AH tew 0 Bete aed aw < saat os ced ons OD dee Met Ot end bee “Nea +O Ae Oe ee ae ho Ge “a ay Oe ee ee Oe ee “ Pe oe | 1H ded He Come er — -4 we an Se WeGetrion a ote eb Se ok eet cg Le ARG Se Gehetideu-ai ter dred in Fae 8 hth ye end Gr Va Ce ae ee ee et ery f « 2 hee. he bette 4, getty Magee te am ‘he thy ging) tL ete deg in Ho Ws ogee eceoae 6: te Lea tr, mel bcd ny ar Rrtet He ae wedew He etek Ga et q he RO ee adele dee a V1 eM Or eas oh ato thnks oth rite olds o) Heth Chee tae Ah AR CE A ae EU Man Mote ak? ea He teat (hae Hy fn tan go SE gece) hp nies Wome g “ag ( ths be tga shoe Li Tle eh aoe {hotetto: Wt Hes hyd i Weis dgetd saoth tron woah he vie oe Rte Wy Hh aw the de tg ee Metis Os ti 00s on ter ; Mote ar Metemtige rhs (4 Uh dettom ap: a re ‘ane tet. Lriphind ‘ WC Tew ed haay Wd Wands of Bata Eo bats HM Ra nls i A de HRA A Wee : Wah chat = er ie GUL Hi Hanslins mates) EE Rea eet Ye aig dene * Hed ded thr hd ithe FF ty rahe Wo i ‘Or teks ts Sane : AE ER A Fra Ws dea Hemet ehe Mohan A oth th soph ate neh 4 Wt li WR oh Garton AL Herb ote it 4 AMER Ata A ag ee ASH Wei ct Gr ate Sih arta gts ae ebay f DOE I ARE RCA EX ta aa bath ee Hatha Hi hth what Harvie thd Wace Stata a tees, a hece mn sete * a true ‘ i) 2a’ " i ate ead ee Selick, Be Sale BORK Me ey Cade Wal a A | +0 ae Gg ata” 4+ % ‘ ed ok Tee | hone lets pee on Aarne bit “ taiteWaliette pels otek " Hs fe Ha Ube ete ee ie, ety OAM Qed et ig ead * Miho W Bw dade a Rohde Neecdcanene ead HM herds ertingl aig 4 &a th a bee yy Um.) 4 ee el Sele +} te bak mi sae ath Lda reat, Cw al teed a ask Hh ane FAA, Be ON Uta ‘ AR ae Bohl We na eg hi, Gah Ht Gone oh tee Wate Gare ip a 0 het UM eras dsGd arts Thew of hell a Re TR eld Biel ae Sch Fy WAR AISED A : i] ra @ aa ae bee ry isi , Ws HES no ie ely ae oe. GAS Srfeliities teins MENA Uh eeciansth cd SAM oe at Ness Wediws Lseall Ahad ded Ane Hite ey Eda weld gs, EAA Se qt wee Hs SME OE aeleontyld eee Mose d aon, gre ae eae Heat: ko deh if ade ew) bib nt Mtl ot glee = / th thi * +. hea hee # \ o mara Pres deh tet: eae 8 Hor tf in debi iA ¢ awl Cb rrt Baek ent oth, eee 4 Hn : véti t PMT beth yy AY Ree Me He NS Wag dat hed ett date f AL oth Coe ber Oy aa feeb nue Woe Oh ees the ty hs 9 # aot! Pts Neen Ge 42 Hs 94 Seah ee thee STC ce ea an Sree iS ~ae Weis Orie be itae cere a ‘ oh ke 4 Nhe ntee’ wae teley AP Seeey Bt prs Cy Se te Maihere dhe ta Wy Wnt niet eats ‘ fa Hod seh 8 Pel rq ; } Chea. SPA a thrid 8 Bane | Wenn toh aoe 4 Sita Wee dee a 4 THe Gea aw j Eaten othe Hiatt CHB dealt ne dete, + wie takes) eo sikh ae FA A Pact y CAA Wee ers Pitot home Ae \tet 8 etd gs Ge gs ewe ay t tae ‘ ‘ Sogn + “4 — bd 44 LEY eee ed ergy ok) PL beet Rapes eC eek SV ee te Ey ae aed i Sans i atta bono he * ¥ ‘ Catal Wat 27 Vo raie ps hey kta hte eet Bi te tanh W rs He ‘ WHanane ane iY aia basi Kee i ¢ rat 4 WHEE 4 t H Hed 4-4 Na ok Yow AA heb a as ba wt 98 J he +4 UVR e's vs a hey easy “ ee tina ‘ Lit havea j cy sak Sab aa eater sisep are Met A v ns 4) er utata ay nc a” Heth WANE HG Wei tite sf eh f' LWW Aton 3 YW me ae Tits ty « € rH Porter. [ Nov. 9, and sulphur in the proportions necessary to the sulphides FeS, Fe.Ss, and FeS:. In each experiment two tubes of the mixture were made, one for the magnet, and one to be fired at a distance from the magnet ; this was done to find out whether there were any persistent differences, other than magnetic, between the two. The mixture for FeS: refused to burn in both the tubes, as indeed might have been expected. The polarity of the Fe.S3; bar was decided, that of the FeS distinct, though very weak. In every case the magnetised bars have remained unmis- takable magnets, but this may be due to part of the soft iron remaining uncombined, or possibly to the formation of small quantities of magnetite from the air between the particles of the powder at starting. The bars were powdered very finely, and their specific gravity taken, with the results shown in the table given in the addenda. It is scarcely necessary to point out that though there is no doubt that one or more of the compounds FS, FesS,, and FeS3 was really formed, we do not know the proportion of each present in each case. Probably the products were purer in the first and last cases. The specific gravities were taken with all care, and the first two places of decimals are reliable; the powders were thoroughly well mixed, and the experiments were made from the same samples of mixture, in similar tubes. The writer hopes to repeat them, and will not comment at present on the differences of density in the magnetised and unmagnetised products. Similar experiments were next tried with the magnetic oxide of iron Fes3Q,. (a) By allowing a considerable quantity of the levigated iron to form a thick chain between the solid conical poles of the electro- magnet, and pressing them together until they formed a dense mass between the poles, and then burning them in szu, in one experiment with the flame of an ordinary Bunsen burner, and in another by that of an oxy-coal gas flame, blowing on the still very hot mass with pure oxygen to complete the oxidation as far as possible; in both cases the resulting mass of magnetic oxide was of very marked polarity. (>) By placing between the poles of the electro-magnet, so that its upper surface was in the strongest part of the field, a charcoal block, having cut in it, parallel to the lines of magnetic force, a shallow groove, which was filled with previously prepared magnetic oxide, and whilst the current traversed the magnet coils, fusing these fragments into one bar, by means of the oxy-gas blow-pipe, afterwards blowing on the still white hot mass with pure oxygen. On cooling the mass and testing it, it proved distinctly polar, but far less so than in the experiments (a). In this experiment the oxide had been heated to a brilliant whiteness, and cooled somewhat quickly by being wetted with water. These experiments seem to prove that it is possible to prepare in 1903.] Some Hxperiments in Magnetism. 9 this way, and quickly, artificial lodestones of some strength, for the products of the (a) experiments were able to pick up small pieces of soft iron. Thinking that in these cases part, at any rate, of the polarity might be due to particles of the levigated iron which had escaped combination with either the sulphur or the oxygen, the writer next attempted to make a magnet out of the particles of levigated iron themselves. This was done by half filling a small test tube with solid paraffin, melting it, loosely corking the tube, and placing it in a horizontal position until the paraffin had solidified. The remaining air space, of semicircular section, was then filled with the levigated iron, the tube again corked, and placed horizontally through the bored poles of the electro-magnet ; the current was switched on, and whilst the tube was in position, with the iron uppermost in the tube, the paraffin below it was melted by cautiously heating it with a Bunsen. As the paraffin melted, the tube being gently tapped meanwhile, the iron particles were for the time capable of small movements, arranging themselves along the lines of force. The paraffin was then allowed to solidity and grow cool, the current being maintained the while, the tube removed and the glass gently peeled off, by first cutting it round with a diamond, leaving a bar of soft iron particles embedded in paraffin, which was a magnet, and has remained so ever since, without showing any signs of losing its magnetism, and sufficiently strong to develop tufts at either pole when dipped into a heap of the levigated iron. As it still seemed possible that the permanent polarity of this magnet might be due to carbon in it, which it certainly contained, though in very small quantity, it seemed well to try a similar experiment with particles of pure electrolytic iron, and to try it with molecules of the iron instead of assemblages of molecules, already under the influence of forces opposing the development of the magnetic polarity, and these considerations at once suggested that the iron should be deposited electrolytically from the solution of one of its compounds in a powerful magnetic field on some “non-magnetic” substance, such as platinum. The photograph shows the arrangement actually used, and, which it may be stated at once, proved entirely successful. The apparatus consisted of the electro-magnet, with two special pole pieces, designed to concentrate the magnetic lines of force into a space of length, the distance between them; of height, the height of the pole pieces as shown, and of thickness one-eighth of an inch. The poles thus tapered to what may be called a blunt vertical line. Between the poles is the glass cell, in the centre of which hangs the strip of platinum foil, previously proved to have no visible effect upon a magnet, so that its plane is in the para- magnetic position, the platinum approaching close to the sides of the 10 Mr. T. C. Porter. [ Nov. 9, glass vessel, but not touching them. Two pieces of platinum foil, about 3 inches long and # inch wide, hang at each end of the glass vessel, facing the central strip on both its sides; the central strip is the cathode, on which the iron is deposited, the other two being the - anodes. The current having been switched on to the coils, and the single Groves’ celi used, to the electrodes, the iron solution is poured in, till it reaches as high as the top of the pole pieces. After three minutes there was a discoloration of the platinum from the very thin film of iron formed upon it, and on testing this, by withdrawing it from the solution and applying it edgeways to a rather heavy compass needle, by no means delicately mounted, it was found to give distinct evidence of polarity, the repulsion and attraction being most clearly visible, and on tearing off a strip of the platinum along its base, and drying it and balancing it on a fine sewing needle, it quickly set itself in the magnetic meridian, and has continued to do so with apparently equal activity ever since, although the film of iron must be exceedingly thin. The main portion of the strip was replaced in the solution, and the deposition of the iron continued, its polarity being tested from time to time. Its magnetic intensity increased, though certainly at a diminishing rate. Finally, when all the iron in the solution had been deposited the platinum was taken out, washed with distilled water, the part covered with iron torn off (so that 1t did not touch any other iron). and put into dry air for preservation. It has remained a magnet ever since, and as yet gives no sign of diminished intensity, though it has been made to vibrate, and hit with a wooden rod repeatedly. The whole experiment has been repeated, with the same results.* If during the deposition the platinum be reversed, side for side, after a certain time the iron deposited shows no polarity, and after that its polarity is reversed. Whether this is due to the reversal of polarity in the molecules first deposited, as well as to the deposition upon them of molecules having opposite polarity, cr to the latter cause only, the writer has not yet determined, nor has it been possible in this paper to give measurements either of the magnetic intensities, or of the thickness of the thinnest film of pure electrolytic iron which is capable of making its polarity felt; nor again is it yet possible to say how long the polarity of these artificial magnets will continue. Answers to these questions, by the very nature of the investigation, must be left for a future communication. The solution from which the iron was deposited is constituted as follows :— 20 grammes FeSO,.(NH4)2S04.6H20, 64 grammes (NH4)20204.H20. Made up to 960 ¢.c. with distilled water. * The writer was unaware at the time that a similar experiment had been made by Beetz, and quoted by Maxwell. 1903.] Some Hxperiments in Magnetism. i This solution is used for the quantitative estimation of iron by electrolysis, and though carbon is present in the oxalate, the deposited iron is soluble in dilute hydrogen chloride, without any visible residue, being pure to all chemical tests, but for possibly occluded hydrogen. ADDENDA: November 24, 1903. (1). It should be stated that when the crystals growing in a powerful magnetic field, and apparently orientated as described and shown in the photograph, become large, and thus approach each others’ surfaces, they are often seen to leave their first position and to take up another; sometimes the crystal will be moved bodily nearer its neighbour, sometimes it is swung round only; but it follows that where the crystals are numerous and close together, the signs of orientation, which were evident during the early stages of their growth, have almost, if not quite, disappeared at a later stage. (2). Some experiments were carried out early in the investigation in weak magnetic fields, and also in a field of as nearly zero intensity as could be managed, by compensating the earth’s field, but no difference between these could be detected with any certainty, though in most cases as many as six crystallisations were carried on at one time, three in the strong field and three in a weak or zero field. (3). When it was found that the direction in which the glass slides used had been rubbed (before they were finally cleaned with acid, alkali, and distilled water) was not in all cases negligible, experiments were made in which the glass had been purposely rubbed hard in different directions, with the result already stated, but in no case was the orientation or arrangement of the crystals the same in appearance as that shown in the photograph, for they formed close together along the lines. The strongest argument, in the writer’s opinion, against a magnetic explanation of the orientation in the photograph, is his repeated failure to obtain the same result under what were apparently precisely similar circumstances, and also the fact that no other substance has given similar indications, at any rate, such evident indications of orientation, though very many have been tried. (4). Since the paper was sent in, a number of experiments have been made on the densities of the substances produced by “ firing” the small “squibs” of the mixtures of iron and sulphur described;in the paper, with the results embodied in the following statement. The quantities taken were in most cases between 4 and 2°5 grammes, and the weighings were certainly correct to milligrammes ; the greatest error possible in any single estimation of density does not exceed + 0°02, and is probably less than 0-005. All the bars fired in the strong magnetic field are magnets, those in the earth’s field are not perceptibly so, though they may possess very weak polarity. 12 Some Hxperiments in Magnetism. [ Nov. 9, Mixture in the proportion of Fe+S. - 8Fe+4s. 2Fe + 38. Magnet. Away from | Magnet. Away from Magnet. Away from magnet. magnet. magnet. oe sf ae eae Fea 1258 eee | Uae 4°455 4°593 | 4:°229 4. 364 4 °358 4179 4°581 4-710 4.°519 4.°301 4-443 4.565 4.°763 4°579 | 4°394 4, °4.20 4°377 4.°506 4-648 4 -407* 4 -703* 4.°653 } | Mean of 8 | Mean of 5 | Mean of 3 | Mean of 3 | Mean of 3 | Mean of 3 experiments.|experiments./experiments.|experiments. experiments experiments. 4600 | 4°637 4381 4 *362 4, -393 4417 Two experiments were made to find out how the density varied in any one bar. The mixtures used were the second and third of the above three ; each was placed in a tube more than twice the length of those used in the experiments already described, and both were fired separately in the strong magnetic field, the tubes passing right through the bored poles of the magnet, the distance between the poles being linch. The rod formed was broken into three portions, called the top, middle, and bottom respectively. The results were as follows :— 3Fe+45 ... top, 4384; middle, 4°533; bottom, 4:305; mean, 4:407. 2Fe+ 3s ... » 4664 4:718 ; 2 4728 3 5 eae My thanks are due to Mr. R. W. Kennedy for help in the deter- mination of the densities. * The tube used in these experiments was much longer than in the others. 