate ; the 316 PROCEEDINGS OF THE AMERICAN ACADEMY. others linear or lance-linear, attenuate: corolla (in a dried state) deep purple, striated with green, 5-6 lines in length ; standard obovate, entire at the base ; alae oblong, strongly auncled upon the upper edge ; keel but slightly curved : stamens alternately shorter, one free. — Collected on rocky hills about Tequila, 1 July, 1893 (no. 4433). Fruiting specimens belonging to this species and collected in the same locality, 24 September, 1S93, have leaflets 2-2^ inches in length, and oblong acuminate pubescent legumes. Eriosema multielorum. A shrub, 3-4 feet high: stems brown- ish, furrowed-angulate, branched ; branches covered with fine silky reflexed appressed pubescence, stipules ovate, acuminate, 2|— 3 lines long, silky pubescent, tardily deciduous: leaves ascending, 3-foliate ; common petioles 1-2 lines long ; leaflets elliptic-oblong, obtuse, often apiculate, pale green, firm in texture, 3-nerved, finely pubescent ahove, appressed silky-villous beneath especially upon the prominent veins, 1^—2 inches long, a third as broad ; the edges revolute : racemes crowded at the ends of the branches, subsessile, densely flowered, 1^—2^ inches long : bracts caducous ovate-rhombic, acuminate, silky, 2-3 lines long, somewhat exceeding the ascending pedicels : calyx finely pubescent, campanulate, rounded at the base, gibbous upon the upper side, parted to the middle ; the segments lanceolate, attenuate ; corolla yellow, 6 lines in length ; standard with a slight bifornicate thickening just below the middle upon the inner surface; vexillar stamen free and with a bulbous thickening at the base : pods (scarcely mature) broadly oblong, shortly and obliquely acuminate, 7-8 lines in length, rufescent silky-villous, 2-seeded. — Collected on rocky hills near Tequila, 7 October, 1893 (no. 4548). This species stands near E. gra-ndiflorum, Seem., but differs in its smaller, more numerous, and densely racemed flowers. Mimosa Tequilana, Wats. Excellent fruiting specimens found by Mr. Pringle on hillsides near Tequila, 26 September, 1893 (no. 4596), furnish the following characters not shown in the type: pods about 6 in a head, flat, thin, oblong, acute, about 3-seeded, f inch long, 2-| lines wide, covered with weak transversely spreading prickles. Begonia palmaris, A. DC. Mr. Pringle has been fortunate in rediscovering this species and securing the hitherto unobserved fertile plant. His excellent specimens, kindly identified by Monsieur Autran with the type in Herb. Boissier, furnish the following supplementary characters : plants apparently dioecious, 2-4 feet high : lower leaves much larger than the upper (hitherto described), becoming 8 or 9 inches broad : petioles 2 inches in length .- lobes of the fertile flowers ROBINSON. — MEXICAN PLANTS. 317 white, broadly ovate, acute, sharply denticulate, smooth or sparingly pubescent upon the outer surface ; the three outer ones 9 lines long, 6 lines broad ; the two inner somewhat smaller : pistil of § Kuesebeckia, A. DC. ; ovary pubescent; capsule 8 lines long, 3-winged ; wings un- equal, broadest near the base. — Fruiting specimens collected on a barranca near Guadalajara, 23 September, 1891 (no. 5178); flowering plants of both sexes were collected in rich soil on shaded rocky slopes of barranca near Guadalajara, 29 July, 1893 (no. 4474). Randia Watsoni. (R. tomentosa, Wats. Proc. Amer. Acad, xxv. 152, not of Blume.) Mr. Pringle's no. 4383, collected on lava beds near Zapotlan, 20 May, 1893, corresponds to the type of Dr. Watson's species except in its somewhat more obtuse leaves, and doubtless should be referred to tliis species. It furnishes the follow- ing additional characters : buds clavate, an inch in length, curved upward: calyx silky, 3 lines in length; segments short, linear; sinuses rounded: corolla tube 9 lines long, hirsute upon the ouler surface; limb abruptly spreading, nearly or quite an inch in breadth ; segments oval : anthers oblong, apiculate, sessile in the oritice of the corolla : stigma 2-lobed. Valeriana Palmeri, Gray. Mr. Pringle's specimens of this plant show that the root terminates in a small oblong tuber-like en- largement about half an inch in length. — Collected on shaded banks near Guadalajara, 2 September. 1893 (no. 4521). Stevia skrrata, Cav., var. ovai.is. Leaves short, oval, scarcely narrowed at the base, 10-12 lines long, 5-7 lines broad. — Collected by Dr. Edward Palmer on the Rio Blanca, Jalisco, July, 1886 (no. 309), and by Mr. Pringle on rocky hills near Guadalajara, 28 August, 1893 (no. 4491). Mr. Pringle's no. 602, from Chihuahua, 1886, is nearly the same. Guardiola rotundifolia. Glabrous : stem rigid, brown, finely furrowed : leaves thickish, ovate-orbicular, closely sessile and clasping, remotely denticulate, cuspidate, glaucous, 1^—2 5- inches in diameter: inflorescences about 3-headed, closely subtended by floral leaves : scales of the involucre 4-5 lines long, finely striated, ovate, narrowed to an obtuse point : achenes 2-J- lines in length, mottled, truncate, contracted below to a thick short stipe: corollas in a dried state greenish. — Col- lected on hills near Tequila, 15 October, 1893 (no. 4571). Near G. platyphytta, Gray, but with larger, more orbicular, less dentate, and more clasping leaves ; fewer larger heads, with broader involucral bracts. Liabum cervinum. Arachnoid-tomentose, 2 feet in height: stem 318 PROCEEDINGS OP THE AMERICAN ACADEMY. simple, leafy to the summit, furrowed, springing from large tuberous- tlnckened and clustered roots: leaves 5-6 inches in length, glabrate above, white woolly beneath except upon the nerves, narrowly connate, deeply and palmately cleft into linear attenuate one-nerved spinulose- denticulate segments; the latter 2-3 lines in breadth, with revolute margins : peduncles and involucres densely fuscous-tomentose : heads corymbose, about 6, discoid, an inch in height: scales of the involucre lance-linear ; the outer shorter : corollas yellow, 5-6 lines long ; the tube hirsute: achene silky; pappus tawny, readily deciduous. — Col- lected on bluffs of barrancas, San Marcos, 9 June, 1893 (no. 4398). Near L. Palmeri, Gray, but differing in its arachnoid and not glandu- lar stem, leafy to the summit, its somewhat larger heads and more imbricated involucre, as well as in the more acute segments of the leaves, which in their mode of forking much resemble stag-horns. Asclepias Jaliscana. Stem simple or sub-simple, erect, setose- hirsute, a foot in height : leaves sessile, oval to ovate-oblong, with rounded apex, sparingly hirsute upon both surfaces, ciliolate. glaucous beneath : flowers large, greenish, in several pedunculate, rather few- flowered umbels ; peduncles 8 lines to 1 inch long, little exceeding the pedicels ; both hirsute : calyx segments lance-linear, half as long as the corolla lobes ; the latter reflexed, 5 lines in length, purplish upon the lower surface, green upon the uj^per : hoods 2^-3 lines long, scarcely at all auriculate at the inner angles ; horn broad and con- spicuously exserted : fruit tomentose, slender, fusiform. — Collected on the Rio Blanco, Jalisco, by Dr. Edward Palmer, in June, 188G (no. 20.) ; then by Mr. Pringle in plains near Guadalajara, June, 1889 (no. 3020), and again in dry soil, on plains and hills near the same city, July, 1893 (no. 4444). This species has the pubescence and much of the habit of A. setosa, Benth., with which it has been hitherto con- founded. It is to be distinguished by its broader and more obtuse leaves, more glaucous beneath, its sub-simple stem, and its somewhat larger flowers with scarcely auriculate hoods and broad horns. Hart- weg's no. 213, Bentham's type, as well as a number of other specimens correctly referred to his species, show that in it the horn is relatively slender and the hoods well auricled at the upper inner angles. Gonolobus sororics, Gray (Proc. Amer. Acad. xxii. 437), de- scribed from fruiting specimens, is now shown in flower by Mr. Pringle's no. 4435, collected on rocky bluffs above Tequila, 3 July, 1893. The following characters may be added: flowers in numerous axillary short-peduncled or subsessile umbels ; pedicels G— 8 lines long: calyx segments lance-linear, obtusish, spreading or reflexed : corolla ROBINSON. — MEXICAN PLANTS. 319 yellow, 5 lines in diameter ; segments ovate-lanceolate, acutish ; orifice covered with a conspicuous dense white beard : anther cells nearly upright : crown of five short obtuse scales, adnate to the base of the corolla. — In foliage and bearded corolla this species approaches G. barbatus, 11 BK., which, however, has larger flowers and is from the coast. Pherotrichis leptogenia. Densely hirsute : stems simple, erect, 1-2 feet in height, from a fusiform root : leaves short-petioled, obtuse ; the lower oval, smaller, subcordate ; the upper ovate to ovate- lanceolate : flowers in successive sessile intra- and super-axillary um- bels ; pedicels 3 lines long : calyx hirsute ; segments lanceolate, acute, 1% lines long: corolla 4-6 lines in breadth; its segments widely spreading, oblong-ovate, obtuse, greenish, striped with yellow, very thinly bearded or glabrate : corona equalling but not exceeding the stigma : membranous appendages of the anthers conspicuous : appen- dage of the stigma conical obtuse. — Collected on hills near Patzcuaro, 30 July, 1892 (no. 5279), and again on dry slopes near Guadalajara, 2o July, 1893 (no. 4466). Possessing the habit of P. Schaffaeri, Gray, and P. Balbisii, Gray, but differing in its somewhat larger, more open and almost glabrous corolla. Mr. Pringle's no. 1148, col- lected in the Mapula Mountains of Chihuahua, 26 October, 1886, is the same, and shows the fruit to be lanceolate in outline, very acute and hirsute, 2|- inches in length. Ehretia cordifolia. A large tree with light brown glabrate branches ; branchlets somewhat hirsute ; lenticels conspicuous : leaves thick, of harsh texture, ovate, obtusish or very shortly acuminate, serrate, cordate or subcordate, scabrous above, paler and soft pubescent beneath, 2-3 inches in length, two thirds as broad ; petioles pubescent, £ inch in length : panicles terminal, only 1-H inches in length : sepals ovate-lanceolate, a line long: fruit oblong, 4-celled, 4 lines in length, 2£ lines in diameter. — Collected in valley, Zapotlan, 19 May, 1893 (no. 4382). Of this species fruiting specimens only are at hand. They possess a strong habital resemblance to E. Mexicana, Wats., but have broader cordate leaves. Ipomoea stans, Cav., var. hirsuta. Very hirsute: leaves broad, sessile: sepals short, suborbicular. Collected by Dr. Edward Palmer on the Rio Blanco, Jalisco, August, 1886 (no. 324), and by Mr. Pringle on plains near Guadalajara, 30 August, 1893 (no. 4488). The typical form is merely puberulent, and has longer relatively nar- rower sepals. Ipomosa perlonga. Stem a little puberulent or glabrous, slightly 820 PROCEEDINGS OF THE AMERICAN ACADEMY. armed with a few small scattered prickles : leaves orbicular, deeply cordate, shortly acuminate, entire, 2-4 inches in diameter, glabrous at maturity, somewhat paler beneath ; petioles li-2i inches in length : peduncles very long, becoming nearly a loot in length, each bearing a regular rather closely 6 to 12-flowered cyme: pedicels thickened up- ward and becoming deflexed in fruit, 3-8 lines long : sepals smooth, ovate-oblong, obtuse, apiculate, in authesis but 2| lines in length, becoming enlarged in fruit: corolla funnel-shaped, 1| inches long; the tube and throat rather slender, white ; the limb bright blue (purple in dried specimen), 1 ' inches in diameter : fruit globose, apiculate. — Collected on hills about Tequila, 15 October, 1893 (no. 4531). Distinguished by its strikingly long-peduncled inflorescences and markedly bicolorous corolla. Evolvdlus pkostratus. Perennial : root perpendicular with horizontal branches : proper stem very short ; branches 4 to 8, pros- trate, simple, 4-G inches in length, silky-villous : leaves imbricated in two rows and somewhat reflexed, suborbicular, rounded at the apex, rounded or subcordate at the base, 4-5 lines in diameter, subsessile, glabrous above, silky-villous beneath : buds and fruit entirely con- cealed beneath the leaves: flowers raised between them; peduncles 1-J lines in length, equalling the calyx, both villous : corolla white, 2 lines in diameter: capsule 2-3-seeded ; seeds dull brown. — First col- lected by Dr. Thomas Coulter in Mexico without exact locality (no. 1011); then by Bourgeau in the Valley of Mexico at Santa Fe, 5 July, 1865-6 (no. 323), wrongly referred to E. holosericeus, HBK. Col- lected by Mr. Pringle on dry thin soil of hills near Guadalajara, 26 July, 1893 (no. 4445). In vegetative habit this species strongly sug- gests Lysimachia nummularia. Bassovia Donnell-Smithii, Coulter (Bot. Gaz. xvi. 145). Mr. Pringle has rediscovered this Guatemalan plant in the barranca of Beltran, 6 June, 1893 (no. 4378). The range is thus considerably extended. His specimens and field notes show that the stem is about 10 feet high, twice regularly dichotomously branched. The calyx is somewhat accrescent, and the fruit is globular, orange-red, many- seeded, and about 5 lines in diameter. Pinguicula parvifolia. Base a small loose bulb, 4 lines in diameter, provided with a number of fine fibrous roots: leaves 3 to 5, elliptic-oblong, at the time of flowering not exceeding 3 lines in length and 1]- lines in breadth ; petioles margined and becoming broad and somewhat scarious below : scape single, strict or slightly curved, very minutely glandular near the summit : the five segments of the ROBINSON. — MEXICAN PLANTS. 321 calyx subequal, lance-linear, acutish, scarcely exceeding 1-$ lines in length : corolla G-8 lines long ; the limb bluish purple ; the segments broadly oblong or somewhat spatulate ; throat yellow as well as the tapering nearly straight or moderately curved spur (2-2^- lines in length). — Collected on mossy gravel bluffs near Guadalajara, 23 Juue, 1893 (no. 4397). An attractive little species distinguished by its paucity of foliage. Vitex pyramidata. Tall shrub, 10-15 feet in height : branchlets, petioles, and inflorescences pulverulent-pubescent: petioles long, deeply sulcate above : leaflets 5, elliptic, mucronate, entire, rounded at the base, becoming decidedly coriaceous with age, green and smoothish above, with sulcate veins and minute reticulation, pale and tomentose beneath, 3-4 inches long, nearly half as broad; petiolules 3 lines long: the sulcate peduncles springing from the upper axils, and bear- ing rather dense compound pyramidal panicles, sometimes subtended by two trifoliate bracts : calyx scarcely a line in length, shortly 5- toothe J : corolla strongly bilabiate, about 6 lines long ; the tube en- larged upward, about equalling the limb; segments rounded, ovate, the two upper suberect, the lowest the largest : stamens exserted, recurved at the summit, anther cells divergent: fruit 6 lines in di- ameter, consisting of a tough exocarp surrounding an irregularly sculp- tured woody endocarp, enclosing four 1-seeded cells. — Tequila, Jalisco, August, 1886, collected by Dr. Edward Palmer ; rediscovered by Mr. Pringle on rocky hillsides about Tequila, 29 June, 1893 (no. 4429). Dr. Palmer's fruiting specimen was referred to V. mollis, HBK., by Dr. Watson (Proc. Amer. Acad. xxii. 444). He states also that the dark brown fruit is eaten by the natives under the name of " ahuilote." Mr. Pringle's flowering specimens show the plant to be amply distinct from V. mollis. The paniculate inflorescence is rare in American species although common in those of the Old World. Cytinds oxylepis. Fuscous and somewhat granular-tomentose, 2-3 inches high : the short scaly stem not greatly enlarged at the base, and not equalling in length the thick clavate inflorescence: scales both of the stem and inflorescence lanceolate or ovate-lanceolate, acu- minate, smooth, spreading, 1£ lines in length: flowers bisexual; seg- ments of the perianth 6-7, ovate, acute, granular-fuscous upon the back, a line in length, united below into a shallow cup : anthers about 6, uniseriate and adnate to the short thick style: stigma obscurely lobed ; ovaries more than half immersed in the thickened axis of the inflorescence ; placentae 5-6. — Collected on lava beds near Zapotlan, 13 and 27 May, 1893 (no. 4373). A fungoid parasite upon woody vol xxix. (x. s. xxi ) 21 322 PROCEEDINGS OF THE AMERICAN ACADEMY. roots of Burseras. The stem is 3-4 lines in diameter and the inflo- rescence is 2-3 times as thick. The species appears to be distinguished from all other plants of the genus by its perfect flowers, as well as by its small and very acute scales. Pedilanthus Pringlei. Stems smooth, alternately few-branched : leaves closely and softly puberulent upon both surfaces, lanceolate, acuminate. 1^-2 inches long, subsessile by au abruptly narrowed base; midrib prominent and white below ; bracts minute, grayish tomentose, caducous; pedicels 1|— 3 lines long: involucres acutish at the base, dark purplish red, 5 lines long; the upper lip quadri-glandular at the base inside, glabrous, abruptly bent, attenuate to a very narrow but truncate entire or slightly retuse apex ; segments of the lower lip finely ciliated, otherwise glabrous ; style slender, dark red, trifid ; cap- sule smooth, 3-3|- lines in length upon a stipe (J inch long) ; seeds ashy, ovoid, apiculate. — Collected on limestone ledges, Las Palmas, San Luis Potosi, 25 July, 1891 (no. 5107). Near P. Tithymaloides, Poir., but with smaller puberulent leaves, darker colored involucre, acutish at the base and with a more slender and attenuate upper lobe. Acalypha hypoGjEA, Wats. (Proc. Am. Acad. xxii. 451). Has been rediscovered by Mr. Pringle on damp slopes near Guadalajara, 28 July, 1893 (no. 4460). His specimens show the following addi- tional characters: stem hirsute, 4-6 inches in height: the largest leaves 15 lines in length: staminate spikes very small. 1-2^ lines long, upon slender axillary often deflexed peduncles (a line in length). Acalypha polystachya, Jacq. A. filifera, Wats. (Proc. Am. Acad. xxii. 451), has been again found by Mr. Pringle in Jalisco (no. 4470). As Dr. Rose notes, this species is not to be distinguished from A. polystachya, Jacq., represented by a rather wooden plate in the Hortus Schonbrunensis. Liparis Galeottiana, Hemsl. This species, insufficiently char- acterized by Richard and Galeotti, Ann. Sci. Nat., ser. 3, iii. 18, as Malaxis Galeottiana, was transferred to Liparis by Mr. Hemsley, Gard. Chron. 1879, i. 559. No exact occurrence of the plant in Mexico was known until Dr. Watson, Proc. Am. Acad. xxii. 454, doubtfully identified with the description a specimen collected by Dr. Palmer on the Rio Blanco, Jalisco. Mr. Pringle has secured the same plant on the Sierre Madre, Chihuahua, 30 September, 1887 (no. 1527), and on moist slopes, Patzcuaro, 18 July, 1892 (no. 5274), and finally on dry granitic hills near Guadalajara, 17 August, 1893 (no. 4512). A por- tion of the last mentioned specimen was sent to Kew, where it was definitely identified with the species in question. Mr. Pringle's speci- ROBINSON. — MEXICAN PLANTS. 323 mens bear out Dr. Watson's amplification of the characters, and confirm its distinctness from L. elliptica, Reichb. f. Isochills DNILATERALE. Glabrous : stems several, slender, sim- ple, 8 inches in height, roughened especially near the base by numer- ous minute warts : leaves linear, slightly retuse, rather rigidly erect, 1^-2 inches long, finely striated : racemes single, simple, terminal, secund, about 7-tlowered : bracts oblong, obtuse, purplish ; pedicels exserted 2-3 lines : flowers large for the genus, 6 lines long, purple, strongly gibbous upon the side toward the labellum, not noticeably fleshy : sepals subequal in length, carinate, acutish : petals oblong, obtusish, slightly broader than the sepals, aud not carinate; labellum lance-oblong, with the sigmoid flexure characteristic of the genus but without lateral lobes : fruit 6 lines in length. — Collected on oaks, Tamasopo Canon, 23 June, 1891 (no. 511 G). This species, with much the habit of I. lineare, R. Br., has larger more one-sided flowers in a looser raceme. Nkmastylis flava. Bulb ovoid, 8 lines in diameter, dark brown, with a tuft of fibrous roots at the base ; stems or rather scapes 1 to 3, slender, simple, leafy only at the base, 8-10 inches high, bearing at the summit a single 2-3-flowered spathe of 3 unequal lance- linear or linear-oblong acuminate bracts (the largest 1^—1 ^ inches in length) : the outer leaves reduced to acute scale-like theaths ; the inner narrowly linear, attenuate, equalling the scape : pedicels about an inch long: flowers 1-1 \ inches in diameter, bright yellow: seg- ments oblanceolate to obovate, acuminate ; the outer 7 lines in length ; the inner two thirds as long and with rounded thickened areolae at the base : filaments connate ; column 1^ lines high ; anthers 2^ lines long, inclining to curl with age: capsule ovoid, smooth, shallowly six-fur- rowed, ^ inch in length. — Collected on gravelly soil near Guadalajara. 23 June and 21 August, 1893 (no. 4400). Near N. bracteolata. Baker, but with single heads, distinctly unequal perianth segments, and, of course, different geographic position. Dioscorea grandifolia, Schlecht. Mr. Pringle's no. 4547, col- lected on talus of cliffs, barranca of Tequila, 4 October, 1893, corre- sponds well to the characterization of this species. The fertile plant, of which I find no description, furnishes the following characters : fruiting racemes about 3 inches in length upon peduncles about an inch long: fruit elliptic in outline, reflexed-spreading, not closely imbricated, 7 lines in length, half as wide. D. Pringlei. Glabrous : stem twining : leaves ovate, acuminate, entire, cordate with an open sinus, 9-11-nerved, somewhat paler 324 PROCEEDINGS OF THE AMERICAN ACADEMY. beneath, 2-3 inches in length, about the same breadth : flowers of the staminate plant in slender compound racemes ; the latter scarcely pedunculate, either solitary in the axds or borne at the successive nodes of the leafless ends of the stem ; branches of the racemes numer- ous, about 5-flowered, the uppermost reduced and fewer-flowered : pedicels \ line long: bracts lance-linear, acuminate: perianth cam- panulate, pale yellow; segments obovate, rounded at the ends, 1^ lines in length : stamens 3 ; filaments slender ; rudiments none : fertile racemes axillary, short-peduncled, in fruit 2-3 inches in length : fruit suborbicular, retuse, 3^ lines in diameter (immature). — Collected on grassy slopes of barranca near Guadalajara, 7 September, 1893 (no. 4526). D. hirsuticaulis. Pubescent-hirsute throughout with crisped white hairs : stem elongated, twining : leaves deltoid-ovate, acu- minate, entire, cordate with an open sinus, about 9-nerved, 2-3 inches long, nearly as broad; petioles 8 lines in length: racemes of the staminate plant slender, simple, about \\ inches in length upon peduncles perhaps half as long, solitary in the axils or borne at the successive nodes of the nearly leafless ends of the stems ; pedicels but \— \ line in length : perianth campanulate, a line in diameter ; segments ovate-oblong, obtuse : stamens only 3, filaments short and straight : inflorescences of the female plant axillary, solitary, in fruit \\ inches in length, \ inch in diameter: fruit stramineous, broadly elliptic, smooth, 4| lines in length, 2>\ lines in breadth, reflexed and very densely imbricated. — Collected under dry ledges, barranca of Tequila, 14 October, 1893 (no. 4572). D. plumifera. Glabrous: stems twining : leaves thin, ovate acu- minate, entire, cordate with a broad sinus, 7-11-veined, slightly paler beneath, 2^—3 inches long, 2 inches broad : staminate flowers yellowish green, 2^- lines in diameter, borne in long slender feathery panicles ; the latter geminate at the axils, 4 inches in length, \ inch in diameter ; the numerous short lateral brandies about 3-flowered ; pedicels |— 1 line long, very slender : bracts lance-linear, acuminate, 1 line in length : segments of the perianth oblong-lanceolate, obtuse : fertile stamens 3, slender, not at all rigid or divaricate, alternating with 3 rudiments: racemes of the fruiting plant short-peduncled, densely flowered, 2 inches in length : fruit (scarcely mature) broadly elliptic, 4 lines long, reflexed, imbricated. — Collected under cliffs, barranca of Tequila, 14 October, 1893 (no. 4530;. D. militaris. A slender vine 2-4 feet in length : stem smooth, from a small tuber : leaves rather distant, 3-lobed and halberd-shaped, ROBINSON. — GALINSOGA. 325 hirsute-pubescent upon both surfaces ; the central lobe lanceolate attenuate to a sharp point, 3-uerved, l£-2£ inches long, 3-7 lines broad; lateral lobes G-8 lines long, somewhat falcate, obtuse; the uppermost leaves reduced, entire ; petioles j inch long : stamiuate racemes slender, loosely flowered, including the peduncle 2-4 inches long : pedicels j line long, equalling or somewhat exceeding the minute subulate bractlets : perianth \\ lines broad; segments of two forms ; the 3 outer spreading, oblong, obtuse ; the inner a little narrower and acutish, ascending, incurved : stamens 3, very short, perhaps one third as long as the segments of the perianth : fruiting raceme 1^-2 inches long upon peduncles of nearly ecpjal length : capsules glabrous, re- flexed, somewhat imbricated, elliptic in outline, 6 lines long, half as broad. — Collected upon the steep sides of wet cliffs of the barranca of Guadalajara, 8 September, 1893 (no. 5434). III. —NOTES UPON THE GENUS GALINSOGA. Few genera have been subject to so much doubt as to proper lim- itation as Galinsoga. This is equally evident from its treatment in I)e Candolle's Prodromus, in which of its supposed six species five are questioned, and from the writings of subsequent authors who have dealt with it. Satisfactory generic limitations can perhaps only be obtained by a monograph including uot only the plants hitherto as- cribed to Galinsoga, but several neighboring genera, for which suffi- cient material could only be found in the larger foreign collections. Something, however, may well be done at more accurate specific and varietal definition of forms growing within our own country. So far as I know, only one species is at present recognized upon the continent of North America north of Mexico, that is G. parviflora, Cav. The telling characteristics of this species, when obtained from the earliest descriptions and figures as well as from the examination of material completely in accord with these, are as follows : stem and branches smoothish or finely and more or less appressed pubescent : rays white or whitish, little exserted: pappus of the disk flowers consisting of spatulate obtusish scales equalling the achene. Of this species several varieties have been suggested. Dr. Gray in the PL Wrightiana, ii. 98, founded two upon the relative development or absence of the pappus in the ray flowers ; namely, var. semicalva, with naked ray achenes, and var. Caracasana, with abortive ray pappus. The writer has examined a large number of specimens of the spesies both from North America and from the most widely separated regions of the world, and finds 326 PROCEEDINGS OF THE AMERICAN ACADEMY. that the ray pappus is never developed as in the disk flowers, but is always more or less abortive, consisting of fewer, shorter, and more slender bristles. Furthermore, the variation in its development seems absolutely independent of any other characters either floral or habital, so that varieties founded upon this character can have no more than formal value. Indeed this is the view which Dr. Gray subsequently took in the Synoptical Flora, i. 303. On the other hand, the form and length of the pappus in the disk flowers seem more worthy of note, and several plants have recently been received at the Gray Herbarium which differ so materially in this regard and in their greater hbpidity from the typical G. parviflora, Cav., as to call attention to the group. In those plants from the Eastern and Middle States, as cited below, the pubescence of the upper interuodes is not only considerably more copi- ous but much more spreading, the rays are bright white and well ex- serted, and the pappus scales of the disk flowers, while in length about equalling the achenes as in G. parviflora, have a very different form, beincr narrower and attenuate to a bristle-like apex. One of these specimens, collected by Mr. H. L. Clark at Pittsburg, Pa., was for- warded to Geneva for comparison with De Candolle's G. parviflora, var. hispida, and has been pronounced by M. Buser identical with that variety. Still a third plant of this genus quite distinct from the pre- ceding has been collected upon waste land about Camden, N. J., by Mr. C. F. Parker. It is characterized by considerable pubescence, distinctly purple rays, and disk pappus but half as long as the achene. Portions of this plant have been forwarded to Geneva and Kew for comparison with Vargasia Caracasana, DC, and Galinsoga hispida, Benth., and it proves identical wiih both, while garden specimens of Regel's G. br achy Stephana show it with scarcely a doubt to be the same. This shows that Dr. Gray was in error in regarding his G. parviflora, var. Caracasana the equivalent of De Candolle's Vargasia Caraca- sana. While Mr. Parker's plant from Camden, the weed-centre of the United States, by no means shows this species of tropical origin to be established in our country, it may well be found that a portion of what has been hitherto referred to G. parviflora is this plant, espe- cially as it is more or less extensively introduced in Eastern Europ3, where it bears the name G. brachystephana, Regel. Summing up the different forms as well as the material at hand permits, we may divide them as follows. * Kays white, pappus of the disk flowers about equalling the achene. G. parviflora, Cav. Smoothish, the upper internodes with a sparing sub-appressed pubescence : pappus of the disk flowers of spatu- ROBINSON. BOTANICAL NOTES. 327 late obtusish scales. — Ic. iii. 41, t. 281 ; DC. Prodr. v. 677 ; Gray, Syn. Fl. i. 303, including vars. semicalva and Caracasana of PI. Wright. Bidens mercurialis folia, &c, Feuillee, Journ. Obs. Phys., &c. ii. 744, t. 32. — Generally distributed from New England to Ore- gon and southward to Mexico. Var. hispida, DC. Prodr. v. 677. Pubescence especially of the upper internodes more copious and not at all appressed : scales of the pappus in the disk flowers attenuate and bristle-tipped : foliage, etc. as in the type. Alleghany City, Pa., 9 August, 1869, Porter ; Milwaukee, Wise, October, 1881, Sherman; Providence, R. I., 5 July, 1892, Bailey § Collins; Pittsburg, Pa., 1893, Clark; Cambridge, Mass., and doubtless widely introduced. The same thing has been recently collected in Central America by Capt. John Donnell Smith, nos. 759, 2352, where it is doubtless indigenous. While in general readily distinguishable by the characters described, this variety occasionally so intergrades with the type that specific distinction seems very un- desirable. * * Rays purplish : pappus of the disk flowers but half as long as the achenes. G. hispida, Benth. Bot. Sulph. 119. The oldest name under the genus. G. brachystephana, Regel in Walper's Rep. vi. 722. Varga- sia Caracasana, DC. Prodr. v. 676. Doubtfully established at Cam- den, N. J., 15 September, 1870, C. F. Parker. So far as the northern and eastern parts of our country are concerned, all these plants are doubtless introduced ; in Texas, New Mexico, and Arizona forms of G. parvijlora may well be indigenous, as Dr. Gray suggests. IV. — MISCELLANEOUS NOTES AND NEW SPECIES. Silene sdbciliata. Perennial: stem strict, terete, glabrous, a foot and a half high, enlarged at the nodes : leaves glaucous, slightly fleshy and finely lepidote, narrowly oblong to linear-oblanceolate, gla- brous on the surfaces but sparingly ciliated, 1^-2 inches long, obtusely pointed with callous tips, narrowed below to winged commonly cili- ated petioles; floral leaves reduced to lance-linear acute bracts: flowers rather distant, pedicellate, forming an elongated racemiform inflores- cence: bractlets lance-linear, ciliated : calyx glabrous, cylindric, 10 lines in length: petals deep red, \\ inches long; the blade elliptic, entire, obtuse ; the appendages lanceolate, entire : fruit and seeds not seen. — Type in the Gray Herbarium, collected by Mr. Charles Wright, but without number or date, and bearing only the locality " Texas and Louisiana." It is to be hoped that this attractive species may be re- 328 PROCEEDINGS OF THE AMERICAN ACADEMY. discovered. It is evidently related to S. regia, Sims, and »S. laciniata, Cav., but differs from the former very much in the form of its leaves, from the latter in its entire petals, and from both in the entire absence of glandular pubescence. From S. Virginica, L., it differs in petals, pubescence, and texture of the leaves. Silexe laciniata, Cav. Prof. W. R. Dudley calls my attention to the fact that the seeds in this species and its broad-leaved variety (S. Greggii) are not infrequently vesicularly crested, as in the Califor- nian S. Parishh, Wats. Arenaria Grcenlandica, Spreng. This attractive species, pos- sessing a wide and interesting north and south distribution, seems worthy of special study. The typical form, with subglobose obtusely pointed capsules and with stems few and decumbent from a spreading rosette of somewhat fleshy leaves, occurs chiefly in Greenland and Labrador. The common form of the mountains of Maine, New Hamp- shire, and Eastern New York becomes rather densely matted, and has many erect stems with very numerous erect less fleshy leaves about the base. Its capsules are ovoid to oblong and more or less acutely pointed. Although in temperate latitudes usually confined to rocky soil of the mountains or higher hills, this species descends nearly or quite to the seashore i» Maine, at Bath and at Mt. Desert. It has also been found at Middletown, Conn. When growing at these lower altitudes, the plant is scarcely at all matted and the segregated few-stemmed individuals have simple or at least less fibrous roots and fewer thicker leaves than in the mountain form, with which however they are connected by frequent intergradations. In these forms also the capsule is ovoid or oblong rather than globose. Until recently the Shawangunk Mountains of New York have passed as the southern limit of this species, but there can be no doubt (see Mem. Torr. Club, iii. 14) that many of the specimens from the higher Mountains of North Carolina, hitherto referred to A. glabra, Michx., are practically identical with the plant of the White Mountains, notwithstanding the fact that the flowers average smaller. The question whether it is best to retain these forms of temperate regions in the same species with those of Greenland and Labrador presents much difficulty. The differences in habit, size of the flowers, and notching of the petals are often striking. Unfortu- nately, however, no one of these characters holds satisfactorily in a large series of specimens, and it appears that the change from the few- stemmed decumbent plant with a basal rosette of leaves and single simple root to the matted plant with fibrous roots and many stems, erect by mutual crowding, is a difference which may well be due exclusively ROBINSON. — BOTANICAL NOTES. 