1903.] Stress and Strain in the Cross-section of a Beam. 13 “On the Distribution of Stress and Strain in the Cross-section of a Beam.” By JoHN Morrow, M.Sc. (Vict.), Lecturer in Engineering, University College, Bristol. Communicated by Professor HELE-SHAW, F.R.S. Received October 27,—Read November 26, 1903. Introduction. Our knowledge of the strains produced in materials by different kinds and combinations of stress rests mainly on theoretical con- siderations. Much accurate experimental work has been done in the observation of direct tensile strains, but little attention has been given to the lateral strain accompanying a simple tensile or com- pressive strain, or to the lateral strain occurring in a bar under bending forces. The latter, indeed, has, perhaps, never before been measured in metal specimens. : For some time past the writer has been making experiments with the object of showing that instruments can be constructed capable of measuring these lateral strains with considerable accuracy. The subject is of great interest to the elastician, as it not only provides a method of determining elastic coefficients, but shows the degree of applicability of mathematical results; and, further, in view of the well-known discrepancies which exist between experiment and theory, the subject is also of no small practical importance. The work here described is confined mainly to experiments on iron beams, and has had for its objects the following :— 1. The design of a comparatively simple instrument which can be used for the accurate measurement of the lateral displacements in the section of a beam ; 2. The determination of the amount and distribution of this strain In iron beams ; and 3. The determination of Poisson’s ratio from the observed lateral displacements. Description of the Apparatus. Whilst making some experiments on the lateral constraction of tie- bars,* it occurred to the author that, with a modified form of the apparatus then used, measurements of the deformation of the section of a beam might be made. After trials of several different forms of instrument, and various ways of applying the load, it was decided that a long bar should be placed in the testing machine, and the readings taken with the mirror apparatus shown in fig. 1. The specimen S is in section * ‘Phil. Mag.,’ Sixth Series, vol. 6, No. 34, p. 417. 14 Mr. J. Morrow. On the Distribution of | (Oct. 27, in the elevation. It is gripped by the two set-screws AA, each of which is screwed through one of the vertical rods C and D. These rods pass respectively through the pieces E and F, and are rigidly attached to them. CE and DF thus form a pair of levers pivoted together at JJ. Any motion of AA, therefore, will be transmitted to the extreme ends of E and F. Thus if the specimen contract, AA Fre. 1. Q Y « y a yi yf a Ean) (i mm | 4 LV ld : a 2) Lh G / | J ELEVATION | g Sm} iD Ti WME 3 ) { yt CO) oS, “G PLAN : Oy approach one another, and E will rise relatively to F; the relative motion of E and F will be a measure of the contraction of the specimen. To obtain the amount of this relative motion, two mirrors, M and N, are used. ‘They are made of specially prepared optically plane glass. The former is supported on three hard steel needle points, two of which rest on F, and the third on E, as shown in the plan. The other 1903.] Stress and Strain in the Cross-section of a Beam. 15 mirror, N, is attached to F on a vertical spindle, about which it can be turned by the fine adjustment screw G. When the instrument is set up, a scale is placed in an upright position in front of the mirrors, and some distance away, and the two images of the scale are brought together in the field of a telescope by means of the screw G. A convenient mark on the N image is then taken as an index, and the reading coinciding with it on the other image is noted each time the load on the beam is altered. From the difference between the readings before and after applying a load, the transverse strain is calculated across the breadth of the beam. The instrument may be placed either below the specimen, when the screws AA are used, or above, when BB perform exactly the same functions. The pressure between the screws and the specimen is maintained by a stiff spring, Q and the weight of the instrument taken up by attaching a thin cord to C, which, passing over a light pulley, carries a balance weight. It will be seen that the instrument is free from strains, which vary with the load on the specimen. The dimension from the axis to the plane of the tilting mirror was 20 em., while that between the centre-line of the set-screws and the axis was 5cem. A magnification of four was thus obtained in the instrument itself, For a long time consistent reading could not be obtained, and it was only after very many trials of the instrument and alterations in its design that the causes of the inconsistency were successively eliminated. The apparatus was very sensivive to vibrations, and it was soon found that accurate readings could. only be obtained when the machinery in the adjacent laboratories was at rest. Vibrations caused by traffic some distance away were quite noticeable. The most troublesome factor was probably the small unavoidable amount of jerk accompanying the application of the load. Occasionally this would cause a very slight displacement of the tilting mirror, which could not be detected on looking in the telescope, but neverthe- less was sufficient to vitiate the accuracy of the readings. Under these circumstances it was impossible ever to depend on a single observation. The load was invariably applied and removed several times, and if the readings were consistent, and returned each time to their original value when the load was taken off, they were accepted as correct. Any imperfection in the adjustment of the instrument or in the manner of appiying the load was always made evident by the erratic nature of the readings. 16 Mr. J. Morrow. On the Distribution of — [Oct. 27, Method of Making the Experiments. For the first series of experiments, a long cast-iron bar, 2°833 em. broad and: 6-452 cm. deep, was placed in the ‘“ Wicksteed ” testing machine, in the Engineering Laboratory at University College, Bristol. It was supported on knife edges at each end of a span of 91:44 cm. (36 inches), and was loaded in the centre in the ordinary way. At a distance of from about 7:5 to 9:0 cm. from the centre of the span lines were carefully scribed on the sides of the beam, marking the places where the set-screws of the strain-measuring instrument were to rest. One of these lines was at the middle of the depth, and the others at distances of 0°635, 1905, and 3°175 cm., both above and below the mid-point. The instrument was attached just clear of the ends of these lines, and 36°80 cm. from one of the planes of supports. Thus the section examined was at a distance from the point of application of the load exceeding the depth of the beam, and it was hoped that this would be sufficient to secure immunity from the effects of local strains due to surface loading. Under these circumstances, each ton applied at the centre would produce a bending couple of 18,694 kilogram-centimetres at the section under observation. The effective distance between the needle points of the tilting mirror, as given by a reading microscope, was 0°6160 cm., and the normal distance of the scale from the mirrors 202°7 cm. The total magnification was, therefore, 2632 for the first experi- ments, but it was, of course, re-determined whenever the position of the scale was altered. The scale was divided to th of an inch, and, with the telescope used, 7),th of these divisions could be readily estimated. The observed numbers are, therefore, given to the nearest ;;7;55 of a centimetre. When taking a series of readings for a certain change of load, the instrument was first applied with the points AA at the lowest mark, and a number of readings taken each time the load was applied or removed. ‘The set-screws were then moved up to the next mark, and the readings continued. When the centre was reached, the instrument was taken off and re-applied above the specimen, with the screws BB at the top marks, and observations were continued down to the centre again. The log sheet was kept as follows :— 1903}. Stress and Strain in the Cross-section of a Beam. Lacy Table I.—Loads 1 to 14 tons. Cast-iron Beam No. 1. Load Scale reading. Diffs. Mean diff. Tons. : 1°0 110-0 1°5 103 °9 Q : 1°0 ELO SE 6'1 1°5 104 °0 6-2 6°15 1:0 110°2 nie ae 104 °0 6-1 1°0 110°1 Position of instrument, 2nd line from top. Index mark, 50. Right-hand scale descends when load is increased. Results of the Hxperiments. The experiments for each load were carried out in the manner described above, and the load was at first advanced by 3-ton intervals. Some difficulty was experienced, however, in obtaining satisfactory results for the first $ ton, and ultimately the zero readings were rejected and a series taken between + and $ ton. This is summarised in Table II. Table I1.—Load 4 to $ ton. Number of. | Distance from Mean scale Lateral strain | mark, centre. differences. | in specimen. | oo | | eis dded | it 3°175 i 3°92 x 10-° | 2 1 +905 Se es al 3 0°635 1:20 1:00 4 0-00 6-008 | 7 O00 5 — 0°635 — 1:10 | — 0-92 6 — 1°905 — 2°95 — 2°47 7 = Solis == 4ncey — 4:08 The minus sign denotes that the lateral strain was a contraction. The values, of course, refer to the elastic state of the bar only, permanent set having been removed as far as possible before the readings were commenced, by alternate application and removal of the load. The experiments were then continued in an exactly similar manner, the loads being increased by $ tons until 2 tons was reached, and the last set of readings being taken between 2 and 2% tons. Fracture took place at 2:46 tons. It showed sound material and occurred about 1 cm. from the middle of the span. VOL. LXXIII. C 18 "Mr. J. Morrow. On the Distribution of [Oct. 27, Table III contains the results of these experiments. At the centre two sets of readings were obtained at each load. These, in general, differed but slightly, and the mean values are given. The slight differences. were probably due to the limits of accuracy having been reached. e Table III. Increments of strain at different loads. (Scale readings.) Position of instrument. + to 4 ton.|4 to 1 ton.| 1 to 13 tons.} 14 to 2 tons.| 2 to 22 tons. 0°30 5°73 | | onl fig 1 4°68 9 63 10°12 2 3-24 5 82 6°15 6°17 ° 3°07 3 1-20 1-92 2-09 2°19 0°84 4 0-00 0:06 0-03 |. - {O81 — Oise 5 = 1-10 —1°88 saan: |) == 2-29 —1-69 6 —2°95 | —5-25 —5°60 —6-00 —3°45 7 — 4°87 —9 -40 —9-00 | —8 55 —3-96 The column headed ‘“ Position of the Instrument” refers to the number of the mark on the side of the beam in line with which the instrument was attached. A minus sign denotes that the lateral strain was a contraction. Distribution of Strain. At the lower loads the maximum lateral strain per unit of load added is on the tensile side of the beam, slightly greater than that on the compressive side. As the load increases, however, the maximum lateral strain per unit of added load in the tensile fibres decreases, whereas that in the compressive fibres constantly increases. Hence, at fracture, this strain is very much greater on the compressive side than on the tensile. The curves in fig. 2 are plotted from the second and sixth columns of Table III, and show the increment of lateral strain at different poe in the depth of the beam. The full line refers to loads of + and 3 ton, and the broken line 2 to 21 tons. For the other loads the curves are given in fig. 3. They atl be intermediate between the two curves of fig. 2 2 ae the strain curves for the tensile side of the beam given in figs. 2 and 3, it will be seen that at the lower loads a straight line is obtained, and that as the load is increased the lines become more and more curved in the direction of decreasing strain in the outer layers. This curvature is first noticed in Numbers 3 and 4, but not till the last curve, Number 5, isit very remarkable. On the compression side at the 1903.] Stress and Strain in the Cross-section of a Beam. 1 Fre. 2. DisTAnce From CENTRE, LATERAL Sraaiw Scade Divisivws Tora Déryrh o« BEAM, Fié. 3. Ry SNS ors, “A s Y RS \ e No. 2 e g 6 “ Ton RX 3 xX 46 fy Tors ‘SS t 4 8 ($62 tons “ete £ Legos e 7 é € SS. 0635 Xs hatercd Strain Sv “s = 72 ew. Seale Divisicas Tora. Derry of Beam W7S lowest load a similar curvature exists (see fig. 2), and although all the readings were carefully checked, the result was practically the same.* In Numbers 2, 3, and 4, however, the line is straight, whilst in Number 5 there is a pronounced curvature in the opposite direction. * It has been suggested that this unexpected curvature might be accounted for by additional strains due to the comparative smallness of the distance from the loaded part. C 2 ‘20 Mr. J. Morrow. On the Distribution of — (Oct. 27, The general conclusions to be drawn from these curves are, firstly, that as the load increases, the increment of lateral strain in the outer tensile layers becomes less, and that this is accompanied by a much increased increment in the less stressed fibres nearer the neutral surface ; and, second, that on the compression side, with higher loads, the increment of lateral strain in the outer fibres increases. To what extent these actions take place before fracture can only be inferred from the experimental results. The greatest changes occurred between 2 and 24 tons, that is with a bending moment at the section under observation of not-more than four-fifths of that which the beam was capable of resisting before fracture at the middle of the span ; and — between these limits the effects noticed would become much more remarkable. The most striking fact brought out is, as will be seen later, that the values of the strains thus obtained are considerably lower than would be expected from theoretical considerations. This fact appeared so important that it was decided to further verify the result by the examination of a second cast-iron beam. Experiments on Cast-tron Beam No. 2. The beam used for this further investigation was 7°578 cm. deep and 3:212 em. broad. The span was 91°44 cm. as before, and the section examined 37 cm. from one of the supports. Readings were taken at the middle points of the depth and 1:2, 2:4, and 3:6 ecm. both above and below these points. The load was varied between $ and 1 ton only. In all other respects the procedure of the previous experiments was repeated. Table IV contains the results in differences of scale readings, the first column showing the distance of the point under observation from the middle of the depth. The curve is given in fig. 4. It will be shown that these results corroborate those obtained from the beams previously described.. Table IV. Position of Differences of instrument. scale readings. 3°6 +7°30 2°4 +5-°00 1-2 + 2°40 0-0 +0°25 ey, —1:97 2°4 —4:34 3°6 ~6°28 1903.] Stress and Strain in the Cross-section of a Beam. 21 Fia. 4. ’Disrance ram CENTRE, LaTeRac STRAIN. 4 & Seace Divisions. . Experiments on a Wrought-von Beam. Toran Derraw o« BéEAm. The following experiments were made on a wrought-iron bar 3°122 cm. broad and 7:614 cm. deep. The span was the same as before, namely 91:44 cm., and the section under observation was 37 cm. from one of the supports. In this case the readings were taken at the middle of the depth and at distances above and below of 1:2, 2°4, and 3°4 cm. The method of procedure was exactly the same as that with the cast-Iron beams. Table V. | : Difference of scale readings. | Increments of strain per 3 ton. Distance from centre eas + to 1 ton. | i} to, 1% tons: 3 to 1 ton. 1 to 13 tons. 3-4 + 4°50 + 5-10 ee Se8Opc1Oe2 .+3°74x10-° 2 °4 +3°52 +3°96 +2 °58 +2°91 L2 +1°90 +2°12 +1°39 + 1°56 0-0 +0:10 +0°12 +0:07 +0:-09 1 —2:°08 —2:12 —1°53 —1-°56 2-4 =373 (| 3-68 —2°74 —2 +70 4 3 4. —4 °87 | —4 8v —3'57 = wae 22 Mr. J. Morrow. On the Distribution of [Oct. 27, The experiments were carried out with two increments of load, namely from 4—1 ton and 1—14 tons. Table V contains the results both in differences of scale readings and in increments of strain per half ton. The curves are plotted in fig. 5, and it will be seen that, whilst on the tension side the two curves practically agree, on the compression, side the increment of strain is greater for the higher load. Fie. 5. > ¢&/ Ton x ¢ & /% Tons ISTAWCE Fhom CEvTAE 72] LATERAL SrRain Tora Deprn of BEAM, <— Relations between Stress and Lateral Strain. In order to appreciate the results of the foregoing experiments, and to compare them with those which would be expected from purely theoretical considerations, it became necessary to investigate the relations which obtain between the lateral strain and the stress producing it, in specimens under direct tensile and compressive forces. To obtain these relations for the material of the first cast-iron beam specimens were cut from the less strained portions of the beam itself. The tensile piece was 1:956 cm. in diameter, and was of uniform cross-section for a length of about 25:0 cm. ‘The ends were screwed, and gripped in the testing machine by the ordinary ball and socket arrangements. The instrument used was very similar in principle to that employed for measuring the lateral strain in the beams. It has been fully explained and illustrated in a paper published in the ‘ Philosophical Magazine ’ (6th series, vol. vi, p. 417, October, 1903). The total magnification was 2744°5, and the load was advanced 4 ton at a time. The lateral strain per 4 ton of load was then calculated by dividing the difference of scale readings by 84540, and the correspon stress was 169°07 kilogrammes per square centimetre. The results are tabulated and reduced in Table VI (Appendix). 1903.] Stress and Strain in the Cross-section of a Bean. 23 The method of applying the load was similar to that described in the paper quoted above, so that the lateral strains do not include that due to permanent set. The specimen was not tested to fracture. It will be seen that the elastic lateral strain is proportional to the stress. ‘The ratio may be expressed Lateral strain per kilogramme per square centimetre = 20°66 x 10-5. The compression piece was cylindrical in shape, 2°436 cm. in diameter, and about 7-5 cm. high. The measurement of change of diameter was made about the middle of the specimen, in order that the effects of any suppression of the dilatation at the ends might be avoided. The load was applied exactly as in the case of the tensile piece. Table VII (Appendix) contains the results. The total magnification was 3002. The ratio of stress to strain is not quite independent of the stress. Near the origin the lateral strain per kilogramme per square centimetre = 20°91 x 10S. In conjunction with my colleague, Mr. E. L. Watkin, M.A., I have made a large number of experiments on these stress-strain relations for cast-iron. We hope to publish an account of them shortly. Some of the specimens used were cast at the same time and from the same ladle as the second beam described in this paper. In Table VIII (Appendix) are embodied the results obtained from two of these, the particular ones chosen giving average and fairly representative values. The tension piece was 1:958 cm. in diameter, and the compression specimen 2°499 cm. diameter and 7:5 cm. high. For comparison with the wrought-iron beam the relations were obtained from separate specimens of similar material. Tn the case of tension the lateral strain has been found to be directly proportional to the stress when within the elastic limit, and the experi- ments described in a previous paper* gave, for the material in question, the value 13°80x10-S for the lateral strain per kilogramme per square centimetre. The wrought-iron compression specimen was similar to those of cast iron, 2°404 cm. diameter and 75 cm. high. The load was applied by + ton intervals up to 2 tons, and then by tons to 4 tons. The lateral dilatations were not quite proportional to the stresses applied. The results are given in Table IX (Appendix). The relations between stress and lateral strain for all three cases are shown graphically in the curves of fig. 6. The origin of the curve for comparison with Cast-iron Beam No. 2 is displaced to the right a distance of 10 x 10-° unit. * See ‘ Phil. Mag.,’ Sixth Series, vol. 6, No, 34, p. 417. 24 Mr. J. Morrow. On the Distribution of [Oct. 277, Distribution of Stress in Beams. Assuming, as in the Coulomb-St. Venant theory, that the rela- tions which hold between direct stress and lateral strain in tensile and compressive tests are, at the lower loads, the same as those which obtain between the stresses parallel to the axis of a beam and the corresponding strains at right-angles to that direction, the distribution of stress over the cross-section of a beam may be inferred from the observed lateral strains. Fig. 6. Sreess, Kes, een Soom, 4 s ; LATERAL STRAIN mad 20 20 40 x 10-> /| e // if 4600 Thus, Table III gives the differences of scale readings measuring the lateral displacements in the first cast-iron beam. In Table X we have the total scale differences due to the different loads. This has been obtained directly from Table III by addition ; and for the zero readings it has been supposed that from O—+4 ton the displacements would follow the same law as those between } and 4 ton. The actual lateral strains are then calculated as in Table XI. They are given 1903]. Stress and Strain in the Cross-section of a Beam. 25 by dividing the numbers in Table X by the magnification and the breadth of the beam, and multiplying the result by the scale unit. By comparing these strains with Tables VI and VII (Appendix), or with the curve in fig. 6, we can find the direct stresses which must have accompanied them. ‘These are given in Table XII for loads up to 14 tons. Proceeding in a similar way for Cast-iron Beam No. 2 and for the wrought-iron beam, we obtain the second column of Table XIII and the last two columns of Table V for the actual strains per 4 ton, and comparing these with fig. 6, we get the stresses given in the Appendix in the third column of Table XIII for the cast-iron beam, and in Table XIV for the wrought-iron beam. The degree in which these stresses agree with those obtained by the ordinary theory of the bending prism, as given by Coulomb and generalised by Saint Venant, may be taken as a measure of the applicability of that theory. Thus Table XIV contains the theoretical values of the stresses calculated for the wrought-iron beam, and it will be seen that the agreement with the experimental results is not very good at the higher stresses. The theoretical stresses for the cast-iron beams are given in Tables XV and XIII (Appendix). Discussion of the Results for Cast Iron. The theoretical stresses are not calculated for loads higher than 14 tons, as above that point some of the values obtained would be greater than the actual tensile strength of the material. At the lower loads the lateral strains, and the stresses inferred from them, are generally lower than those obtained theoretically. The amount of this discrepancy can best be seen by comparing Tables XII and XV. It is also noteworthy that as the load is increased the results show a distinct shifting of the position of the neutral axis from the tensile towards the compressive side of the beam. For loads over 14 tons not only is the theoretically calculated maximum stress greater than the ultimate stress the material will stand before fracture, but the values obtained by measuring the strains are also considerably greater than the maximum lateral strain which occurs in a tensile specimen. The strain at fracture of a tensile specimen is about 35x 1075. This is nearly equalled by the lateral strain noted at 2} tons load on the beam, whilst the maximum lateral strain at fracture would be greatly in excess of this value. Saint Venant* assumes that, since all materials are capable of a small * See Saint Venant’s ‘ Navier’ (Paris, 1864), p. 178. 26 Mr. J. Morrow. On the Distribution of [Oct. 27, amount of flow when nearing the point of rupture, the curve of distribution of stress over the cross-section of the beam becomes parallel to the plane of the section when rupture commences. This suggests that the higher lateral strains may be due to the amount of plasticity of the material just before fracture. In a tensile test no such plasticity can be observed. This, however, does not invalidate the assumption, as it would be necessary to check the plastic elongation, in order to measure it before complete rupture occurred. . One effect of the permanent set in the direct strains may be to produce a more or less profound change in the distribution of stress over the section. This might give rise to the existence of lateral strains, which, though perhaps elastic in their nature, would not be included in the observed values of the experiments. It has been pointed out* that, when the load on an overstrained beam is reduced, initial stresses may result. ‘The strain in the outer layer might then be reduced by an amount greater than the real elastic strain due to the load removed. It appears, however, to be far more probable that the departures from theory which have been noticed are due to the inapplicability of the theory employed ; that is, mainly to the facts that (1) the Saint Venant solution implies different conditions of loading and end-fixing,T and (2) it is based on an assumption of the absolute proportionality ot stress to strain. In this connection it should be remarked that a different distribu- tion of stress and strain over the section would not necessarily alter the deflection or curvature of the centre line of a beam due to a given bending couple. t The author hopes to deal more fully with these and other questions in a future paper. Note on the Determination of Poisson’s Ratio from Bending Haperiments. The instrument described in this paper provides a method not hitherto available for the determination of the values of the “ stretch- squeeze ” ratio for different materials. The method appears at present only suitable for those materials and stresses to which the ordinary theory of bending may be applied. It depends on the relation between the lateral strain at any section, and the deflection of the centre line produced by the applied bending couple. Consider a beam, of uniform rectangular section, supported at the * See ‘Encycl. Brit.,’ vol. 22, article ‘‘ Strength of Materials.” + See paper by L. N. G. Filon, ‘ Phil. Trans.,’ A, vol. 201, p. 83. t~ See Wiedemann’s ‘Annalen der Physik und Chemie,’ vol. 52, 1894, paper by W. Voigt, p. 536. 1903.] Stress and Strain in the Cross-section of a Beam. 27 ends, and subjected to bending by applying a transverse force at the centre of its span. Let the origin be at one end and let z be taken in the direction of the length, y = the distance from the neutral surface, P = the stress at (a, y), W = the load applied at the centre, 5 = the deflection of the centre line at «, K = Young’s modulus for the material, oc = Poisson’s ratio, d = the length, breadth, and depth of the beam. If at any section, x //2, the change produced in the breadth ot the beam be measured when the load is applied, and, at the same time, the deflection of the centre line of the beam at that point be observed, then the angle turned through by the side of the beam and the deflection, 6, is given by Wi WW Sr ie ee 16EI 12EI Combining these we obtain an expression from which o can be at once determined, namely, angle turned through by side of beam 12065 312 — 49? The left-hand side of this equation is obtained from the mean straight line through the points representing the lateral strain in the beam. In conclusion the author must express his thanks to Messrs. F. C. Prentice and C. M. Rushton for assistance rendered at various times during the preparation of this paper. 28 Mr. J. Morrow. On the Distribution of — [Oct. 27, APPENDIX, Table VI (see fig. 6).—Tension and Lateral Contraction of Test Piece cut from Cast-iron Beam No. 1. 7 : ‘i vee, | emacs | al | cee (ema antin | ivisi (EEE CD a gs). (scale divisions). sq. cm. Tons. = 2°95 | 2°95 3°49 x 107° 169 1 2°95 5 90 6°98 338 le 2-94, 8 84 10 -46 507 2 2-95 11°79 13 95 676 25 7 IF) 14°76 17°46 845 | 3 2-96 172 20°96 1014 3% 2°95 20 °67 24°45 1183 | Table VII (see fig. 6).—Lateral Dilatation of Compression Specimen cut from Cast-iron Beam No. 1. iE, EOP ISLS of Total lateral Laterai strain | Stress, kilos. per oad. strain (scale strain (scale (clastic only) Sa ea divisions). divisions). q ue 1 Tons. S 2°60 2°60 2°26x10~° 109 1 2°56 5°16 4°48 218 13 2 63 1°79 6°76 327 2 2°63 10°42 9-05 436 24 2-60 13 -02 11°31 545 3 Pe 15°75 13 °68 654. 35 Qi ba 18 -50 16 :06 763 4 2°67 21-17 18 °38 872 4k 2-70 23-87 20°73 981 5 ZAB 26 °60 23 °10 1090 os 2°70 29 °30 25°44: 1199 6 2°75 32°05 27°83 | 1308 6h 2-80 34,85 30 °26 | 1417 | 1903.] Stress and Strain in the Cross-section of a Beam. 29 Table VIII (see fig. 6).—Stress and Lateral Strain in Cast Iron for comparison with Beam No. 2. Tension. Compression. eee Lateral strain _ Stress Lateral strain LoS TNL (elastic only) pe osebel (elastic only) sq. cm. J? sq. cm. y)- 169 3°37 x10~-° 207 3°95 x 107° 337 6-80 414. 8°12 506 10°25 621 12 -48 675 13 -65 828 16°83 843 17 :08 1035 21°28 1012 20°45 1243 25°70 1181 23 85 1450 30°14 1350 27 °32 1657 34°53 1518 30 °75 1864: 38 °98 1687 | 34°25 2071 43 °45 2278 AT 85 2485 52°30 Table IX (see fig. 6).—Stress and Lateral Strain in Wrought Iron under Compression. Increments Total strain, | Hlactie Stress, Load. of strain scale divisions | ),4.. : kilos. per Pies : ateral strain. (scale divisions). | (elastic only). sq. cm. Tons OF 1°80 1:80 1°58 x 107° 112 iL 1°84 3°64: 3°20 224 1s 1°85 5°49 4°83 336 2 1-90 239 6°50 4.48 3 3°80 11°19 9 °85 672 4, 3°82 15°01 13 °21 895 Table X.—Cast-iron Beam No. 1. meta Total strains at different loads. (Scale divisions.) Point in depth : of beam. 0 to 4 ton. ietony te tons: 2 tons. 2+ tons 1 9-36 18 :99 29 *11 39 °41 45°14 2 6°48 12 -30 18°45 24 °62 27 -69 3 2 40 4°32 6°41 8 60 9°44 4, 0-00 0:06 0:09 — 0:03 — 0°34 i) —2°20 — 4°08 — 6°21 — 8°50 —10°19 6 — 5°90 —11°15 —16°75 — 22°75 — 26 °20 a —9 °74, —19°14 — 23°14 — 36 °69 — 40°65 Mr. J. Morrow. On the Distribution of | Oct. 27, 30 Table XI.—Cast-iron Beam No. 1. Point Total strains at different loads. in depth of beam 4 ton. 1 ton. 14 tons. 2 tons. 2+ tons. | 1 7-83 x 1075 15°89 x 107° 24°36 x 107° 32°98 x 107° 37°78 x 1072 2 5°42 10°29 15°44. 20°60 23°17 asa 201 3°62 5°36 7:20 7:90 4, 0:00 0:05 0°08 — 0°03 — 0°28 5 — 1°84 — 341 — 5°20 — Fal — 833 6 — 4,94, — 9°33 — 14°02 —19°04 — 21°93 7h —815 —16°02 — 23°55 — 30°70 — 34°02 Table XII (Table XI converted to Stresses). Dae Stresses in kilos. per sq. em. depth Se Bsa > ton. 1 ton. 13 tons. ates Sse Ss _— | | — 1 — 375 — 760 — 1152 2 — 259 —492 — 738 3 — 96 —173 — 257 | 4, 0 - 2 - 3 5 90 165 252 6 239 452 679 7 395 775 1140 — sign indicates compressive stress. Table XIII.—Comparison of Theoretical Distribution of Stress with that obtained from the Lateral Strains in Cast-iron Beam No. 2. Stresses corresponding to observed : lateral strains. Calculated Distance from . stress, kilos. per centre in cm. sq. cm. Lateral strain. Stress. 3°6 Size Ome — 270 — 290 2 °4 3°57 —187 — 194 4 asl — 90 — 97 0=0 0°18 — 9 00 1:2 —1°41 69 97 | 2-4. —3°10 153 194 3°6 — 4°48 221 290 1903.] Stress and Strain in the Cross-section of a Beam. oil Table XIV.—Comparison of Theoretical Distribution of Stress with that obtained from the Lateral Strains in the Wrought-iron Beam. Stresses corresponding to Distance observed lateral strains. Caleulated from centre stress, kilos. in cm. per sq. cm. 3 to 1 ton. 1 to 13 tons. 3°4 = 231 — 253 — 278 2 °4, —181 —199 — —196 1°2 — 99 —107 — 98 0°'0 = © — 6 O 1:2 lil ‘113 98 2 °4: 198 196 196 3°4 259 2595 278 Table XV.—Theoretically-calculated Stresses in Cast-iron Beam No. 1. | ah Calculated stresses in kilos. per sq. em. Distance from centre ete ikl “jn aa % ton. 1 ton. 13 tons. 3°175 — 460 . —920 — 1380 1-905 — 276 —552 — 828 0-635 — 92 184, — 246 0-00 00 00 00 0 -A35 92 184 276 1-905 276 552 828 3°175 460 920 1380 32 Dr. S. M. Copeman and Mr. F. G. Parsons. __[ Dee. 1, “ Observations on the Sex of Mice.—Preliminary Paper.” By S. Monckton CoPpEMAN, M.A., M.D., F.R.S., and F. G. Parsons, F.R.C.S. Received December 1, 1903,—Read January 28, 1904. In this communication we record the results obtained, during a period of fifteen months, from the breeding of fancy mice. The experi- ments were commenced with the object of determining the extent, if any, to which the relative proportion of the sexes is capable of being influenced by varying conditions of age, nutrition, inter-breeding, etc. The work is still in progress, but in the hope of obtaining help and criticism from other observers, we think it desirable to put on record | the experimental work that has already been carried out. Our paper consists of two parts: (a) a list of the various crossings and their results, a record which we believe to be perfectly trust- worthy ; and (0) a series of conclusions at which we have arrived after careful study of the figures. As these conclusions form the most generally interesting portion of the paper, we have decided to place them first, especially as the mere statistics are only likely to be of use in the criticism of our deductions or in furnishing material by the aid of which others, not at first apparent, may perhaps be formulated. In using these tables an explanation of our symbols may be necessary ; it should, for instance, be noticed that all the bucks are indicated by small Greek letters—the does by Roman capitals. When a numeral follows the letter representing a particular doe, it shows that the bearer is the daughter of the doe whose indication is the letter alone. Thus, B? represents the second daughter which we kept of B, while B?? refers to the third daughter kept of B?, and, conse- quently, the granddaughter of B. This method, of course, gives no clue to the male ancestry of a mouse, but this can always be ascer- tained by referring to the record of the particular animal. If we translate one line taken at random from the doe’s record it will, perhaps, make our system clear : C1 (to «). Sept. 22/02. 26,4 2 (2 34m. J at least 4 m.) This means that on September 22, 1902, C! (a daughter of C) bore two males and four females to the buck «, and that at the time of conception (some 20 days before) the mother was 34 months old, while the father was at least 4 months. On looking at the top of the paragraph devoted to C1in the doe’s record, her percentage will be seen, while all that is known of the ancestry of ¢ will be found at the top of the paragraph devoted to him in the buck’s record. 1903,] Ubservations on the Sex of Mice. Jo Part [.