329 to the greater leDgth of the season. As to the size of the flowers, a series of specimens collected at Mt. Desert by Mr. Redfield and Mr. Rand conclusively shows that this is largely a function of the season, the autumnal flowers being much smaller than the earlier ones, indeed far smaller than those of the southern form. The shape of the cap- sule gave promise of furnishing a good character, but the difference between globose and ovoid is not a sharp one, and on examination it proves that subglobose capsules occasionally occur in the White Moun- tain plant, and even in the few Greenland plants which could be ex- amined the capsules were found to vary to ovoid, so that no sharp distinction can be founded upon this character. Dalea neglecta. Branches slender, terete, glabrous, glandular- punctate : leaves smooth, 1^ — 2 inches in length ; leaflets 5-7 pairs and odd one, 4—5 lines long, elliptic or oblanceolate, petiolate, obso- letely crenate, rounded at the apex, acute at the base, smooth and veinless above, glaucous, glandular-punctate and 1-nerved beneath: peduncles very slender, equalling or exceeding the curved loosely- flowered spike (H inches in length) : flowers 2^—3 lines long, spread- ing or reflexed upon very short pedicels : calyx turbinate, strongly ribbed with yellow glands and covered with very short upwardly ap- pressed or rather incurved hairs ; teeth lance-linear, acute, inflexed, nearly equalling the tube : corolla in dried specimens bright purple. — Collected at Guanajuato, Mexico, by Prof. Alfred Duges (no. 2576). Habit nearly of D. nutans, Willd. The peculiar pubescence of the calyx appears rather exceptional in the genus, being in most species straight, silky or lanate. Saxifraga Pennyslvanica, L. A specimen of this species, col- lected at Royalton, Vt., with deep red petals, has recently been received at the Gray Herbarium from Miss Emily P. Robinson of Manchester, N. H. A hasty search in the literature of the species has failed to show any record of this variation, the petals being always described as greenish or yellowish green. As the anthers are bright orange, the dark red or crimson petals give to the flowers a much more striking contrast of color, doubtless correlated with insect pollination. So far as observed, the specimen in question presented no other differences from the typical form. Aster paucicapitatus. Stems several, simple, flexuous, leafy to the summit, ribbed, somewhat pubescent, 10 inches to lh feet high: leaves elliptic-oblong, obtuse or obtusish, mucronulate, sessile by a scarcely narrowed base, finely and somewhat glandularly pubescent, 9-16 lines long, 3-5 lines broad, erect or ascending: heads usually 330 PROCEEDINGS OF THE AMERICAN ACADEMY. single, terminal, short-peduncled, less frequently 3-4, corymbose, in- cluding the rays 1^-1^ inches in diameter: involucre of few subequal lance-linear attenuate finely pubescent and ciliolated scales ; the latter loosely imbricated in 2-3 series, rather firm and broadly white-mar- gined near the base, 4 lines long ; the very acute tips often purplish : rays 12-18, white or pink, 6 lines in length, \\ lines in breadth, 3-toothed at the apex ; disk flowers including the silky achenes 5 lines in length : pappus rather copious, of unequal bristles but not distinctly double. — A. Engelmanni, Gray, var. (?) paucicapitatus, Robinson, Proc. Am. Acad. xxvi. 176. — Collected by C. V. Piper in the Olympic Mountains, Washington, August and September, 1890 (nos. 926, 934). Never satisfied with the earlier and somewhat provisional disposition of this plant, I have returned to its study, and conclude that it is specifically distinct from A. Engelmanni, through its much less imbricated involucre of subequal scales. It should stand in the genus near A. Xylorhiza, Torr. & Gray. Dioscorea Dugesii. Stem slender, climbing, angulate, sparingly pubescent with fine brown hairs : leaves ovate, cordate, sharply acumi- nate, thin, 9-nerved, pellucid-lineolate, nearly or quite smooth, 2^ inches long, 2 inches broad; petioles pubescent, 1^ inches loug : staminate flowers 1|- lines in diameter, borne in short slender simple solitary axillary pubescent racemes 1^—2 inches in length: peduncles but 2 lines long ; pedicels scattered, ^ line in length, 1-3-flowered : bracts subulate, considerably exceeding the pedicels : segments of the perianth linear-oblong or linear-lanceolate, obtuse, with crisped margins : sta- mens six, equal, inserted on the perianth near its base, half as long as the segments. (Fertile plant not seen.) — Collected by Prof. Alfred Duges at Guanajuato, 1880 (no. 37). A slender species, somewhat resembling D. remotiflora, Kunth. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 331 XV. ON THE INHERITANCE OF ACQUIRED CHARACTERS IN ANIMALS WITH A COMPLETE METAMORPHOSIS. By Alpheus S. Packard, M. D Presented January 10, 1S94. I. The Physical Basis of Heredity. Before discussing our subject it may be well to give a brief histori- cal sketch of the present views as to the physical basis of heredity. As early as 1849. Owen, in his "Comparative Anatomy," suggested that there was a physical basis for heredity. Herbert Spencer, in his "Principles of Biology" (1866), based the phenomena of heredity on the supposed presence of " physiological units," which he conceived to be immensely more complex than chemical units or molecules. (Vol. I. p. 183.) But Darwin in 1868 brought the question to the front in his " Hypothesis of Pangenesis," which was disproved by experiments on the effects of transfusion of blood by Francis Galton,* who in 1875 published a theory of heredity which in some ways approached that of Jaeger. Galton also claimed that acquired characters are only " faintly heri- table," and he endeavors to explain the almost complete non-trans- mission of acquired modifications, f The first, however, to suggest an objective and scientific basis appears to have been Dr. G. Jaeger, of Germany, who in 1876 * Galton in 1875 suggested that each individual may properly be conceived as consisting of two parts, one of which is latent, and only known to us by its effects on posterity, while the other is patent, and constitutes the person manifest to our sense. "He also claimed that we are made up bit by bit of inherited structures, like a new building composed of the fragments of an old one, — one element from this progenitor, another from that, although such elements are usually transmitted in groups." — A Theory of Heredity, Journal of the Anthropological Institute, 1875. See also Contemporary Review, December, 1875. t Contemporary Review, 1875, Vol. XXIII. p. 95. 332 PROCEEDINGS OF THE AMERICAN ACADEMY. made the following statement, as quoted by J. A. Thompson in his " History and Theory of Heredity." * "Through a great series of generations the germinal protoplasm re- tains its specific properties, dividing in every reproduction into an ontogenetic portion, out of which the individual is built up, and a phylogenetic portion, which is reserved to form the reproductive material of the mature offspring. This reservation of the phylo- genetic material I described as the continuity of the germ protoplasm. Eucapsuled in the ontogenetic material the phylogenetic protoplasm is sheltered from external influence, and retains its specific and em- bryonic characters." f In 1880 M. Nussbaum $ substituted a new hypothesis for Darwin's pangenesis. According to the view of this observer, the germinal cells from which the sexual products are derived are separated off from the other cells of the embryo very early, and undergo little alteration. Hence he concluded that some of the original germ substance is di- rectly abstracted from the egg, and preserved without essential altera- tion to become, by giving rise to the sexual elements, the germ substance of another generation. Nussbaum also expressed his disbelief in the transmission of acquired characters. This belief, held by Darwin as well as by Lamarck, and almost universally adopted by medical men, had not before been called in question. In 1884 Nageli § took the ground that there are in every living cell two substances ; one the nutritive plasm, and the other his hypothetical " idioplasma." This view was indorsed by Kolliker, who claimed that the sharp distinction between body and germ cells does not exist. * Proc. Royal Society Edinburgh, 1889, pp. 91-116. See also Brooks, " The Law of Heredity," 1883; Osborn, "The Cartwright Lectures," 1892; " Present Problems in Evolution and Heredity," Medical Record, New York, 1892; "The Present Problem of Heredity," Atlantic Monthly, March, 1891. t Lehrbuch der algemeinen Zoologie, Leipzig, 1878, II. Abtheilung. In a previous book, published at an earlier date than the one quoted by Thompson, Zoologische Briefe, Wien, 1876, Jaeger writes thus : " Hier muss ich noch einmal den Gegensatz zwischen Darwin's Theory von der Pangenesis und meiner Theorie von der Continuitat des Keimprotoplasmas hervorheben " (p. 326). See also Weismann, " The Germ Plasm," p. 200. The author appears to have overlooked the statement of Jaeger in his Zoologische Briefe, wherein he explicitly, as shown by our quotation, refers to the "continuity of the germ protoplasm." | Die Differenzirung des Geschlechts in Thierreich. Archiv fiir mikroskop. Anatomie, Bd. XVIII., 1880. § Nageli, Mechanisch-physiologische Theorie der Abstammungslehre. Miin- chen u. Leipzig, 1884. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 333 In 1887 Minot* in a brief note suggested that Niigeli's hypothetical idioplasm is probably identical with the nuclear chromatin of mor- phologists ; and that heredity is due to the transfer from parent to off- spring of the nuclear substance. Maupas,f in two memoirs published in 1889 and 1890, reaffirmed and extended this view, concluding that the chromatin of all cells is the bearer of heredity. Meanwhile in 1885 appeared Weismann's epoch-making essays on heredity, his view being somewhat in the line of Jaeger's theory, but greatly expanded, and with many new and original suggestions. He stated that it was impossible to prove the existence of gemmules, and substitutes for pangenesis the now famous doctrine of " the con- tinuity of the germ plasm," as affording a more rational basis than pangenesis or Brooks's modification of it. Weismaun says : " The nature of heredity is based upon the transmission of nuclear substance with a specific nucleoplasm of the germ cell, to which I have given the name of germ plasm" (p. 180). As stated by Mr. E. B. Poulton,J an able commentator and exponent of Weismann's views, — "The word 'continuity ' expresses the theory that heredity depends on the fact that a minute quantity of this germ plasm is reserved unchanged during the development of the individual, and afterwards grows and gives rise to the germ cells. Hence the germ plasm is continuous from one generation to another in an unending succession, and from it the germ cells of each generation are produced. " Parent and offspring resemble each other because both arise from the same substance, which develops rather later in the case of the offspring. Hence everything which is predetermined in the germ cell, every blastogenic character, may be transmitted, while somatogenic characters cannot be transmitted." We will quote Weismann's definition of acquired and blastogenic characters : " We maintain that the ' somatogenic ' characters cannot be transmitted, or, rather, that those who assert that they can be trans- mitted must furnish the requisite proofs. The somatogenic characters not only include the effects of mutilation, but the changes which follow * Science, New York, VIII. 125. t Sur la Multiplication des Infusoires Cilie's. Archiv de Zoologie expe'rimen- tale, se'r. 2, VI. 165-273; Le Rajeunissement Knryogamique chez les Cilie's, VII. 149-517. See also Hartog, Quart. Journal Microscop. Science, December, 1891, and Osborn, loc. cit., pp. 54-56. \ Theories of Heredity. Reprinted from the Midland Naturalist, November, 1889. 334 PROCEEDINGS OP THE AMERICAN ACADEMY. from increased or diminished performance of function, and those which are directly due to nutrition and any of the other external influences which act upon the body. Among the blastogertic characters, we include not only all the changes produced by natural selection opera- ting upon variations in the germ, but all other characters which result from this latter cause." (p. 413.) "Weismann remarks that Niigeli has shown that even in so minute a space as -j^ott °f a curjic millimeter such an enormous number (400,000,000) of " micellae " * may be present that the most diverse and complicated arrangements become possible. It therefore follows that the molecular structure of the germ plasm in the germ cells of an individual must be distinguished from that of another individual by certain differences, although these may be but small ; and it also follows that the germ plasm of any species must differ from that of all other species. (Weismann, p. 191.) It also follows, the author con- tends, that the molecular structure of the germ plasm in all higher * The existence of such primary elements as these, which are supposed to form the basis of organization of the protoplasm of cells, as well as the physical basis of heredity, is insisted on by nearly all of the biologists who have written on this subject. Professor Whitman, in an able article entitled '.' The Inadequacy of the Cell Theory," states that Ernst Briieke in 1861 first contended for the organization of the cell, and the existence of " smallest parts " as the basis of this organization, quoting him as follows. " We must therefore ascribe to living cells, in addition to the molecular structure of the organic compounds which they contain, still another and otherwise complicated struc- ture; and this it is that we designate by the name organization." (Elementar- organismen, p. 387. Wiener Sitzungsberichte, October 10, 1861, Band XLIV. Heft 2, p. 381.) Whitman then goes on to say that " we have seen similar ideas reappear in the 'physiological units' of Herbert Spencer, the 'gemmules' of Darwin, the 'micellae' of Niigeli, the ' plastiilules ' of Elsberg and Haeckel, the ' inotagmata' [plasomes] of Wiesner, the ' idioblasts ' of Oscar Hertwig, and the ' biophores ' of Weismann." Whitman contends that the secret of organization, growth, development, lies not in cell formation, "but in those ultimate elements of living matter for which idiosomes seems to me an appropriate name." He adds: "All growth, assimilation, reproduction, and regeneration may be supposed to have their seat in these fundamental elements. They make up all living matter, are the bearers of heredity, and the real builders of the organism." (Journal of Mor- phology, VIII. 639, 658, Boston, 1893. Compare also Weismann's "The Germ Plasm," Introduction.) Here should be quoted the striking remark of Herbert Spencer, " that sperm cells and germ cells are essentially nothing more than vehicles in which are contained small groups of the physiological units in :* fit state for obeying their proclivity towards the structural arrangement of the species they belong to." (Piinciphs of Biology, I. 254.) PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 335 animals must be very complex, and at the same time this complexity- must gradually diminish during ontogeny. In his latest work Weismanii * thus states his mature and appar- ently final views : "All the phenomena of heredity depend on minute vital units which we have called biophors, and of which living matter is composed ; these are capable of assimilation, growth, and multiplica- tion by division." (p. 450.) He further discusses the nature and mode of action of these hypothetical bodies, which are contained in the nu- cleus, the latter serving as the " bearer of the biophors controlling the character of the cell." How these biophors are grouped into deter- minants, and how the latter form aggregates called ids, the nuclear rods (chromosomes) being aggregates of ids, called idants, is set forth in a very circumstantial way. He then states in the summary of his work : " The germ jrfttsm, or hereditary substance of the Metazoa and Metaphyta, therefore, consists of a larger or smaller number of idants, which in turn are composed of ids ; each id has a definite and special architecture, as it is composed of determinants, each of which plays a perfectly definite part in development." (p. 453.) Weisman's reasons for not accepting the doctrine of transmission of acquired characters would appear to be purely hypothetical and a priori, as will be seen by the following extracts : " It is self-evident from the theory of heredity here propounded that only those charac- ters are transmissible which have been controlled — i. e. produced — by determinants of tbe germ, and that consequently only those varia- tions are hereditary which result from the modification of several or many determinants in the germ plasm, and not those which have arisen subsequently in consequence of some influence exerted upon the cells of the body. In other words, it follows from this theory that somatogenic or acquired characters cannot be transmitted. This, however, does not imply that external influences are incapable of pro- ducing hereditary variations; on the contrary, they always give rise to such variations when they are capable of modifying the determinants of tbe germ plasm. Climatic influences, for example, may very well produce permanent variations, by slowly causing gradually increasing alterations to occur in certain determinants in the course of genera- tions. An apparent transmission of somatogenic modifications may even take place under certain circumstances, by the climatic influence affecting certain determinants of the germ plasm at the same time, * The Germ Plasm. A Theory of Heredity. Translated by W. N. Parker New York, 1893. 336 PROCEEDINGS OF THE AMERICAN ACADEMY. and when they are about to pass to that part of the body which tliey have to control. This is inchoated by the climatic variations of the butterfly Pulyommatus phlceas." He then adds : — " The primary cause of variation is always the effect of external influences. Were it possible for growth to take place under abso- lutely constant external influences, variation would not occur; but as this is impossible, all growth is connected with smaller or greater de- viations from the inherited developmental tendency. " When these deviations only affect the soma, they give rise to tem- porary non-hereditary variations ; but when they occur in the germ plasm, they are transmitted to the next generation, and cause corre- sponding hereditary variations in the body." (pp. 462, 463.) That the physical seat of heredity does exist in the nucleus has been wellnigh demonstrated, if not quite, by some remarkable ex- periments by Boveri at the Naples Zoological Station, so that what was a mere hypothesis has apparently become a matter of fact. Boveri's results appeared in 1889,* and a translation of his short paper has been published by Prof. T. H. Morgan in " The American Naturalist" for March, 1893. Five years ago, by accident, the brothers Hertwig discovered that in consequence of shaking, certain eggs of sea-urchins fell to pieces ; some of these pieces contained nuclei and others not. It was found that the non-nucleated pieces could be artificially fertilized as well as those containing nuclei, and that the bits of yolk underwent what is called segmentation. Boveri, taking the hint suggested by these happy accidents, made the astonishing discovery that the enucleated bits of eggs could be fer- tilized, and that such bits developed into larvae or young sea-urchins as completely formed as those growing from ordinary entire fertilized eggs. The further experiments to prove the seat of heredity were to hy- bridize the fragments of eggs of one genus of sea-urchins with the sperm cells of another genus, and to rear them far enough along in life to determine whether the young showed the qualities of both species or one only. By cross fertilizing the enucleated egg fragments of Sphserechinus with the male germs of Echinus, Boveri produced an almost exact middle form, standing half-way between the two parents. He found however that a portion of the cross-bred larva? agreed entirely * Ein gpsclilcchtlidi erzougter Organismus oline miitterliche Eigenschaften. Sitzung der Gesellschaft fur Morphologie und Physiologie zur Miinclien. Sitzung am 16 Juli, 1889. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 337 with the simple ordinary larva of Echinus, which he thinks must have been produced chiefly from enucleated fragments. This seemed to be proved by the fact that he could distinguish in a preserved and colored larva whether or not it had originated from a nucleated or enucleated eog by the size of its nuclei, which are considerably smaller in the larva? derived from the enucleated bits of eggs. Hence all doubt seemed removed, and Boveri claims that he has proved that, by cross fertilization of whole eggs or bits of eggs having nuclei, larva? are formed that are half-way between the larval forms of the parent species. On the other hand, larva? arising from the enucleated bits of eggs have entirely the characteristics of the parent (male) species. Hence, if his experiments are correct, he demonstrates the law that the nucleus alone is the bearer of hereditary qualities. Thus the hy- pothesis that the substance of the nucleus of reproductive cells is the physical basis of heredity seems provisionally at least placed on a foundation of fact. On the other hand reference should be made to the recent papers of Driesch and of O. Hertwig,* who from researches on the phenomena of cleavage and the formation of organs in the embryo of the frog , regard the egg as isotropic, its first cells as qualitatively alike, the development of the embryo being the result of an epigenetic forma- tion of organs, the process being one of interrelation of the cleavage cells. Hence in place of the mosaic theory of Roux and the germ plasm theory of Weismann, Hertwig substitutes the theory of the controlling inter-adjustments of the embryonic cells and later of the tissues and organs. o IT. Heredity of Characters acquired during the Life- time of the Individual. It would appear that many, if not most, of our leading anatomists and cytologists agree that there may be a physical basis for heredity, and that this basis is afforded by the germ plasm of the nucleus, a portion of which is continuous in succeeding generations. They do however disagree as to whether acquired or " somatogenic" char- acters can be transmitted by heredity, and whether the contents of the nuclei of germ cells are influenced or not by whatever affects the body in general * 0. Hertwig, Archiv fur Mikroskop. Anatomie, 22 December, 1893, XLII. 662-794. See Abstract by E. A. Andrews in the American Naturalist, March, 1892, pp. 272-278. vol. xxix. (n. s. xxi.) 22 338 PROCEEDINGS OP THE AMERICAN ACADEMY. Perhaps the ablest objector to this phase of Weismann's theory of heredity is Kolliker,* who (1) denies that tliere is any fundamental difference between body and germ plasm ; and (2) claims that in the various cellular changes the characters of the original germ plasm may be either wholly retained, or degenerate, or be wholly lost. Sir William Turner f and other medical men also favor the theory of the transmission of acquired characters. He suggests that the more subtle processes of generation may be transmitted where mutilations may not. Detmer J likewise opposes Weismann's view on the following grounds : — 1. The intimate histological influences of external conditions on the organism. 2. The importance of correlation in allowing an influence to satu- rate from one part to another, and thus to the sexual cells. 3. The suggestiveness of the persistence of certain phenomena (in plants) after the inciting conditions have ceased. The criticisms of Kolliker have been ably discussed by Weismann in his Essays upon Heredity. The views which have had weight with us, and which seem to oppose Weismann's theory that acquired traits cannot be transmitted, are the following : — 1. The laws of correlation («) of growth, and (b) of organs in the mature organism. If one part or organ of the body is removed, aborted, or changed, the rest may, in certain cases, be either temporarily or permanently affected by the change. 2. Whatever affects the body in general would tend to affect the germ plasm, since the tissues and cells of the ovaries and testes are supplied with blood, and are innervated like other parts and organs of the body ; hence the plasm of the nuclei of these cells, though it may exist in a temporarily indifferent state is nourished, or at least preserved from degeneration, and is thus influenced by whatever affects the body. 3. The operation of castration in either sex, as is well known, re- * Das Karyoplasma unci die Vererbung. Zeitschrift fiir wissens. Zoologie, XL. iv. 228, 1886, and Anat. Anzeiger, III., 1888. f Report of the British Association for the Advancement of Science for 1889, pp. 756-771. 1890. t Zum Problem der Vererbung. Archiv fiir die ges. Physiologie, XL. 1887. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 389 suits in a profound modification of the physical, intellectual, and moral nature of the suhject operated upon. 4. It has not yet been satisfactorily disproved that new characters, or the tendency to the heredity of such characters, are not the result of a change of external environment, however slight. This appears to be the primary cause of all changes in organisms. 5. As blastogenic or congenital characters are not invariably trans- mitted, with much less reason may somatogenic or acquired characters be invariably transmitted, especially at the present da}'. 6. The transmission of acquired characters may have been more frequent and regular in early geological ages, during the period of the origin of family, ordinal, and class ancestral types, and when such forms were more plastic than now owing to more wide-spread and rapid changes in the physical geography of the earth's surface than occur at present. During paheozoic times somatogenic characters may have greatly preponderated over blastogenic characters ; for at certain critical periods iu geological history there were wide-spread extinc- tions of certain species of plants and animals, followed or accompanied by profound modification of others, which led to the origination of new types. Hence a study of the origin and subsequent modification and disappearances of organs in series of extinct animals will afford weighty facts. 7. If congenital characters are the only ones which can be inherited, they must have in the beginning originated from those acquired during the lifetime of the individual, or if not in the first, in the second or third, or a later generation. 8. Can we always draw the line between congenital and acquired characters ? It seems to us to be often not only very difficult, but well- nigh impracticable, except in animals with a metamorphosis. 9. The results of the cultivation of fruits and of the domestication of animals, as well as the experiments of Brown-Ssquard, Bert, and others, strongly suggest that the characters acquired during the life- time of such organisms are capable of transmission. 10. If there were no such thing as the transmission of characters, either anatomical, physiological, or mental, originating during the life- time of an organism, how should we have any evolution resulting in the different groups of organisms ? Does not the denial of the fact of transmission of acquired features either in the past or present cut away the support for either phase of evolution, whether Lamarckism or Darwinism? If the processes of heredity have to he begun over again with the 340 PROCEEDINGS OF THE AMERICAN ACADEMY. birth of each individual, then we shall have to invent a new term for what is ordinarily understood to be progressive or continuous evolu- tion. The assumed peculiar property of the germ plasm is its con- tinuity from one generation to another, and its capability of receiving with each new generation the impress resulting from a change in the environment, or tendency to such change. Were this not so, then the offspring would be simply a repetition of the parents, instead of being like them with a difference, and there would be neither any fixed variation nor any individuality in organisms. Take the subject of human education. Does it wholly depend on the permanence of the intellectual enviroment, or is there an inherited capacity or aptitude for learning which runs in families or strains, and which is the result of the education of one or several generations, whether the training be for business, for the learned professions, or even for criminal pursuits ? Unless we are much mistaken, all human progress in learning, or in the arts and sciences, is based on the conception that in the long run mankind will increase in mental intelligence and capacity for learning. The history of science shows that a new department of learning may arise and each succeeding generation work more easily on the founda- tion laid by the previous generation. The work does not have to be begun de novo, but some degree of capacity for the new cult is inher- ited by successive generations ; certainly the intellectual environment may be said to change with each generation. All progress in humanity appears to be due, not only, in the first place, to our maintaining the present intellectual environment, with the manifold and many-sided stimuli of our present social structure, but also to the unceasing efforts of the leaders in advanced thought in many different departments of mental training and effort to open up new fields of research in natural, physical, and mental science, and their applications, to gain new and higher points of view in sociology and morals as well as in statecraft, and in short to perfect and hasten the development of the ideal man. Unless this progress, which is an historic fact, has been due not only at the outset, but all through human history thus far, to this principle of the inheritance of mental traits, causing the intellectual efforts of one generation to pass down and thus to have finally a cumulative effect, how could there be any progress in human society ? * * Herbert Spencer states in his Principles of Biology: "Certain powers which mankind have gained in the course of civilization cannot, I think, he accounted for, without admitting the inheritance of acquired modifications." Vol. I. p. 249. PACKARD. — INHERITANCE OP ACQUIRED CHARACTERS. 341 Oil the one hand, let us imagine a cessation of the operation of this principle. Suppose till the forces and stimuli of modern society to be removed, and the human organism to live like blind beetles in a cave, or a savage tribe isolated in the midst of an otherwise uninhabited con- tinent, with a total uniformity of conditions, physical, social, and moral, the effects of disuse would at once set in. Heredity without this vivify- ing principle of cumulative transmission, as it might be called, would be retrogressive in its action, aud the race would by reversion return to the status of prehistoric times. Or, on the other hand, if the present intellectual environment were maintained without the cumulative ac- tion of the principle of inheritance of acquired characters, the social organism would become stagnant, and the race would be semi-fossil- ized, or in a state of arrested development, like the Chinese. As we have already suggested in the beginning, blastogenic or ac- cpuired characters may have greatly preponderated over the somato- genic, and in fact the former or acquired characters may have constituted the fundamental elements of heredity iu general when life forms had only got as far as the Monera and lowest Protophytes. Then as the life forms became more differentiated there may have ensued a corre- sponding specialization into both blastogenic and somatogenic charac- ters. It seems most probable, as Kolliker suggested, that there is no fundamental difference between the body and germ plasm, and such a difference if it exists may be incapable of physical demonstration. Apropos of the view that whatever affects the body iu general must have some effect, however slight, on the germ plasm in it, we would cite the following facts and considerations. Mr. Herbert Spencer iu a powerful article in the Contemporary Review for March, 1803, entitled "The Inadequacy of 'Natural Selection,' " after quoting the facts regarding Lord Morton's hybrid between a male quagga and a chestnut mare seven eighths Arabian, and the results of crossing English and French breeds of sheep, and Giles's "sow and her produce," as fatal to Weismann's theory of the non-transmission of acquired characters, contends that these facts demonstrate " that the somewhat different units of a foreign germ plasm permeating the organism permeate also the subsequently formed reproductive cells, and affect the structures of the individuals arising from them." He then quotes Professor Sedgwick's letter to himself, dated December 27, 1892, referring to the continuity of the cells composing the tissues of animals, so that the protoplasm of the whole body is continuous, in which he states " that the connections between the cells of adults 342 PROCEEDINGS OP THE AMERICAN ACADEMY. are not secondary connections, but primary, dating from the time when the embryo was a unicellular structure." Hence Spencer maintains that " the alleged independence of the reproductive cells does not exist." Thus the soma is a "continuous mass of vacuolated proto- plasm, and the reproductive cells are nothing more than portions of it separated some little time before they are required to perform their functions." In his " Monograph of the Development of Peripatus Capensis," Mr. Adam Sedgwick, F. R. S., Reader in Animal Morphology at Cambridge, writes as follows : — "All the cells of the ovum, ectodermal as well as endodermal, are connected together by a fine protoplasmic reticulum." (p. 41.) " The continuity of the various cells of the segmenting ovum is primary, and not secondary ; i. e. in the cleavage the segments do not completely separate from one another. But are we justified in speak- ing of cells at all in this case? The fully segmented ovum is a syncy- tium, and there are not and have not been at any stage cell limits." (p. 41.) He then states in his letter to Mr. Spencer : — " It is becoming more and more clear every day that the cells com- posing the tissues of animals are not isolated units, but that they are connected with one another. I need only refer to the connection known to exist between connective tissue cells, cartilage cells, epithe- lial cells, etc. And not only may the cells of one tissue be continuous with each other, but they may also be continuous with the cells of other tissues." (pp. 47, 48). " Finally, if the protoplasm of the body is primitively a syncytium, and the ovum until maturity a part of that syncytium, the separation of the generative products does not differ essentially from the internal gemmation of a Protozoon, and the inheritance by the offspring of peculiarities first appearing in the parent, though not explained, is ren- dered less mysterious ; for the protoplasm of the whole body being continuous, change in the molecular constitution of any part of it would naturally be expected to spread, in time, through the whole mass." (p. 49.) " Mr. Sedgwick's subsequent investigations confirm these conclu- sions. In a letter of December 27, 1892, passages which he allows me to publish run as follows : — '• ' All the embryological studies that I have made since that to which you refer confirm me more and more in the view that the con- nections between the cells of adults are not secondary connections, PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 343 but primary, dating from the time when the embryo was a unicellular structure. . . . My own investigations ou tbis subject have been con- fined to the Artbropoda, Elasmobrauchii, and Aves. I have thorougbly examined the development of at least one kind of each of these groups, and I have never been able to detect a stage in which the cells were not continuous with each other ; and I have studied innumerable stages from the beginning of cleavage onwards.'" As regards plants, De Vries * and other botanists believe that all or the greater number of cells in the plant body contain the total heredity characters of the species in a latent condition. f In this connection should be noted an observation of Maupas, who saw the cytoplasm of an infusorian pouring into the nucleus until its bulk was increased eight times. Hoffman observed the transmission of acquired characters in the poppy, etc., as the result of deficient nutrition. | In 1890 Van Bemmeliu § gave a useful and very exhaustive account of the doctrine of heredity, with especial reference to the question of heredity of acquired characters, stating the views of various patholo- gists, anthropologists, and physiologists. In conclusion he makes some objections to the view that the germ plasm is independent of external influences. I am indebted to Prof. G. W. Field for the statement that MM. Charrin and Phisalix || cultivated Bacillus pyrocyaneus for several successive generations at 42.5° C. with the result that it lost its chromogenic property. This non-chromogenic character apparently thus acquired was retained upon cultivation under most favorable cir- cumstances, and it seemed to show no tendency to recover the chromo- genic property. "Laurent modified the chromogenic function of the Kiel water bacillus by exposing it to direct sunlight for a limited time. The suppression of this peculiarity was transmitted from generation to generation, so that a perfect albinotic variety was formed. The color property was also lost when cultivated at blood heat, and was not * Intracelliilare Pangenesis, Jena, 1889. t Osborn, loc. cit., p. 62. t Biol. Centralblatt, 1887, p. 667; Botan. Zeitung, 1887, pp. 260, 772, 773. § De Erfelijkheid van verworven Eigenschappen. 