—GENERALISATIONS. The first question as to which we are desirous of obtaining informa- tion concerns the possibility of the male or female parent, in any particular instance, exerting a marked influence in the direction of a preponderance of male or female offspring. In seeking an answer to this question it is necessary to state that out of the total number of 493 young produced in the course of our experiments, 258, or 52:3 per cent., were of the male sex, and 235, or 47°7 per cent., of the female sex. Taking the buck’s descendants first, we have the following records :— Be cae CoN ata 71 48 percent.) 77 (52 per cent.) res. NORCO Boye) (0 (Ooi. >.) 1) Deru 1) 32535 ee.) eS Ome a) 39 (43) ae) eer. De IB eu 835 (57 i) Ce 6 2 Pe antec S5 Genie: yo 17s ae) Cee if . 6 tes Levee PTO Oh) Ik ReMi 0 | Cia) 258 239 _ The cases of & 0 and x may be left out of consideration, as the numbers of their offspring are so few. Of the rest, « and « closely approach the normal, but 6, 6 and 7 have male offspring in excess, in connection with which fact it must be remembered that 6 was the son of 8. On turning to the record of (, we find that in the case of all five does with which he was mated, the male offspring was in excess, and in the case of 65 also when put to five different does in succession, more males were produced than females in every instance. n was put to seven does—five times males were in excess, once the sexes were equal, while only once (with C!?-?-!) were there more females than males produced. y and « on the other hand had female offspring in excess of male to a rather marked degree, and in each case the record is taken from more than fifty young. With y the females were in excess of the males in four litters out of nine, while in two others the sexes were equal, so that in only three out of nine litters were there more males than females. In the case of « the females were in excess in five litters out of nine, while in one other the sexes were equal. In this case too there were more males than VOL. LXXIII. D 34 Dr. 8. M. Copeman and Mr. F. G. Parsons. [Dec. A, females in only three litters out of nine. When it is remembered that all these bucks were put to at least five different does selected at random, it certainly does appear that some bucks have a tendency to beget more male, and others more female offspring. ‘This, we believe, is the experience of many breeders of animals. It may be objected that these bucks were not placed with exactly the same series of does in each instance, and we regret that this is so, but by looking through the records of 6 (a male producer) and « (a female producer) it will be noted that they both had young by the three does B?, C! and D!, with the following results :— é. é. ——“— _—— oa g72 B?inshae sete DI 4 2 a meee wee Phra Sens IDeA TS coe: nee ae eG Shp ce 14 9 Sul so that their tendency to produce an excess of male or female offspring would appear to have had no relation to the particular does with which they were mated. With regard to 6 (a male producer) and y (a female producer), both had young by the four does A B D and F with the following results :— B. Ne ~S é Ce: TENORS bah See 4 3 4 0 Bisa ener 1 4 7 9 (in 3 litters) De yee See S xing) tae eee ce Oe 4 3 4 3 12 A 2 oe This is not nearly so satisfactory a result for our contention, as with the same does y (the female producer) actually produced a larger proportion of males than did / (the male producer). We shall, therefore, content ourselves by saying that 6 and 7 are instances of bucks which tend to produce an excess of male, and « of one producing excess of female offspring. On looking through the record of the does the point that is most striking is the behaviour of C! and her descendants :— 1903. ] Observations on the Sex of Mice. 35 3. oF, Cl produced 9 17 C12 ¥ 3 11 (1-2-1 ¥: D) 3 (Que 33 2 5 (61-2-2-1 Ss 5 10 21 46 (31 per cent.) (69 per cent.) On the other hand C1? had 17 male and 6 female young, but she was a daughter of « not of a. There is, however, an influence at work in this family which may possibly account for the excess of females over males, and it is that the same buck « was largely responsible for the results, as each of the does was put to him; thus C!:?-2:'- was his daughter on the male side, and his great, great granddaughter on the female. It has already been noted that a was a buck who produced practically the normal proportion of young, so that his influence alone is not likely to account for the excess of females. But the practice of inbreeding a buck with his daughter, granddaughter, &c., for several generations may perhaps account for an excess of female offspring ; this theory being strengthened by the fact that C1-?-2-!- was the only doe with which the male-producing buck 7 had more females than males in a litter. The clue is one which we are now following up and which we would suggest to other breeders as worthy of further investigation. On the whole, our statistics seem to point to the fact that certain bucks and does tend to produce a preponderance of one sex, but that the influence is greater in the male parent; also that a doe which is the result of prolonged inbreeding is more likely to produce female than male offspring. The next point inquired into refers'to the possibility of the number of young in a litter exerting any influence on the proportion of the sexes. If there is any basis of fact in the theory that the amount of nourishment an embryo receives affects the determination of its sex, we should expect that large litters would show a predominance of one sex and small litters a predominance of the other. Up to the present we have neglected the young which were eaten by their mothers before their sex had been determined, because we have no reason to believe that the mother preferred to eat male or female young, and, in taking large numbers, we have presumed that as many male as female young would be eaten in this way. We have, however, kept a record of the total number in each litter when first seen; often an hour or two, and never more than 12 hours after the birth. A doe hardly ever eats the whole of a young one at once, as she apparently prefers to first eat the viscera and brains of several, leaving their carcases for a future meal. So that, even after 12 hours from the birth, it is easy to see, by- the D 2 36 Dr. 8. M. Copeman and Mr. F. G. Parsons. [Deert, remains, how many young have been eaten. The record of these litters, with the proportions of the surviving young, are as follows :— Number in | Surviving Surviving Vs : AIT Dive. hitter. males. females. BIE De ——-| — _____. ____ bo ox = bo bo mm bo bo Gb bo iW) (oo) MH DOAONANNWANRDONNONNNODMDOWOAANONNOMANAOMADAWNBGGTMAAVA DEA ww SD e WAHNW AOR WWARH AMO D | WBOWNOIMFWNRFAONEFPNFPNWANNRFPEBRWKENKB EB WH WN NN wWwWhBBBE NWNWNRHWWWOKRERERNWEROWWWKRHEENTOANWUANBHEDHWOWWEK BROOD ON eS od iS) Orc & CO CO C1 1903. | Observations on the Sex of Mice. Bik ’ Number in| Surviving Surviving ene Be. litter. males. females. oe 55 1gee 6 4; 2 BOW. . B?? a 5 2 57 13 5) dh 1 Sona hee Bee 10 6 A 59.. cl 7 iL 5 1 O46 Aloe is 8 2 6 Oe iste chs 8 5 3 Gor F 8 6 2 63. - 7 6 il 64. LD 5 2 1 65 De 5 5 GOR aie. D*3 4 2 2 Ci, ahs as Chat 5 2 3 68 (Opie of 2 5 69 @y323 10 2 6 2 ae Gq 3 1 5 72. z 7 5 2 73.. H 8 6 2 487% 231 211 45 | This table shows that 487 young were produced in 73 litters, thus giving an average of 6°7 for a litter. There are 45 litters of 7 or over, and 28 of under 7. These we may speak of as large and small litters respectively. Ifthe number of young in the 45 large litters is added up it amounts to 356, and of these 164 (46 per cent.) were males ; 155 (43-5 per cent.) females; while 37 (10°5 per cent.) were eaten and their sex undetermined. Similarly if the number of young in the 28 small litters is taken we geta total of 131; of these 67 (51 per cent.) were males; 56 (43:7 per cent.) females; while 8 (6 per cent.) were eaten. ‘Tabulated these results are as follows :— gS 2 Eaten per cent. per cent. per cent. Large litters) (ever 6). ......2.. 46 43°5 10°5 Small litters (6 or under) ... 51 43 6 This result suggests two conclusions—firstly, that in a large litter a greater percentage of young is eaten by the mother, whichis, perhaps, what one would expect, and, secondly, that in small litters there is a slightly greater percentage of males than in large ones. Other things being equal, one might fairly suppose that, in a small litter, each individual embryo would be better nourished than in a large litter, and this * The records of the six does with which the buck 6 was placed in a large eage have not been added in, because we have no means of Enowing how many were eaten, 38 Dr. 8S. M. Copeman and Mr F.G. Parsons. _—‘[Dee. 1, supposition is strengthened by our experience that, in small litters, the young are individually of greater size than we have found to be the case in larger families. The difference between the percentages of the large and small litters is not great enough to enable any general conclusion to be drawn from it, but, so far as it goes, it suggests that ample nourishment is more likely to result in an excess of male rather than of female offspring. The next point to be discussed is whether the age of either parent affects the proportion of sexes in the young. As an aid to the deter- mination of this question the records of all the does of two months or under at the time of conception, may be compared with those of all the does of 6 months or over. We have records of 21 litters produced by does up to 2 months of age. These give a total of 108 young, of which 55 (51 per cent.) are males, and 53 (49 per cent.) females. With does over 6 months old we have also records of 21 litters, with a total of 134 young; of these 74 (55 per cent.) are males and 60 (45 per cent.) females. These figures show that there is an increase in the proportion of males to females in the progeny of those does over 6 months of age. It will now be worth while noticing that the does at intermediate ages, that is, from 2$—54 months (inclusive), produced 27 litters, giving a total of 173 young, of which 85 (49 per cent.) were males, and 88 (51 per cent.) females. Tabulated, we arrive at the following results :— 3 ? per cent. per cent. Does up to and including 2 months ...... 51 49 », between 24 and 54 months............ 49 51 7. MOlO MOMURS GNe@nOMet ae eerie er 55 45 This suggests that, in does over 6 months old, the proportion of males to females increases, but we are unable to deduce anything from this knowledge at present, for the reason that we do not know the duration of the breeding period of a doe, nor at what stage the young are likely to receive the greatest amount of nourishment. Moreover, the difference in the proportion of the sexes among the young is not avery great one, and, doubtless, there are many other influences at work, the effect of which it is difficult to eliminate. One of these is the tendency, of which proof has been adduced, that certain bucks produce a preponderance of male or female offspring, but this tendency, to a certain extent, has been neutralised by the fact that our statistics have been drawn from the pairing of ten bucks with twenty-eight does. It is unfortunate that the buck of which we possess most records should have been paired so extensively with does under 6 months of age, as had he been put with an equal number of old and young does, it would have been interesting to have determined 19038. ] Observations on the Sex of Mice. 39 whether the percentages of male and female young would have varied. Some little information may, however, be obtained from the record of 6, who when he was mated with five does of an average age of 7 months, produced 65 per cent. of male offspring, while with six does averaging 2 months old the male offspring was only 53 per cent. Though the numbers here are small (eighty-three young in all), the record certainly supports the suggestion that an adult doe is more likely to bring forth an excess of male offspring than a very young one. With regard to the effect of the age of the buck we are unfortu- nately unable to give any definite opinion, since the bucks from which we chiefly bred were apparently fully grown when purchased, though we had no means of actually determining their age. It will be seen, on looking at the record of «a, that in his later offspring females pre- dominated, but there three factors at least are concerned: (1) The advanced age of the buck; (2) the predisposition of C! and her descendants to produce females; and (3) the possibility that the excessive inbreeding to which these mice were subjected may have led to an increase of females. The last factor we propose to consider at the present time is that of external temperature. The mice were kept in an unheated greenhouse in which the temperature usually ranged between 80° and 100° F. during the day-time in summer, while in winter it often descended several degrees below freezing point. It may be interesting to contrast the records’ of the young born in July, August and September with those born in December, January, February and March, and both of these with the total records for the whole 15 months over which our experiments extended. During the 3 hot months 136 young were born, of which 75 (55 per cent.) were males, and 61 (45 per cent.) females, while during the 4 cold months 127 young were born, of which 65 (51 per cent.) were males and 62 (49 per cent.) females. Be ?. Mince: hot months. 2. jn... 55 per cent. 45 per cent. ome Coldsmonmthiey oo. :ss.os5a0: 51 bi 49 Total records for 15 months... 53°3 ,, ASE TE oe These results do not seem to indicate that temperature or time of year exerts any marked effect on the proportion of the sexes in the young. In conclusion it must be confessed that we have learnt comparatively little from this 15 months’ experimental work on mice, but we are hopeful that the labour has not been entirely expended in vain, and that at least certain clues have been obtained which may usefully be followed up, both by ourselves and other breeders. So far as our experiments have gone, the chief points of interest would seem to be :— AN Dr. 8. M. Copeman and Mr. F. G. Parsons. [ Deer a; i that the murat of males born is slightly larger than that of females. 2. That certain males beget a markedly large proportion of male, and others of female offspring. 3. That there is some evidence that this tendency is hereditary. 4, That certain does tend to bear an excess of either male or female offspring, but the evidence of this is not so conclusive as in the case of the male. 5. That mice bear inbreeding between a male and his offspring for five generations without loss of fertility or apparent bodily degenera- tion—this inbreeding in our one series of experiments being attended with a large excess of female offspring. 6. That the average number of young in a litter, judged from seventy-three litters, is 6°7. 7. That in large litters more of the young are likely to be eaten by the mother than in small ones. 8. That in large litters the proportion of females is greater than in small ones. 9. That more males are produced by does over 6 months than is the case with does under that age. 10. That the temperature and time of year at which impregnation occurs seem to exert little or no influence on the proportion of male and female offspring. Of course the larger the number of experiments the greater will be the likelihood of obtaining reliable statistics, so that it will be interest- ing to determine whether another year’s breeding confirms or neutralises the results now recorded, but we think it desirable to publish our observations at this stage, for two reasons, firstly, to invite criticism on our methods and suggestions for future work, and, secondly, to indicate to other breeders clues which would appear worth while following up. It should, perhaps, be mentioned that, in Soh instance, careful record - has been ke of the colour of the individual mice mated together, and also of that of their progeny. These results we have handed over to Mr. Bateson, by whom they have been utilised in connection with his investigation of Mendel’s theory of inheritance of parental characteristics. * * “ Zool: Soc: bres M1903) vol. 25 pos: 1903. | Observations on the Sex of Mice. Part I].—-REcoRDSs. A. Does’ Records. — 1. Doe A (bought on April 18, 1902, under breeding age). Gin eae as Oe be lo May 9/02 4 0 About 2m. At least 4 m. 5) Che June 22/02 4 3 Ss Poe, kr Obee aL ice Oct 2/02) 92) 0 oy Flamer poles 5 tees 2. Doe B (bought April 18, 1902, under breeding age). Geet ea: oy 3. More, 24... May 11/02 4 3 About2 m. Atleast 4 m. Boe. Tielke, U0) Py Ee eae 5 Sept 8/02 9 3 2 yh ceammve esl 0M, na) eee Ocn 18/020 2 3 oe Ge ae ek: Now 11 Oo) ch EY an ee 15 13 3. Doe C (bought April 18, 1902, under breeding age). 3-1 2- oe a: TOCA ene May 17/02 2 2 About2m. At least 4 m. | Cer anee July 2/02 3 -— Bo ae Wier? aoe 5) 4. Doe D (bought April 18, 1902, under breeding age). b-1 2. ? : 3. OTS 36.415 May 22/02 1 3 About 2 m. At least 4 m. Rees July 1/02 Sn SALES Ah ever Behan vosdties Aug. 20/02 4 4 Be seen y 2h Tine ey teas Ove miOiOny 249 4 So 6 yy BF Pa Geh id 815 Nov. 25/02 to= i eet least, 10m WS) We 5. Doe E (bought April 18, 1902, under breeding age). Soy: a 3. ene May 26/02 4 4 About2m. At least 44m. 4 4 42 Dr. S. M. Copeman and Mr. F, G. Parsons. _‘[ Dee. 1, 6. Doe F (bought April 18, 1902, under breeding age). 