's Gravenhage. p. 279. (See abstract by V. Haecker in Biolog. Centralblatt, Bd. X. pp. 641-652, 686- 694. Also a brief abstract in Zoologischer Jahresbericht fur 1890, Alg. Biologie u. Entwick., p. 23.) || See Comptes Rendus, 1892, CXIV. 1565-1508. 844 PROCEEDINGS OF THE AMERICAN ACADEMY. regained when continued cultivation was carried on at lower tern- peratures." * In still more recent experiments Gley and Charriu f have vacci- nated a certain number of male rabbits against the effects of the bacillus of blue pus (Bacillus pyocyaneus) by injecting into them atten- uated cultures. These males then mated with females while in heat. The greater number of the offspring died before birth or soon after ; those which lived were atrophied, more or less deformed, but in some cases endowed with an immunity from the effects of the bacillus of blue pus. "It is evident," says Cuenot, in commenting on this experiment, " that the attenuated infection communicated to the fathers by vacci- nation had profoundly deranged the structure of the male germ plasm ; hence the numerous abortions and malformations of their descendants. It is exactly the same as in the case of females alone which have before fecundation received injections of poisons: they either abort, or the young do not grow and die at an early age." That the effects of alcoholism and other forms of intoxication with poisons, etc., produced on individuals are inherited by the next gen- eration is allowed by M. Cuenot. Whether the effect upon the system saturates through and affects the germ plasm or not does not affect the fact that such lesions are acquired during the lifetime of the indi- vidual. As he says : — " Intoxication of the organism which is not fatal likewise reacts on the germ plasm, which may undergo profound modifications. Alcoholism, for example, which exaggerates in a way so characteris- tic the diatheses of parents (insanity, cirrhosis, etc.) also alters the sexual cells. The children of parents both affected with alcoholism, when born, are sickly, unhealthy, presenting a special predisposition to consumption and to nervous troubles." It is possible, adds Cuenot, that the celebrated observations of Brown-Sequard, confirmed by Dupuy and Obersteiner, on the heredity of epilepsy produced in guinea-pigs, is also explained by a partial in- * Bacteriology in its General Relations. By H. L. Russell. Amer. Naturalist, December, 1893, p. 1060. t Gley et Charrin, Influences here'ditaires experimentales. Comptes Rendus de l'Acade'mie des Sciences, Paris, CX VII., 1893. Quoted from L. Cuenot in " Revue gen. des Sciences, pures et appliquees," 15 Fe'v., 1894. This and other cases are cited by M. Cuenot in favor of Weismannism, but we think they directly prove the contrary ; they illustrate the direct action of a change of environment during the lifetime of the individual, the changes being inherited by the suc- ceeding generation. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 845 toxication of the germ plasm. The guinea-pigs are rendered epileptic by different procedures, i. e. section of the spinal cord, of the sciatic nerve, of the great sympathetic, etc. The youug to which they give birth after these operations often present (17 times in 30) a general feebleness, and various nervous affections ; motor paralysis of the fore or hind legs, or trophic paralyses, which result in the loss of toes, of the cornea, etc. ;* in 32 youug born of epileptic parents two have shown symptoms of epilepsy. " The transmission of nervous troubles cannot then leave the shadow of a doubt, although in no case has the mutilation which has been the cause of it in the parents been repro- duced in the descendants.'" Cueuot then adds, " We are still in the presence of an infection of the germ plasm, due perhaps to bacteria as in the case of syphilis (an opinion sustained with much force by Weismann), perhaps also to poisons secreted by epileptic parents, which carries us back to the case of alcoholism. " Another case is the result of experiments by Paul Bert,f who at- tempted to acclimatize some Daphniae in salinated water by adding from day to day a little salt in the water of an aquarium. At the end of 45 days, when it contained 1.5% of salt, all the Daphniae had died, but the eggs contained in their brood-sac survived, and the new generation of Daphniae to which they had given birth flourished per- fectly well in the same medium. " This experiment," adds Cuenot, "shows with admirable clearness that the germ plasm has, owing to the modification, become accustomed to the salt, causing it to produce a generation so different from the preceding." We should interpret these facts as showing that the Crustacean had been so profoundly affected in the lifetime of the individual as to produce young per- fectly adapted to a changed environment ; the germ plasm may have been the vehicle, all the same, but the experiment is a case in favor of the Neo-Lamarckian principle. Other authors who have advocated the views that acquired char- acters may be transmitted are Cope,| Ryder,§ Oshorn, Vines, || * It appears that, if the young are born with tails, an even more impor- tant lesion, the loss of toes and of the cornea, etc., results. Why should not these cases strongly confirm the Lamarckian principle of the inheritance of characters acquired during the lifetime of the individual '? t P. Bert, Sur la Cause de la Mort des Animaux d'Eau douce qu'on plonge dans l'Eau de Mer. Comptes Rendus de 1'Acade'mie des Sciences, XCVII., Paris, 1883. \ Origin of the Fittest, 1887, and minor papers. § A Physiological Hypothesis of Heredity and Variation. Amer. Naturalist. January, 1890, pp. 85-92. || Lectures on the Physiology of Plants. Cambridge, 188G. 346 PROCEEDINGS OP THE AMERICAN ACADEMY. Eiineiy* Geddes and Thompson, f Brown-Sequard, j and Giard,§ as well as Henslow, Lloyd Morgan, and others. Eimer's work is entirely based on the view that the inheritance of acquired characters is a fundamental law of organic growth, though he allows " that the permanent action of external conditions on the body of the organism in most cases is not immediately perceptible. From physiological principles this is not in general possible." He thinks that more time than even Darwin supposed to be necessary must be invoked. He adds : " Every character which must have been formed through the activity of the organism is an acquired character. All characters, therefore, which have been developed by exertion, are ac- quired, and these characters are inherited from generation to genera- tion. The same holds for all organs atrophied through disuse ; the degree of atrophy is acquired and inherited. In the first class we see especially the action of direct adaptation, in the second the results of the cessation of this action. A third class of acquired characters are to be traced simply to the immediate action of the environment on the organism, and originally, at the commencement of their appearance, all characters must have belonged to this class." (p. 87.) At present it seems the better course, now that the hypothesis has been so fully discussed, to wait for more facts, and to very thoroughly test the cases which seem to favor or to oppose the doctrine. Some of the discussions held on this subject have savored of the metaphysics of the Middle Ages, and quite artificial distinctions, with the invoking of " natural selection " by authors whose natural selec- tion is quite a different doctrine from the natural selection of Darwin ; and other occult causes have been given undue prominence. Meanwhile we feel justified from the facts now known in holding the view that characters acquired in the lifetime of the individual, as the result of functional activity in certain regions of the body or in certain organs, may under favorable conditions be more or less cora- * Organic Evolution as the Result of the Inheritance of Acquired Characters, etc. Translated by J. T. Cunningham. London, 1890. t Evolution of Sex, 1890. See also Dr. O. von Rath's Criticisms of some Cases of Apparent Transmission of Mutilations. Translated by Prof. H. B. Ward, Amer. Naturalist, with the bibliography at the end of the article, Jan- uary, 1894. | Faits nouveaux e'tablissant l'extreme Fre'quence de la Transmission par Here'dite d'Etats organiques morbides, produits accidentellement chez des Ascendants. (Comptes Rendus de l'Acade'mie des Sciences, 13 Mars, 1882.) § L' Here'dite des Modifications somatiques. Revue Scientifique, Tom. XL VI. No. 23, 6 Dec, 1890. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 347 pletely transmitted ; or at least the tendency to such transmission, if latent in one generation, may appear in a succeeding one. And in the earlier geological ages this principle may have been much more active than at. present. The hypothesis seems to be a good working one to account for phenomena which cannot be otherwise explained, and should not iu consequence of adverse, though often very able and candid criticism, be set aside. On the contrary, if as the result of Weismann's criticisms it be only provisionally adopted, should it not lead to further experiments in the laboratory, and to further and more thorough studies of the metamorphosis of animals, with a view to ascertain how far they are correlated with changes of habit and function ? The lines of future investigation in this field appear to lie mainly in four directions : — 1. In the domain of comparative cytology. 2. In the study of the life histories or metamorphoses of animals. 3. In the further observation of the facts of heredity as observed in the raising of plants, in the breeding of domestic animals, and that of the different races of mankind, and especially by laboratory experi- ments like those of Semper, Bert, and others, in changing the sur- roundings of organisms. 4. In the study of the gradual modification and specialization of some organs in forms now extinct, with the degeneration and loss of others, a subject so fully worked out by Professor Osborn as regards the teeth of mammals. III. Inheritance at corresponding Periods of Life. But perhaps the crucial cases will be found to occur in animals with a complicated metamorphosis, because in such instances we can draw the line between characters which are congenital and those which are acquired. As regards the characters which appear in post-embryonic life, it is not difficult to see that they have originated in response to stimuli brought about by changes in the environment. My attention has been turned to this subject while studying the complicated life histories of some of the Bomhycine moths, in which there are usually five distinct larval stages, and sometimes as many as nine, not to mention the pupal and imaginal stages. Now in each and all of these stages the organism is as a rule adapted to some more or less temporary change in its environment. It seems, the more closely we observe the habits of some of these caterpillars, almost capable of demonstration that the different temporary colors, 348 PROCEEDINGS OP THE AMERICAN ACADEMY. markings, organs, and structures developed, and for the time being useful only to become at a later stage useless and therefore discarded, as new and dissimilar conditions of life arose, — it seems almost self- evident that such markings and structures were the result of the re- sponses of an organism in its most plastic time of life to changes in its habits, such changes being due either to changes in its surround- ings, or to the effort to repel the attacks of insects, birds, etc. However such stages arose, they are at the present epoch trans- mitted from parent to offspring with wonderful certainty. Among the Arctians, and in other caterpillars, the number of moults is known to vary, either from artificial breeding or from other unknown causes, possibly lack of nutrition. This form of heredity was called by Dar- win* ''Inheritance at corresponding Periods of Life," and by Haeckelf " Homochronous Transmission." Darwin thus describes the phenome- non : " When the embryo leads an independent life, that is, becomes a larva, it has to be adapted to the surrounding conditions in its struc- ture and instincts, iudepeudently of those of its parents ; and the principle of inheritance at corresponding periods of life renders this possible." (p. 51.) Again: "On this principle of inheritance at corresponding periods, we can understand how it is that most animals display from the germ to maturity such a marvellous succession of characters." (p. 60.) Examples of this law are the complicated metamorphosis of certain free, but more especially the parasitic worms, notably the fluke worms and the Cestodes, the complicated metamorphoses of the Echinoderms, of the Mollusca, the Crustacea, and the metamorphic insects, and more especially such insects as the Meloidae, Rhipiphoridai, and Stylopitke, in which there is a hypermetamorphosis. It is not altogether improbable that the phenomena of alternation of generations is primarily due to changes in surroundings, and hence of habits, resulting in new needs which were met by adaptation to the new surroundings, the different stages being finally fixed by homochronous heredity. Take the case of the Hydroids, where the generation of fixed hydra-like individuals gives rise by budding to the free-swimming, egg-producing medusa form. The hydra-like indi- viduals are the result of direct inheritance, while the medusa is prob- * The Variation of Animals and Plants under Domestication, II. 51. t History of Creation, I. 217, 218. See also Giard, wlio, referring to the laws of heredity, remarks: " Plusieurs de ces lois, et en particulier la lot de l'heredite homoehrone, fournissent aussi, nous le verrons, de bons arguments en faveur du principe de Lamarck." Revue So., 6 Dec, 1890. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 340 ably a secondary product, the result of adaptation to a free-swimming mode of life; all its organs in a so much higher scale of perfection than those of the generalized hydra-form ancestor having resulted from the manifold stimuli of a changeable environment. The same may be said of the Amelia, with its free Scyphistoma larval stage.* The alternation of generations in the Trematodes is apparently like- wise the result of adaptation to a change of hosts bringing about homo- chronous heredity. It is evident that alternation of generations is the extreme of a series, beginning (1) with simple direct growth; (2) an incomplete metamorphosis (the lower winged insects) ; (3) genuine metamorphosis ; (4) the hypermetamorphosis of Meloidre, Rhipi- phoridae, and Stylopidae ; the oth and last term being the cases of alternation of generations. It seems difficult to account for these sets of individuals or generations, unless we resort to the principle of inheritance of characters acquired during the lifetime of the ancestral forms, which gave rise to these interrupted or alternate series of forms. On the other hand Weismann in order to account for alternation of generations carries us out of the sphere of observation and induction into speculative regions, and assumes that >k two kinds of germ plasm exist \n those species in xohich alternation of generations occurs, both of which are present in the egg cell as well as in the bud, though only one of them is active at a time and controls ontogeny, while the other remains inactive. The alternating activity of these two germ plasms causes the alternations of generations." | Are not over-nutrition and changes in station and habits the more appreciable and potent causes ? The complicated metamorphosis of the Crustacea is the result of the adaptation to variations in the environment. It is not improbable that the Nauplius of the primitive Crustacea, the Branchiopoda (including the Phyllopoda), was a secondary and not a primary form. The proof would seem to be the non-existence of any adult Arthropod with such a form and structure as that of the Nauplius. The earliest Crustacean was probably a naked, shelless Cladoceran or Phyllopod-like form, with a few or many definite segments, bearing unjointed or imperfectly jointed lamellate swimming legs, derived from the flattened parapodia * In his suggestive book entitled " A Theory of Development and Heredity " (1803), Prof. H. B. Orr accounts for alternation of generations by secondary changes of the environment which favored the hydroid stage and the perfect medusa stage, and at the same time tended to eliminate the intermediate stages. In other cases the secondary changes of environment destroye 1 the hydroid stage, as we find Medusae without any hydroid stage, (p. 22-3 ) t The Germ Plasm, p. 457. 350 PROCEEDINGS OP THE AMERICAN ACADEMY. of some ancestral Annelid worm. The cylindrical rowing appendages of the Nauplius appear to be secondary and adaptive characters, fitting it for its free-swimming surface life. The Zoea larva of the Decapoda, with its body composed of head and abdomen alone, without thoracic segments and appendages, is also an adaptive stage, a differentiated or farther advanced Nauplius. The Megalops stage of the Brachyura, or crabs, induced at the end of the free-swimming life of the zoea, and intermediate between the zoea and crab stage, the thorax and thoracic appendages being present, is likewise an adaptation to the transition period connecting the free- swimming or surface life and the creeping and bottom life of the adult. These different stages, the result of adaptation, are signal examples of the inheritance of characters at corresponding periods of life, and would appear to have beeu originally the result of the inheritance of characters originated or acquired during the life of the individual ; i. e. the ancestor of the existing decapodous Crustacea. On the other hand, in groups where a metamorphosis is the rule, there are exceptional forms in which development is abbreviated or direct. Such cases are the direct development of the starfish, Lepty- chaster kerguelenensis Smith, while Pteraster militaris is viviparous ; and the direct development of Anochanus sinensis and of Hemiaster caver- nosas among Echinoids. The lobster and crayfish are exceptions to other macrurous Crustacea, in which there is a complicated metamor- phosis, their development being condensed or abbreviated, and limited to embryonic life. So with the crabs of several species living in the Black Sea, whose direct development was traced by Rathke. In such exceptional cases as these, the phenomenon of direct de- velopment will undoubtedly be found to be caused by some change in the conditions of existence. Attention should also be here drawn to the fact that the term con- genital is an elastic one. Mammals, at least the placental ones, are only born after a long uterine life, but fish and tadpoles, as well as the higher nesting birds, are born in a more premature condition, and congenital characters in these animals have quite a different significance from those of mammals. So it is in a less degree with the larvae of insects, the degree of inequality in the perfection of the larva being very great in different groups, especially in the parasitic Hymenoptera and Coleoptera. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 351 IV. HOMOCHRONOUS HEREDITY IN INSECTS WITH A HTPER- METAMORPHOSIS. The hypermetamorphoses of the Meloidos. Rhipiphoridse, and Stylo- pidne very strikingly illustrate the principle we are endeavoring to emphasize and establish. The facts are given in the writings of New- port, Fabre, Westwood, Siebold, Valery-Mayet, Riley, and others. In Meloe' the freshly hatched larva, or " triungulin." is an active Campodea-like larva, which runs about and climbs up flowers, from which it creeps upon the body of bees, such as Anthophora, who carries it to her cells, wherein her eggs are situated. The triungulin feeds upon and destroys the eggs of its hostess. Meanwhile its in- active life in the bee's cell reacts upon the organism ; after moulting, the second larval form is attained, and now the body is thick, cylin- drical, soft, and fleshy, and it resembles a lamellicorn larva, with three pairs of rather long thoracic legs. This second larva feeds upon the honey stored up for the young or larval bees. After another moult, there is another entire change in the body; it is motionless, the head is mask-like without movable appendages, and the feet are represented by six tubercles. This is called the semi-pupa or pseudo-pupal stage. This form moults, and changes to a third larval form, when apparently, as the result of its rich concentrated food, it is overgrown, thick-bodied, without legs, and resembles a larval bee. After thus passing through three larval stages, each remarkably different in structure and in the manner of taking food, it transforms into a pupa of the ordinary coleopterous shape. The history of Sitaris, as worked out by Fabre and more recently by Valery-Mayet, is a similar story of two strikingly different adapta- tional larval forms succeeding the triungulin or primitive larval stage. The first larva is in general like that of Meloe, the second is thick, oval, fleshy, soft-bodied, and with minute legs, evidently of no use, the larva feeding on the honey stored by its host. The pseudo-pupal stage is still more maggot-like than in the corresponding stage of Meloe, and the third larva is thick-bodied, with short thoracic legs. In the complicated life history of another Cantharid, Epicauta vit- tr/ta, as worked out by Dr. C. V. Riley, we have the same acquisition of new habits and forms after the first larval stage, which evidently were at the outset the result of an adaptation to a change of food and surrroundings. The female Epicauta lays its eggs in the same warm, sunny situation as that chosen by locusts (Caloptenus) for depositing their eggs. On hatching, the active minute carnivorous triungulin, 352 PROCEEDINGS OF THE AMERICAN ACADEMY. ever on the search for eggs, on happening npon a locust egg gnaws into it, and then sucks the contents. A second egg is attacked and its contents exhausted, when, owing to its comparatively inactive habits and rich nourishing food after a period of inactivity and rest, the skin splits along its back, and at about the eighth day from beginning to take food the second larva appears, with much smaller and shorter legs, a much smaller head, and with reduced mouth-parts. This is the Cara- bicloid stage of Riley. After feeding for about a week in the egg a second moult occurs, and the change of form is slight, though the mouth- parts and legs are still more rudimentary, and the body assumes " the clumsy aspect of the typical lamellicorn larva." This Riley denomi- nates the Scarabaeidoid stage of the second larva. After six or seven days there is another transformation, the skin being cast, and the insect passes into another stage, " the ultimate stage of the second larva." The larva, immersed in its rich nutritious food, grows rapidly, and after about a week leaves the now addled and decaying locust eggs and burrows into the clear sand, where it lies on its side in a smooth cell or cavity, and where it undergoes an incomplete ecdysis, the skin not being completely shed, and assumes the semi- pupa stage, or coarctate larval stage of Riley. In the spring the partly loose skin 'is rent on the top of the head and thorax, and then crawls out of it the " third larva," which only differs from the ultimate stage of the second larva " in the somewhat reduced size and greater whiteness." The insect in this stage is said to be rather active, aud burrows about in the ground, but food is not essential, and in a few days it transforms into the true pupa state. These habits and the corresponding hypermetamorphosis are proba- bly common to all the Meloidae, though the life history of the other species has yet to be traced. In the genus Hornia described by Riley, the wings of the imago are more reduced than in any other of the family, both sexes having the elytra as rudimentary as in the European female glow-worm (Lampyris noctiluca). These, with the simple tarsal claws and the enlarged heavy abdomen, as Riley remarks, " show it to be a degra- dation al form." Its host is Anthophora, and the beetle itself lives permanently in the sealed cells of the bee, and Riley thinks it is subterranean, seldom if ever leaving the bee gallery. The triungulin is unknown, but the ultimate stage of the second larva, as well as the coarctate larva, is like those of the family in general, the final transformations taking place within the two unrent skins, in this respect the insect approach- ing Sitaris. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 353 It appears, then, that as the result of its semi-parasitic mode of life the Campodea-form or triungulin larva of these iusects which have free biting mouth-parts like the larva? of Carabida? and other carnivorous beetles, instead of continuing to lead an active life and feed on other insects, living or dead, and then like other beetles directly transform- ing into the normal pupa, moult as many as five times, there being six distinct stages, before the true pupal stage is entered upon. So that there are in all eight stages including the imagiual or last stage. One cannot avoid drawing the very obvious conclusion that the five extra stages, constituting this hypermetamorphosis, as it is so well styled, were structural episodes, so to speak, due to the peculiar para- sitic mode of life, and were evidently in adaptation to the remarkable changes of environment, so unlike those to which the members of other families of Coleoptera, the Stylopida? excepted, have been subjected. The fat overgrown body and the atrophied limbs and mouth-parts are with little doubt due to the abundant supply of rich food, the proto- plasm of the egg of its host, in which the insect during the feeding time of its life is immersed. Since it is well known that parthenogenesis is due to over, or at least to abundant nutrition, or to a generous diet and favoring temperature, there is little reason to doubt that the greatly altered and abnormally fat or bloated body of the insect in these supernumerary stages is the result of a continuous supply of rich pabulum, which the insect can imbibe with little or no effort. The life history of the Stylopida? is after the same general fashion, though we do not as yet know many of the most important details. The females are viviparous, hatching within the body of the parent, as I once found as many as 300 of the very minute triungulin larva? issuing in every direction from the body of what I have regarded as the female of Stylops childreni in a stylopized Andrena caught in the last of April. The larva? differ notably from those of the Meloida? in the feet heing bulbous and without claws, yet it is in general Cam- podea-like and in essential features a triungulin. The intestine ends in a blind sac, as in the larva? of bees, and this would indicate that its food is honey. The complete life history of no Stylopid is com- pletely known. It is probable that, hatched in June from eggs fer- tilized in April, the larva? crawl up on the bodies of bees and wasps; finally, after a series of larval stages as yet unknown,* penetrating * Westwood in his excellent account of tins group remarks " Ilenee, as well as from the account given by Jurine, it is evident tliat the pupa of the Stylops is enclosed in a distinct skin, and is also in that state enveloped by the skin of vol. xxix. (n. s. xxi.) 23 354 PROCEEDINGS OF THE AMERICAN ACADEMY. within the abdomen of its host before the latter hibernates, and living there through the winter. The females, owing to their parasitic life, retain the larval form, while the free males are winged, not leading in the adult stage a parasitic life, though passing their larval and pupal stages in the body of their host, and are so unlike ordinary beetles as to be referred by good authorities to a distinct order (Strepsiptera). The triungulin stage of these insects corresponds in general to the form of the larval Staphylinidae and allied families, such as the Tene- brionida?, which are active in their habits, running about and obtain- ing their food in a haphazard way, often necessarily suffering long fasts. In these external-feeding, less active coleopterous larva?, like the phytophagous species, which have an uninterrupted supply of nutritious food, we see that the body is thick and fleshy. So also in the larvae of the Scarabaeidas, Ptinida?, and the wood-boring groups. In internal feeders, like the larval weevils and Scolytid;c, which live nearly motionless in seeds, fruits, and the sap-wood of plants and trees, with a coustant supply of nourishing, often rich food, the eruciform body is soft, thick, and more or less oval-cylindrieal. So it is with the larva? of Hymenoptera, especially in the parasitic forms, and in the ants, wasps, and bees, which are nearly if not quite motionless, at least not walking about after their food. Now the change from the active triungulin stage to the series of secon- dary nearly legless, sedentary, inactive stages is plainly enough due to the change of station and to the change of food. From being an independent, active, roving triungulin, the young insect becomes a lodger or boarder, fed at the expense of its host, and the lack of bodily exertion, coupled with the presence of more liquid food than is actually needed for its bare existence, at once induces rotundity of body and a loss of power in the limbs, followed by their partial or total atrophy. That this process of degeneration may even occur in one and the same stage of larval existence is very well illustrated by what we know of the life history of the wasp parasite of Europe, Rhipiphorus paradoxus. Thanks to the very careful and patient observations of Dr. T. A. Chapman, we have a nearly complete life history of this beetle, the representative of a family in many respects connecting the Meloidoe and Stylopidae.* Where Rhipiphorus lays her eggs is un- file larva, contrary to the suggestion of Mr. Kelly." Class. Insects, II. 297. This is all we know about the supernumerary larval stages. * Some Fncts towards a Life History of Rhipiphorus paradoxus. Annals and Magazine of Natural History for October, 1870. PACKARD. INHERITANCE OF ACQUIRED CHARACTERS. OOJ known. Dr. Chapman however found a solitary specimen of the young larva in the triuiigulin stage. He describes it as "a little black bexapod, about ^ti inch (.5 mm.) in length, and T.] () inch in breadth, broadest about the fourth segment, and tapering to a point at the tail; a triangular head with a pair of three-jointed antennae nearly as long as the width of the head, with legs very like those of Meloe larvae ; the tibiae ending in two or three claws, which are supported and even obscured by a large transparent pulvillus or sucker of about twice their length ; this was marked by faint striae radiating from the ex- tremity of the tibia?, giving it much the aspect of a lobe of a fly's proboscis. Each abdominal segment had a very short lateral spine pointing backwards; the last segment terminated by a large double sucker similar to those of the legs; and the little animal frequently stood up on this, and pawed the air with its feet, as if in search of some fresh object to lay hold of." This almost microscopic larva finds a wasp grub and bores into its body, probably entering at a point near the back of the first or second segment behind the head. Dr. Chapman succeeded in finding the larva of the beetle within that of the wasp, before the latter had spun up. " Assuming that the wasp larva lives six days in its last skin be- fore spinning up, I should guess that the youngest of these had still two or three days' feeding to do. The Rhipiphorus larvae were but a little way beneath the skin of the back, about the fourth and fifth segments [counting the head as the first], and indifferently on either side. The smallest of these was T\j inch in length, and, except its smaller size, was precisely like the larger ones I am about to refer to, having the same head, legs, plates, etc. These were of the same size as those of the larger larvae, the difference in size of the latter being due to the expansion of the intermediate colorless integument." After the wasp grub has spun the silken covering of its cell the larva of Rhipiphorus may be still detected in some of them, being rendered visible by its black legs and dark dorsal and ventral plates. " On extracting this larva, it bears a general resemblance in size and outline to the youngest larva of Rhipiphorus that I had found feeding externally on the wasp grub, but with the very notable exception of the already mentioned black marks. These are, in fact, a corneous head, six-jointed legs, and a dorsal and ventral series of plates. I immediately recognized the head and legs as identical with those of the little black mite already described, but presenting a ludicrous ap- pearance in being widely separated from each other by the white skin of the larva. I have no doubt that the dorsal and ventral series of black 85G PROCEEDINGS OF THE AMERICAN ACADEMY. marks are the corresponding plates of the mite-like larva floated away from each other by the expansion of the intervening membrane. By measurement also they agree exactly in size, although the larva extracted from the wasp grub is ten times the length and six times the width of the little Meloe-like larva. In length it is £ inch (4.5 mm.) and o's inch in breadth." The remarkable changes thus described in the larva of this beetle after it has begun its parasitic life within the body of its host are especially noteworthy because the great increase in size and difference in shape, as well as in habits, all take place before the insect has moulted. The rapid development in size, and consequent distention of the body and the separation of the sclerites of the segments behind the head, are paralleled, as Chapman says, by the greatly swollen abdomi- nal region of the body in Sarcopsylla penetrans and in the female of the Termitida?. In those insects this distention is due to the enlarge- ment of the ovaries and of the egg« contained within them, but in the Rhipiphorus it is due to the comparative inactivity of the larva, and to its being gorged with an unending supply of rich food, the blood and fat of its host. It follows, then, that if a sedentary life, and over, or at least abundant nutrition, will have this effect within the short period covered by the single first larval stage of the Rhipiphorus, it is reasonable to infer that the hypermetatnorphosis is also due to the same factors. Chapman then goes on to say, that finally, within six hours of the time of spinning up of the wasp grub, the Rhipiphorus larva at the end of Stage I., which is " usually in motion, and for its situation might be called tolerably active, is seen to lay hold of the interior of the skin with its anterior legs, and keeps biting and scratching with its strong and sharp jaws until it is able to thrust through its head, when, in less than a quarter of an hour, it completely emerges by a vermiform movement ; and at the same time it casts a skin, together with the black head, legs, plates, etc." The larva, now in its second stage, passes forward and seizes hold of the upper or lateral aspect of the prothoracic segment of the wasp grub. On emerging it becomes shorter and thicker, "and very soon loses length by that curving forward of its head which is so marked in the full-grown larva, and which does not exist before its emergence." The larva is now found "lying like a collar immediately under the head of the wasp grub, and is attached to it by the head, though not very firmly. At this stage the feeding of the young Rhipiphorus is rather sucking than eating. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 357 When about 6 mm. in length it moults a second time, and the full- grown larva closely though superficially resembles a Crabro or Pem- phredou larva, the small head being bent over forwards. By the time it is ready to pupate it has wholly eaten the wasp larva, and the temperature of the cell being high, a larva 5 mm. long grows large enough in two days to fill the top of the cell of its host, and the larva is ready to pupate about a week after hatching, so that its develop- ment is very rapid. The beetles themselves do not live in the cells. Chapman thinks they hibernate, and that the eggs are laid in the spring or summer. We thus have in this insect three larval stages, the triungulin, and two later stages, the great differences between the first and the last two being apparently due to their parasitic mode of life, the larva spending its second stage within its host, involving an existence in a cell with a high temperature, an uninterrupted supply of rich, stimulat- ing food, and a comparatively sedentary mode of life compared with that of the triungulin at the beginning of its existence. It is quite obvious that the hypermetamorphosis is primarily due to a great change in its surroundings, i. e. the parasitic mode of life of the beetle, habits of very rare occurrence iu the Coleoptera, numerous iu species as they are. In the Proctotrypidae there is also a hypermetamorphosis.* In a species of Platygaster which is parasitic in the larva of Cecido- myia, the first larva (Cyclops stage) is of a remarkable shape, not like an insect, but rudely resembling a parasitic Copepod crustacean. In this condition it clings to the inside of its host by means of its hook- like jaws, moving about, as Gauiu says, like a Cestodes embryo with its well known six hooks. In this stage it has no nervous, vascular, or respiratory system, and the digestive canal is a blind one. After moulting the insect entirely changes its form, it is thick oval cylindrical, nearly motionless, with no appendages, but with a digestive canal aud a nervous and vascular system. After a second moult the third and last larval stage is attained, and the insect is of the ordinary appearance of ichneumon larv;e. * Metschnikoff, Embryologische Studien an Insecten. Zeitsclirift fiir wis. sens. Zoologie, XVI. 389-500, 1806. Ganin, Beitriige zur Erkenntniss der Entwickelungsgeschichte bei den In- secten. Zeitsclirift fur wissens. Zoologie, XIX. 381-451, 1869. Ayers, On the Development of CEcanthus nivens and its Parasite, Teleas. Memoirs Bost. Soc. Nat. Hist , III. 225-281, 1884. For an abstract of the work of Ganin, see Balfour's Comparative Embry ology, I. 345-348; also Packard's Our Common Insects, 1873, pp. 161-107. 358 PROCEEDINGS OF THE AMERICAN ACADEMY. Not less striking is the life history of Polynema, which lays its eggs in those of a small dragon-fly (Agrion virgo). The first larval stage is most remarkable. It hatches as a microscopic immovable being, entirely unlike any insect, with scarcely a trace of organization, being merely a flask-shaped sac of cells. After remaining in this state five or six days it moults. The second stage, or Histriobdella-like form, as Ganiu names it, is more like that leach-like worm than an insect. The third larval form is very bizarre, though more as in insects, having rudimentary antenna?, mouth-parts, legs, and ovipositor. In this condition it lives from six to seven days before pupating. The strange history of another egg parasite (Ophioneurus) agrees in some respects with that of the foregoing forms. It is when hatched of an oval shape, with scarcely any organs, and differs from the genera already mentioned in remaining within its egg membrane, and not assuming their strange shapes. From the cylindrical sac-like non- segmented larva resembling the second larva of Platygaster it passes directly into the pupa state. A fourth form, Teleas, is an egg parasite of Gerris, and in America one species oviposits in the eggs of CEcanthus. The spindle-shaped larva iu its first stage roughly resembles a trocho- sphere of a worm rather than the larva of an insect so high in the scale as a Hymenopter. It is active, but after moulting the second larva is oval, still without segments. Dr. Ayers gives a profusion of details and figures of the first and second stages of our Teleas, the second strongly resembling the Cyclops stage of Ganin. He describes three stages, and though he did not complete the life history of the insect he thinks it changes to an ovoid flattened form which succeeds the Cyclops stage in other Pterornalida?, and that there are at least four ecdyses. It is difficult to account for these strange larval forms, unless we suppose that the embryos, by their rich abundant food, have undergoLe a premature development, the growth of the body walls being greatly accelerated, the insects so to speak having been, under the stimulus of over-nutrition and their unusual environment, and perhaps also the high temperature of the egg, hurried into vermian existence on a plane scarcely higher than that of an active ciliated gastrula. Further observations, difficult though they will be, are needed to enable us to account for the singular prematurity of the embryo of these parasites. That these stages are reversional and a direct inher- itance from the vermian ancestors of these insects is not probable, but PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 359 the forms are evidently the result of adaptation in response to a series of stimuli whose nature is in part appreciable but in part unknown. It may be noted, however, that the appearance of a primitive band in the second larval stage indicates tbe origin of these forms, as well as that of insects in general, from a Peripatus-like, and again from an earlier leach-like Annelid ancestor. Hence the first larval or Cyclops sta^e is due to a precocious development caused by the unusual environment, and is simply adaptational, and not of phylogenetic significance. *&* V. Ox the Inheritance of Acquired Characters in Lepidoptera. Perhaps in no other group of animals may we study the subject of the inheritance of acquired characters with more success than in the Lepidoptera. In these insects the four stages of existence, the egg, larva, pupa, and imago, are definite and fixed, and during each of the last three the organism is, so to speak, a differeut creature, with dis- tinct and separate shape and structure external and internal, and during each stage leads a different life. Family, generic, and specific charac- ters are inherited at each of these stages, and at each there is a com- bination of congenital and acquired characteristics, some. of both classes of which, i. e. the least marked, are difficult to separate from one another. The following is an attempt at a rough grouping of such features at the last three stages. We omit the egg stage, for though they more or less vary in shape and ornamentation, this is perhaps due more to difference in the struc- ture of the lining of the oviduct of the female than to the action of external circumstances on the egg after it has been laid. Yet this should be said with some reservation, because we are not aware that any one has discussed the probable mode of origin of the specific dif- ferences in the shape and color of the eggs of birds, or the shape and markings of the eggs of insects ; though undoubtedly the agency of external causes together with natural selection has something to do with the variation.* * It has seemed to us that the relation of specific and generic characteristics in the eggs of insects is a most difficult problem. Yet it should be observed that, while the differences in ornamentation and shape are primarily due to the impression on the shell received from the lining of the oviduct, yet the wonder- ful diversity we see in the eggs of insects is often readily seen to be correlated with the external conditions in which they exist, after having been deposited by the parent. As regards the eggs of birds, the thick solid shell and conico 860 PROCEEDINGS OF THE AMERICAN ACADEMY. In the larval histories already published we have endeavored, where they have been observed with sufficient completeness, to discriminate between the congenital and the acquired characters. 1. Larval State. — A. In this state we have the inheritance of congenital characteristics. B. Inheritance of what were originally acquired characters, the results of attacks of enemies. Examples are the tubercles armed with spines aud sometimes with a singular kind of easily detached spines which are hence called caltrops (Empretia, etc.), stripes, spots: all ap- parently inherited at different periods of larval life; the least impor- tant specific and varietal characters probably having been originally acquired during the life of an individual. 2. Pupal State. — A. Cocoon : the absence or presence of a cocoon was doubtless originally due to differing external conditions : while the dense, perfect cocoon is characteristic of the spinning moths (At- tacidce, Lasiocampidce, etc.) ; the Ceratocampidaa make none at all, but, like the Sphinges, the larvre simply bury themselves in the earth before pupation. In the ArctiidaB and the Liparidoe the cocoon is chiefly composed of the barbed larval hairs, with a little silk to fasten them more firmly together: ir. the Geometridre certain larvae spin a loose thin web. In such cases the spinning of a cocoon is intimately associated with a change of larval habits, and is with little doubt an acquired habit, originally formed by a single individual. B. The shape of the pupa is often dependent on the presence or absence of a cocoon. In the NotodontidaB the cremaster is often absent in genera (Gluphisia, which spins a very slight cocoon, and Lophodonta which spins no cocoon) which do not spin a cocoon, oval shape of the murre's eggs seem due to the unprotected manner in which they are left on rocks and shelves from which they are liable to fall. Here might be cited the suggestive essay of Prof. John Ryder, " The Mechanical Genesis of theForm of the Fowl's Fgg," (Amer. Phil. Soc. Philadelphia, 1893, XXXI. 203-209,) in which he attempts to show that " the configuration of the outline of the hen's egg is determined by mechanical means, while the egg mem- branes and shell are in process of formation within the oviduct." We may con- trast with the murre's egg that of the robin, in which the shell is thin and uniform in color, since it is protected from harm by being contained in a nest; so also the color of murres' eggs may be due to the action of protective mimicry, the spots assimilating them to lichen-grown or variously tinted rocks, by which they escape the observation of their natural enemies, the fox, the mink, and other egg-devouring animals. So the eggs of Chrysopa, of lice, of many bugs, etc., are in shape and mode of attachment beautifully adapted to prevent them from being seen by egg-destroying animals. PACKARD. — INHERITANCE OP ACQUIRED CHARACTERS. 361 and are closely allied to those which do. In Cerura there is no spine on the rudimentary cremaster, because the pupa lies in a very dense cocoon. The cremaster affords excellent generic and specific charac- ters. In the subterranean pupa of Datana it is present, and is of use in aiding the pupa to reach the surface of the ground. It is very large and acute in the subterranean pupae of Ceratocampidae and Sphinges. It is evident that in the presence or absence of a cre- master, and in its shape, and in the number of hooks and their shape, we have a set of very plastic characters, whose variability and plasti- city are due to the varying habits of the pupa, whether living above or under ground, whether protected by a very thin loose netlike cocoon, or by a solid double one like that of Cerura or of the silk-worms. Also whether the thread is continuous and can be readily reeled, as in Bomhjx mori, or whether the thread is often interrupted at the ante- rior end, as in Platysamia cecropia, is a feature which was probably the result of a slight chauge of circumstances, and may have been inaugurated as the result of variation in a single individual, duriujj a single lifetime, becoming eventually fixed by homochrouic inheritance. III. Imago Stage. — It is easier to select what may have been ac- quired characters in caterpillars than in butterflies and moths, and yet the last have a complicated series of what may originally have been acquired characters. It should be borne in mind that, while caterpil- lars live for weeks and even months, are subject to frequent moults, are active and dependent on a proper supply of their food, usually this or that plant, butterflies and moths perish directly after mating, taking little or no food. Of course, acquired characters are most marked in the parts which are most used, as the maxillae, wings, and external genital armature. The absence of maxillae, or their very rudimentary condition, in Bombycine moths, is with little doubt a recently acquired character. The very arbitrary distribution in Lepidoptera of scent-organs (andro- conia, etc.) is apparently a character recently acquired. The won- derful variations in the markings of the wings, due to a variety of slight causes, may often arise during an individual's lifetime and become a matter of inheritance, the result of sudden changes in tem- perature, moistness or dryness, and changes in food of the larva. By subjecting individual pupae to prolonged cold or heat, varieties, and a greater or less number of broods, may be produced artificially, and this may illustrate how seasonal varieties have arisen in nature. Many species are only separated by differences in the male genital armature. These, as is well known, are subject to great individual 362 PROCEEDINGS OF THE AMERICAN ACADEMY. variation, and why should not the characters peculiar to a distinct variety, or even species, arise during the lifetime of two individuals when mated ? Many individuals die without being mated ; an unusu- ally vigorous polygamous butterfly may have some new congenital extra development of hooks and processes, and by frequent use develop the muscles controlling these to the extent of providing an acquired character, which may be, if useful, inherited in the next and succeed- ing generations. But an especially interesting and fruitful field of investigation would be a study of wingless Lepidoptera, such as the canker-worm, the autumn moths allied to it, the tussock moths (Orgyia), and especially the sac-bearers or Psychidae. The loss of wings in these cases seems originally to have been due to disuse in certain individuals more sluggish than others, and with little doubt has been the result of inheritance of what were originally acquired characters. It is easy to imagine how this has been induced by a study of a series of forms beginning with certain European genera in which the wings of the female are very small, and passing to those in which they become simple pads, as in the Orgyia, and ending with those, such as Anisopteryx, in which their reduction is still further carried out. And then examples like those should be com- pared with certain of the Ephemera?, whose hind wings are so much reduced ; with Pezzotettix and other Orthoptera with aborted wings, and certain Hemiptera in which the wings are aborted, ending with the great order of Diptera, comprising a vast number of species in which the hind wings have not only undergone a great reduction, but have been transformed through change of function into balancers, with their extraordinary sense-organs. It is not difficult to see that the disuse of wings may have begun in the life of a single individual, which, losing its wings, having perhaps inherited a tendency to thi3 lesion through corpulency and other bodily changes, became inactive, averse to flight, and finally transmitted the peculiarity or the tendency to such peculiarity to its offspring.* In a paper in the Proceedings of the Boston Society of Natural History (XXIV. 482) on the life history of Drepana arcuata, I have described the different stages of this moth, and at the end recapitu- lated the congenital characters, and finally given a synopsis of the chief * L. Knatz (Archiv fur Naturgeschichte, LVII. 49-74, 1 pi., 18G1) mentions 183 instances of reduced wings, and states that the reduction in wings is ac- companied by an enlarged abdomen and an increase in the size of the ovaries, with greater fertility. Journ. Royal Microscopical Society for 1801, p. 4G2. PACKARD. INHERITANCE OF ACQUIRED CHARACTERS. 363 steps in the evolution of the adaptational characters which appear after the first exuviation. It seems very probable that these later features were the result of the action of external stimuli, both physi- cal and biological, and that they were acquired not only during the lifetime of the larva, but at certain distinct stages or periods during the growth of the creature. The changes are both colorational and structural, aud during the different stages the larva was adapted for different surroundings, and thus at each important stage was virtually for the time being a distinct animal. During the pupa stage special aud unusual structural adaptations arose ; the cremaster being unusually developed, and also a pair of cephalic hooks, serving to entangle the head in the web of the cocoon, so that the pupa cannot be thrown out of the curled leaf which remains in the fir.-t brood on the tree. These I regard as characters acquired by the insect after birth, and in response to the exigencies of life at different stages. The reader is also referred to the conclusions given in that paper. Acquired Characters in the Notodontidoe. — In the systematic por- tion of my work on this group I have given a number of life histories of the family from papers recently published, and with more or less detail pointed out the later adaptational, as distinguished from the congenital characters. I have called attention, in late articles, to the varying shape of the tubercles and setre in the larvre of the Bom- byces and other of the higher Lepidoptera, and to their probable mode of origin, and why they appear on certain segments in preference to others. The attention of the reader is called to the summary or re- capitulation of changes especially in the life history of Datana inte- gerrima, Ajuitelodes torrefacla, Symmerista albifrons, Macrurocampa marthesia, and several species of Centra, while there is a summary of the steps in the assumption of the adaptive characters at the different larval stages of Schizura. The steps in the evolution of what may be regarded as acquired characters in Schizura. and in Dasylophia anguina, Hyparpax, Heterocampa, etc., are readily seen by an exami- nation of the plates in the monograph referred to. The Notodontians are remarkable in general for the humps, tuber- cles, and spines of their larvae, some of which are congenital, while others appear at different stages after birth. Still some larva; of this group are entirely without them, and remain so throughout their larval life. And this is an argument that the various processes of the cuticle or outgrowths of the entire integument are characters origi- nally acquired during the post-embryonic life of the young insect. 364 PROCEEDINGS OF THE AMERICAN ACADEMY. Take, for example, the larval Nadata gibbosa ; this, like the cater- pillar of Gluphisia and of Lophodonta is a smooth-bodied larva, ornamented with lines, but entirely unarmed. The life history tf N. gibbosa shows that it is born with a smooth body, without any rudiments of tubercles or enlarged bristles, while no traces of the yellowish subdorsal lines appear until at the end of the second stage. This form is therefore a primitive one, and this fact would seem to demonstrate that the humps, tubercles, and spines, so frequently ob- served in the group, arose within recent geological times, and were acquired during the postembryonic stages of the larvae of different genera, in response to various changes in the surroundings of different species, these finally becoming fixed and regularly transmitted along various lines of development, definite when the changes in the envi- ronment are definite, and resulting in a series of forms constituting the present genera of the family. One of the most notable cases in the family is that of the loss at about the middle of the larval life of the remarkable antlers of Hete- rocampa biitndata. During the three earliest stages the larva bears on the prothoracic segment a pair of enormous autlers, each with four tines. At the second moult these are discarded, and in the last two stages are represented by a pair of conical, rounded, polished, piliferous knobs. The rest of the partly grown body of the larva is smooth. After casting its horns the larva assumes a new ?et of color- ational markings, so that in its last two stages it is a total lv different creature in appearance from the earlier stages. I have also observed the wonderful changes undergone by the caterpillar of Heterocampa guttivvtta, representing five stages, nearly every one of which presents notable differences. In the first, directly after hatching, the reddish larva has not only a pair of enormous antlers with four tines on the first thoracic segment, but a pair of long antler-like spines on abdominal segments 1 to 6. and also 8 and 9 ; those on segments 1 and 8 being about three times as large as the others. It is certainly one of the most singular larvae of the family. Now this bizarre armature is entirely discarded at the first moult, excepting' that the prothoracic antlers are represented by a pair of knob-like tubercles, the other segments, however, showing no trace of the former existence of spines. Also, while the body was not striped in Stage I., it is now paler red with a more brownish tint, and is marked with four yellowish stripes. At the end of this stage the lines become effaced, and the body grows more yellowish on the sides. In the third PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 365 stage the tubercles still persist, but the markings differ very much, as reddish dorsal patches appear in the middle and near the end of the body, and there are anticipations of the markings of the fully grown caterpillar. In the present stage the insect closely resembles the mature larva, having bright crimson markings on the thoracic seg- ments, and on the third and fourth and on the fifth and sixth abdomi- nal segments, these bright spots becoming somewhat less decided and conspicuous in the final stage. Now it seems natural to suppose that the disappearance of the armature of this insect with the first moult was due to the lack of need for it by the caterpillar, which gradually became adapted to a life on the under side of an oak leaf, where it assumed a simple spindle- shaped body, extended when at rest along the midrib, in which posi- tion we have found such caterpillars, its body glaucous green, and so marked with yellowish lines and reddish spots, as well as with dashes and lines, as to be wonderfully assimilated to the greenish, reddish, and whitish hues and shades of the leaf under which it was sheltered. It also seems reasonable to suppose that these adaptational colora- tional features were acquired by the ancestors of the present forms during the different stages succeeding the first ecdysis. And thus we are warranted in assuming that this and multitudes of other cases of adaptation to the change in habits and modes of life, and special situa- tions, were acquired originally at different periods after birth during an earlier geological period than this, when the ancestors were fewer in number and more plastic than now. Otherwise, how can we have had the differentiation of a few ancestral forms into the present series of genera, subfamilies, and families, represented by such a great number of species ? Indeed, it seems diffcult to account for the evolution of the vast hordes of existing species of insects unless we assume that there was going on throughout the entire process the rise and perfecting of post- natal acquired characters, such characters becoming fixed by heredity, and reappearing with unerring certainty at different stages in the life of the individual ; while in some animals whose postnatal metamor- phosis through some environmental change became suppressed, we have the more salient stages epitomized during the life of the embryo.* * For many additional facts in the ontogeny of the Bombycine moths bear- ing on this subject the reader may be referred to my papers in the Proceedings of the Boston Society of Natural History, XXIV. 510-560, 1890 ; Journal of the 366 PROCEEDINGS OF THE AMERICAN ACADEMY. In his Chapter XIII. of "The Germ Plasm," Weismann gives an in- teresting example of the inheritance of climatic variation in butterflies. After an account of his recent experiments on Polyommatus Phlceas, he states: " Both experiments, however, prove the correctness of the old assumption of lepidopterists that the action of heat on a single gen- eration is capable of giving the German form of a blackish tint; and since, moreover, it is clear that the development of a single generation at a lower temperature can render the color of the Neapolitan butter- fly less black, it appears that the two varieties may have originated owing to a gradual cumulative influence of the climate, the slight effects of one summer or winter having been transmitted and added to from generation to generation. This would then seem to be an instance of the transmission of acquired characters." He then adds, that he does not believe that this is the correct interpretation of the facts; on the contrary he insists: "The theory of determinants will I believe supply a very simple explanation of this apparently compli- cated case, which I consider of great value, because it confirms this theory. Instead of supporting the doctrine of the transmission of somatogenic characters, this example shows how such a process may apparently be brought about, and on what it depends. A somatogenic character is not in this case inherited, but the modifying influence — the temperature — "ffects the primary constituents of the wings in each individual, — i. e. a part of the soma — as well as the germ plasm contained in the germ cells of the animal. It modifies the same determinants in the rudiments of the wings of the young chrysalis as in the germ cells, namely, those of the wing scales," etc. We certainly prefer the more simple explanation first given, but only to be rejected, by Weismann, since it appears to be really based on observed facts, and to be a natural and logical induction from such facts, and is thus a more scientific explanation. The same process of reasoning will apply to the inheritance of acquired characters in the ontoo-eny of the Lepidoptera and that of other groups, such as we have endeavored to set forth in this essay. It is, moreover, a simple and natural inference, such as in the case of the butterflies experimented on by Weismann, and would be the New York Entomological Society, I. 22-28, 57-76; Proceedings Amer. Phil. Soc. Philad., February, 1893, pp. 83-108, March, 1893, pp. 139-192; Proc. Amer. Acad. Arts and Sciences, Boston, 1893, XXVIII. 55-92 ; Annals New York Acad. Sei., May, 1893, VIII. 41-92. Also Monograph of the Bomhycine Moths of North America, Part I., in course of publication in the Memoirs of the National Academy of Sciences, Vol. VII. PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 307 first explanation which would suggest itself to any observer. On the other hand, the extremely complicated speculative and a priori second explanation of Weismann, based as it is on pure assumptions, does not carry conviction, and thus is not an efficient working hypothesis to explain inheritance at corresponding periods of life. It is noticeable that in his writings Weismann does not touch upon homochronous heredity, though his earlier work, "Studies in the The- ory of Descent" (187G), which is largely based on Lamarckian view.-, afterward abandoned by the author, is a storehouse of the most sug- gestive facts. In seeking to explain the causes of a metamorphosis in animals, one is compelled to go back to the primary factors of organic evolu- tion, such as the change of environment, whether the factors be cos- mical (gravity), physical changes in temperature, effects of increased or diminished light and shads, under or over nutrition, and the changes resulting from the presence or absence of enemies, or of isolation. The action of these factors, whether direct or indirect, is obvious, when we try to explain the origin or causes of the more marked meta- morphoses of animals. Then come in the other Lamarckian factors of use and disuse, new needs resulting in new modes of life, habits, or functions, which bring about the origination, development, and perfection of new organs, as in new species and genera, etc., or which in metamorphic forms m ly result in a greater increase in the number of, and an exaggeration of the features characterizing the stages of larval life. VI. The Adequacy of Neolamarckism:. It is not to be denied that in many instances all through the cease- less operation of these fundamental factors there is going on a process of sifting or of selection of forms best adapted to their surroundings, and best fitted to survive, but this factor, though important, is quite subordinate to the initial causes of variation, and of metamorphic changes. Neolamarckism,* as we understand this doctrine, has for its founda- * In 1885, in the Introduction to the " Standard Natural History," we proposed the term Neolamarckianism, or Lamarckism in its modern form, to designate the series of factors of organic evolution, and we take the liberty to quote the pas- sage in which the word first occurs. We may add that the briefer form, Neo- lamarckism, is the more preferable. "In the United States a number of naturalists have advocated what may be called Neo-Lamarckian views of evolution, especially the conception that in 368 PROCEEDINGS OF THE AMERICAN ACADEMY. tion a combination of the factors suggested by the Buflfbn and Geof- froy St. Hilaire school, which insisted on the direct action of the milieu, and of Lamarck, who relied on the indirect action of the en- vironment, adding the important factors of need and of change of habits resulting either in the atrophy or in the development of organs by disuse or use, with the addition of the hereditary transmission of characters acquired in the lifetime of the individual. Lamarck's views, owing to the early date of his work, which was published in 1809, before the foundation of the scieuces of embry- ology, cytology, palaeontology, zoogeography, and in short ail that distinguishes modern biology, were necessarily somewhat crude, though the fundamental factors he suggested are those still invoked by all thinkers of Lamarckiau tendencies. Keolamarckism gathers up and makes use of the factors both of the St. Hilaire and Lamarckian schools, as containing the more funda- mental causes of variation, and adds those of geographical isolation or segregation (Wagner and Gulick), the effects of gravity, the effects of currents of air and of water, of fixed or sedentary as opposed to active modes of life, the results of strains aud impacts (Ryder, Cope, and Osborn), the principle of change of function as inducing some cases rapid evolution may occur. The present writer, contrary to pure Darwinians, believes that many species, but more especially types of genera and families, have been produced by changes in the environment acting often with more or less rapidity on the organism, resulting at times in a new genus, or even a family type. Natural selection, acting through thousands, and some- times millions, of generations of animals and plants, often operates too slowly ; there are gaps which have been, so to speak, intentionally left by Nature. Moreover, natural selection was, as used by some writers, more an idea than a vera causa. Natural selection also begins with the assumption of a tendency to variation, and presupposes a world already tenanted by vast numbers of animals, among which a struggle for existence was going on, and the few were victorious over the many. But the entire inadequacy of Darwinism to account for the primitive origin of life forms, for the original diversity in the dif- ferent branches of the tree of life forms, the interdependence of the creation of ancient faunas and floras on geological revolutions, and consequent sudden changes in the environment of organisms, lias convinced us that Darwinism is but one of a number of factors of a true evolution theory ; that it comes in play only as the last term of a series of evolutionary agencies or causes ; and that it rather accounts, as first suggested by the Duke of Argyll, for the pres- ervation of forms than for their origination. We may, in fact, compare Dar- winism to the apex of a pyramid, the larger mass of the pyramid representing the complex of theories necessary to account for the world of life as it has been and now is. In other words, we bel'eve in a modified and greatly extended Lamarckianism, or what may be called Neo-Lamarckianism." PACKARD. — INHERITANCE OF ACQUIRED CHARACTERS. 369 the formation of new structures (Dohru), the effects of parasitism, commensalism, and of symbiosis, in short the biological environment ; together with geological extinction, natural and sexual selection, and hybridity. It is to be observed that the Neolamarckian in relying mainly on these factors does not overlook the value of natural selection as a guiding principle, and which began to act as soon as the world became stocked with the initial forms of life, but he simply seeks to assign this principle to its proper position in the hierarchy of factors. Natural Selection, as the writer from the first has insisted, is not a vera causa, an initial or impelling cause in the origination of new species and genera. It does not start the ball in motion ; it only so to speak guides its motions down this or that incline. It is the expres- sion, like that of " the survival of the fittest" of Herbert Spencer, of the results of the combined operation of the more fundamental fac- tors. In certain cases we cannot see any room for its action ; in some others we cannot at present explain the origin of species in any other way. Its action increased in proportion as the world became more and more crowded with diverse forms, and when the struggle for existence had become more unceasing and intense. It certainly cannot account for the origination of the different branches, classes, or orders of or- ganized beings. It in the main simply corresponds to artificial selec- tion ; in the latter case, man selects forms already produced by domes- tication, the latter affording sports and varieties due to change in the surroundings, that is, of soil, climate, food, and other physical features, as well as education. In the case also of heredity, which began to operate as soon as the earliest life forms appeared, we have at the outset to invoke the prin- ciple of the heredity of acquired characters during the lifetime of the lowest organisms. Finally, it is noticeable that when one is overmastered by the dogma of natural selection he is apt, perhaps unconsciously, to give up all effort to work out the factors of evolution, or to seek to work out this or that cause of variation. Trusting too implicitly to the supposed vera causa, one may close his eyes to the effects of change of environ- ment or to the necessity of constant attempts to discover the real cause of this or that variation, the reduction or increase in size of this or that organ ; or become insensible to the value of experiments. Were the dogma of natural selection to become universally accepted, further progress would cease, and biology would tend to relapse into a stage of atrophy and degeneration. On the other hand a revival of vol. xxix. (n. s. xxi.) 24 370 PROCEEDINGS OF THE AMERICAN ACADEMY. Lamarckism in its modern form, and a critical and doubting attitude towards natural selection as an efficient cause, will keep alive discus- sion and investigation, and especially, if resort be had to experimenta- tion, will carry up to a higher plane the status of philosophical biology. TABER. — LINEAR TRANSFORMATIONS. 371 XVI. ON THE GROUP OF AUTOMORPHIC LINEAR TRANS- FORMATIONS OF A BILINEAR FORM. By Henry Taber. Presented February 14, 1894. § 1. Alternate Bilinear Form. 1. In the Philosophical Transactions for 1858, Cayley gave the following identity between two matrices, viz. : — (fi - Y) (fi + Y)"1 n (fi - Y)-1 (Q + Y) = n. From this identity Cayley derives the general solution of the matrical equation Q = fi, in which f2 is a skew symmetric matrix and denotes the transverse of . Thus for <£ Cayley gives the expres- sion (O-Y)"1 (12 + Y), in which Y is an arbitrary skew symmetric matrix, but such that | O - Y 1 + 0. From this expression may be derived J= (ft-Y) (fi + Y)-1; therefore, substituting, the above equation is satisfied identically. As shown by Cayley, the solution of this equation is equivalent to the determination of the automorphic linear transformation of the alternate bilinear form A A (Q) (arlt ar2, . . . ) (tfu y2, . . . ), the z's and y's being cogrediant. Cayley's expression gives every solution with non-vanishing deter- minant of the equation <£ O <£ = fi, except those of which —1 is a latent root (root of the characteristic equation)- If | & | ^ 0' and 872 PROCEEDINGS OF THE AMERICAN ACADEMY. if satisfies the above equation, | <£ J tj: 0. In what follows it will be assumed that the determinant of the skew symmetric matrix fi is not zero. Since the product of any two matrices satisfying the equation Q = U is again a solution of this equation, the matrices satisfying this equation form a group. This group can be generated by the solutions of the equation given by Cayley's expression.* For every solution of the equation 12 <£ = O is given by the product of four of Cayley's expressions. See (2) and (10). The group of solutions of the equation <£ fi = Q can be generated by the matrices of the group which differ infinitesimally from the matrix unity ; for, see (2) and (5), every solution given by Cayley's expression can be formed by the product into itself an infinite num- ber of times of a solution (given also by Cayley's expression) which differs infinitesimally from the matrix unity. The group of matrices satisfying the equation <£ CI = fi is identi- cal with the group of linear substitutions which automorphically trans- form the alternate bilinear form A A (Q) (xu x2, , , . ) (y1} y2, . . . ), in which the a?'s and y's are cogrediant. Therefore, this group can be generated by the infinitesimal transformations of the group, namely, those differing infinitesimally from the identical substitution. 2. The proof of these theorems relating to the group of solutions of the equation O = Q may be made to depend upon the proof in the special case in which the skew symmetric matrix Q is also orthogonal (i. e. if CI2 = — 1). Thus, let 12