3. . ? 3. Novas er.. June! 4/02 0 1 About2 m. At least 45 m. Fi Cane July 23/02 4 38 SME hs eS 8 Ray, anes Sept. 19/02 4. 3 3 Oi 5305: ee ARE Cena Oct. 30/0296 CE ser sa Oe 12 18 7. Doe B! (B+, born May 11, 1902). Cees fe 3. 2a TO ee ts. July 14/02 2m. At least 6 m. 8. Doe B? (B+a, born May 11, 1902). g. 2 ?. g- Tovah. July 12/02 2 1 2m. At least 6 PCM ck Sept. 25/02 4 2 44,, Piesk o Pe) ae Kebs w2/03qm2e Le roe. 53. Owes Pc es May 30/08 4 412 ,, eee 1238 Oe ee 3. flow eee July 29/02 2 5 2m. At least 64 m. a Ure Octi Toe 4 5a", ae sats hee: Nov. 13/0262 2° 3152,, «2m: 8 13 Eames Vert BOO ose July 25/02 3 2 2m. At least 65 m athe mee Sept..27/02 4" i Oe ae, » aa Oe See Nov. 25/0236 2 i ueGe sO eaeee ae a Feb. 2/03 - 3 2. 8%, de io) — ~I 1903.] Observations on the Sex of Mice. 11. Doe C? (C+, born May 17, 1902). UNO?) Or, 3. sew Aug. 1/02 1 1 2 m. At least 64m. egieac Aug. 22/02 5. 1 22,, 3 ees Ae Sent. 22/02 2 4 32,, io dale Sis Mec 0/027) 33: OF mer 35m: oe ee Sy ue 9s, 78 HS is! os tae Mave20(Oa5 9 — 39 12> 5° 2% 5. 14 15 12. Doe D! (D+, born May 22, 1902). hee ee sak ce 3. Pe July 31/02 2 5 dm. At least 65 m. he Ori 5/028" 4 (31, ee a Say Now. 18/02 > 5) 3 5 m. 25 m: GEA: ane O 28 aD 2 t).'5, Ee a ar epeHOS 4 4 18h 5 1O"LS 13. Doe D? (D+a, born May 22, 1902). Oe ter 3. nea ioe Ole ot 2 me) At least. ( mm, ae Sept.-30/020°54 4° 32. ,, at) Sega ae Or 2300" FE 8 Ae epee sr oe Dec a/O2 23) 3 6S 40 3k m 12 14 14. Doe D? (B+, born May 22, 1902). SEO! ON ey Be eat Aug. 92/02" 3 3 2m) Atleast 7 m. ee Sept. 29/02 1 4 3 ,, bas Nd eee Decw2/025 3.0 OF 5,9 6 m: 15. Doe B?! (B? +a, born July 12, 1902). Oe OF 3. Paes Sept. 11/02 2 3 2m. At least 8 m. seer INove /An02 Eb 32 ae. wee HOUR: Ben sane ealOs- 6s 3 ~* bi 5,0 OF Mm. Una Mareen OS a te 10>;,) At least, 16 m. ante July) 8/038 -4 2 12), >» 5, 43 44 Dr. 8. M. Copeman and Mr. F. G. Parsons. | Deeray 16. Doe B?? (B2+ 7, born May 30, 1902). one oF 3. 2 bw.’ At least Gum 17. Doe B?! (B+, born July 29, 1902). Greate é Q. Le Lod =o Oct. 5,02 4 1 2m. At least 9m, 4 il 18. Doe B?? (B? +a, born July 29, 1902). Eee Ae, CE uss TOW eek Oct 20/028 Bow 4 24m. Io m 6) 4 19. Doe Ct? (C!+a, born July 25, 1902). é. F. Qe 3. Ove, cee: Oct) 2/02 earl ahr iw.. At least’ 9m A Si JanwlO/0>ha enero 5 m. Jf aia a lel To To 20. Doe C3 (Ci +, born September 27, 1902). 6. Ff. ?. he Sha se Heb: 16/031) (5. p3un fet an Gane ide: May 20/03 6 2 7 OME ONE ...:.. - dune 22/03 Geol Ss At least 5 m. Lee NO 21. Doe D1! (D! +a, born July 31, 1902). Ga: f . 3. Eat dh Dec a/O2t ara 34m. At least 6 m. 2) Gil 22. Doe D?3 (D?+a, born August 7, 1902). Shouiwcee 2. PRC Sit Dias cl Oct. 26/02 5 — 2m. .At least 94 m. 5 1 aie 1903. ] Observations on the Sex of Mice. 23. Doe D3 (D? +a, born August 2, 1902). Sh oF 3. Mowe. 2 Oca 23/027 22 Ze. ome 29 24. Doe Cl?! (C!24 a, born October 2, 1902). ay Ss os) 3 dy M3) 25. Doe Cl? (C!#+ 4, born October 2, 1902). Oe P cy 7 26. Doe C1-2-?1 (Cl-2-2 +a, born December 27, 1902.) anole 9. a Mor ase... May 13/03 2 6 4m. At least 16 m. Ben th: Mane 2403 uee tii 5, Ja Vas 5 10 27. Doe G (age and ancestry lost, but under 3 months). Greely e « Be 3. ON Seen May 22/03 1 2 About3m. 8% m. pe otis June 29/03 5 2 eae GA leAste osm 6 4 28. Doe H (age and ancestry lost, but under 3 months). 6. &. ?. 3. Gh ener July 17/03 6 2 About 3m. At least 6 m. O56 2 Be wh Deer 22)02n 2 3 2m. At least 114 m. Bai Deer 2 02 25 2m. At least 114 m. 45 46 Dr. 8. M. Copeman and Mr. F. G. Parsons. 1. a (bought April 18, 1902, at least 4 months old). B. Bucks’ Record. By: TAGES. s.. sear May 9/02 fest Me... May 11/02 By Ole ach 3. May 17/02 So ada e oan May 22/02 > a Nov. 25/02 sot ela cess oe May 26/02 ie Ser wNaMs S June 4/02 AD ce ee Oct. 30/02 S50) uD ha nena ae July 14/02 5 ll? 5 acten ene July 12/02 0 Bt eae July 29/02 Pi) Ci age July 25,02 San CRB Rees a Aug. 1/02 Be le eee aa Aug. 22/02 DL eee July 31/02 ey 1)? eee ee Aug. 7/02 peg)? ores eee Aug. 2/02 = Sel Oat ta omnia Sept. 11/02 ll Oa ee Nov. 17/02 ap Oa mg May 21/03 ean lepeee ys venae Oct. 5/02 aa DL Aas a Oct. 2/02 Pipe Onion ne wee Jan. 10/03 ~ Denes Oct. 26/02 Chel cece sss) Wee 22/02 Ore ecotan IWee, 27/02 OIA iat, Sane May 12/03 By Afar Hoe de June 22/02 sh ME cee eee July 16/02 pf IO 8 ares etre July 2/02 Bre oD Bec nie aa July 1/02 ye ene WEEE July 23/02 ~J =| WHY TNH FR WHE DWE DTH WH DD NYH wh TE pK RY _— =| a Suton O Hh tw ow Ome ee NS ee ene ee Benes | to 2. 6B (bought May 30, 1902, at least 4 months old). [Deert, 1903. | Observations on the Sea of Mice. 47 3. y (B+a, born May 11, 1902). 3 Rye Oct. 2/02 4 spila genta D tease Sept. 8/02 3 BD ea sesh tei Oct. 18/02 2 WS Ce hea Nov. 11/02 2 a} Wal aaa Rete Aug. 20/02 4 ee oR aera auc iss Con WOO 4 le eee Sept. 19/02 4 2 sik Oat er Nov. 25/02 2 =f ull OE arene Dec. 12/02 3 28 4. §(B+, born July 16, 1902). This buck, when 4 weeks old, was placed in a large cage with the following six does, all of which were under breeding age: B!, F', hee? Cit) Whe young of course were all mixed, but alto- gether 27-g¢ and 24 2 were born between November 9, 1902, and Januaryg17, 1903. ‘the buck was then put separately to five does with the following result :— Ge [Sra seen anne Feb. 2/03 2 PO ies 8 sce Feb. 2/03 3 ee Dveiot ave Jan. 13/03 5 fal aesaN re a Feb, 24/03 4 Oya ee ceeiaee Jan. 15/03 6 Pe Oh eae es Sls Feb. 6/03 5 21 Previous record with young GOES TW Sas tod oe 25% 27 48 5. « (Bought September 6, 1902, at least By Bet ey. Sept. 25/02 4 Nokal BSCR Male Oct. 1/02 4 si OR ee nario Sept. 27/02 1 Oe an Sept. 22/02 2 Pain ee Oct. 5/02 3 LOE a a Sept. 30/02 4 ssi l) ie antaaas yu Oct. 23/02 4 a 0 Meer ape Sept. 29/02 1 dele Bae ioe a Dec. 5/02 3; 25 4 months old.) (SU) S| Pm co He He HE op OL bO TO 48 Observations on the Sex of Mice. 6. ¢ (ancestry lost, about 2 months old). By AGr eo :: ..csceee May 20/03 6. 6 2 7. 7 (bought May 1, 1903, at least 4 months). By. Cle ee June 22/03 5 be. ee May 30/03 Rn Orcas a June 24/03 3 Gace. eee June 29/03 BA bu eee July 8/03 Pe OR ee ec: July 17/03 bt INE AS ce July 27/03 8. @ (C?+a, born August 22, 1902). By 13? csin eek ie Nov. 13/02 a D2 Pease Nov. 18/02 9. 1 (C?+a, born August 22, 1902). By KO2. ee) ce eee Dec. 20/02 A Oe” s Sie Feb. 4/03 Sa Daten meat Dec. 21/02 sy aillsy) eam Oct. 30/02 5 US eae Oct. 28/02 baa i ag 2 May 22/03 10. « (C?+., born March 1. 1903). By (©? eee May 20/03 3. 2. 6 4 4 4 3 4 By 2 4 2 6 Y 5 2 33 17 Ny 2: 2 3 = 3 7 6 a. Q. 3 3 3 3 3 3 6 4 yy 2 a ds | 2, 18 17 ae Q. 0 3 ae oe eT 1903.] The Acquirement of Secondary Sexual Characters. 49 “Observations upon the Acquirement of Secondary Sexual Characters, indicating the Formation of an Internal Secretion by the Testicle.’"* By 8. G. SHATTOCK and C. G. SELIGMANN. Communicated by Professor J. R. BRADFoRD, F.R.S. Received December 14, 1903,—Read January 28, 1904. The Problem Stated. The question taken up in the present communication may be concisely stated as follows :— The most prominent and obvious function of the testicle is the formation of the sperm. Under normal circumstances this is dis- charged ; it constitutes, that is to say, an external secretion. In spermatogenesis the male attributes culminate. There is, how- ever, another element in maleness, of a different kind, less essential, yet in many cases well pronounced, viz., the acquirement of certain external characters which distinguish the male from the female in many groups of living forms. That the development of such secondary characters is related {to some function of the testicle, appears from the results which follow castration when carried out before the advent of sexual maturity. On what, then, does the production of these characters depend ? It is conceivable that the result may be due to a nervous reflex arising out of the physical function’ of the sexual mechanism. This view our observations seem to us to disprove. The genesis of the external male characters must, in our opinion, be transferred from the influence of the nervous system to the realm of chemistry. It depends, with more probability, upon the formation of a second secretion by the testicle, the absorption of which into the circulation induces the metabolic changes that reveal themselves as secondary sexual characters. The suggestion that such an internal secretion might be elaborated by the “interstitial cells,” which le in groups between the tubuli, was put forward by one of us (S. G. 8.)7 in 1897. The experiments to be recorded were, in fact, primarily designed with the object of eliminating any part that might be played by the tubuli in this connection, and so of determining whether any function could be ascribed to the cells named. They consisted in ligation of the vasa deferentia in the young of- certain forms in which the male exhibits marked secondary characters. * Towards the expenses of this research a grant was made by the Council of the Royal College of Surgeons; and by the British Medical Association, on the recommendation of the Scientific Grants Committee of the Association. + ‘British Medical Journal,’ Feb. 20, 1897. VOL, LXXITI. 50 Messrs. 8. G. Shattock and C. G. Seigmann. [Dee. 14, It appeared possible that the epithelium of the testicular tubuli would, under these circumstances, on proliferating, undergo degenera- tion and atrophy from the pressure due to its own accumulation, whereas the interstitial cells of the stroma might remain intact. This result, however, did not: ensue, but others, which we venture to record as bearing on the problem under consideration. The forms selected were a breed of sheep (Herdwick), the male of which is furnished with long recurved horns, of which the female is quite destitute, and the common fowl. Observations wpon Sheep. We owe to the kindness of two friends the opportunity of observing many castrated sheep, as well as a certain number of others on which some form of obliteration of the vas deferens had been practised. Besides the horned (Herdwick) sheep already referred to, we made observations upon the hornless Southdown, in which the results, though less striking, are none the less constant. The results of occlusion of the vasa deferentia in the Herdwick breed have to be compared with those following castration, and both with the normal standard. Lambs of the same age were selected, and the procedures mentioned were carried out at about the same time. The occlusion of the vas deferens was effected a short way above the testicle by the application of a silk ligature in two places and division of the duct between. The animals were examined at different periods during their growth, and were killed when fully developed at ages of from 10—14 months. In those castrated either no horns appeared externally, and on preparation the skull exhibited only two low osseous tubercles or horn- cores, or very diminutive horns were produced, and beneath them a slightly more prominent core than in the first case. In the ewe of the Herdwick breed there is no external trace of horn, nor does the prepared skull show any osseous core. As contrasted with the results of castration, those of vasotomy are very striking. The horns attain their full size, and the skull its complete male characters, so that the head in no way differs from that of the normal or intact ram. The form of the skull is modified by castration, not by double vasotomy, the modification in question being obviously correlated with the absence of horns. The skull of the castrated sheep, or wether, is less rugged, and the bones thinner, but besides such general differences the plane of the os frontis is continued backwards behind the orbits at a very obtuse angle. In the intact ram, and equally after vasotomy, the plane of the 1903.] Lhe Acquirement of Secondary Sexwal Characters. 51 frontal behind the orbit lies almost at a right angle with the inter- orbital portion of the bone, the horn-cores arising from the upper or horizontal area. Although the skull generally is thicker in either case than in the wether, this alone does not account for the difference in external form; the cranial cavity presents a corresponding extension in the frontal region. In the configuration of its skull, as in the absence of horns, the castrated animal precisely resembles the hornless ewe of the breed. We have studied the effects of the same procedures upon sheep of a well-known pedigree Southdown herd. The result in such animals is less striking than in the Herdwick, partly because each sex is hornless, and partly because amongst Southdowns individual variations in the form of the head are not uncommon: thus, whilst the head of the wether usually offers a marked contrast to that of the ram, in certain cases the characters of the two so nearly approximate that even an expert may find it difficult to distinguish between them, the ram under such circumstances being commonly called “ wether-headed.” In the Southdown there is not (as in the Herdwick) any marked difference produced by castration in the form of the forehead, the angle between the pre- and post-orbital portions of the frontal bone being equally obtuse in the vasotomised sheep and the wether, and this, for the reason that both are equally destitute of horns. That the occlusion of the vas had been complete in all the cases observed, was proved by a careful dissection of the testicles after the animals were killed. Seeing that the full development of male characters proceeds in spite of double vasotomy, it becomes interesting to inquire into the condition of the testicles and into the sexual physiology of the animals themselves. To take the latter first. A Southdown, the subject of double vasotomy when a lamb, and kept apart until full grown from any female, was turned loose with a couple of maiden ewes; he at once copulated, erection and intromission being complete. The two ewes were not again admitted to the flock, but were kept apart, with the result that neither afterwards proved to be in lamb. ‘This animal was killed 18 months after the vasotomy. ‘The testicles had, from the first, grown symmetrically, and had reached the normal size; dissection revealed a complete interruption of each vas close above the gland. In certain cases one of the testicles underwent a marked diminution, 1.¢., 1t not merely failed to grow but rapidly wasted. In the other cases both organs attained the full dimensions. These differences are to be ascribed to differences in the condition of the blood-supply ; when the vas is cleanly isolated and divided after ligation without the inclusion of vessels, or without the subsequent occurrence of thrombosis, no atrophy of the gland ensues. When atrophy of one testicle arises, E 2 52 Messrs. 8. G. Shattock and C. G. Seligmann. [Dee. 14, the other suffices smgly to bring about the full development of the male characters. A careful dissection, carried out in all the cases of vasotomy examined and cited, showed that the vas had been completely occluded; not only was its continuity interrupted, but the noose of the ligature was demonstrated on the end of each segment of the divided duct. The epididymis after occlusion of the vas may become notably larger than normal ; this is especially obvious in the lower end or globus minor, and is to be ascribed to its over-distention with the secretion trans- mitted from the body of the gland. © Microscopic sections of the testis of the normal adult Herdwick sheep and of that from the vasotomised animal of the same age and killed at the same date, show similar histological pictures. The tubuli are filled with epithelial cells, and in nearly every one spermatogenesis is In progress.* Laupervments wpon Fouls. Kven still more striking are the results of double vasotomy in the cockerel of the common fowl. In the fully grown cock the exposure of the vasa deferentia and their ligation is not particularly difficult, but in the young bird it is otherwise ; and the results cited are limited to those cases where careful dissection afterwards proved that this difficult procedure had been successfully carried out, the continuity of the duct being found interrupted, and the noose of the ligature discovered at the site of operation. The method of proceeding was as follows :— The vas is exposed is its course over the kidney by a curved incision carried through the lateral wall of the abdomen; the duct having been ligatured as near to the testicle as possible, is then cut across a short distance below the ligated spot, no ligature being placed on the lower segment. The vas of the other side is afterwards similarly dealt with through a second incision carried through the corresponding side of the abdomen. Owing to the difficulties of this operation, vasotomy was in some cases performed on one side only, the testicle of the other being removed. The anesthetic used was chloroform, and the material of the ligatures, silk. The results of double vasotomy, or of one-sided vasotomy combined with one-sided castration, are in all cases alike. When carried out upon the young, immature, bird, or cockerel, the development of the mele characters proceeds without any notable interruption, and reaches its full degree. The birds used in the experiments were so young that it needed an, * It may be incidentally remarked that whilst in the castrated lamb the prostate fails to grow, in those submitted to vasectomy the gland comes to equal in size that of the intact ram. The same is true of the vesicule seminales. If one testicle is removed and the vas of the other ligatured and cut across, the prostate and vesicule acquire the full size, and this without asymmetry. a a re Se 1903.