cf> uT , O' = — g , this becomes CO ^nfi}/ = n', * This has been proved otherwise by Frobenius Crelle, 1878, who shows that every solution of the equation

= co- 1 \p co is the pro- duct of four of Cayley's expressions. Further, if ^ is a solution of the equation <££&<£ = Q, given by Cayley's expression, then i/^ = co c/> co- ! is a solution of the equation ip 12' {{/ = f2' given by Cayley's expression. Therefore, if if/ is equal to the product into itself an infinite number of times of a solution of the equation ^12'^ = 17', differing iufinitesimally from the matrix unity, cf> is equal to the product into itself an infinite number of times of a solution of the equation ^fi^fi, differing infinitesimally from the matrix unity.* 3. Let now c/> be any solution of the equation 12 4> = Q, in which il = —17, and 172 = —1 (and therefore 12 O = 1). The complete de- termination of may be obtained by the consideration of the identity e-aene9(i _ q^ oo (0 nv in which een denotes the exponential series 2r- — p- , convergent for any matrix. Thus, let — 17 ^ be a polynomial in c/>, satisfying the equation = e-ne. "We then have c/> = een; but the identity gives = een. Since — 170 is a polynomial in c/>, #£2 is a polynomial in c/>; and since c/>-1 = Or1 Q, $0, = Q- '(17 0)12 is also a polynomial in <£. Consequently, 017 and 0f2 are commutative. Therefore, e(B-9)a = eene-en _ e0fi.(een)-i _ ^-i = i. * Moreover, if (p is a solution of the equation

). 374 PROCEEDINGS OF THE AMERICAN ACADEMY. 4. Conversely, if 0 is any matrix for which 6 to 0 = 0to6 and e(e-0)n __ ^ tben <£ _ . eeo js a solution of the equation £to = to. In particular, if 6 is symmetric, e9n is a solution of this equation, and so also is e» , in which n is any positive integer. But (e«en)" = een. Therefore, every expression e0a in which 6 is symmetric is a solution of the equation <£Q = to, and can be generated by the product into itself an infinite number of times of a matrix which is a solution of this equation, and differs only infinitesimally from the matrix unity. Let $ = eeil in which 0 is symmetric. If the positive integer n is sufficiently great, no odd multiple of it \/—l is a latent root of - $ to ; n and therefore —1 is not a latent root of e*Bil. But <$> = (e»en)n. Therefore, <£ is the nth power of a matrix given by Cayley's ex- pression. 5. Every solution of the equation <£fi<£ = to, given by Cayley's expression, can be put in the form e9n in which 6 is symmetric. Thus, let = (to — Y)-1 (O + Y) = to-1 . (i - yd-1)-1 (i + Yto-1) . to, in which Y is symmetric. If now &' =f(—Yto~1) is a polynomial in — YI2-1, satisfying the equation e»' = 1 — YO-1, then &" =f(Yto~1) will satisfy the equation ed" = 1 + YO"1.* * Since |

' + d")n But —d/ + &" is a polynomial in odd powers of Yfi-1; and since Q-1 (Y£2-1)2r + 1 for any positive integer r is symmetric, Cl~x{— &' + &") CI is the product of a symmetric matrix into Q. Whence it follows that every solution of the equation <£ CI = CI given by Cayley's expression is the product into itself an infinite number of times of a solution which differs only infinitesimally from the matrix unity.* 6. Assuming that 0 Cl is a polynomial in <£ satisfying the equation $n = Q, let 0 = ei(« + *)n and 0 = gi(»-«)n. Then, by (3), = <£<£<> = <£o0; and, by (3) and (4), — — "-" (f)Ci(f> = ci, 4>0ci4>0 = ci, <£02 = l. By (4), $ is the product of matrices given by Cayley's expression. But the matrix 0 may be so chosen that no integer multiple of 7r V— 1 other than zero is a latent root of £ (6 + 0) 12, in which case <^> is given by Cayley's expression.! * In the three preceding sections (3), (4), (5), the assumption ft2 = —1 has not been employed in the demonstration. The theorems therein given, there- fore, hold for any skew symmetric matrix whose determinant does not vanish. t It is readily shown that

0 = 0. Let © = O 0, then © =: — ®, ®fi® = fi; moreover, ©2 = —1. Consequently, © is given by Cayley's expression ; and therefore is the product of three of Cayley's expressions. 8. If cf> is orthogonal, it is commutative with ft ; therefore 6 and 0 are commutative with ft. Consequently, <£ = e-hn(9 + d) — e- .'(9 + 0)0 _ <£-l From the last equation follows ft (f)Q = cfiQ ft. Let now ® = ft<£0; then © = —0, ©ft© == ft, and ®2 = — 1. Consequently, © is given by Cayley's expression. We have = ft_1®<£; therefore any orthogonal matrix satisfying the equation <£ft is orthogonal, and is also a solution of this equation, and consequently can be thus expressed. Whence it follows that <£ is given by the product of two of Cayley's expressions.* 9. This theorem holds if ft2 dp — 1. In this case, if <£ ft cf> = ft and is orthogonal, let w be any fourth root of — ft2 expressible in powers of ft2. Then v, ft to «i to — to <£ to =ft. to * The two matrices given by Cayley's expression whose product is equal to (p caji both be taken orthogonal. For in the equation — il_1 &

n4>=n in which fl- = —1 is generated by the totality of such expressions for all symmetric matrices T commutative with fl and such that | fl — Y | dp 0. TABER. — LINEAR TRANSFORMATIONS. 377 But (j> is commutative with 12, and consequently with w. Therefore, « Q 12 £~a ^ = ,75 to (U and by the last theorem we may put ♦-e-*r©+*.)G-*rs+*) = co(12 — a)Y'aW)-I(fi + a.Y'a(o) (12 — uY'fiOiy^n + wY'fiw)^-1 = W(0-Ytt)-1((2+Ya)(0-Yp)-1(0+Yp)a.-1, in which Y'a, Y'p, and therefore Ya = uY'.w, Yp = wY'^w, are sym- metric. Whence it follows that «£ = a)"1 4>w = (12 - Y,,)-1 (12 + Ya) (12 - Yp)-» (12 + Yp) is the product of two of Cayley's expressions.* 10. If is neither symmetric nor orthogonal, let © = 12 c£0 ; whence it follows that 0 = — ©, and 0 12 ® = 12. Let 6' 12 be a polynomial in © satisfying the equation 0 = e0'a ; and let * It is not proved that the two factors of are orthogonal. But it can be readily proved that is equal to the product of three matrices given by Cay- ley's expression, each of which is orthogonal; and therefore the sub-group of orthogonal solutions of the equation is orthogonal, we may put (p = , where

being polynomials in ft = ft, and 0 ft = 12 is the product of three solutions of the equation 0ft0 = ft given by Cayley's ex- pression, each of which is orthogonal. 378 PROCEEDINGS OF THE AMERICAN ACADEMY. Then ^ _ u © ft © = ft, ®0 O 0O = ft, ©o2 = 1 ; and, if 6' is properly chosen, —1 is not a latent root of ©, which is then given by Cayley's expression. Since © is skew symmetric, 6' ft, B' ft, ft 0', ft 6', are commutative. Therefore, o © ©a = © © Whence follows (®0©0)2 = ©02©02=1. Let now ft®0 = r; then f = — 1\ and r ft T = ft. We also have and therefore r4 = l. A symmetric matrix rn of non-zero determinant can be found to satisfy the equation r w T = m . And if &? denotes any symmetric square root of nr, the last equation may be written (^)_1r^. bt* T (st*)"1 = 1; which, if ij/ = iu*T (nr*)-1, becomes $ijs = 1. Therefore, T= (tjl)-1 ifr tat, in which if/ is orthogonal. But, since T ft V = ft, if/ is commutative with (n7=)-Ift (zu?)~l. Therefore, by (9) ip is equal to the product of two expressions of the form [(^)-!ft (nr*)"1 - Y']-1 [O^)-1 ft (cr*)-1 + Y'] = 57* (ft - bt*Y' ari)-1 w* • (bt*)-1 (" + srfY'sri) (bj*)-1 = st* (ft - Y)-1 (ft + Y) (w*)-1, in which Y' and therefore Y==E7£Ysr*, is symmetric. Whence it follows that r = (zzi)~l \p vfi is the product of two of Cayley's expressions. Therefore, <£ = 0O <£ = n-1 © 0 = ft"1 ®0 0 <£ = — r®<£ is equal to the negative of the product of four of Cayley's expressions. Since — is also a solution of the equation ft <£ = ft, it is also ex- pressible thus. Whence it follows that $ is given by the product of four of Cayley's expressions. TABER. — LINEAR TRANSFORMATIONS. 379 11. If <^> is real, 6 may be so chosen that and 0 are real ; and then, if 12 is real, © = 12~~ 1 <£0 is real. But then, since 0 is real and skew symmetric, — 1 is not a latent root of ©, which is therefore given by Cayley's expression. Whence it follows that any real matrix satisfying the equation <£ 12<£ = 12, in which 12 is real, is the product of 12" * into two of Cayley's expressions. But 12-1 cf> is then also a real solution of this equation, and can therefore be thus expressed. Consequently, 0 is equal to the product of two of Cayley's expressions.* In each of the factors of 0 given by Cayley's expression, the sym- metric matrix may be taken real. 12. These theorems may be extended to the real solutions of the equation 0 12 0 = 12, in which 12 is any real skew symmetric matrix whose determinant does not vanish. For if 12 is real, the latent roots of — 122 are all positive. Therefore, if w denotes a fourth root of — 122 expressible in powers of 122, w may be taken real. If now 0 is a real solution of the above equation, we have 12 , 12 w (t>(a—io}(f>ur1 = 2 T' 2 to to in which w 0 w x and -^ are real. Therefore, by the preceding theorem, = u> (12 - Y,)"1 (12 + Yft) (12 - Y^)-1 (12 + Y^) to~\ in which Y'a, Y'p, and therefore Ya = to Y'a to, Y^ = to Y'p w, are real symmetric matrices. Thus it appears that the group of real solutions of the equation 012 0 = 12 can be generated by the Cayleyan solutions of this group, that is, by the totality of expressions of the form (12 — Y)-1 (12 + Y), in which Y is a real symmetric matrix such that | 12 — Y | £ 0. § 2. Bilinear Form, neither Symmetric nor Alternate. 1. The problem to determine the most general solution of the equa- tion 0 12 0 = 12, in which 12 is neither symmetric nor skew symmetric, may, provided neither ± 1 is a latent root of 1212-1, be reduced to the solution of two algebraic equations of the nth degree. * See ante, p. 178. 380 PROCEEDINGS OF THE AMERICAN ACADEMY. Thus, let 2fi0 = n + fi, 2 Oj = O — n. Then, if (f> satisfies the above equation, we have, as shown by Cayley, (j>n0 = n0, n1 = n1; and conversely. If neither ± 1 is a latent root of Cl Cl~\ then | fi0 | + 0, | fij | £ 0. Since I Cl0 | + 0, ty, has a square root expressible by Sylvester's for- mula in powers of Cl0. For the determination of this square root the solution of the algebraic equation | Cl0 — g | = 0 is requisite. If Q01 denotes any symmetric square root of fi0, the equation $£)0(£ = fi0 may be written which, if i^ = fi0* <£ (fi02)_1, becomes ($= 1. Therefore, the most general expression for the matrix $ satisfying the above equation is Cl0~ 2 i/r fi0l, in which if; is an arbitrary orthogonal matrix. If $ Cl (f> = Cl, we must also have <£ fix = Clv Therefore, ci0i ^ n0_ ^ Qi n0~ 2 $ ci0* = fii , But (9) § 1 i/r is given by the product of the three matrices. (n-i o, n0- 1 - y'0)-' (q,-* Qi «<>-* + Y'a) = iy (fix - cy Y'afyr1 (fii + n0* Y'* Clx = Clx , generated by the totality of expressions TABER. — LINEAR TRANSFORMATIONS. 381 (fi1-Y)-1(fi1 + Y), in which the symmetric matrix Y is subject to the condition that f}0"~ 2 Y ft0 ' * is commutative with fy,-*!^ i)0~*, and is such that the determinant of | Q — Y | ^ 0. To find the most general skew symmetric matrix O0-iYfi0~i commutative with Q~*QiQ0~s requires in general the solution of an algebraic equation of the nth degree.* Again, since i{/ is the product of three orthogonal matrices of which — 1 is not a latent root and which are commutative with Cl1} the group of solutions of the equation $ Q (/> = fi is a sub-group of the group of solutions of the equation $ O0 <£ = fi generated by the totality of expressions n-i (i - y)-1 (i + y) n0i = (Oo - iy Y of Class III. Crawford H. Toy, ) Rumford Committee. Wolcott Gibbs, Benjamin O. Petrce, John Trowbridge, Edward C. Pickering, Erasmus D. Leavitt, Charles R. Cross, Amos E. Dolbear. OF ARTS AND SCIENCES. 399 C. M. Warren Committee. Francis H. Storer, Samuel Cabot, Thomas M. Drown, Henry B. Hill, Charles L. Jackson, Leonard P. Kinnicutt, Arthur M. Comey. Member of the Committee of Finance. Augustus Lowell. The President appointed the following Standing Com- mittees : — Committee of Publication. Charles L. Jackson, William G. Farlow, Horace E. Scudder. Committee on the Library. Barrett Wendell, Samuel H. Scudder. William E. Story, Auditing Committee. • Henry G. Denny, Augustus Lowell. The following papers were presented by title : — On the Occlusion of Gases by the Oxides of Metals. By Theodore William Richards and Elliot Folger Rogers. On the Compound Acetates and Formiates of Cupri- ammonium. By Theodore W. Richards and Hubert Grover Shaw. A Revision of the Atomic Weight of Barium. Second Paper : The Analysis of Baric Chloride. By Theodore William Richards. The President appointed a Committee on the Custody of the Hall, consisting of the Recording Secretary, the Librarian, and Edward J. Young. 400 PROCEEDINGS OF THE AMERICAN ACADEMY Eight hundred and sixtieth Meeting. June 14, 1893. — Monthly Meeting. The Academy met at the Physiological Lecture Room of the Harvard Medical School, Boston. The President in the chair. Harold C. Ernst explained the methods of investigation pursued in the study of bacteriology, illustrating his remarks with diagrams and lantern slides. The attention of the Academy was then invited to an exhibit consisting of two main sections : — 1. Cultures of bacteria, including pathogenic and non-pathogenic organisms, in various nutrient media, each one, so far as possible, grown upon the standard media, in tube and plate culture, and the various cultures of the same organism grouped together and labelled with the common name ; so that each one could be seen growing upon one or more of the following materials : nutrient gelatine (plain or with glycerine), nutrient agar-agar (plain or with glycerine), potato, bouillon, blood serum, milk, peptone, bread paste. Thus an oppor- tunity was afforded for a comparative study of a number of varieties of bacteria under similar conditions. 2. Preparations placed under the microscope, showing the various appearances of the bacteria in tissues and in pure culture. At the close of the lecture the President, after paying a just tribute to Dr. Ernst's interesting communication, pro- posed a vote of thanks, which was unanimously adopted. Eight hundred and sixty-first Meeting. October 11, 1893. — Stated Meeting. The President in the chair. The Corresponding Secretary read letters from J. Bryce, L. Hermann, and F. A. Kekule, acknowledging their elec- tion as Foreign Honorary Members ; from G. K. Gilbert and F. R. Hutton, acknowledging their election as Associate OF ARTS AND SCIENCES. 401 Fellows ; and from Sylvester R. Koehler, accepting Fellow- ship. The following gentlemen were elected members of the Academy : — Edward Salisbury Dana, of New Haven, to be an Asso- ciate Fellow in Class II., Section 1. Samuel Louis Penfield, of New Haven, to be an Associate Fellow in Class II., Section 1, in place of the late Frederick A. Genth. Jean Baptiste Edouard Bornet, of Paris, to be a Foreign Honorary Member in Class II., Section 2, in place of the late Alphonse L. P. P. de Candolle. In accordance with the recommendation of the Committee on the Custody of the Hall, it was Voted, To grant the use of the Hall of the Academy to the Colonial Society of Massachusetts on the third Wednesday of December, 1893, January, February, March, and April, 1894, after 3 o'clock P. M. O. W. Huntington presented a paper entitled, " Further Observations upon the Occurrence of Diamonds in Meteor- ites." On the motion of F. H. Storer, it was Voted, That a Committee be appointed to consider changes in the statutes relating to the methods of electing officers. On the motion of the Corresponding Secretary, it was Voted, To meet on adjournment on the second Wednesday in November. Eight hundred and sixty-second Meeting. November 8, 1893. — Adjourned Stated Meeting. The Academy met at the Engineering Building of the Massachusetts Institute of Technology, Boston. The President in the chair. The President announced the death of Moses Gerrv it Farmer, Resident Fellow, and of Benjamin Jowett, Foreign Honorary Member. VOL. XXIX. (n. S. XXI.) 26 402 PROCEEDINGS OF THE AMERICAN ACADEMY The Corresponding Secretary read letters from the Rum- ford Historical Association of Woburn, acknowledging the receipt of the bronze copy of the Rumford Medal presented by the Academy ; from E. S. Dana and S. L. Penfield, ac- knowledging election as Associate Fellows ; from the Anthro- pological Society of Washington, announcing the conditions of competition for its citizenship prizes ; from the Natural History and Medical Society of Bonn, inviting the Academy to participate in its seventy-fifth anniversary festival ; and from the Belgian Chemical Association, announcing an In- ternational Congress of Applied Chemistry, and inviting the Academ}7 to send delegates. The chairman of the C. M. Warren Committee made a brief report, recommending the appropriation of three hundred dollars to Charles F. Mabery for investigations on the Amer- ican sulphur petroleums. This appropriation was voted by the Academy. The President appointed a Committee on Methods of Elect- ing Officers, consisting of the following gentlemen : Augustus Lowell, Chairman, F. H. Storer, Barrett Wendell, W. G. Farlow, and W. R. Livermore. The Recording Secretary gave a brief account of the circumstances attending the award of the Massachusetts Charitable Mechanic Association's Grand Gold Medal at the exhibition in 1881 to Albert H. Emery. This medal, "for the exhibit most conducive to human welfare," was, at the request of the Association, bestowed by a committee of the Academy acting as judges. Gaetano Lanza then explained and exhibited the Emery testing machine of 300,000 pounds' capacity, recently acquired by the Institute. This machine contains all the essential features of the 800,000 pounds testing machine built by Albert H. Emery, and now at the Watertown Arsenal. The operation of this machine consisted in testing the compressive strength of a Virginia white-oak beam 13| inches X 8£ inch X 15 feet, which failed when a load of 176,000 pounds was applied. Cross Breaking. — The next experiment consisted in breaking a yellow-pine beam 17 feet 6 inches between the supports, weighing 617 OF ARTS AND SCIENCES. 403 pounds, and having a cross-section of 16^ inches X 6 inches. The load was in the centre, suspended from a yoke weighing 102 pounds. A micrometer was placed on tha beam, and the deflections were measured as follows : — Loads. Micrometer Readings. Difference. 2,000 .05595 .24160 7,000 .29755 The modulus of elasticity corresponding to these deflections was 1,778,200 pounds per square inch. The beam broke by tension and longitudinal shear, at a load of 38,000 pounds. The corresponding calculated modulus of rupture, including the weight of the beam, was 7,409 pounds per square inch. The maximum intensity of longitudinal shear was 287.9 pounds per square inch. Torsion Testing Machine. — A two-inch Norway iron shaft, 6 feet long between the holders, was twisted 16 turns without breaking. This specimen, left over night with a twisting moment on it of 76,680 inch pounds, broke the next morning ; the maximum twisting moment being 82,080 inch pounds. The shaft increased ^ of an inch in diameter. Rope Testing Machine. — A Manila rope, 4^ inches in circum- ference, 100 inches long, and held by two eye splices, broke at the centre under a load of 15,000 pounds. The rope stretched 24 inches before breaking. The other machines and appliances on exhibition were as fol- lows : — The 150 horse-power triple expansion engine. The 16 horse-power Harris-Corliss engine. A rotary pump, a plunger pump, and a pulsometer. In the department of hydraulics was seen the discharge through a circular orifice 1 inch in diameter in a thin plate. The Blake duplex pump, having a steam cylinder 16 inches in diameter, a water cylinder, 10| inches in diameter and 12-inch stroke, discharged water through three nozzles over a weir four feet wide. The following paper was read by title : — New Genera and Species of Laboulbeniaceae, with a Sj'iiop- sis of the Known Species. By Roland Thaxter. 404 PROCEEDINGS OP THE AMERICAN ACADEMY Eight hundred and sixty-third Meeting. January 10, 1894. — Stated Meeting The Vice-President in the chair. The chair announced the death of Francis Parkman and Henry Warren Paine, Resident Fellows, and of Charles Merivale, Foreign Honorary Member. The Corresponding Secretary read letters from W. K. Brooks, acknowledging his election as Associate Fellow ; from E. Bornet, acknowledging his election as Foreign Hon- orary Member ; and from the Corresponding Secretary of the Colonial Society of Massachusetts, thanking the Academy for affording the use of its Hall for the Society's meetings during the present season. Barrett Wendell, on behalf of the Committee on Methods of Electing Officers, submitted a report. On the motion of W. R. Livermore, it was Voted, That this report be laid on the table. The chairman of the Rumford Committee presented a report, and, in accordance with a recommendation contained therein, it was Voted, To appropriate five hundred dollars from the income of the Rumford Fund to William H. Pickering for investiga- tions on optical constants at different altitudes. Granville Stanley Hall, of Worcester, was elected a Resi- dent Fellow in Class III., Section 1. C. L. Jackson read abstracts of the following papers : — On the Formation of Volatile Compounds of Arsenic from Arsenical Wall Papers. By Charles R. Sanger. Chronic Arsenical Poisoning from Wall Papers and Fabrics. By Charles R. Sanger. Remarks on these papers were made by F. H. Storer, E. S. Wood, and W. G. Fallow. The following papers were presented by title : — On the Automorphic Linear Transformation of Bilinear Forms. By Henry Taber. Contributions from the Gray Herbarium of Harvard Uni- versity. By B. L. Robinson. I. Monograph of the North OF ARTS AND SCIENCES. 405 American Alsinece. II. New and hitherto imperfectly known Plants collected in Mexico by C. G. Pringle in 1892 and 1893. III. Notes upon the Genus Galinsoga. On an Apparatus for the Measurement of Coefficients of Self-induction and the Investigation of the Phenomena of Alternating Currents. By Frank A. Laws. Presented by Charles R. Cross. Eight hundred aud sixty-fourth Meeting. February 14, 1894. — Monthly Meeting. The Academy met at the Walker Building of the Massa- chusetts Institute of Technology, Boston. The President in the chair. Elihu Thomson showed and described a number of phe- nomena produced by electrical discharges of high potential and of high frequency. Work in this field began with Henry, and a considerable amount of experimentation in the same field has recently been carried on by Tesla, and others. To obtain high frequency discharges, the current of an alternating current dynamo of 1,000 volts was reduced to 100 volts by suitable lighting transformers. This low pressure current was passed through the primary of a coil resembling a Ruhmkorff" without break-piece, and raised to an alternating current of from 20,000 to 30,000 volts in the secondary, such current having still the original frequency of 120, that is 120 complete waves per second. This high pressure current was led directly into a large condenser', the metallic foils of which were insulated by a heavy mica sheet and contained in boxes filled with oil. The condenser, charged to 20,000 or 30,000 volts, in discharging jumped spaces or gaps between balls of brass through which spaces a strong air jet was kept blowing (Figure 1). The discharge circuit of the condenser includes these gaps, and may also include various coils, or plain wires, or paths designed to be trav- ersed by the condenser discharge. If the condenser discharge takes place over a short path, of say otdy a few feet total length, the fre- quency or rate of alternation of the current composing the discharge may be several hundred thousand per second. This rate is so high 406 PROCEEDINGS OP THE AMERICAN ACADEMY that even a short length of conductor opposes the flow of current by its self-induction so greatly that lamps are lighted in a shunt to it, cards punctured, etc. (Figure 2). In fact, in one experiment an incan- descent lamp was lighted in shunt to about 18 inches of \ inch copper rod, and the lamp was brought to full brilliancy (Figure 3). More- over, it was distinctly shown that these currents are comparatively harmless, and may be passed through the body in an amount sufficient to light fully ordinary incandescent lamps (Figures 4 and 5) ; and the speaker stated that from the extremely small sensation experienced he thought it would be easily possible greatly to increase the current, and probably to an extent sufficient to light a small arc lamp. The discharge circuit in this case had the form of a U of heavy copper rod, and it could be grasped so as to short-circuit it between points of such potential that insulation of considerable thickness would be readily pierced ; but there was no disagreeable sensation, and very little sensation at all experienced. The extraordinary energy of the induction of these dischai'ges in the U-shaped conductor was shown by bringing near to it a rectangle of heavy copper wire the circuit of which was completed through a large incandescent lamp requiring more than one third of a horse- power of electrical energy to light it. It was found that the incandes- cent lamp could be brought far beyond its normal brilliancy on placing the rectangle a few inches away from the U (Figures 6 and 7). By causing the condenser discharges to pass the air gaps and also a large open coil of about ten turns immersed in oil, the inductive effects of such coil on a parallel secondary coil having about 460 turns were exhibited. This secondary coil of 4 GO turns was wound on a hard rubber frame as a single layer, and the turns kept apart by a small space, and together with the primary immersed in a box or bath of oil of high insulating power (Figure 8). The ends of the finer and longer secondary were carried to the terminals rising out of the oil at the ends of the box or bath, and between these terminals torrents of sparks of about 30 inches in length were produced, accompanied by a loud roar- ing or crashing sound. Lateral discharges from the terminals were shown to be capable of being received through the body without injury, although of sufficient energy of current to light fully an incandescent lamp (Figure 9), and in fact to melt off the ends of fine wires, set fire to sticks of wood, etc. In this connection it was shown also that water was for these discharges a comparatively fair insulator, inasmuch as the sparks could be made to pass under water with substantially the same appearance as when they passed through air. Disruptive sparks as well as brush discharges were produced under water*. Fig. 1. High frequency circuit. * f> Condenser. 1=^-- Spark gaps with air *J) jets. 30,000 volt secondary. 75 volt primary. Card. .J) 0 Fig. 2. High frequency loop, with shunt puncturing card. Fig. 3. High frequency loop, shunted by lamp. Fig. 4. Fig. 5. High frequency loop, shunted by person. High frequency loop, shunted by lamp and person. Fig 6. High frequency round loop, with circular turn and lamp lighted inductively. Fig. 7. r r ^\ -*\ 1 c J High frequency loop, induction into adjacent circuit or loop, with large lamp between terminals thereof. Fig. 8. Long sparks B §K>KJN™=s£§@> Diagram of vat with oil, immersed in which is the high frequency primary of few turns with secondary of a layer of finer wire of relatively 30 to 40 times the primary turns. Sparks obtained from prolongations of secondary outward or discharge terminals, A, B. Fig. 9. Lamp lighted by sparks from one terminal. Discharge traversing person's body. Fig. 10. Geissler tube, a, lighted by electrostatic field except at neutral position, OF ARTS AND SCIENCES. 407 As exhibiting the electrostatic activity of the rapidly alternating electrostatic field around the apparatus, a long Geissler tube without metallic terminals was used, and brilliantly illuminated even at dis- tances several feet away from the outside of the vat containing the induction apparatus. The tube, however, failed to light in one posi- tion ; that is, in the centre plane dividing the apparatus into halves, which during action are electrified respectively positively and nega- tively at each alternation (Figure 10). At the close of the experiments a vote of thanks was extended to the speaker. S. H. Sc udder exhibited a medal struck to celebrate the one hundredth anniversary of the birthday of Chevreul. C. L. Jackson presented the following papers by title : — On Mucophenoxychloric Acid. By H. E. Sawyer. On Certain Bromine Derivatives of Resorcine. By C. L. Jackson and F. L. Dunlap. The Smithville Meteoric Iron. By Oliver W. Hun- tington. Eight hundred and sixty-fifth Meeting. March 14, 1894. — Stated Meeting. The President in the chair. The chair announced the death of Pierre J. Van Beneden and of Sir James Fitzjames Stephen, Foreign Honorary Members. The Corresponding Secretary read the following letters : from Edouard Van Beneden, announcing the death of his father, P. J. Van Beneden ; from the Council of the Scientific Alliance of New York, calling attention to the proposed establishment of a fund for the endowment of research in honor of the late John Strong Newberry, and soliciting sub- scriptions thereto ; from the Royal Academ}' of Physical and Mathematical Science of Naples, announcing a prize of 1,000 lire for the best memoir on prime numbers ; also from the Executive Committee of the Eighth International Congress 408 PROCEEDINGS OP THE AMERICAN ACADEMY of Hygiene and Demography, and from the Tenth Inter- national Congress of Orientalists, inviting the Academy to take part in the work of those Congresses. On the motion of C. S. Minot, it was Voted, That the President be authorized to appoint dele- gates to either or both of these Congresses. The following gentlemen were elected members of the Academy : — Francis Mathews Green, of Boston, to be a Resident Fellow in Class I., Section 2. Wallace Clement Sabine, of Cambridge, to be a Resident Fellow in Class I., Section 3. John Donnell Smith, of Baltimore, to be an Associate Fellow in Class II., Section 2, in place of the late George Vasey. The chairman of the C. M. Warren Committee presented a report, and, in accordance with a recommendation contained therein, it was Voted, To appropriate from the income of the C. M. War- ren Fund the sum of two hundred dollars ($200) to aid Charles F. Mabery, of Cleveland, Ohio, in making a sys- tematic study of the petroleums of Ohio and Canada, with especial reference to the separation of the individual hydro- carbons therein contained. On the motion of S. H. Scudder, it was Voted, To take from the table the report of the Committee on Methods of Electing Officers. A protracted discussion ensued, in which W. R. Livermore, A. Lowell, F. H. Storer, C. R. Cross, S. C. Chandler, C. S. Minot, H. W. Williams, S. H. Scudder, A. Hyatt, and G. Hay took part. On the motion of A. Lowell, it was Voted, That it be the duty of the Council annually to prepare a list of candidates for the offices of President, Vice- President, Corresponding Secretary, Recording Secretary, Treasurer, Librarian, Councillors, and members of the Stand- ing Committees, and to cause this list to be sent by mail to all the Resident Fellows of the Academy three weeks before OF ARTS AND SCIENCES. 409 the Annual Meeting. Whenever the Council shall receive a nomination for any office, signed by not less than ten of the Resident Fellows of the Academy, it shall be the duty of the Council to publish such nomination with its own. On the motion of the Corresponding Secretary, it was Voted, To meet on adjournment on the second Wednesday in April. The following paper was presented by title : — On the Inheritance of Acquired Characters in Animals with a Complete Metamorphosis. By A. S. Packard. Eight hundred and sixty-sixth Meeting. April 11, 1894. — Adjourned Stated Meeting. The Academy met at the Walker Building of the Massa- chusetts Institute of Technology, Boston. The President in the chair. The chair announced the death of Charles E. Brown- Sequard, of Paris, a Foreign Honorary Member. On the motion of John Trowbridge, on behalf of the Rumford Committee, it was Voted, To appropriate from the income of the Rumford Fund the sum of three hundred dollars ($300) to F. A. Laws in aid of his investigations on the thermal conductivity of metals. Voted, To appropriate from the income of the Rumford Fund the sum of one thousand dollars ($1,000) to be ex- pended in aid of investigations in light and heat. Payments from this sum to be made on the order of the Chairman of the Committee. On motion of the Corresponding Secretary, it was Voted, That the Committee of Publication be authorized to use for illustrations a portion of the regular annual appro- priation for publications not exceeding fifty dollars (850) for each paper. Larger appropriations of money from the general funds for illustrations to be made only by vote of 410 PROCEEDINGS OF THE AMERICAN ACADEMY. the Academy at a stated meeting, on recommendation of the Committee of Publication, approved by the Treasurer. The Corresponding Secretary gave notice of a proposed change in the statutes : to amend Chapter IV. of the Stat- utes by changing the numbers of Sections 4, 5, and 6, to 5, 6, and 7, respectively, and to insert a new section as follows : — .. 