| The Acquirement of Secondary Sexual, Characters. De ww expert to determine their sex: examination, moreover, of the testicles removed from cockerels subjected to the combined castration-vasotomy just referred to, as well as of those removed from birds of the same brood, showed that no spermatogenesis had arisen at the age selected for operation. We may adduce examples in order to give the full erounds for the general statement set forth with regard to these experiments. Double Vasotomy.— Impure “ Plymouth Rock,’ 7—8 weeks old. Nine months after the operation the head was male in type; neck- hackles well developed; tail beginning to assume male characters ; spurs indicated. ‘Twelve months after the date of operation the spurs were stout, though short; head thoroughly male; neck- and saddle- hackles moderately well developed; tail short, male in kind, with sickle feathers. The bird remained in the same condition, and was killed 12 months alter the date of the operation. At the autopsy, the testicles were found to be of full size (about that of a pigeon’s egg), and in their general aspect quite normal. In connection with the right there was a spermatocele about as large as a haricot bean ; this, on being punctured, gave exit to a whitish fluid which microscopically showed numerous spermatozoa, some of them motile. The superior segment of the divided vas, or that in connection with the testicle, was dilated; the upper end of the lower segment was traceable into scar tissue in which it terminated. On the left side there was no spermatocele in connection with the gland, but the tubuli of the epididymis were abnormally evident. On each side the noose of the silk ligature was found zn situ on the upper segment of the vas, above the level of the lower border of the testicle. As a second instance we may recount the following :— Double Vasotomy.—< Plymouth Rock,” about 8 weeks old. In the summer of the year following the operation the head and neck-hackles were typically male, saddle-hackles fairly so ; tail short, carried almost vertically, contained a number of short curved feathers; spurs short and stout. In the winter of the same year the neck- and saddle- hackles were typically male ; tail short, bushy, feathers curved ; spurs long and sharp. The bird was killed in the spring of the following year. Dissection showed the left testicle to be of full size, 4 cm. in longer diameter ; in connection with the upper end of the epididymis is a retention cyst filled with white secretion, and about 1 cm. in diameter; the epididymis is, as a whole, enlarged from distension. The upper end of the lower segment of the vas terminates a short distance above the lower border of the testicle. The ligature lies on sitw on the end of the epididymal segment of the vas, which is separated by a distinct interval from the other. The right testicle is slightly smaller than the left, the epididymis 54 Messrs. S. G. Shattock and C. G. Seligmann. [Dee. 14, distended, and the continuity of the vas interrupted ; the ligature hes mm situ on the lower end of the upper segment of the duct. As an example showing the results of unilateral vasotomy combined with unilateral castration, we may select the following :— “Buff Orpington,” about 8 weeks old. Nine months after the operation the bird was thoroughly male, the comb and wattles being well developed, as well as the neck-hackles and sickle-feathers of the tail. Twelve months after the date of operation the spurs were sharp. On being put with a hen the bird immediately copulated, although it had had no previous opportunity of approaching one. Eighteen months after the operation it was killed. On dissection the right testicle was found to be of full size, about 3-5 cm. in the longer diameter ; the epididymis was slightly distended. A scraping from the divided body of the gland revealed the presence of spermatozoa. The lower segment of the vas was found to taper off and end quite dis- tinctly about a quarter of an inch below the level of the testicle. ‘The noose of the ligature was covered with a thin layer of connective tissue, and lay on the posterior surface of the organ. The position of the ligature may be explained by the general growth of the gland; this growth would naturally lead to an extension in all directions, and that in the downward direction would, relatively to the testicle, raise the site of the ligature. On the left side no trace of testicle was found. Microscopic examination of the body of the testis from the case of double vasotomy first cited shows the tubuli to be full of cells, and spermatogenesis in high activity, all the typical histological pictures being present. The same holds true of the right testicle from the case of combined vasotomy and castration last detailed. These results offer a striking contrast to those following a double castration when carried out upon the immature bird. Double castration was performed through a lateral incision on each side, the testicle being exposed to view, and afterwards carefully disconnected from its attachments with fine forceps, and withdrawn. In ideal experiments the gland was withdrawn entire; in others rupture occurred during the process of detachment, the organ being then removed piecemeal. Lesults of incomplete caponisation.—In certain of our experiments it happened that the testicle gave way during its detachment, and that minute fragments were unintentionally left behind. Sometimes such remnants, as told by subsequent dissection, were left in their normal position ; at others they were dislocated and transplanted upon the adjacent viscera, or abdominal wall. Under such circumstances the — cockerel assumed in different degrees the character of the male. The actual number of gland remnants left at such imperfect opera- tions, and the position of the grafts resulting from their displacement, varied considerably. ‘Thus, in one case the dissection of the fully 1903.] The Acquwirement of Secondary Sexual Characters. 25 grown bird, which had been castrated when from 6—8 weeks old, showed on the left side a spherodial mass of testicular substance, 2 em. in diameter, lying in front of the upper part of the kidney, and into the lower end of which the vas deferens is directly traceable. Hang- ing in the mid line from a loose “mesorchium” is a spheroidal graft 1-5 cm. in diameter. On the right side there is a bi-lobed mass 2°3 em. in the chief vertical diameter, with the lower end of which the right vas is directly connected ; closely adherent to the front of the upper lobe of this, though slightly movabie over it, is a spheroidal mass 0°6 cm. in diameter. A further oval nodule 0-7 cm. in chief diameter is closely adherent to the surface of a coil of the small intestine in the neighbour- hood of the liver; a scraping from this graft when cut through in the recent state showed large numbers of spermatozoa. Lastly there is a grait of about the same dimensions intimately adherent to the under surface of the liver itself. The external characters acquired by this bird were fully male throughout. It may be remarked, in passing, that such grafts do not bear classifying with glandular tumours or adenomata, since they do not grow independently of the general requirements of the body. For the whole sum of a series of such grafts and hyperplastic remnants does not exceed the volume of the two fully developed testicles. In this the remnants behave like those of thyroid tissue left experimentally aiter partial excision of the thyroid gland; or as do the dormant accessory thyroids after the complete removal of a goitre, when the accessory gland after attaining a certain size ceases to increase further - or the process, again, resembles the reproduction and hyperplasia of hepatic tissue which follows partial excision of the liver, of a fourth or even half its bulk.* In the most perfect cases of reproduction, each gland attains its full normal size. A bird was castrated when quite young, 6—8 weeks old. Six months later, the comb and wattles presented a medium degree of development ; the spurs were very small. Nine months after the date of operation, the spurs were still small, and the general male characters ill developed. ‘Twelve months after the operation, the spurs were short but stout. Seventeen months after the operation, the comb and wattles were thoroughly male, the neck- and saddle-hackles fully developed, and the spurs long, stout and sharp. The bird was killed 21 months after the date of the operation. Each testis was found to be of normal form and full size; the epididymis well pronounced, and without retention cysts. Each vas was in every respect normal and filled with white secretion, which microscopically showed countless actively moving spermatozoa. The history, as above given, shows a marked delay in the development of the male characters, * Ponfick, ‘ Centralblatt f. Med. Wiss.,’ 1894; Von Meister, ‘ Centralblatt f. Allg. Path. und Anat. Path.,’ 1891. 2) Oa Messrs. 8. G. Shattock and C. G. Seligmann. [Dec. 14, and indicates that these developed pari passu with the reproduction of the testicles, until they ultimately became fully pronounced. That a comparatively small volume of testicular tissue will suffice to bring about the development of male characters appears from the following result, in which the bird grew to be fully male with the slight exception that the neck-hackles were somewhat less closely set than is normally the case. “‘ Buff Orpington,” of about 8 weeks, at which time double castration was performed. Eight months after the operation, the comb was well developed and bright in colour ; the plumage in general, somewhat pale and sparse; neck-hackles moderately developed; spurs small. Eleven months after the operation, the comb and wattles were well developed; neck-hackles moderate ; saddle-hackles fairly male; tail feathers beginning to take the male curve; spurs grown to the normal male extent. The bird was killed 17 months after the operation, its condi- tion being as last noted. Dissection shows on the ieft side no trace of testicle in its normal position, but an inch and a half lower down, and three quarters of an inch anterior te this spot, there is an oval graft 2°5 cm. in chief diameter, loosely connected with the lateral wall of the abdomen. Above it, separated by a distance of 1:5 cm. and intimately incorporated with the peritoneum, is a second graft 0°5 cm. in chief diameter: and behind or dorsally to this isa further minute nodule 0-2 cm. in ee and likewise inseparably adherent to the peritoneum. The vas is extremely fine and traceable to the vacant, original, site of the testicle. On the right side in the situation of the testis there are two small flattened sano. the larger, lower, of which, is 0°8 cm. in chief, vertical diameter. Into the lower end of the inferior the vas, diminished in size and empty of secretion, is directly traceable. A third nodule which lay about 1 cm. anteriorly to these and slightly lower in the abdominal cavity was removed for microscopic purposes : scrapings from its divided surface disclosed the presence of spermatozoa. Histologically the largest graft (that on the left side of the abdomen) shows closely applied tubuli of full size, every one of which presents the histological pictures typical of active spermatogenesis. The lumen of the tubuli contains free spermatozoa. All the cell nuclei are throughout perfectly stained with nuclear dyes, proving that the tissue is living and not in an obsolete or necrotic condition. The amount ot inter-tubular stroma is very small, and supports well formed arterioles and other vessels. The much greater size ef the dislocated graft on the left side ot the abdomen, and its high state of activity, suggest that it is the chief element ae 1903.] The Acquirement of Secondary Sexual Characters. D7 concerned in the production of the male characters. This graft is strictly ductless, and is, moreover, entirely disconnected from its proper nervous relations. But much smaller grafts than any of these may be met with in imperfect castration, and in such circumstances the male characters are correspondingly ill-pronounced. One must in fact regard the external character of maleness as a quantity which varies proportionally with the amount of gland-tissue present. As an example of a minimal development of such characters associated with a correspondingly small amount of gland-tissue, we may adduce the following observation :— A cockerel (impure breed of Plymouth Rock) was castrated when about 6 weeks old. The bird was killed 10 months after the date of the operation, when it exhibited the following characters. The head presented no male development of comb er wattles. As indications of maleness, however, are the full development of the neck-hackles, a certain development of saddle-hackles, the presence of a few straggling badly curved feathers amongth those of the tail, and the growth of short blunt spurs on the legs. It may be noted that the occurrence of spurs in the hens of this breed is not known, except in the case of old birds. The bird took no notice of the hens with which it was habitually kept. On dissection, no trace of either testicle was discovered at the normal site, and no graft, with the exception of a minute nodule the size of a hempseed, which was adherent to the surface of one of the coils of intestine. Microscopic examination of this minute nodule proved it to consist. throughout of testicular tubuli distended with epithelial cells and large numbers of spermatozoa, spermatogenesis being in active progress. Conclusions. From the fact that in the young of the Herdwick sheep and fowl, occlusion of the vasa deferentia does not inhibit the full acquirement of secondary male characters, it is clear, in the first place, that the discharge of the sperm is not in any way the factor responsible for the production of the characters referred to. This conclusion admits of being extended to mean that the produc- tion of secondary characters is not due to metabolic changes set up by a nervous reflex arising out of the mere physical function of the sexual mechanism. This is made still more forcible by the results of incomplete caponisation in those cases where the grafts were found in situations far removed from the normal, and altogether disconnected from the nerve supply proper to the testicle in its natural position and connections. Such grafts, devoid as they are of any channels communicating externally, and consisting as they do, of tubuli only, are virtually AS aS =~), 58 The Acquirement of Secondary Sexual Characters. {Dee. 14, ductless glands, and the metabolic results arising from their function must, as in analogous cases elsewhere, be attributed to the elaboration of an internal secretion and its absorption into the general circulation. What particular cell elements are concerned in the production of such a secretion cannot as vet be stated. Various possibilities arise which demand the test of further experiment. The function of spermatogenesis, although not itself the whole or sufficient cause, may be the initial factor of a dual or even a more complex process. It is quite within the bounds of possibility that certain of the epithelial cells within the tubuli may produce a pro-secretin such as is produced within the intestinal epithelium; that the chemical changes accompanying spermatogenesis in other of the cells of the tubule may lead to the conversion of this pro-secretin into a secretin, much as the acid chyme does in the case of the pro-secretin present in the intestinal cells ; and that the secretin so formed may, without being shed into the lumen of the tubule, be transferred to the lymph spaces, and thus eventually reach the general circulation, and incite those metabolic changes in distant parts of the body which disclose themselves as secondary sexual characters. The intimate connection that arises in the process of spermatogenesis between the spermatoblasts and the “sustentacular” cells is a phenomenon not yet explained; this phenomenon possibly coincides with the interaction suggested. In regard to the interstitial cells of the stroma, they have characters so unmistakeably glandular that some secreting function, probably a sexual one, must be assigned to them, and they may, of course, take a part in the elaboration of such a secretion as that suggested. But the great variation in the proportion of such cells present in different forms of mammals makes it difficult to formulate any hypothesis to test by way of experiment, and we are not as yet na position to make any statement inregard to them. We have to acknowledge our indebtedness to Mr. George Jonas, of Duxford, for much technical information; to Mr. Marcus Van Raalte for generous help in defraying portion of the expense incurred by the work; and to Mr. C. S. Wallace, Mr. H. J. Marriage, and Dr. H. C. Jonas, for assistance in various ways and on various occasions. Preparations illustrating the various observations referred to are now in the museum of the Royal College of Surgeons, London. 1903. | The Retrocalcarine Region of the Cortex Cerebri. a9 “The Morphology of the Retrocalearine Region of the Cortex Cerebri.”- By G. Exiiot SmitH, M.A. M.D., Fellow of St. John’s College, Cambridge, Professor of Anatomy, Egyptian Government School of Medicine, Cairo. Communicated by Professor A. MACALISTER, F.R.S. Received December 1, 1903,—Read January 28, 1904. Although many writers, amongst whom MHenschen, Vialet and Ramon y Cajal may be specially mentioned, have devoted a con- siderable amount of attention to the study of the white streak in the occipital cortex cerebri, which Gennari first described in 1776 as ‘“lineola. albidior admodum eleganter,” no one, so far as I am aware, has ever used this feature as a guide to the identification of homologous areas and sulci in different brains. On the contrary, it has been employed as evidence that the furrows on the surface of the hemi- sphere have no value for the orientation of physiological areas, seeing that it occupies the edges of two adjoining gyri and the floor of the sulcus between them.* In this preliminary note I hope to demonstrate that Gennari’s stria is a sure criterion for the identifica- tion of three or perhaps four sulci. I began the study of its distribu- tion in the brain of Man and that of the Apes to test the accuracy of the homology which I had suggested between the sulcus occipitalis lunatus (mihi) in the former and the so-called “ Affenspalte ” of the latter.— In the course of these investigations I found that the stria-bearing cortex (or “area striata occipitalis,” as it may be called) presented such definite relations to the calcarine sulcus in the Human brain, that it could be used as a ready and sure means of determining the homologue of this furrow in the Apes and the other Mammalia. The accompanying diagram of the mesial surface of the right occipital region of the brain of an adult male Egyptian Fellah will make this relationship clear. The drawing is schematic in as much as all the submerged gyri are represented as though they were exposed on the surface ; the distribution of the stria Gennari is shown by the punctate shading. The furrow commonly known as the “fissura calcarina” consisted, in this case, of four separate elements, of which the most anterior one alone was strictly entitled to Huxley’s term “sulcus calearinus.” ‘The other three furrows (7!, 7? and 7?) represent the sulcus for which in a * Oscar Vogt, ““Zur anatomischen Gliederung des Cortex Cerebri,” ‘Journal f. Psychologie und Neurologie,’ vol. 2, part 4, 1908, p. 168; see especially Plate 11, fig. 1. + “The so-called ‘ Affenspalte ’ inthe Human (Egyptian) Brain,” ‘ Anatomischer Anzeiger,’ 1903, pp. 74-83. 60 _ Prof. G. Elliot Smith. Zhe Morphology of — [Dee. 1, previous work | have suggested the name “ retrocalcarinus.”* In that memoir I have emphasised the fundamental distinction between the calcarine and the retrocalcarine sulci and have discarded the customary Fie. 1.—Diagram of the Mesial Aspect of the Occipital Region of the Right Cerebral Hemisphere of a Male Egyptian Fellah. The area striata is shaded. e . md eo Pe ies Ole. PRAECUNEUS heal str Genn. ~ x- -- } 7 [IES sulc. cale,_ % a z ‘ eh \ 0 — a str Genn. oH) Q=== Sulc. calfat.— 3 ee } = = Sulcus retrocalcarinus anterior; v7 = Sulcus retrocalcarinus intercalatus ; v* = Sulcus retrocalearinus verticalis ; « = Sulcus limitans dorsalis arez striate ; y = Sulcus‘limitans ventralis arex striate. The fossa parieto-occipitalis consists of the depression in which the three sulci—incisura parieto-occipitalis, suleus paracalcarinus, and sulcus limitans precunei—are submerged. application of the term calcarine to the latter furrow, which shares so few features in common with the former, and is of such subsidiary importance. The distribution of the area striata indicates the same distinction in a striking manner. For, whereas the sulcus calcarinus * “Qn the Morphology of the Brain in the Mammalia,” ‘Linn. Soc. Trans.,’ 2nd Series, Zoology, vol. 8, 1902, p. 386. 1903.] the Retrocalcarine Region of the Cortex Cerebrr. 61 represents the anterior boundary line of the stria-bearing cortex, the sulcus retrocalcarinus lies wholly within that area (fig. 1). If a section be made through any part of the true calcarine sulcus (except its extremities), the ventro-caudal lip of the furrow will be found to contain the stripe of Gennari, which stops sharply opposite the bottom of the sulcus and does not invade its dorso-cephalic lip (fig. 1, A). I have demonstrated this relationship in a large series of Egyptian and Soudanese brains. It is clear that the true calcarine sulcus accurately marks the line of separation of the area striata and the gyrus fornicatus. In a previous work* I was led to the conclusion that the calcarine sulcus on the mesial surface and the suprasylvian sulcus on the lateral surface of the cerebral hemisphere are probably due to some causes other than the mere general expansion of the neopallium, because they occur with such remarkable constancy in the most diversely-specialised Mammalian Orders, in which the mechanical conditions moulding the growing cortex must be far from uniform. In the case of the suprasylvian sulcus I stated reasons for believing (p. 410) that these other causes were to be found in the unequal rates of growth of the receptive area for auditory Impressions and a more dorsal region performing some other function. The true calcarine sulcus is also probably the result of the dis- proportionate expansion of two neighbouring areas of different physiological significance. If we admit the validity of the teaching of Henschen, Vialet and others, 7 who regard the visual centre as being exactly limited to the region distinguished by the stripe of Gennari, we can conclude that the true calcarine sulcus is caused by the unequal rates of growth of the visual area and the cortex in front of it, which performs some other function. If the brain of any large Carnivore or Ungulate be examined (in sections either of the fresh brain or after staining according to Weigert’s method), the area striata will be found to be limited anteriorly on the mesial surface of the hemisphere by the retrosplenial part of Krueg’s “splenial sulcus.” The stria extends into the caudal lip of the sulcus and ceases abruptly opposite the bottom of the furrow. ‘This affords a striking confirmation of the view, which I put forward last yeart on general morphological grounds, that the retrosplenial part of Krueg’s splenial sulcus in the Ungulata, Carnivora, and other Mammalia is the strict homologue of the carcarine sulcus (in the restricted sense just explained) of the Human brain.§ * Op. cit., supra, ‘ Linn. Soc. Trans.,’ 1903. + See 8S. E. Henschen, ‘‘ Revue Critique de la Doctrine sur le Centre Cortical de la Vision,’ XIIle Congrés International de Médicine, p. 130, Paris, 1900, trans- lated by Dr. Dor. { Op. cit., supra. § Op. cit., ‘Linn. Soc. Trans.,’ 1903, p. 376 et seq. 62 Prof. G. Eliot Smith. Zhe Morphology of — [Dee. 1, If a section be made through any part of the retrocalearine suleus (or sulci) the stria Gennari will be found to occupy both lips of the furrow (fig. 1 B). Moreover, it extends upward for a short distance into the cuneus and downward for a similar extent into the gyrus lingualis. It sometimes happens (as in the specimen which I have represented in fig. 1) that superior (7) and inferior (y) limiting furrows of this area striata make their appearance. When the posterior end of the retrocalcarine sulcus is bifid, or when it is represented by a separate vertical furrow (fig. 1, 7°), the area striata becomes extended so’ as to completely surround the depression. In some cases‘ the retrocalcarine sulcus may be placed close to the superior margin of the hemisphere; its anterior part may be less than 1 cm. distant from the fossa parieto-occipitalis, in fact, this sulcus or its separate constituents may occupy in different human hemispheres any position within the triangle bounded below by the tentorial margin, above by the dorsal edge, and in front by the fossa parieto-occipitalis. But wherever this retrocalcarine sulcus or any of its sub-divisions may happen to be placed, it will be found to be invariably within the area struta. The obvious inference is that the sulcus retrocalcarinus is produced by the folding of the visual cortex itself. The variability of its form and constitution points to the conclusion that it is the result of the mechanical conditions to which a limited cortical area must be subjected when it begins to expand. In his great monograph on the ‘‘Surface Anatomy of the Cerebral Hemispheres,”* Cunningham has expressed the opinion that there is no “posterior calcarine fissure” (sulcus retrocalcarmus mihi) in the Apes, the so-called “‘calcarine fissure” being the representative of the “stem” only (z.¢, the sulcus calcarinus mihi) of the Human brain (p. 42). I have already strongly opposed this interpretation on general morphological grounds, because the retrocalcarine sulcus of the Human brain is often apparently formed by the backward prolongation of the true calcarine sulcus, as Gustaf Retzins has demonstrated. The distribution of the area striata in the Apes enables us to settle this matter decisively. I have examined sections of more than 50 hemispheres of the genera Cebus, Cercopithecus, Macacus, Papio, Cynopithecus, Semnopithecus, Hylobates and Sima to determine the distribution of the stria Gennari, and have found that it oceupies both. lips of the greater portion of the sulcus commonly called calcarine. This shows that the greater part of this furrow represents the sulcus retrocalcarinus of the Human brain. The examination oi foetal brains of Semnopithecus seems to indicate that the retrocalcarine sulcus is very precocious in those brains in which the area striata is of great extent. It may develop con- * © Royal Irish Academy Memoirs,’ 1892. + ‘Das Menschenhrin,’ 1896. 1903. ] the Retrocalcarine Region of the Cortex: Cerebri. 63 temporaneously with or perhaps even before the true calcarine sulcus, or, as seems to be the commonest method, in most Apes (certainly in Hylobates) one great fossa develops as the representative of the con- joint calcarine and retrocalcarine sulci. It is certain that the mere chronological order in which sulci develop in different brains is of little value in the identification of their homologies, so that we cannot place such implicit reliance on this method as Cunningham* seems to do when he regards it as the basis on which the accurate comparison of the sulci in the Human and Simian brain must rest. In many Apes the area striata, which forms the walls of the retro- calearine sulcus, may not reach the surface of the cuneus or the lingual gyrus. In the brains of certain large Ungulates and Carnivores there is a furrow behind the calcarine (splenial) sulcus, which occupies a position analogous to that of the retrocalcarine in the Primates. In the case of the Camel’s brain, I suggested the definite homology of this furrow with the primate retrocalcarine.t In a series of hemispheres of Camelus dromedarius, 1 have found that the distribution of the area striata closely follows this sulcus. When, as often happens, this retrocalearine sulcus bends upward to the dorso-caudal angle of the mesial surface, the stria-bearing cortex has a similar peculiar distribu- tion. This places beyond all doubt the question of the identity of the retrocalcarine sulci in the camel and the primates. It has been stated by Henschen{ that the stripe of Gennari does not extend on to the lateral aspect of the Human brain. This is true in some, but by no means all, cases. The exact extent of the caudal prolongation of the stria is subject to a wide range of variation. As the area striata is traced backward alongside the suleus retrocalcarinus, it will be found to expand near the caudal pole*of the hemisphere (fig. 1), but in some cases it does not reach the true lateral aspect. Im more than half of the specimens which I have examined, however, it does extend on to the lateral surface (fig. 2) and reaches almost (in some cases quite) as far forward as the sulcus ocecipitalis lunatus (or “ Affenspalte”). I have chosen to repre- sent as fig. 2 a case in which the sulcus lunatus is placed far back, because this presumably presents the nearest resemblance to the common European type of brain. But it often happens in the Egyptian brain that a large occipital operculum is present, as in the Apes, and in these cases the stria Gennari is prolonged far forward on the lateral aspect into the operculum. It has been known for many years that the stria extends into the occipital operculum in the Apes, but I have been unable to find any = Op. cit. Devi + Op. cit., ‘Linn. Soc. Trans.,’ p. 389, fig. 55. * Op. cit., supra. 64 Prof. G. Elhot Smith. Zhe Morphology of — (Dee. J, reference to a more detailed account of its distribution. In most of the brains of Apes that I have examined, the area striata extended forward on the surface of the operculum, but often failed to reach as far as its anterior free margin; in some cases it ceased abruptly at a point 4 or 5 mm. behind the edge of the operculum. In the Cercopithecidz (in which the operculum is relatively biggest) the stria Gennari extended upward as far as the dorso-mesial edge of the hemisphere, and laterally (or ventrally) as far as the sulcus occipitalis inferior. The sulcus occipitalis superior les in the midst of the area Fig. 2.—Horizontal Section through the Posterior Part of the right Cerebral Hemisphere of an Egyptian. t ) t y = Sulenus retrocalearinus. The thick line represents the stria Gennari. It is quite exceptional for the suleus collateralis to be prolonged into the area striata as in this diagram. striata. Thus, if the whole stria-bearing cortical region were spread out in one plane, it would present a racquet-like shape—the handle of the racquet corresponding to the retrocalcarine cortex and the expanded part to the occipital operculum. In the Human brain the ‘handle of the racquet” becomes greatly enlarged at the expense of the rest of the area striata. But its form is subject to a very wide range of variation, whick I shall describe in detail in a forthcoming memoir. [ Addendum, January 11.—The so-called “calcarine fissure ” in the Apes is a complete involution of the whole of the mesial part of the area striata. The so-called ‘“calearine fissure” (of most writers) in 1903. ] the Retrocalcarine Region of the Cortex Cerebrr. 65 the Human brain usually consists of an anterior part, which is an anterior limiting sulcus of the area striata, and a posterior part, which is a mere indentation (or indentations) of part of the mesial area striata, therefore it is not exact to speak of these similarly named furrows as being strictly homologous. | “On the Acoustic Shadow of a Sphere.” By Lorp RAYLEIGH, O.M., F.RS. With an Appendix, giving the Values of Legendre’s Functions from P, to P,, at Intervals of Five Degrees. By Professor A. LopcE. Received December 28, 1905,—Read January 21, 1904. (Abstract.) The problem here considered is that of the intensity of sound at the various points of a rigid and fixed sphere on which plane waves impinge, or reciprocally the intensity at a distance,in various directions due to a source of sound situate upon the surface of the sphere. The analytical solution is readily given, but in the interpretation everything depends upon the ratio of the wave-length (27/k) to the circumference (2zrc). If ke be small, the sphere has little effect. In my book on the »“Theory of Sound,” § 328, I have considered (but only-for certain special directions) the case of kc= 2. The extension to various directions is now given; and the calculation is pushed to the case of ke = 10, about as far as is practicable. For this purpose the values of Legendre’s Functions up to Poo are required. ke= 10: 6. 4 (F2 + G2). 6. 4(F?