4. The C. M. Warren Committee, of seven Fellows, to be chosen by ballot, who shall consider and report on all applications for appro- priations from the income of the C. M. "Warren Fund, and generally see to the due and proper execution of this trust." On the motion of the Corresponding Secretary, it was Voted, To refer this proposed change in the Statutes to a committee. The President appointed the Recording Secretary as this committee. On motion of the Librarian, it was Voted, To present a copy of the Life and Works of Count Rum ford, published by the American Academy, to the Rum- ford Kitchen. The following papers were read by title : — Further New and imperfectly known Plants collected in Mexico by C. G. Pringle in the Summer of 1893. By B. L. Robinson and J. M. Greenman. On the Determination of Sulphur in Volatile Organic Compounds. By Charles F. Mabery. Edward C. Pickering described the work in astronomical photography of the Harvard College Observatory. Photographs were thrown upon the screen, some of which illus- trated the following subjects : — The two principal stations of the Observatory at Cambridge and at Arequipa. Photographs of the double star £ Ursoe Majoris, the planet Saturn showing its ring, and the occultation of a star by the Moon, taken by George P. Bond in 1857. Various members of the solar system, including Venus as a crescent, Vesta showing the method of discovering asteroids by pho- tography, Jupiter, Saturn, Neptune and its satellite. Swift's comet of 1892. An eclipse of Jupiter's satellite. Eclipsed Moon, showing the OF ARTS AND SCIENCES. 411 method of searching for a luuar satellite. Photographs of the Moon, taken at the Harvard station in Southern California, enlarging the image in the telescope ; the results were compared with those obtained elsewhere. Great spiral Nebula in Orion, invisible to the eye, but covering a large part of that constellation. Method employed at Are- quipa for making the 13-inch Boyden telescope follow the apparent motion of a star. Instead of a finder, images in the photographic telescope are used. The telescope is made to follow a star by means of an eyepiece outside of the photographic field. The entire plate is free to rotate around this eyepiece as an axis. The plate is then turned at intervals so that another star is kept in the axis of a second eyepiece. Sources of error, such as variations in flexure, in refrac- tion, and in adjustment of the polar axis, which cannot be eliminated by the ordinary method with a finder, are thus corrected. Photographs of the clusters k Crucis, w Centauri, and 47 Tucana?. Chart and spectrum plates of the vicinity of the variable star rj Carina?, showing stars of the fifth type and other objects having peculiar spectra. Dis- tribution of stars of the fifth type. Identity of the spectra of a Aurigae and the Sun. Spectrum of (3 Auriga?, showing the lines single and double. Spectrum of a new variable star in Scorpius, show- ing the method of discovering variable stars of long period by means of the bright hydrogen lines in their spectra. Charts showing this star bright and faint. Light curve of this and of other variables. Spectrum plates showing the presence and absence of the new star in Norma. Charts showing the absence of this star before July, 1893, and its presence since then. Identity of the spectra of Nova Norma? and Nova Auriga?. Spectra of Nova Auriga? and o Ceti, showing the difference between the spectra of new stars and variable stars of long period. Method of detecting by inspection stars having a large parallax, proper motion, or variation in light ; pairs of photographs are taken at intervals of six months, with the film side of the plate out aud in ; superposing plates taken six months apart in opposite posi- tions so that the gelatine films are in contact, small deviations are perceptible ; a motion corresponding to a parallax of 0".6 was thus readily visible across the room. AMERICAN ACADEMY OF ARTS AND SCIENCES. Report op the Council. — Presented May 9, 1894. BIOGRAPHICAL NOTICES. Moses Gerry Farmer By Amos E. Dolbear. Hermann August Hagen Samuel Henshaw. James Russell Lowell Horace E. Scudder. Henry Warren Paine Nathaniel Holmes. Francis Parkman Barrett Wendell. Henry Warren Torrey William Everett. Marquis de Caligny William Watson. Benjamin Jowett William W. Goodwin. Charles Merivale Silas M. Macvane. Notices of Wheatland, Van Beneden, Brown-Sequard, Marignac, and Stephen are necessarily deferred to the next volume ; while those of Lowell and Caligny, deferred from last year, are given below. REPORT OF THE COUNCIL. Since the Annual Meeting of the 10th of May, 1893, the Academy has lost by death eleven members ; — five Fellows, Moses Gerry Farmer, Hermann August Hagen, Henry War- ren Paine, Francis Parkman, and Henry Warren Torrey ; and six Foreign Honorary Members, Pierre Joseph Van Beneden, Charles Edward Brown-Sequard, Benjamin Jowett, Jean Charles Galinard de Marign&c, Charles Merivale, and Sir James Fitz James Stephen. RESIDENT FELLOWS. MOSES GERRISH FARMER. Moses Gerrish Farmer was born in Boscawen,New Hampshire, in 1820. In his youth he had strong predilections towards scientific matters, and displayed some ability as a mathematician and as a musi- cian, being able to play with skill upon the organ and some other instru- ments. In 1837 he was sent to Phillips Academy, where it seems he did as so many others having marked ability are reported to have done at school, namely, chosen to do something else than what was required of them, and here he was admonished that he was " disappointing the best hopes of his friends." He contrived, however, to enter Dartmouth College in 1840, but left before completing the college course. He began teaching at Elliot, Maine, and soon after had charge of the Bel- knap School in Dover, New Hampshire, where he remained until 1847. He became interested in electrical phenomena in 1845, and from that time on he devoted himself to enlarging the sphere of electrical indus- tries, and in perfecting the apparatus employed. It is to be remembered that the first telegraphic line was built and operated in 1844, and on its success being demonstrated its development was very rapid, and Farmer became an operator and inspector in 1847, moving to South 41G MOSES GERRISH FARMER. Framingham, Massachusetts. He had already invented an electro- magnetic engine and an electric locomotive, and had given numerous lectures on electrical subjects in the towns and cities about. It was at this time he devised the fire alarm system which was soon adopted by the city of Boston, he superintending its construction. This work has been ranked as the beginning of electrical engineering in the world. The automatic signalling apparatus, the control of powerful mechanical apparatus at a distance, the closed electrical circuits to be worked by interruptions, the adjustment of such circuits to a properly constructed battery, the devising of magneto-electric machines for producing powerful currents and driving such machines by water power, the protection of the conductors by placing them in pipes, the protection of the station itself from lightning by special ingenious apparatus, with many more things essential for the proper working and efficiency of the system, were devised by him, — for there were no such devices in existence and there was no exj)erience. He was at work in an alto- gether new field, and his work was so well done that his system was adopted presently by all the cities in the country. Mr. Farmer's 'work at Framingham required much Sabbath atten- tion, and as he had conscientious scruples against such work he aban- doned the position and moved to Salem in 1848. Here he devised several improvements in telegraphic apparatus, and became superin- tendent of the telegraphic line from Boston to Burlington, Vermont. Some of his inventions made within the next ten years have since been indispensable in doing telegraphic work. What is known as the auto- matic repeater is one with which he had succeeded before 1855 in sending four messages simultaneously over one wire, — what has since been called the quadruplex system. He invented too the first auto- matic printing telegraph, now known as the ticker. In 1855 he succeeded in electrically depositing aluminum, and constructed for the Dudley Astronomical Observatory at Albany a chronograph and an electrical clock. In 1856 he made an electric gyroscope so as to run continuously at uniform speed, and he read in that year a paper on Multiplex Telegraphy before the American Association for the Advancement of Science. About this time he began investigations in the problem of electric lighting and advanced so far as to have his parlor in Salem lighted by an incandescent electric lamp in 1859. The supply of electricity was furnished by a galvanic battery, and he satisfied himself that with such a source of electricity electric lighting was not feasible. MOSES GERRISH FARMER. 417 From 1864 to 1868 he experimented much with thermo-electric devices, and discovered an alloy to be used for such a battery that had much higher efficiency than any others which had been employed, — one that compared favorably with the best we now have, except that it is rather too brittle for commercial use. With such a source, however, he coated steel and iron wire with copper for the sake of combining high conductivity with great tensile strength, and this was developed into a commercial enterprise of considerable magnitude. It was aban- doned for some years, but its superior qualities for telephonic work have again made demand for it. It appears that the first written description of what is now known as the self-exciting dynamo, in which the dynamo current is sent through its own field coils to strengthen its magnetism, was by Mr. Farmer in a letter to an eminent English electrician. The same idea occurred to several others about the same time, notably Siemens and Varley, but its immense importance did not so impress any of them as to put the idea into a practical form, and it remained for others unknown before in the electrical field to put it into the commercial shape so familiar now. In 1872 he was appointed Professor of Electrical Science at the Torpedo Station at Newport, Rhode Island, a position which was quite to his taste, and where he remained for nine years. During the first fifty years of his life he was in a constant struggle with poverty and sickness. Like Agassiz, he had no time to make money. He was so fertile in new things, he cared but little for the worth of his inventions; though he took out many patents, he profited little from them. His .work was mostly of a fundamental sort, and few persons who have applied for patents on electrical devices for the past twenty years have not found that Farmer had preceded them in their territory, and there were few who had new ideas on any elec- trical matter who could not find in some of his numerous note-books the identical things already specified, and oftentimes the experimental work done, but not published, for he does not appear to have written much. Mr. Farmer's relation to the electric lighting industry has not been generally known, but it appears that in 1868 he had a dynamo made with which he lighted forty incandescent electric lamps, in multiple arrangement, as is now the practice, and automatically regulated. His lamps were of iridium, which he found to possess the proper electri- cal qualities, but that metal not being found in an available form for commercial use, the system now so common was by him devised vol. xxix. (n. s xxi.) 27 418 MOSES GERRISH FARMER. and carried out, lacking only the vacuum carbon lamp for completion. A fire in 1868 destroyed this first dynamo. The armature was the only portion that was rescued, and this was exhibited at the World's Fair in Chicago as part of an historic exhibit which he was preparing of his own work, when he was suddenly prostrated and died. For a number of years he had been unable to walk, owing to a paralytic attack, and it became impossible for him to do any work. Love for electrical science remained with him to the last, and it was a real pleasure to hear him tell of his attempts, his successes and his failures, though he was not one to introduce personal matters in conversation. There is one event in his life which is worth chronicling, and which may not find its way into other notices. It may be remembered that a certain Dr. Gary, in 1878, had for a while on exhibition in Boston a machine which purported to be a perpetual motion. It was made with permanent magnets, and had a rotating armature so mounted as to change its polarity at certain points. Many saw this machine, but most thought there was some fraud about it. Professor Farmer went to see it, and asked permission to examine it, and was told he might take it to pieces and reconstruct it himself if he liked, which he did, and he told the writer the machine actually started up and ran without further attention. He offered to buy it, but the inventor did not want to sell. As Professor Farmer was an expert mechanician and elec- trician, it would seem unlikely that a piece of trickery should not have been discovered when subjected to such a critical examination in his own hands. To me it seems more likely that in some way there was a draught of energy for the propulsion from some source not hitherto recognized, able for a while to supply a small amount, for the machine would not run long at a time ; but that it should run at all is the won- der, and Professor Farmer testifies that it did. During the last years of his life he lived in Elliot, Maine, able to do but little on account of his paralysis. He could ride about somewhat and was an occasional visitor to Boston, but the younger race of elec- tricians know of him only as a name. His work was chiefly done when electrical nomenclature had no existence, when there were no standards, when the whole field was new and telegraphy was growing, as the electric lighting industry has grown during the past ten years, and there was a great demand for electrical facilities for that kind of work. Mechanical aptitude was as much needed as an ally ao it is now, and this combination of talents Professor Farmer possessed in a hich decree. 1894. A. E. Dolbkak. HERMANN AUGUST HAGEN. 419 HERMANN AUGUST HAGEN. Hermann August Hag en was bom at Kouigsberg, East Prussia, on May 30, 1817. I lis mother was Anna Dorothea Linck and his father Carl Heinrieh Hagen, Professor of Political Economy, Technol- ogy, and Agriculture at the University of Konigsberg. His education was obtained at the Gymnasium Collegium Friedericianum and at the Kneiphoiisehe Gymnasium ; and after graduating from the latter in 183G he studied medicine at the University of Konigsberg. Hageu's early interest in natural history, stimulated by his cele- brated instructors, von Baer, Rathke, and von Siebold, was especially directed towards entomology by his father and his grandfather, the latter, Carl Gottfried Hagen, a Professor of Natural History in Konigs- berg and author of " Chloris Borussica," a small volume published in 1819. It is said that the attention of young Hagen was called to the Odonata, or dragon-flies, "because by chance the first specimen he caught proved to be an undescribed insect of that order." In 1839 in company with Professor Rathke he visited Norway. Sweden, and Denmark, and here, though he paid some attention to the habits and structure of marine animals, he studied chiefly in the principal entomological collections and libraries. He received the degree of Doctor of Medicine from the University of Konigsberg in 1840; his thesis being entitled " Synonymia Libellu- larum Europaearum," (Regimontii, 1840, 8°, pp. 84,) and indicating thus early the exactness which was so marked a feature of all his bibliographical woik. After his graduation in medicine, in 1840, Dr. Hagen studied in Berlin, Vienna, and Paris, and, returning to Konigsberg in 1843, set- tled there as a practising physician and surgeon. At the surgical hospital, where for several years he was first assistant, he performed a large part of the operations ; among the needy his services, always in demand, were given with that ready tenderness so characteristic of the sympathetic side of his nature. His native city also claimed a large amount of his time during the years 1863-67, when as Vice-President of the City Council and a member of the School Board he was required to report upon a great number of subjects demanding much painstaking research. Though educated as a physician, like so many of the older German men of science, Hagen, during all the years of his medical practice and of his civic duties, published continuou-ly. His first entomological 420 HERMANN AUGUST HAGEN. paper, a List of the Dragon-flies of East Prussia, was printed in the Preussische Provinzialblatter for January, 1839, when he was still at the University ; * from that date until incapacitated in 1890 his con- tributions to science number several hundreds. Some of these deserve special notice. In 1842, when studying in Paris, Dr. Hagen met Baron de Selys-Longchamps of Liege, and the friendship then begun resulted in the production through their joint studies of the well known "Revue des Odonates d'Europe" (1850), "Monographic des Calopterygines" (1854), and "Mono- graphic des Gomphines " (1857). The co-operation of De Selys and Hagen was not limited, however, to the works published conjointly ; the writings of both had the advantage of a most generous inter- change of notes, specimens, and drawings for a period of more than forty years. The study of the insects found in amber and of the fossil Neurop- tera early and continuously engaged his attention, his well known accuracy and thoroughness both in description and in delineation pro- ducing some of the best work yet done in palaeo-entomology. The absence of generalizations, apparent in all bis work, is conspicuous here ; he was slow in framing hypotheses, ready and critical in testing their validity. His " Monographie der Termiten," published in the Linngea Ento- mologica for 1855, 1858, and 1860, is a masterpiece of original work, supplemented by a most exhaustive and thorough resume of all previous studies in this most difficult of all families of insects ; it will always remain a classic. One of Dr. Haven's earliest and most useful contributions to American entomology is the " Synopsis of North American Neu- roptera," prepared at the special request of the Smithsonian Institution, and published by the Institution in 1861. With a keen insight into the systematic relations of forms, and quick to perceive specific differ- ences, he possessed the rare power of stating these differences clearly and concisely. Probably his best known work, certainly the one more frequently consulted by entomologists and zoologists than any other, is the " Bibliotheca Entomologica : die Litteratur tuber das ganze Gebiet der Entomologie bis zum Jahre 1862," published in two volumes in 1862 and 1863. Modelled somewhat after the earlier work of Percheron, it is to day one of the most complete and accurately prepared of scien- * Two earlier genealogical studies were printed in 1834. ..- HERMANN AUGUST HAGEN. 421 tific bibliographies. The titles are arranged alphabetically under the names of the authors, and under each author's name in chronological sequence ; entries are made under initials and pseudonyms when these alone are known, and anonymous works are arranged by subjects. An excellent and very detailed synoptic register affords easy reference, and adds greatly to the usefulness of the work. Like all his work, this was prepared with the most laborious painstaking; the libraries of Germany, France, Belgium, Holland, and England were examined, and an amount of drudgery that would have discouraged many was overcome by hard and persevering work. From the study of various papers written by Dr. Th. Bail, Dr. Hagen became convinced that the ravages of various injurious insects could be readily cheeked by the application of the yeast fungus, and his paper on the " Destruction of Obnoxious Insects, Phylloxera, Potato Beetle, Cotton Worm, Colorado Grasshopper, and Greenhouse Pests, by the Application of the Yeast Fungus," (Cambridge, 1879,) though based on the mistaken assumption that the beer yeast fungus (Sacc/iarontyces cerevisice) enters an insect's body and produces a dis- ease, caused much discussion. Metschnikoff was thus led to examine several minute parasitic fungi, and was able to prove experimentally the deadly character of " green muscardine " (Isaria destructor) to insects exposed to infection. In his efforts to cultivate the " green muscardine" apart from the insects themselves, Metschnikoff used beer-mash successfully, producing thus a poison which once set ar work multiplied and spread spontaneously, and it is probable that we have here the true explanation of the value of the application of yeast to plants affected by insects. In 18G7 Dr. Hagen left Konigsberg and accepted the invitation of Professor Louis Agassiz to take charge of the entomological department of the Museum of Comparative Zoology at Cambridge ; in 1870 as Pro- fessor of Entomology in Harvard College he commenced his Univer- sity career. Dr. Hagen entered upon his duties at the Museum with great zeal ; and his detailed plan for the arrangement of the collec- tions, though somewhat modified, is, and is likely to remain, the basis for the future. Deeply interested ia everything relating to museum work, as his appreciation of series of specimens, his care for their preservation and for the accuracy of their localities, and many minor details, clearly indicate, it is in this collection as well as in his writings that his contributions to science are to be found. Here alone we can fully realize the extent of his discoveries, the keenness of his insight, his skill at preparation and dissection and with the pencil. I lis devo- 422 HERMANN AUGUST HAGEN. tion to the Museum knew no bounds ; all personal interests were secondary. In 1876 he refused a most flattering and urgent invitation to take charge of the great entomological collections of the Konigliches Museum fur Naturkunde in Berlin, and the time that might have been given to original work was lavished upon the care and arrangement of the collections, which grew rapidly both in size and value. The bio- logical collection, or that illustrating the life history of the species, is a prominent specialty of the Cambridge Museum. In this are preserved specimens showing every condition of an insect's life, the eggs, larvae in all stages from those just hatched to those full grown, their barrows, nests, partially devoured leaves, etc., the work of both larvae and adults, the frass or excrements often of great importance, pupal stages, adults of both sexes, and the parasitic and predaceous enemies also in all stages of development. Dr. Hagen's influence upon the formation of such biological collections has been very great ; few were in existence at the time when, almost unaided, he created that at Cambridge, and the care and elaborateness with which the whole is labelled makes it not only a worthy model, but most truly a monument to persistent and well directed industry. His lectures, given at rare intervals to advanced students, contained much genuine and exact knowledge, and his many acts of kindness and words of wise counsel will not soon be forgotten by those who enjoyed the facilities of the department under his charge. Most of Dr. Hagen's journeys were undertaken for study among collections and in libraries. In the summer of 1882, however, accepting the generous and thoughtful invitation of Professor Raphael Pumpelly, at that time Director of the Northern Transcontinental Survey, he visited California, Oregon, Washington, and Montana. The object of the survey was to collect data concerning insects injurious to vegeta- tion, both of the field and of the forest. The greater part of the time was spent in the Yakima and Columbia regions of Washington ; many important entomological discoveries were made, some with a direct economic bearing, and large collections of insects were obtained from a most interesting locality. I had the good fortune to be one of the party accompanying Dr. Hagen on this occasion, and saw everywhere his delight at the great natural beauties of the country, and if the many discomforts, inseparable from such a trip to a man at the age of sixty-five, were borne with not a little impatience, yet his enthusiasm and determination assured his companions that the expedition would be carried to the end he had planned for it. In after years his expe- riences during the summer of 1882 were among the most enjoyable of his memories. JAMES RUSSELL LOWELL. 423 Dr. Hagen was a man of marked character, simple and sympathetic, and if at times somewhat hot and hasty in temper and impatient of opposition, he had also one of the warmest of hearts and most gener- ous of dispositions. His unostentatious hospitality was enjoyed by many entomologists, who found his life in Cambridge quiet, contented, and happy. Of Dr. Hagen's domestic life it is sufficient to record here that in 1851 he married Johanna Maria Elise Gerhards, who survives him. Dr. Hagen received the honorary degree of Doctor of Philosophy from the University of Konigsberg in 18G3 ; Harvard made him a Doctor of Science in 1887. The renewal of his medical degree on the 17th of October, 1890, the date of his graduation fifty years previously, after the custom of German universities, gave him great pleasure. He was elected a Fellow of the American Academy of Arts and Sciences, November 11, 1868, and served on the Council in 1877-78. He was also a member of a goodly number of scientific associations, and most of the entomological societies the world over were glad to enroll him as an honorary member. Stricken with paralysis in September, 1890, Dr. Hagen lingered for more than three years ; his painful sufferings being lightened by all that affectionate and devoted care could do. fie died at Cambridge, November 9, 1893, and was buried in the grounds of Harvard University at Mount Auburn, near his associate, Louis Francois de Pourtales. 1894. Samuel Henshaw. JAMES RUSSELL LOWELL. James Russell Lowell was born at Elmwood, Cambridge, Mas- sachusetts, February 22, 1819 ; he died at the same place, August 12, 189 1. He was the youugest of a family of five, two daughters and three sons, born of Charles and Harriet Spence Lowell. His father at the time of Lowell's birth was thirty-seven years old and lived till 1861, when his son was forty-two. He was minister to the West Church, Boston, and his son has drawn his portrait in a letter to C. F. Briggs, written in 1844: " He is Dr. Primrose in the comparative degree, the very simplest and charmingest of sexagenarians, and not without a great deal of the truest magnanimity." The Lowells traced their descent from Percival Lowell of Bristol, England, who settled in Newbury, Massachusetts, in 1639, and in the generations just preceding that of James Russell Lowell three of the family besides his father had 424 JAMES RUSSELL LOWELL. specially honorable names : Francis Cabot Lowell, who gave a great impetus to New England manufactures, from whom the city of Lowell took its name ; Judge John Lowell, the author of the section in the Bill of Rights which wrote the death warrant of slavery in Massa- chusetts ; and John Lowell, Jr., whose wise and far-sighted provision gave his native city that powerful centre of intellectual influence, the Lowell Institute. Mrs. Harriet Spence Lowell, a native of Portsmouth, New Hamp- shire, was of Scotch origin. She is described as having "a great memory, an extraordinary aptitude for language, and a passionate fondness for ancient songs and ballads." It pleased her to fancy her- self descended from the hero of one of the most famous ballads, Sir Patrick .Spens, and at any rate she made a genuine link in the Poetic Succession. In a letter to his mother, written in 1837, Lowell says: " I am engaged in several poetical effusions, one of which I have dedicated to you, who have always been the patron and encourager of my youthful muse." Elmwood in the days of Lowell's boyhood was in a more distinctly rural neighborhood than now, but it never has wholly lost its charm of seclusion. In his paper, "My Garden Acquaintance," in many of his poems, such as " An Indian Summer Reverie," " To the Dandelion," " Under the Willows," " Al Fresco," and in many passages in his let- ters, he bears witness to the intimacy which he enjoyed with that phase of nature which we may call homely and friendly. He once expressed to me his delight in Poussin's landscapes, and in his descriptive poetry it is noticeable that the large, solemn, or expansive scenes of nature make no such appeal to his interest as those nearer vistas which come close to human life and connect themselves with the familiar experience of honie-keeping wits. Lowell's school days were spent in his own neighborhood. Mr. William Wells, an Englishman and unsuccessful publisher, opened a classical school in one of the spacious Tory Row houses near Elm- wood, and, bringing with him English public school thoroughness and severity, gave the boy a drilling in Latin which his quick appropria- tion of strong influences turned into a familiar possession. Possibly the heavy hand of the schoolmaster, by its repression, gave greater buoyancy to the spirit of the student when the comparative freedom of college followed. Lowell was sixteen when he entered Harvard College with the ciass which graduated in 1838. In "An Indian Summer Reverie," he says : JAMES RUSSELL LOWELL. 425 "Though lightly prized the ribboned parchments three, Yet collegisse juvat, I am glad That here what colleging was mine I had, — It linked another tie, dear native town, with thee ! " Whether or no there was a reaction from the discipline of school days, it is certain that the independence which characterized Lowell through- out his life found expression now, not in insubordination, but in a frank pursuit of those courses of study and lines of reading which four years of academic leisure and the tolerable equipment of the college and home library put in his power. " Never," says Lowell in his essay, " A Great Public Character," when speaking of college life, — " Never were we ourselves so capable of the various great things we have never done " ; and however much he may have been generalizing for college youth, he recalled well his own spiritual experience ; with an impulse which outwardly was wayward, he obeyed that law of his being which his growing consciousness of intellectual power disclosed to him. In his penetrating discrimination between talent and genius, he says profoundly : " The man of talents possesses them like so many tools, does his job with them, and there an end ; but the man of genius is possessed by it, and it makes him into a book or a life according to its whim. Talent takes the existing moulds and makes its castings, better or worse, of richer or baser metal, according to knack and opportunity ; but genius is always shaping new ones and runs the man in them, so that there is always that human feel in the results which give us a kindred thrill. What it will make, we can only conjecture, contented always with knowing the infinite balance of possibility against which it can draw at pleasure." His was a singularly self- centred nature, and he was always true to that large ideal which was his consciousness of greatness projected in history and literature ; but there was a whimsical uncertainty in his mind as to the precise direc- tion in which his genius would at any time take him. It is interesting to observe this self-centred nature in its early struggle after equipoise. So far as outward activity was concerned, he took a degree in law, but confessed to an aversion from the practice, and for a while busied himself in a counting-room. His vacillation of mind regarding his vocation, his apparent fickleness of purpose, the conflict going on between his nature craving expression and the world with its imperious demands, the stirring within him of large designs, and the happy contentment in the pleasures of the day, all seek outlet in his natural yet uneasy letters. He was finding himself in these early days, as many another young man, and there are glimpses all 426 JAMES RUSSELL LOWELL. through Lowell's letters of this restlessness, this subtle sense of one's self which in weaker natures hardens into a mordant self-conscious- ness. Now and then he turns upon himself in a sort of mingled pride and shame, as if at once aware of his power and angry that he has it not wholly at his beck. But for the most part one is aware of a nature singularly at one with life, and finding its greatest satisfaction in getting at the world through the reflection of the world in literature. No one would deny that Lowell was eminently a man of books, but it would be a wholly inadequate phrase which described him as a bookish man. That he was at home in a library his early letters show, but they show also how even then he read through his books into life, and interpreted history and literature by means of an innate spiritual faculty which was independent of intellectual authority. It is this criticism at first hand, this swift, direct penetration of the reality, which mark emphatically what I have characterized as Lowell's self-centred nature. He has told us that his brain required a long brooding time ere it could hatch anything. He was speaking of the matter of expres- sion ; but the phrase is a fit one for his habitual temper. The super- ficial charge of indolence could apply only to his apparent disregard of bustling activity. His nature was of the sort that knows the power of stillness, and though he upbraids himself in his letters for his unproductiveness at times, he had plainly the instinct which waits on opportunity. His faculty of observation was very strong, but it was no stronger than his power of assimilation ; and thus it was that when opportunity came he had not hurriedly to adjust himself to the situation. It was while he was engaged with his books and his friends, profess- ing law but practising literature in the way of poetical and prose contributions to the magazines, that he was roused out of his dreams by the prick of necessity in the sudden loss by his father of much of his property, and by the impulse given to his own moral force by the coming into his life of Maria White. He became engaged to this lady in the fall of 1840, and the next twelve years of his life were pro- foundly affected by her influence. Herself a poet of delicate power, she brought an intelligent sympathy with his work ; it was, however, her strong moral enthusiasm, her lofty conception of purity and justice, which kindled his spirit and gave force and direction to a character which was ready to respond and yet might otherwise have delayed active expression. They were not married until 1844, but they were not far apart in their homes, and during these years Lowell was mak- ing those early ventures in literature, and first raids upon political and JAMES RUSSELL LOWELL. 427 moral evil, which foretold the direction of his later work, and gave some hint of its abundance. In 1841 he collected the poems which he had written and sometimes contributed to periodicals into a volume entitled " A Year's Life," and inscribed in a veiled dedication to his future wife. In hopes of better- ing his fortune, and in obedience to the instinct which most young men of letters have, he undertook with Robert Carter the publication of a literary journal, "The Pioneer," which died under their inexperienced hands in three numbers. He began also to turn his studies in dra- matic and early poetic literature to account, and after printing a portion in Nathan Hale's " Miscellany " published, in 1844, " Conversations on some of the Old Poets." In the same year he again collected his poetic work into a volume of " Poems." The difference between the two volumes of poems, though separated by three years only, is marked. Few of the verses from "A Year's Life" are included in the poet's final collection of his writings, few are omitted from " Poems." One poem in the earlier volume, " Irene," is conspicuous as a poetic portrait of the figure of peace which had come into his somewhat turbulent spiritual life, but the volume as a whole is characterized by vague sentimentalism and restless beating of half-grown wings. Three years later, some of this same immaturity is discoverable, but with the poems which wander in somewhat unmeaning ways are those spirited adventures like " Rhcecus," "The Shepherd of King Admetus," and " Prometheus," which denote the growing consciousness of positive poetic power, and also those stirring Sonnets to Wendell Phillips and J. R. Giddings, and the lines entitled " A Glance behind the Curtain," which disclose a new passion leaping up as the champion of truth and righteousness. It is noticeable, too, that in the first volume there is no trace of humor and scarcely any singular felicity of phrase; in the second, wit and humor begin to play a little on the surface. In " Conversations," where the familiar form gives freer scope, there is a gayety of speech which inti- mates the spontaneity of the man and anticipates the rich fruitage of later years. In all these books, however, there is good evidence of the rapid growth which was taking place in Lowell's intellectual and moral life, a coming to his own which it would take only some strong occasion to make sure. This occasion was the Mexican War, with the greater contest which flamed up with it over the encroachments of slavery. Lowell and his wife, who brought a fervid antislavery temper as part of her marriage portion, were both contributors to the " Liberty Bell," and Lowell was a frequent- contributor to the " Antislavery Standard," 428 JAMES RUSSELL LOWELL. and was iudeed for a while a corresponding editor; but in June, 1846, there appeared one day in the Boston Courier a letter from Mr. Ezekiel Biglow of Jalaam to the Hon. Joseph T. Buckingham, editor of the Boston Courier, enclosing a poem of his son, Mr. Hosea Biglow. It was no new thing to seek to arrest the public attention with the vernacular applied to public affairs. Major Jack Downing and Sam Slick had been notable examples, and they had many imitators ; but the reader who laughed over the racy narrative of the unlettered Ezekiel, and then took up Hosea's poem and caught the gust of Yankee wrath and humor blown fresh in his face, knew that he was in with the appearance of something new in American literature. A score of years afterward, when introducing the Second Series of "The Biglow Papers," Lowell confessed that when he wrote this letter and poem he had no definite plan, and no intention of ever writing another. It was struck out from him by the revolt of his nature at the iniquity of slavery and the war into which slavery was dragging the nation. But he adds, " The success of my experiment soon began not only to astonish me, but to make me feel the responsi- bility of knowing that I held in my band a weapon, instead of the mere fencing stick I had supposed. ... If I put on the cap and bells, and made myself one of the court fools of King Demos, it was less to make his Majesty laugh than to win a passage to his royal ears for certain serious things which I had deeply at heart." "The Biglow Papers " not only gave Lowell to himself and opened the flood gates of his patriotism and his noble indignation ; they gave him a public, and thus furnished the complement which every author demands. " Very far," he says, in the same Introduction, " from being a popular author under my own name, so far, indeed, as to be almost unread, I found the verses of my pseudonym copied every- where ; I saw them pinned up in workshops ; I heard them quoted and their authorship debated." The force which he displayed in these satires made his book at once a powerful ally of a sentiment which heretofore had been ridiculed ; it turned the tables and put Anti- slavery, which had been fighting sturdily on foot with pikes, into the saddle, and gave it a flashing sabre. For Lowell himself it won an accolade from King Demos. He rose up a knight, and thenceforth possessed a freedom which was a freedom of nature, not a simple badge of service in a single cause. His patriotism and moral fervor found other vents in later life, and he never laid down the sword which he then took up, but it is significant of the stability of his genius that he was not misled by the sudden distinction which came JAMES RUSSELL LOWELL. 429 to him into a limitation of his powers. It was shortly after this that he wrote, in one of those poetic ahsences from his every -day life, which were to overtake him more than once afterward, his " Vision of Sir Launfal," and the exuberance of his nature together with his keen power of criticism found expression about the same time in his witty '• Fable for Critics." A third volume of Poems appeared in the same year, 1848, as the last named. A year in Europe, 1851-52, with his wife, whose health was then precarious, stimulated his scholarly interests, and gave substance to his study of Dante and Italian literature. In October, 1853, his wife died, and in 1855 he was chosen successor to Mr. Longfellow as Smith Professor of the French and Spanish Languages and Litera- tures, and Professor of Belles Lett res in Harvard College. He spent two years in Europe in further preparation for the duties of his office, and in 1857 was again established in Cambridge and installed in his academic chair. He married also at this time his second wife, Miss Frances Dunlap, of Portland, Maine. Lowell was now in his thirty-ninth year. As a scholar, in his pro- fessional work, he had acquired a versatile knowledge of the Romance languages and was an adept in old French and Provencal poetry ; he had given a course of twelve lectures on English Poetry before the Lowell Institute in Boston which had made a strong impression on the community, and his work on the series of British Poets in connec- tion with Professor Child, especially his biographical sketch of Keats, had been recognized as of a high order. In poetry he had published the volumes already mentioned. In general literature he had printed in magazines the papers which he afterward collected into his volume " Fireside Travels." It was not long after he entered on his college duties that "The Atlantic Monthly" was started, and the editorship given to him. For the details of the office he had little aptitude, although he looked keenly after nice points of literary finish in the proof-reading ; he was relieved of much of the detail by his active assistant, Mr. F. H. Underwood, to whom the inception of the maga- zine was largely due. But the Atlantic afforded a good outlet for his literary production, and though he held the editorship but a little more than two years he stamped the magazine with the impress of his high ideals in literature and criticism ; his selection of articles wa9 judicious, his own contributions and criticism were full of life, and he was most generous in his critical aid to contributors. In 1862 he was associated with Mr. Charles Eliot Norton in the conduct of "The North Ameri- can Review," and continued in this charge for ten years. In 1877 430 JAMES RUSSELL LOWELL. he was appointed by President Hayes Minister to the Court of Spain. These twenty years, from 1857 to 1877 were the most productive period of Lowell's literary activity. He was in the maturity of his mental power, he held a convenient position in University life, his home relations were congenial and stimulating, and his collegiate work as well as his editorial charge successively of the Atlantic and North American gave him a needed impulse to literary effort. During this period appeared the most of that body of literary history and criticism which marks him as the most distinguished of American critics. Brought together in his writings under the general head of " Literary Essays," these papers are the rich deposit of a mind at once sympathetic and discriminating, capable of enjoying to the full the varied manifestations of life in literature, and yet an intuitive judge and penetrating critic. While this broad stream of literary criticism was flowing, there was another expression of Lowell's nature, never divorced from this love of letters, — a criticism of life, especially as it took form in con- temporaneous American history. The period which I have named covered the preparation for the war for the Union, that war itself, and the reconstruction era afterward, and the expression of Lowell's nature in its attitude toward the whole period was manifold. The volume of " Political Essays" contains the incisive papers which stung the irresolute and time-serving, and inspirited the ardent lovers of truth and liberty. It is impossible to read these papers now without admiration for the political sagacity of the writer, — a sagacity before the event, not after. Every page bears witness to the sanity with which he regarded contemporaneous affairs, when madness seemed the most natural temper in the world, and his insight of human nature was that of a poet who did not regard his power of vision as exclud- ing the necessity of paying taxes. History has been supplying foot- notes to these pages, with the result, not of correcting the text, but of confirming it. In this same period also he wrote and published the Second Series of the Biglow Papers, and used his satire and his moral indignation with a depth of feeling which surpassed that shown in the first series, a little to the detriment thereby, it may be, of the brilliancy of the general effect. In truth, strong as was Lowell's power of invective, his passion of patriotism found this vent too narrow ; there was a large, constructive imagination at work on the great theme of national life, which found fuller expression in the Odes which the Centennial JAMES RUSSELL LOWELL. 4ol and Commemorative occasions called out. Lowell seized the occasions with a spirit which scarcely needed them, and merely employed them as fit opportunities for casting in large moulds the great thoughts and feelings which rose out of the life of a man conscious of his in- heritance in a nohle patrimony. It was at the close of this period, in which he had done incalcula- ble service to the Republic, that Lowell was called on to represent the country, first at Madrid and afterward at London. Eight years were thus spent by him in the foreign service of the country. His sole participation in practical politics, as the term is, up to this time had been to attend a national convention once as delegate, and to have his name used as Presidential Elector. To the minds of many of his countrymen he seemed doubtless a dilettante in politics. Special preparation in diplomacy he had not, but he had what was mote fundamental, a large nature enriched by a familiar intercourse with great minds, and so sane, so sound in its judgment, that whether he was engaged in determining a reading in an Elizabethan dramatist or in deciding to which country an Irish colossus belonged, he was bring- ing his whole nature to the bench. No one can read Lowell's despatches from Madrid and Loudon without being struck by his sagacity, his readiness in emergencies, his interest in and quick per- ception of the political situation in the country where he was resident, and his unerring knowledge as a man of the world. Nor could Lowell lay aside in his official communications the art and the wit which were native to him. " I asked Lord Lyons," he writes in one letter, " whether he did not think suzerainty might be defined as • leaving to a man the privilege of carrying the saddle and bridle after you have stolen his horse.' He assented." But though Lowell's studies and experience had given him a preparation for dealing with diplomatic questions, the firmness with which he held his political faith afforded as sure a preparation for that more significant embassy which he bore from the American people to the English. Not long after his return, he published a little volume containing the more important speeches which he had made while in England. Most of them had to do with litera- ture, but the title address in the volume, " Democracy," was an epigrammatic confession of political faith as hopeful as it was wise and keen. A few years rater he gave another address to his own countrymen on "The Place of the Independent in Politics." It was a noble apologia, not without a trace of discouragement at the apparently sluggish movement of the recent years, but with 432 HENRY WARREN PAINE. that faith in the substauce of his countrymen which gave him the right to use words of honest scorn and warning. What impresses one especially in reading this address, remembering the thoughtless gibes which had been flung at this patriot, is the perfect self-respect with which he defines his position, the entire absence of petty retalia- tion upon his aspersers, the kindliness of nature, the charity, in a word, which is the finest outcome of a strong political faith. It must have been galling to Lowell to find himself tauuted with being un-American. He could afford to meet such a charge with silence, but he answered it with something better than silence when he reprinted in a volume his scattered political essays. The public life of Mr. Lowell made him more of a figure before the world. He received honors from societies and universities ; he was decorated by the highest honors which Harvard could pay offi- cially, and Oxford and Cambridge, St. Andrews and Edinburgh, and Bologna, gave gowns. He established warm personal relations with Englishmen, and after his release from public office he made several visits to England. There, too, was buried his wife, who died in 1885. But the closing years of his life in his own country, though touched with domestic loneliness and diminished by growing physical infirmities that predicted his death, were rich also with the continued expression of his large personality. He delivered the public address in commemoration of the 250th anniversary of the founding of Har- vard University, he gave a course of lectures on the Old English Dramatists before the Lowell Institute, he collected a volume of his poems, he spoke and wrote on public affairs, and the year before his death revised, rearranged, and carefully edited a definitive series of his writings in ten volumes. 1894. Horace E. Scudder. HENRY WARREN PAINE. The death of Henry Warren Paine, LL.D., took place at his residence in Cambridge on the 2Gth of December, 1893. He had been a Resident Fellow of the American Academy in Class III. Section 1, Philosophy and Jurisprudence, from the year 1871. A membership so long continued may bear witness to the interest taken by our late associate in the advancement of knowledge and liberal culture beyoud the immediate sphere of his own active life and achievements in the special field of the practical jurist. He was born at Winslow in the State of Maine, August 30, 1810, HENRY WARREN PAINE. 433 a son of Lemuel and Jane Thomson (Warren) Paine. His father was a lawyer by profession, and in later years engaged in literary pursuits. His mother was a niece of General Joseph Warren, of Bunker Hill memory. His early education was well cared for, and he made rapid progress in his studies, taking little hand in the ordinary sports of youth. When duly fitted, he entered the College (now Colby University) at Waterville, Maine, in 1826, and graduated with the highest honors in the Class of 1830. In his Senior year he was also Principal of the Waterville Academy, and served one year as Tutor in the College. He then began the study of law in the office of his uncle, Samuel S. Warren, and continued it with William Clark, of Hallowell, and then for one year in the Law School of Harvard University, and was admitted to the Bar in Maine in 1834. In the course of twenty years of practice at Hallowell, Mr. Paine had attained a distinguished position in the profession, had been for five years Attorney for Kennebec County, and had served for several years (between 1836 and 1853) as a Representative of Hallowell in the State Legislature. Tempting offers of further political promotion, and even a seat upon the Supreme Bench of the State, were declined. It seems to have been his purpose to adhere strictly to the line of the profession he had chosen. Chief Justice Appleton spoke of him as "a profound and learned lawyer, as well as an accomplished advo- cate." It is evident that by the year 1853 he had determined to seek a wider field for his exertions in the metropolis of New England. Mason, Webster, Fletcher, Choate, and many others, had done the like before him. These were indeed perilous examples to be followed, unless a man felt quite sure of his own strength. Perhaps he had heard of the remark of Mr. Webster, that " there was always room enough up above." In 1854 he had removed his office to Boston, and established his residence in Cambridge. Not long afterwards he was engaged with Rufus Choate and F. O. J. Smith in the impeach- ment trial of Judge Woodbury Davis at Portland, on which occasion it was remarked that " Paine furnished the logic, Choate the rhetoric, and Smith the slang." For the rest of his life Mr. Paine steadily pursued his professional avocations in Boston, where he soon acquired both a large business and a high standing at the Bar. It is said that his charges were always moderate. With a practice sufficiently lucrative for all his needs, he showed no grasping eagerness for excessive wealth. Indeed, a great lawyer is apt to think that, if a man desires to be very rich, be should quit the law and go into the oil business. For many years, vol. xxix. (n. s. xxi.) 28 434 HENRY WARREN PAINE. almost the only relief he could find, or could allow himself, from pressing employments and assiduous labors, was in the quiet enjoy- ment of his elegant home, with his books and private studies, or in social intercourse with his friends. Those who knew him best speak highly of his social qualities, his kindness of heart, his agreeable conversation, and the conservative character of his views in general. In public affairs, while he was a firm supporter of constitutional gov- ernment and law, his sympathies leaned to the side of the Democratic policy; in religion his affiliations were with the liberal Unitarian Church. The proceedings of the Bar meetings in Boston upon occasion of his death may furnish the best evidence of the high estimation in which the character and professional career of Henry W. Paine were held by his brethren of the bench and bar. They signalize his emi- nent ability as a lawyer and advocate, his learning in the law, his literary attainments, his wit, uniform courtesy of manner, and con- stant regard for the honor and dignity of the profession ; and they further assure us that his sterling integrity and skill were so generally appreciated by the community at large as to reflect credit on the call- ing, and tended to strengthen the confidence of the public in the faith- ful administration of justice in the courts. It is said that the place of Chief Justice of this State was on one occasion tendered to Mr. Paine, an honor which he felt constrained to decline, for reasons doubtless satisfactory to himself. No one could be more fully aware of the exhausting labors and weighty responsibilities of high judicial station. Mr. Justice Nelson once observed that, what with hearings and consultations by day and the writing of opinions in the night, the best if not the only time he could have, he had often wished he had been some plain farmer, who could sleep soundly when his day's work was done. In 1854 the degree of LL. D. was conferred upon Mr. Paine by Colby University, of which he was a Trustee. He was also a member of the " Historical Society" in Maine. From 1870 to 1882 he served as one of the Board of Overseers of Harvard University ; and from 1872 to 1883 he was a Lecturer on the Law of Real Property in the Law School of the Boston University, and such were his powers of mind and memory, and such his knowledge of the subject, that he was able to deliver his lectures orally, and to the entire satisfaction of his hearers, upon this abstruse and difficult branch of the law. In later years, when a slight deafness began to interfere with his appearance in court, his office practice was rather increased than inter- FRANCIS PARKMAN. 435 rupted. Many cases of importance were brought before him as ref eree, or Master in Chancery, and his opinions and advice were much valued and often sought. At the age of about seventy-five his memory began to fail, but still, when once reminded of the lost facts, he would discourse as intelligently as ever upon the topic in hand; and he sometimes remarked that he had been all his life a close student, had never hunted, fished, or swum, had never drunk, snuffed, smoked, or chewed, and had been so continuously occupied with his life-work, to the neglect of needful rest and recreation, that he feared he was now to suffer the penalty of an over-tasked brain. His daily visits to his office were continued almost to the last, and he read his news- paper as usual but a day or two before his death at the quite venerable age of eighty-three.. In 1837 Mr. Paine married Miss Lucy E. Coffin of Newburyport, Massachusetts, who is said to have beeu a lady of rare accomplish- ments, and was in after life worthily active in charities and good works. She died on March 1G, 1887. They left an only daughter, Miss J. W. Paine, of 66 Sparks Street, Cambridge. In the "Bay State Monthly" for November, 1885, may be found an appreciative sketch of the life of Henry W. Paine, with an excel- lent portrait, written by his early pupil and life-long friend, Professor William Mathews, LL. D. 1894. Nathaniel Holmes. FRANCIS PARKMAN. Francis Parkman was born in Boston, on the 16th of September, 1823. The son of the Reverend Dr. Francis Parkman, an eminent Unitarian minister, and of Catharine, daughter of Nathaniel Hall, of Medford, he belonged by birth and tradition to the gentry of New England, whose right to a certain consideration and dignity was iu those days still recognized. " His childhood," to use his own words, "was neither healthful nor buoyant. His boyhood, though for a time active, was not robust, and at the age of eleven or twelve he conceived a vehement liking for pursuits a devotion to which at that time of life far oftener indicates a bodily defect than a mental superiority." The chief pursuit in question was chemistry, in which he dabbled with no good result. At fifteen or sixteen, however, his tastes took a new turn and this time a permanent. " He became enamored of the 436 FRANCIS PARKMAN. woods, — a fancy which soon gained full control over the course of the literary pursuits to which he was also addicted." In 1840 he entered Harvard College. Here, he writes, "before the end of the Sophomore year my various schemes had crystallized into a plan of writing the story of what was then known as the ' Old French War,' — that is, the war that ended in the conquest of Canada, — for here, as it seemed to me, the forest drama was more stirring and the forest stage more thronged with appropriate actors than in any other passage of our history." — This was about 1842. — "It was not till some years later," he goes on, " that I enlarged the plan to include the whole course of the American conflict between France and England, or, in other words, the history of the American forest ; for this was the light in which I regarded it. My theme fascinated me, and I was haunted with wilderness images day and night." The purpose thus conceived he always adhered to. In carrying it out, however, he met and surmounted obstacles that to almost any other man would have been fatal. In the Proceedings of the Massa- chusetts Historical Society for November, 1893,* is printed the char- acteristically impersonal autobiographic fragment from which my quo- tations are taken. Written with no purpose of confession, but rather as a pathological document of possible value to the future, it may best not be quoted in detail ; it is accessible to whoever cares to read. With the calmness of a scientific narrative, it tells the story of an obscure, almost unique malady — physical, mental, nervous, by turns — which pursued him from early youth to the end. Among the first specific symptoms was a weakness of sight, which persisted throughout life, and which rapidly grew so serious that, years afterwards, he notes with satisfaction that he can at length permit himself to read, on the average, five minutes at one time. At intervals, one of which ex- tended through four years, he found himself unable to bear the slightest mental concentration. Another trouble was a difficulty of the knees, which occasionally crippled him; and his temperament was remarkable for physical activity. In the earlier stages of his trouble he had striven to conquer it by physical exercise. This resulted in a muscular power which even his prolonged illness never destroyed. An athletic boy, who knew him in his sixty-ninth year, lately expressed an admiring hope that he might some day grow strong enough to pull a boat like Mr. Parkman. With a moral strength only shadowed by this lasting strength of muscle, he adhered to his youthful purpose through this whole lifetime of suffering. * Mass. Hist. Soc. Proceedings, Second Series, Vol. VIII. pp. 350-360. FRANCIS PARKMAN. 437 During his college course lie had to go abroad for his health. Tie returned in time to take his degree in 1844. Fur a little while he studied law. Then, his eyes failing him, he started, in search of health and of material for his history alike, on that exhausting jour- ney to the then savage West, which he has recorded in " The Oregon Trail." The physical strain served only, in the end, to confirm his trouble. The book in which he tells the story was dictated, chiefly from memory, to the kinsman who was his comrade on the journey. It was published in 1847. In 1850 he was married to Catharine Scollay, daughter of Dr. Jacob Bigelow, of Boston. In 1851 appeared the first fruit of the historical purpose which he had already cherished for nine years. In this book — "The Conspiracy of Pontiac" — is virtually sketched the whole plan of the historical work which occupied him for forty years to come. A few years later he published a novel, " Vassall Morton," which he is said subsequently to have regretted. It was produced at moments when his health forbade him to work seriously at the task he really cared for. In parts, perhaps, it was more nearly autobi- ographic than he meant it to be. A cheerful reticence about himself was one of his marked traits. " Vassall Morton," they say, he dis- liked to hear mentioned ; it has generally been forgotten. Another of his avocations, however, will always be remembered. In 1851 he bought a small place on the edge of Jamaica Pond, where he lived, for part of the year at least, until his death. Here, as his strength per- mitted, he devoted himself to horticulture, with such results that his name is almost as familiar to lovers of flowers as to lovers of books. It was here that he died. The garden he so cared for has already become a part of the great park system of Boston. In 1858 Mrs. Parkman died, leaving two daughters. In the same year he went abroad for his health. It was not until 1865 that his next book appeared, — " The Pioneers of France in the New World." From this time his power of production increased. His malady, per- haps, was beginning to relax. In 1867 came " The Jesuits in North America," in 1869 "The Discovery of the Great West," in 1874 "The Old Regime in Canada," in 1877 " Frontenac," in 1884 "Mont- calm and Wolfe," in 1802 his final book, " The Half-Century of Con- flict." Though the conflict referred to in this title is of course that between France and England for the continent of America, the title, by a happy accident, has a peculiar felicity. Just half a century had elapsed since 1842, when he first conceived the historical plan which at length he had finished. On the 8th of November, 1893, he died, after a very short illness, at Jamaica Plain. 438 FRANCIS PARKMAN. Remarkable as this mere statement of his achievement is, it by no means includes the whole story. The many journeys and researches demanded for the collection of his historical material might be inferred from the results of his historical work. Fie found time and energy also for much other activity. To name only a part of this, he was for thirteen years a Fellow of the Corporation of Harvard College, and for six years an Overseer ; he was an active member and Vice-Pres- ident of the Massachusetts Historical Society ; for its first six years he was President of the St. Botolph Club ; and he wrote not a few articles for newspapers and magazines, on matters of public import. In 1855, he became a Fellow of the American Academy. He was rarely, if ever, able to attend its meetings; but no name on its lists was more cordially honored. II. Mr. Parkman was the last and in many respects the ripest of the romantic historians who for more than half a century gave distinction to the literary life of New England. Younger than the others, surviving them all in spite of his prolonged years of illness, and doing his best work toward the end of his life, he seems to-day a far more modern ligure than Prescott, or Ticknor, or Motley. More than theirs, too, his work concerned our own country. The chief centre of his interest, from the beginning, was the frontier of that British civilization in America from which has sprung the United States. From the first lines of "The Conspiracy of Pontiac," one may say, his work tended unerringly toward the closing words of "Montcalm and Wolfe": The United States "has tamed the* savage continent, peopled the solitude, gathered wealth untold, waxed potent, imposing, redoubtable ; and now it remains for her to prove, if she can, that the rule of the masses is consistent with the highest growth of the individual ; that democracy can give the world a civilization as mature and pregnant, ideas as energetic and vitalizing, and types of manhood as lofty and strong, as anv of the systems which it boasts to supplant." This constantly national purpose, none the less profoundly patriotic that with the open sincerity which has always been the virtue most cherished by men of Harvard he disdained to neglect or to deny our errors and our dangers, makes his work peculiarly ours. The literary sensitiveness, too, with which his style changed from what now seems the somewhat excessive floridity, or at least the figurative formality, of half a century ago, to the direct, fluent simplicity of the best modern English, makes him above most men of letters steadily contemporary. FRANCIS PARKMAN. 439 Besides this, his unswerving tenacity of purpose makes his work sin- gularly complete. In the first six chapters of " The Conspiracy of Pontine/' as I have said, he sketched what may broadly be called the whole scheme of his historical writing. For forty years of enthusi- astic study, in the course of which he sought out every available authority, he busied himself in finishing, on the grand scale, the pic- ture thus sketched. In its own way, then, his work probably stands among the most permanent that has been done by American hands. Perhaps its most salient trait is its unbroken vitality. His imagina- tion was very vivid. To him men were always alive, — thinking, feel- ino-. acting, stirriiisr, in the midst of a living Nature. To him a docu- mentof whatever kind — a state paper, a Jesuit " Relation," the diary of a Provincial soldier, the record of a Yankee church — was merely the symbol of a fact which had once been as real as his own hardships among the Western Indians, or as the lifetime of physical suffering which never bent his will. In turning from "The Oregon Trail" — the single volume which records experience of his own — to the series of volumes which record the experience of men who have been dead for generations, one feels strangely little difference. Both alike are records of actual human existence. This constant vitality is generally recognized. By those who know his work well, indeed, it is by and by assumed, in a mood akin to that in which the great generalizations of human wisdom are accepted by posterity as commonplace. If much remarked, it is spoken of as notable in view of the maladies which kept him so long a cripple or an invalid. These, it is said, in no way impaired his scholarly and artistic vigor. To a great extent, the remark is true. More vivid writing than his is hard to find ; nothing could be further from what he called " the pallid and emasculate scholarship of which New England has already had too much." For all this vitality, there is an aspect of his work which thorough criticism cannot neglect. Here and there one sometimes hears from people who cordially admire his writings an occasional expression of regret that he did not devote his exceptional powers to the execution of a task in itself more important. After all, these critics say, he has only told us — incomparably, to be sure — how European mis- sionaries and pioneers penetrated and tamed the American wilderness, and how in the end the provinces that used to be French became sub- ject to the Crown of England. Full of vigorous interest as all this is, it sometimes seems — in this age of grave constitutional and philo- sophic study — just a shade puerile. Boys like to read it. Nowadays 440 FRANCIS PARKMAN. this is often reason enough for grown people to think it a bit the less worth their attention. In this criticism there is some apparent truth. Undoubtedly Mr. Parkman's first youthful purpose was, in his own words, to write " the history of the American forest," which incidentally should include the long struggle between France and England. Undoubtedly his love for the woods pervaded his fancy to the last. Undoubtedly, in comparison with much that has happened on earth, these matters seem at first glance rather picturesque than notable, of romantic interest rather than of historic. To assert that they are really so, however, is not to understand them. A little consideration reveals them in a dif- ferent light. Historical phenomena of any kind must be the result of historical forces ; and though here the historical phenomena may sometimes seem trivial, the historical forces that underlie ihem prove before long to be of prime importance. In the first place, we have European civilization inevitably, unwittingly overpowering the bar- barism of savage America. In the second place, far nearer to ourselves, we have the absolutism of the old French monarchy struggling to the death for the dominion of a conquered continent with that firmest known system of human rights, — the common law of England. All this, too, we have implied in Mr. Parkman's own pages ; to feel it so that we may philosophize about it to our hearts' content, we need only turn to him. The matters he deals with, then, are really matters grave enough for anybody. The fact that the seriousness of his work is not to all readers instantly apparent, however, is in itself significant. Here, and only here, I think, is revealed the superficiality of that commonplace criti- cism which declares that his illness in no way affected his achieve- ment. Whoever knew him, at least in his later years, must have felt that the man himself was as far removed as possible from that delight- ful but unimportant personage, the mere teller of stories. You could not talk with him for five minutes without feeling that he not only knew things, but thought about them too ; that, to a rare degree, he was a critic of life. In his historical work, however, this trait, though by no means absent if one will but search for it, is not quite obvious. In his later books, to be sure, it is more apparent than in his earlier ; implicitly, after a while, one finds it everywhere ; but to find it one must sometimes search. To a superficial reader, in fact, the luxuriant profusion of his detail — a trait which would naturally result from the circumstances under which bis illness compelled him to work — must sometimes obscure the principles which any FRANCIS PARKMAN. 441 one who knew liim knows that he constantly realized to underlie the facts and the people he has made so vividly real. A deliberate practice of his, too, gives color to the superficial criticism. He was very sparing of generalization, of philosophic comment. For this he had good reason. To philosophize with certainty of convic- tion means to think long and hard ; to philosophize flippantly means not to realize the responsibility which lies on whoever dares to leave written records behind him. This responsibility Mr. Parkman fully realized ; from the beginning of his life to the end his infirmities forbade him sustained intensity of thought. To them, I believe, and only to them, we may attribute our misfortune in that this gentlest and ripest of our historical writers has not left us books that should instantly show him beyond cpuestion the gravest of our historical thinkers. III. Mr. Parkman's personality was so marked that any memory of it is worth recording. I shall ask no further justification for telling here what I remember of him. Before I ever saw him, or ever read a line of his writing, I had heard, like any man of my age, a good deal about him. In the first place, he was one of the New England historians ; and somehow these writers, I think, were generally held by the local public opinion of their time rather more profoundly respectable than auything else on earth. In the second place, though the actual details of his illness were not generally known until the publication of his autobiography, the fact of his illness was of course apparent. I have no earlier memory than of tales about him, very properly presented to my child- ish mind in the light of a good old-fashioned example. In spite of his illness, he had written his books, and was going to write more. In spite of his lameness, he was sometimes to be seen, walking with two canes, but still with a brisk step, resting at intervals for a fresh start. Mentally and physically, then, he was indomitable. The fruit of these anecdotes in my mind was that inarticulate sentiment of awful respect so familiar to the traditional youth of New England. When I first saw him, I was still an undergraduate in Harvard College. The circumstances of this first sight vanished from my memory long ago. In this very fact there is something characteristic of the man ; locally notable as he was, and notable too in personal appearance, he was very unobtrusive in address, and in general com- pany he was little given to talk. All I can now recall of the begin- 442 FRANCIS PARKMAN. ning of the time when to me he changed from a personage to a personality is that, at certain intervals, I began to meet him here and there ; that his strongly featured, smooth-shaven, thin face looked as if made for an expression of severity ; and that there was always about his eyes and his mouth an expression of alert, kindly interest in what- ever was doing, which did away with the notion of severity altogether. Remarkably self-contained, you felt him after a while ; but, to a still more remarkable degree, not a bit self-centred. For all his firmness of aspect, and all the reputation by which this firmness was more than justified, you found yourself insensibly growing to think of him chiefly as a keenly interested observer. Alert, observing, kindly he looked, whether in a private house, or in the full dignity of the Commence- ment stage at Harvard, where he would sit with the other Fellows of the Corporation behind that impressive rail on the stage of Sanders Theatre, which in any less dissenting atmosphere would so inevitably remind one of an altar. Perhaps the most vivid memory I have of him in those days is of how he sat one Class Day evening in a doorway of Holworthy Hall. He had come to Class Day with one of his daughters ; he did not wish to hurry her home. Very likely, he was pretty well tired out; but he did not look so. He sat there, in the dim light of the lanterns, listening to the singing of the Glee Club, leaning forward a little, resting one hand on his cane, talking very little, but just watching, with his kindly, half-amused look, the swarms of young people who were passing. Somehow the memory of his figure has clung ever since to that doorway of Holworthy. It was not a bit the figure of a grave, heroic historian ; it was just that of a quiet, kindly New England gentleman. Several years later I first chanced on the historian that was in him. Meanwhile I had grown to know him better, meeting him always with the pleasure that comes to one from a cordially friendly greeting, but never getting further. In the early autumn of 1880 he was abroad, collecting material, I believe, for his last books. One day I happened to meet him in the Luxembourg Gallery ; and by mere chance to stop with him before Delaroche's familiar picture of the death of the Due de Guise. My notion of the historical circumstance therein portrayed was deplorably vague. I happened to say so. Thereupon Mr. Parkman began to tell the story with the vividness so familiar in his writings. All the essential details of time and place were at his tongue's end, — picturesque little touches, too, of what this personage and that said or wrote. It was such a bit of story-telling as now and then you hear from an eyewitness, — just as vividly real as FRANCIS PARKMAN. 443 if he had lived at the court of Henri III. If it had lain in the direct course of his professional study, it would have been remarkable ; lying, as it did, a little beyond his actual province of work, it was astonishing. You began to wonder whether, in that stirring imagina- tion of his, he had stored away all the picturesque facts of recorded history. Very likely his accurate knowledge of the Guise times was a matter of chance. It was a chance, though, which happened to be very impressive, — all the more because of the extraordinary power of oral composition which had come to him from the frequent necessity of dictating rather than actually writing his books. For several years after this, I have but two distinct memories of him, — one general, the other individual. The first is of his presidency of the St. Botolph Club in Boston. Founded early in 1880, for the purpose of bringing together men genuinely interested in literature and art, it contained in its earlier years many of the most interesting men in New England. Mr. Parkman was its first President. His health did not allow him often to attend the larger meetings of the club, which at that time occurred every Saturday evening. What one remembers, then, is the pleasant stir which would go through the room when now and then he would unexpectedly appear. He would gen- erally sit on a sofa, leaning back perhaps a shade more languidly than one commonly remembers him. He would talk, with more animation than usual, to whoever chanced to be near at hand. He would greet whoever approached with a cheeriness which made one feel as if the pleasure of meeting were mutual. Here, as everywhere else, it was the man whose presence that you felt, and not the historian. Little by little you began to wonder whether the grave, heroic personage of your youthful . tradition could possibly be identical with this alert, kindly, quietly sympathetic human being. My other distinct memory of him, in these years, is perhaps too complacently personal for record. One afternoon in the first months of my married life, when I happened to be alone at home, his card was brought to me. A moment later he came up stairs to the little library where I was sitting. He did not stay long, he said very little. I was a bit- conscious, I suppose, and I rather think there was a moment or two of conscious silence. The fact that he came in person, though, to wish God speed to a pair of young friends just beginning their mar- ried life seems to me worth remembering. Just that sort of friendly kindness is not so common in New England as to be commonplace. It was only during the last two or three summers of his life that I saw him with anything approaching intimacy. In " The Half-Century 444 FRANCIS PARKMAN. of Conflict" he mentions that he wrote at least part of his story of Sir William Pepperell's conquest of Louisbourg in the old house that in Pepperell's clay was the county seat of Penning Wentworth, Governor of New Hampshire. Then he goes on to write some pleasant descriptions of the country about the mouth of the Piscata- qua. The old Wentworth house, some few years ago, passed into the possession of Mr. Parkman's son in law. Mr. Parkman found the region pleasant and restful. He slept there more soundly than else- where. So he would come and stay there for weeks at a time ; and there as a neighbor, I saw him constantly. The old house is a large, rambling structure, absolutely without the pseudo-classic dignity and proportion which distinguish most of the architecture of Colonial New England, and for that very reason it is perhaps the most romantic, unexpected mansion which has survived here from the time of George II. It lies rather low on a point of land which separates Little Harbor — the smaller of the two outlets of the Piscataqua — from a tidal creek that runs two or three miles inland, to lose itself in salt meadows. When the tide is low, creek and harbor alike are almost dry ; but when the tide is high there is no bet- ter water anywhere for rowing and sailing ; and the constant variety of aspect which comes from this daily contrast combines with the romantic air of the old house, half hidden by lilac bushes almost as old as itself, to give the region a charm peculiarly its own. In this environment, so pleasantly fitting for the historian whose chief work dealt with the colonial times of whose memory the region is full, Mr. Parkman was seen oftenest, perhaps, in a character as remote as possible from the humdrum dignity of history. While he was rarely able to walk much, his arms and chest were generally in prime condition. On crutches, or at best with the help of a cane, he would come briskly down to the pier, where a small fleet of small boats was constantly lying ; he would seat himself in a queer, non- descript little row-boat made to suit him by a local genius ; and he would pull himself about for hours at a time, or perhaps anchor him- self and fish contentedly, as long as the tide served, for small cod or perch. His personal appearance at these moments may be inferred from an incident that amused him one day. At some little distance from home he observed a fisherman hauling in lobster-pots, and row- ing alongside to watch his luck, was much impressed by the man's stal- wart honesty in carefully measuring every lobster with a foot rule, and throwing overboard all that were not indubitably of full legal size. After a friendly interview, the couple parted company. A few days FRANCIS PARKMAN. 445 later, Mr. Parkman discovered that the exceptional stalwartness of this fisherman's honesty was at least partly due to- the fact that the man had mistaken him for an official inspector of the local fisheries. When one met Mr. Parkman thus taking his ease, one grew aware of a certain boyish freshness of feeling and nature in him. If he caught sight of you on the shore, or in a passing boat, he would wave his hand with a jolly sort of greeting, or perhaps, if you were near enough, would shout a friendly word or two. As I have said before, too, the vigor with which he would send his boat through the water, paying scant respect to the swift tidal currents of the Piscataqua, won the instant, lasting admiration of athletic boys. You felt instinctively that the man was enjoying this simple open-air pleasure as keenly as if he were a child of ten ; that the mere fun of pulling himself about the rockbound little harbor, and of playing with the far from sportive fish still to be caught there, was enough to make the days when he could do it worth living. Or perhaps you would yourself be rowing past the old Wentworth place, and would find him sitting on the pier, with the lilacs, and the great chimneys with their wilderness of rambling roof, behind him. You would stop to pass the time of day, as the saying is, to talk for a few minutes of whatever might turn up. As likely as not, the subject would be the last new novel or story that was really amusing. For Mr. Parkman liked to be amused, and found few things more amusing than a good, rattling story read aloud to him. Here, without any affectation of literary doctrine, his taste was romantic. If I remember rightly, he had small patience with that considerable body of modern fiction which gravely claims the right to bore you. If he had ever seemed self-conscious enough to warrant such a surmise, you might sometimes have ventured to wonder whether the consciousness that, as a serious historian, he had never presumed to bore anybody, might not have whetted his indignation at solemn scribblers of pretentious make-believe. Such an idea, though, could never have occurred to you in his presence. The normal impersonality, the animated ob- jectivity of his talk, the frank, idiomatic raciness of his phrase, the wholesomeness of his nature, made you forget that he had ever written anything. You thought of him, by and by, just as a remarkably friendly human being. You forgot even that he was not exactly of an age with you. Like his own literary style, which kept pace so sensitively with the best literary feeling of his day, the man himself was steadily contemporary. There were times, too, far fewer than one would wish, when one saw him even more intimately. In the midst of all this vigorous 446 FRANCIS PARKMAN. open-air life, he was really an invalid still. If you happened to call at the house when he was staying there, the chances were that he would not appear. Sometimes, doubtless, he was at work. Oftener I take it, he was resting, or struggling with the malady which so often made even the excitement of meeting anybody the source of acute suffering. It is only since one has read his autobiography, however, that one can realize what his suffering was ; in real life one never caught a glimpse of it. Either one did not see him at all, and inferred that he was a bit indisposed ; or perhaps one found him very silent ; beyond this, there was nothing to suggest that he was not as strong, as well, as the best of us. Now and then, however, — generally when his lameness bothered him more than usual, — he liked you to come and sit with him a little while. I saw him thus alone perhaps half a dozen times. The last time of all I remember best. It was a few weeks before his seventieth birthday. I had been taking tea near by, and, hearing that he was laid up for the moment, ventured over to see him while I smoked my cigar. I found him quite by himself, and seemingly glad of a visitor ; for,* in the growing twilight of that August evening, he talked more than I ever remember to have heard him talk at once before. As always, in his talk with me, he had something to say about Harvard. Like any man of his time and traditions, he had more doubts than a few concerning the quality of youth which is now grow- ing up even in that most respectable institution of learning. Just at that moment, however, he was full of interest in a new book of college stories, which told most sympathetically of Harvard life as a con- temporary undergraduate sees it; and to Mr. Parkman's mind — as well as to mine — they had two very reassuring merits : they could not possibly bore you, and they could not possibly have been written by anybody but a gentleman. So our last words about Harvard were cheerful ones: however languishing the finer traditions sometimes seem, they show themselves there, ever and again, as freshly vital as ever. Then, somehow, the talk turned upon the Puritans. Some little time before, I had written a short life of Cotton Mather. Little as I had found in Mather's faith or practice which I could literally or actually share, I had found in the strength of his conservative enthu- siasm something which commanded my heartiest sympathy. Now what in the end I found most sympathetic in Mr. Parkman him- self was the enthusiastic strength of his conservative feelings. No one ever seemed to me more heartily to hate the folly of abstractions, more prudently to dread the sacrifice of what we know, for all its FRANCIS PARKMAN. 447 faults and errors, to be present good, for the sake of conceivable, but in no wise certain, future benefit. The trait I have in mind, which pervaded his serious talk, animates his well known writings about Woman's Suffrage. When it came to talking of the Mathers, however, I found, as I had found before, that he could not in any way share my sympathy for the conquered orthodoxy of New England. Their narrowness, their pettiness, their limits, their tyranny, and above all their absurdities, were to him almost the sum of their character. Sewall's Diary either amused or provoked him. What seemed to me the noble side of it — the passionate eagerness to preserve unaltered what the Puritans believed to be their divinely sanctioned system of faith and government — Mr. Parkman could not quite appreciate or care for. He did not like the Puritans ; and what he did not like he could fervently condemn. He could not feel, as I felt, that what I liked best in him — his wholesome conservatism of impulse — differed from what I liked in them only as the nineteenth century differs from the seventeenth. In truth, I take it, this instinctive dislike of Puri- tan dogma and character marked him as a man of an older generation than at first he seemed. He was old enough to remember the days wrhen, in New England, orthodox Calvinistic bigotry was actually dangerous to the freedom of thought in which both he and I believed. So to him the Puritans were once for all what Secessionists are to tho surviving Republicans of the early days. One more passage in the talk of that evening I remember. Speak- ing somehow about the time of the Rebellion, he said very simply that to his mind one of the chiefly deplorable things about our civil war was that it had disturbed, had dimmed, the concrete ideal of character and conduct traditionally kept before the youth of America. Up to the war, the one great ideal figure had always been the figure of Washington, — a figure alike worthy for its moral dignity and for its personal. Since the war, he said, we have heard too little of Washington ; that is almost the worst thing the war has done for us. For the less, as Americans, we think of Washington, the worse for America. So he talked on ; and by and by I took leave of him in the dusk. I never saw him again. Too personal for record these slight remembrances may seem. Yet, if they serve to tell how the solemnly traditional personage called " Parkman" could merge insensibly into the friend whose loss came as a personal grief, they may perhaps serve, in time to come, better to preserve his memory than if they had been left unwritten. 1894. Barrett Wexdell. 448 HENRY WARREN TORREY. HENRY WARREN TORREY Henry Warren Torrey, Professor Emeritus of Ancient and Modern History in the University at Cambridge, was born in Roxbury, Massachusetts, November 11, 1814. He was the son of John Torrey and Marcia Otis, daughter of Henry and Mary War- ren of Plymouth. He was educated in the Adams Grammar School in Boston, where he received a Franklin Medal at ten years old, losing his father in the same year (1824). He was a pupil of the Boston Latin School under Master B. A. Gould, completing the five years' course in four years, from 1825 to 1829, and gradu- ating from Harvard College in 1833, where he stood second in rank, the late Professor Bowen being first. He was usher at the Latin School under Master F. P. Leverett, whom he assisted greatly in the compilation of his Latin Lexicon. The English-Latin part of this book was entirely prepared by him, and the part of it from A to C was an original work, where Mr. Torrey had in sober earnest attempted that almost impossible task, the preparation of a book really worth the name, to assist in translating good English into good Latin. At this stage of the work he was forced to desist, and for the remainder to confine himself to a recasting of "Ainsworth," a task singularly uncon- genial to Mr. Torrey, whose fine mind was one of the first to recog- nize the worthlessness of Ainsworth's Dictionary, — a work to be recast only by being cast into the fire. His service on the two Lexicons had practically ruined his eyes; but this loss did not deter him from studying law at New Bedford in the office of his accomplished and ever genial uncle, the Hon. Charles Henry Warren. He was admitted to the bar, but never practised, on account of his eyes. He kept school in Providence for a year and a half, and made a voyage to the Azores with Hon. William W. Swain. This was one of the many vain efforts of that warm-hearted and excellent man to restore health to his son Robert, whose lovely character, tried in the furnace of scarcely intermitted suffering, only seemed to approach nearer to the an- gels with every hour that brought him to his end at the age of twenty-one. Mr. Torrey was again instructor in the Latin School in 1842, and Tutor in History and Instructor in Elocution at Harvard Col- lege from 1844 to 1848. For the next eight years he kept a girls' school at No. 5 Hamilton Place in Boston, with his sister, which HENRY WARREN TORREY. 449 attained the highest reputation. In 1856 the McLean Professor- ship of Ancient and Modern History, for many years vacant through the meanest of political intrigues, was conferred on him. He immediately took a year of European travel, and returned to discharge the duties of his professorship in 1857. He was engaged in its active labors till 1866, and was then appointed Professor Emeritus. While holding this title he was chosen on the Board of Overseers, and died on the anniversary of the death of Wash- ington, December 14, 1893. He was elected to the American Academy, November 12, 1856. Mr. Torrey's active life was that of a teacher. It is of little moment to speculate whether he might not have distinguished himself in some other line. He was emphatically an instructor, and not a few of his pupils are 3ret living who feel that he was the best instructor they ever had, — the one to whom they owe most. He was a teacher according to the old fashion; that is, being a scholar and a student himself, he expected his pupils to be scholars and students on the lines he laid down for them. He impressed upon all of them the idea that he who will nut work cannot learn ; and that both working and learning must be done in the line of duty, — that of subjecting one's own mind and energy to the control of those in rightful authority. He never swerved from this rule, yet he administered it with such unvarj'ing sym- pathy and kindness that no pupil who was worth teaching ever felt any sense of coercion or harsh pressure, or was other than stimulated and braced by his kindly insistence. As soon as his pupils had done their work his began. He supplied to them an amount of knowledge, not only on the specific subjects they were studying, but on a score of others, which they never could have acquired without repeating his own indefatigable labor. From first to last this knowledge wras unerringly accurate. A mistake of fact was to Professor Torrev simply an untruth ; and to state a name or a date incorrectly was to falsify. Whatever was not right was wrong. But he did not end with facts. He combined his vast stock of information into an organic whole by the spirit of a philosophy which soared far above the mere earth of accumulation which makes up so much of human learning. The effect of this con- structive power on his pupils was startling, if such a word can be applied to so precise and gentle a character. A student might come before him, fancying he knew a good deal of history for a vol xxix. (n. s. xxi.) 29 450 HENRY WARREN TORREY. young man; he might have studied it, read it, or even, as a boy does, written it; then would Mr. Torrey reveal to him, not only how little all this information was, but how infinitely greater and better in kind was the Philosophy of History, — the relation of nation to nation, of period to period, and of all to the mighty scheme of Providence which he saw working through all time. Yet in all this setting forth the great subject in its true perspec- tive, and making the amount of his pupils' knowledge appear very small, there was absolutely nothing unkind, no undervaluing a single real acquirement. On the contrary, he made the value of every pupil's attainments far greater than it would have been had they gone on without him in their own boyish way and spirit. This power of lighting up the specific subjects that he treated was nowhere better shown than in his instruction on the Constitu- tion of the United States. Not a few of his pupils have been called upon to serve in positions of authority under that august instrument ; and none of them would deny that they received from his instruc- tions an understanding of their practical duties which no other teacher, public or private, could give. But more than one teacher is learned and stimulating, and yet is entirely without two things which were the very life and soul of Mr. Torrey's instruction : first, his affectionate heart, which went out to every pupil who would accept it as to a personal friend; and, secondly, his supreme allegiance to duty, wherewith he tried, in that absolute simplicity which was his nature, to inspire his pupils, as might an elder brother who knew that if they did not follow the way of right they never could accomplish anything worth doing. That nobody could really be a good scholar, or a good historian, without being a good man, was the core of all his teaching; and it was worth while to be under him, if only to learn this one thing. Such was the man in the class-room, clinging to the old- fashioned discipline and methods, which he might well think the best, because his own glowing personality had warmed them into a vital force that needed no further kindling from modern devices. There was a certain veil of formality about his pres- ence which might at first deter one from penetrating into a more intimate knowledge. Yet something about his ready smile and his eager manner, never cold though always courteous, encouraged his pupils to bring to him personal requests and difficulties, which were invariably received half-way, and proved the ready key to HENRY WARREN TORRE Y. 451 outside acquaintance. The threshold once passed, it seemed amaz- ing that one ever could have considered him formal. No heartier or more genial friendship ever glowed in a human breast. He had been brought up under one of the most brilliant wits of his day. the late Judge Warren, already named; and an inherited playful- ness, softened by an ever-present sense of propriety and modesty, made his conversation inexpressibly delightful. It was so to all; but it was most peculiarly so to such as having been his pupils became after an interval his colleagues in instruction. To them his house and his heart were open in a combination of the elder and the equal hard to describe. There was nothing that mind or heart desired to fill the void caused by exhausting daily work, to heal the blows received from daily stupidity, waywardness, and ingratitude, which a younger teacher could not find in his com- pany,— sympathizing with one's annoyances, confirming one's pur- poses, elevating one from depressing doubts. It had been much to sit at his feet and obey him; it was infinitely more to sit by his side and love him; and the world is darker since his presence lias been withdrawn. .Having given years of such public and private work to the College, he saw with pain that methods and aims with which he found it hard to sympathize were in the ascendant; but there was no want of loyalty in the efforts which he made to accept them; and though he distrusted seriously the indefinite extension of an elect- ive system, that s}rstem received no better assistance than he gave it in his advanced courses. The best of the old was equal to the best of the new. It was touching to see how, on not a few occa- sions, his faultless pen, from which flowed the purest of old classi- cal English, was employed to give adequate form to matter which certainly did not originate with his suggestions. His nature was sensitive; but it was not the sensitiveness of a morbid or selfish constitution, — it was the sensitiveness of his Pilgrim ancestors, — the sensitiveness of a conscience which loved and courted the gaze of Heaven, and shrank from every transgres- sion of absolute right as an insult to the God of truth. His un- tiring industry, his capacious and penetrating intellect, his ardent energy, were all directed and chastened by his ever living sense of Christian duty; and his work, which may pass unchallenged the severest tests of practical humanity, was devised, carried on, and completed in strictest view of the higher scrutiny which belongs to an eternal standard and an unseen Judge. 1894. William Everett. 452 MARQUIS OF CALIGNY. FOREIGN HONORARY MEMBERS. THE MARQUIS OF CALIGNY. An ancient family, which since 1660 had produced a succession of military engineers, became extinct by the death of the Marquis of Caligny ; several of its members were esteemed by Vauban, and their works are yet well known to civil and military engineers. Anatole Francois Hue, Marquis de Caligny, was born at Valognes (Manche), May 31, 1811. The son of Bernard-Henri-Louis Hue, Marquis de Caligny, and Eugenie-Marie-Lconore-Avice de Fer- manville. He was related to many of the illustrious families of France, and popular tradition affirmed that his descent could be traced from Charlemagne ; however this may be, it is certain that he was the last of a distinguished group of hydraulic engineers, among whom may be mentioned Louis Roland de Caligny, founder of the port of Cherbourg. De Caligny entered the College of St. L6 in 1822, where he first met Leverrier, who became his intimate friend. This friendship, based on a similarity of tastes, remained constant and unalterable, in spite of the great astronomer's changes of opinion. Caligny remained firmly Legitimist and Catholic, for him the political changes subse- quent to 1789 were as if they did not exist ; with this difference, the two were as brothers. At the age of sixteen, he began the study of philosophy at Valognes, in order to be near his family, who resided in the Chateau de Flottemanville, a league from that town. The next year he took up his residence at the Chateau, where he conceived the first ideas of his inventions. Following the advice of the Cherbourg engineers, he removed to Paris in 1836. Here he recommenced his experimental investigations, which were to be pursued with little interruption for half a century. He soon made the acquaintance of Coriolis, who, appreciating his researches, advised him to embody them in a report to the Insti- tute. In conformity with this suggestion, he prepared a memoir on " Hydraulic Machines with Oscillating Liquid Columns," which was presented to the Academy and referred to a committee consisting of MM. Poncelet,* Coriolis, Gambey, and Seguier, who reported so * Poncelet was his predecessor in the American Academy. MARQUIS OF CALIGNY. 453 favorably upon it, that the Academy awarded him the Montyon Prize* in Mechanics, which had just been established. CaJigny's in- vention is thus described in the report of the committee: " The happy idea which distinguishes this machine, and characterizes it as a veritable invention, is the discharge through a vertical tube, after an upward oscillation, without the loss of any living force except that consumed by friction ; that is to say, by depressing very slightly the centre of gravity of the fluid column to be discharged. Without doubt the machiue appears simple both in conception and construction, but this simplicity only enhances its merits." From this time Caligny's success was assured. He was the dis- coverer of a new branch of hydraulics, namely, that relating to oscillating liquids ; and he published from time to time accounts of his researches, which were remarkable both for the methods employed and the results obtained. To these researches he devoted his life, never attaching himself to any administration, nor taking part in any private enterprise; he even refused to take patents for his inventions, prefer- ring to devote himself with indefatigable zeal to the disinterested pursuit of science. Ou several occasions his rights of priority were disputed, but they were always easily established by means of his publications in the transactions of the French Academy. His investigations were brought together in 1883, in two volumes entitled, " Theoretical and Exper- imental Researches on the Oscillations of Water and Hydraulic Machines with Liquid Oscillating Columns." Prizes were awarded to his machines at all the International Exhibitions, and he received numerous diplomas from different European Academies. The following succinct account of his first apparatus is thus given by M. Boussinesq : — " From the bottom of a reservoir containing water at rest, a long horizontal or slightly inclined conduit issues, having its further end terminated by a vertical pipe of the same diameter as the conduit, and rising some distance above the level of the reservoir. It is proposed to raise and discharge the water from the top of this pipe. For this purpose, a valve movable around a horizontal axis, manipulated from without, separates the empty vertical pipe from the conduit filled with water under pressure. The valve, being at a given moment released, is raised by the water which begins to rise in the pipe, and that in the conduit also, but only little by little, on account of its great mass. * This prize consists of a gold medal valued at 500 francs. 454 MARQUIS OF CALIGNY. Its living force attains its maximum at the instant when the liquid in the pipe rises to the reservoir level. This living force is capable, except for the slight loss due to friction, of carrying the ascending liquid column as high above the reservoir level as its point of depart- ure was below it. As only a part of the force has been consumed at the moment when the liquid begins to pour out above, the discharge continues until the whole column in motion has been brought to rest. But at this moment the valve falls by its own weight, and opens at the same time a short horizontal tube which it closed while it was raised, thus offering a free exit for all the water in the vertical pipe, which thus rapidly escapes into a second reservoir just above the level of the valve. The liquid contained in the horizontal conduit during this time comes to rest, and then begins to oscillate by raising the valve ; then a new period of asceusiou with discharge begins." If Caligny's experiments had not been realized for half a century, might not this description have been taken as a project for perpetual motion ? In 1868, the Administration of Roads and Bridges built, on a work- in