+@). lig A eens a 3°8300 NO Sire ey eroe 1°06117 LS i eae 3°8176 NOW ee eas 0°56815 PRON ists. occsine 3°7148 Thoma tte eta as 0°27890 1 ES, gene ee aaa 3°4978 ANS OR Mecano. ae 0°13338 BOM ys coos. : 3°1098 GD veer 0:09492 A Rees eae 2°4984 WO cal cme cince O12591 EO) aac One aca yey ke) TIAA ina eee eee 0:°69395 TSO ear oes: 1°09263 The table gives the intensity in directions making angles 6 with the radius which passes through the source. On the same scale the intensity would be unity were the sphere removed. The most - interesting feature is the existence of a fairly good shadow between 135° and 170°, and the subsequent rise of intensity in the neighbourhood VOL LXXIII. F 66 Dr. E. F. Bashford and Mr. J. A. Murray. [Jan. 12, of 180°. This corresponds in some degree with the bright spot in the centre of the optical shadow of a circular dise. The problem which arises when both the source and the point of observation are situated upon the sphere is more difficult. It is treated less completely, but some results of interest are obtained for the case ot kG — 10: The Appendix by Professor A. Lodge contains tables of Legendre’s Functions up to Po for angles ranging at intervals of 5°, accom- panied by a statement of the method of calculation. It is believed that these values may prove useful in other physical investigations. “The Significance of the Zoological Distribution, the Nature of the Mitoses, and the Transmissibility of Cancer.” By E. F. BasH¥FoRD, M.D., and J. A. Murray, M.B., BSe. Communi- cated by Professor J. Rose Braprorp, F.R.S. Received January 12,—Read January 21, 1904. [PuatE 2. | The object of this communication is to relate some results of the work conducted under the immediate direction of the Executive Committee of the Cancer Research Fund during the past year. We believe that these results will convince others of the important practical assistance which biologists generally can give in the further elucidation of certain problems of cancer which must be settled before preventive and curative measures can be devised. It will also be made evident that the elucidation of cancer is something more than a problem of human pathology. We shall adduce evidence tending to show that the wide zoological distribution, the character of the mitoses, and the transmissibility of cancer, are nearly related phenomena with a common basis. The fundamental significance of ascertaining the extent of the zoological distribution of cancer was recognised by the Cancer Research Fund from the first, and determined the prosecution of definite lines of inquiry, not only with the object of eliciting new facts in regard to the zoological distribution itself, but also with the object of dis- covering cancer in animals well adapted to cytological and experi- mental observations. Loological Distribution. Within the past year specimens of malignant new growths have accumulated from all the domesticated animals and from many other 1904. ] On the Zoological Distribution, etc., of Cancer. 67 vertebrates. The appended List shows the abundance of the material which has thus been examined. The List includes also specimens which we have been privileged to examine through the courtesy of investi- gators abroad, who, subsequently to the imauguration of the Cancer Research Fund, have published descriptions of malignant tumours in the lower vertebrates. It is noteworthy that such growths have been obtained, not only in domestic animals, but also in animals living in a state of nature: wild mouse, codfish, gurnard. The clinical, pathological, anatomical, and microscopical characters of these new growths are identical with those found in man in all essential features, although the animals themselves are drawn from the different classes of the vertebrate phylum. A detailed histological description of the various tumours examined will not be attempted here. Only the general significance of the observations in relation to the incidence of cancer in man will be emphasised. The great diversity of the habitat, food, and conditions of life generally of the forms in which malignant new growths occur relegates such external conditions to a subsidiary réle in determining the inci- dence of the disease, and shows that the essential factors must be sought in the potentialities which reside in the cells constituting the living body. The list of specimens, while giving no safe basis of deduction as to the relative incidence of cancer in the different classes of vertebrates, ‘or of the comparative susceptibility of the different sites of the body, is extremely suggestive. The large number of epitheliomata obtained in the Bees and dog indicates very clearly that malignant new growths are recognised according to the ease with which animals can be examined, and the length of time they are kept under observation. In the same way, numerous malignant new growths have been discovered in the internal organs of cows during the inspection at abattoirs. Stated generally, it may be said that malignant new growths are known to occur chiefly in animals which are habitually examined with care, and are unrecorded in forms which are difficult to examine or do not reach old age in considerable numbers. The figures are not sufficiently extensive to determine accurately the age incidence of the various types of new growths in different animals, but a relatively higher incidence in old age is apparent. The Phenomena of Cell-division in Malignant New Growths. The progressive increase in size of malignant tumours is due to the division and increase in size of their constituent cells. The process of cell-division is usually indirect, mitotic division of the nuclei preceding the division of the protoplasm. The protoplasm division is frequently F 2 681): Dr. E. F. Bashford and Mr. J. A. Murray. [Jan. 12, omitted, and multi-nucleated cells are formed, and these may subse- quently enter on mitosis, giving rise to pluripolar figures. Amitosis or direct nuclear division also occurs, but its full significance has not yet been determined. It is, however, evident that the occurrence of amitosis does not signify degeneration. The amount of chromatin entering into the equatorial plate of the mitoses of malignant new growths had long been recognised as subject to variation (hyper- chromatosis, hypochromatosis, of von Hausemann, 1893), but a new light has been thrown on this phenomenon by a paper communicated to the Royal Society on December 10, 1903, by J. B. Farmer, F.R.S., J. EH. S. Moore, F.L.S., and C. E. Walker, entitled “ Resemblances exhibited between the Cells of Malignant Growths in Man and those of Normal Reproductive Tissues.” These observers found that while the growing margin of carcinomata and sarcomata presented mitoses similar to those found in other tissues in repair and inflammation, certain cells in the deeper layers, after a slight increase in size, entered on mitosis with ring chromo- somes similar to those found in the heterotype division of spore mother-cells of plants and spermatocytes of animals, and like these, with chromosomes in number only half that characteristic of the mitoses of somatic cells. Mitoses similar in character to the somatic divisions, but with reduced number of chromosomes, were also seen (homotype), corresponding to the divisions in the sexual generation of plants and the second ripening divisions of animals. From these observations the authors concluded that malignant new growths were virtually reproductive tissue arising in abnormal situations and possessed of an independence and power of growth like that of the testis in the mammalian body. This striking sequence of characteristic mitoses had been found in all malignant tumours examined, and was absent in those of benign character. We at once determined to communicate with the authors, who with great courtesy afforded us an early opportunity to examine their preparations. It was then decided to determine how far similar phenomena were characteristic of the malignant new growths occurring in animals. The result has been a complete confirmation of Farmer, Moore, and Waiker’s observations in tumours from the trout (Mr. Gilruth’s and Miss Plehn’s cases of adeno-carcinoma), the mouse (two cases of adeno-carcinoma, Jensen’s epithelioma), and the dog (mixed cell sarcoma, round cell sarcoma, squamous cell carcinoma). In the columnar cell carcinoma of the trout the phenomena were especially distinct, the small number of chromo- somes* (24, 12), the striking contrast between the long slender * The achromatic figure has always been carefully studied as a control to the observations made on the chromatin of the mitosis. When the chromosomes have been counted, this has been done on Sections 5—10 yu thick, mounted in series. 1904. ] On the Zoological Distribution, ete., of Cancer. 69 chromosomes of the somatic mitoses, and the rings of the heterotype division, being of diagrammatic clearness. Homotype mitoses occurred, but were very few in number. Mitoses in the stroma are relatively scanty in the tumours at our disposal, but such as have Fie. 1.—Adeno-carcinoma of Trout. Fie. 2.—Epithelioma of Mouse. Somatic Hometype amphiaster. Reduced prophase and amphiaster. number of chromosomes, arranged transversely in spindle, and showing longitudinal splitting. Fie. 3.—Epithelioma of Mouse. Heterotype mitosis, late prophase. Ring loop and bivalent chromosomes. been seen are somatic in character. Farmer, Moore, and Walker record a similar result. In the mouse it has been possible to compare the mitoses in the testis, and those occurring in the irritation produced by iodine, with the result that the significance of the mitoses in cancer has been further confirmed. 70 Dr. EK. F. Bashford and Mr. J. A. Murray. [Jan. 12, The following points are of great importance in these observations. A complicated sequence of cell-changes has been found to be charac- teristic of carcinoma and sarcoma alike. This sequence is the same as that which initiates the origin of the sexual generation in plants from the asexual, and is terminal in the history of the sexual cells in animals. It must be noted also that all the cells of the malignant new growths do not undergo the reducing division, a certain number differentiate in the direction of the tissue among which they have arisen, and in the secondary growths when present, somatic mitoses occur in the growing margin, which it will subsequently be shown is also a feature in the growth of cancer when transferred to a new host. Fia. 4.—-eterotype amphiaster, chromo- Fie. 5.—Heterotype amphiaster. Ring somes arranged longitudinally on chromosomes, 12 in number. Adeno- spindle. Reduced number. All the carcinoma of Mouse. (The nucleus chromosomes are not figured. was contained in two consecutive sections). The Transnussibility of Malignant New Growths from one Ammal to Another. The transmission of cancer from man to animals, or from one animal to another of different species, has never been successfully performed. Successful transplantation experiments have been made, however, from animals suffermg from malignant new growths to others of the same species. The most exhaustive observations in this connection have been made by Jensen and Borrel on mice. Professor C. O. Jensen, of Copenhagen, most generously placed at the disposal of the Cancer Research Fund a portion of one of his experimental tumours,* and with this, and another tumour occurring naturally in a tame mouse, similar experiments have been performed. We have thus been able to confirm Jensen’s observations by microscopical examination of the tissues at the site of inoculation at short intervals, and have found that the new tumours which develop arise from the actual cells introduced. While many of these degenerate, a few * Sent by post under such precautions as to preserve sterility. The trans- plantations were effected by the Cancer Research Fund five days after the tumour was posted in Copenhagen. 1904. ] On the Zoological Distribution, ete., of Cancer. VE remain of normal appearance, and these gradually increase in number. In the earliest transplantations mitotic division is absent, and it is not till later, when a considerable mass has arisen, that mitoses appear. The earliest mitoses we have been able to observe have been of the somatic type. Transplantation is, in fact, identical with the process of metastasis as it occurs in the individual providing the tumour. It is remarkable, however, that the tumour of Jensen’s experiments does not produce metastases naturally, and its malignancy is only evidenced by its progressive growth, and the undifferentiated character of the cells. The process is in no sense an infection, the tissues of the new host not participating in the formation of the new parenchyma. In this interpretation we are in agreement with Jensen and differ from Borrel, who conceives the results to be due to the agency of a wrus CUNCETCUL. The origin of the stroma has not been accurately determined. The power of growth of this tumour is remarkable. In every mouse in which the transplantation succeeds, the new growth may attain a weight equal to that of the animal itself, and over 400 such trans- missions have been effected by Jensen in Copenhagen and the Cancer Research Fund in London. A mass of tumour, 16 Ibs. in weight, has thus actually arisen from the original one, and that without participa- tion of the cells of the various hosts and without manifest change in structure. This great power of growth is a phenomenon un- paralleled in the mammalia, and indicates the potentialities in cases in which widespread dissemination has occurred before death in a human patient. The experimental transmission of carcinoma shows that we must distinguish between the problem of the genesis of a malignant new growth, and that of the conditions which permit of its continued existence. While the conditions leading to the initiation of malignant tumours are relatively infrequent, we have examined upwards of 1000 tame mice, and have discovered two with cancer; once begun, this proliferative activity can, under favourable conditions, persist for a long time unaltered, and can give rise to masses of tissue of great size, having no relation to the restrictions which limit the growth of adult organisms in a large proportion of healthy animals. The phenomena of cell division, indicating a similarity to the normal reproductive tissues, may help to explain the nature of this great power of multiplication, but leave the problem of cancer genesis practically untouched. They give, however, important indications of the character of the processes on whose elucidation the solution of the question depends. The wide zoological distribution of malignant new growths—its limits are not yet determined—indicates that the cause of cancer is to be sought in a disturbance of those phenomena of 72 Dr. E. F. Bashford and Mr. J. A. Murray. [Jan. 12, reproduction and cell-life, which are common to the forms in which it occurs. Our observations on animals show that malignant new growths are always local in origin and of themselves produce no evident constitu- tional disturbances whatsoever. ‘These facts are in full accord with accumulated clinical experience in man. In connection with diagnosis and statistics we have already emphasised the importance of the absence of specific symptoms. The evidence we have advanced that cancer is an irregular and localised manifestation of a process, other- wise natural to the life-cycle of all organisms, probably explains why it is that malignant new growths and their extensive secondary deposits, gud cancer, are devoid of a specific symptomatology. We desire to add the accompanying dated note because we find that conclusions which have been drawn by others are attributed to us. [Nore.—We find that the guarded terms in which the points of importance are emphasised may lead to a misconception of our inter- pretation of the facts. The cells which have undergone the reducing division are not responsible for the active invasion of surrounding tissues, nor for the production of metastases; the cells dividing somatically are responsible for both. The number of heterotype mitoses may not stand in any relation to the degree of malignancy and their absence is only presumptive evidence of the benign character of a tumour. We postulate nothing as to the future of the cells which have undergone the reducing division, though we believe the latter to be a terminal phase in the life cycle of cancer cells as it is in the history of sexual cells in animals. The local origin, and the expansive and infiltrating growth of cancer in its relation to surrounding tissues, while respecting its own proper elements, is the criterion of its malignancy. This stamps it as belonging to a new cycle comparable in its entirety to the whole organism which it is invading, rather than to any one of its tissues, reproductive or otherwise. We intentionally restricted our original statement to recording the facts, and only such general conclusions as could be irrefutably drawn from them.—January 25, 1904]. We cannot here make full acknowledgment to those who have assisted our inquiry in this country, but our indebtedness may be expressed to those observers abroad who have recorded isolated instances of cancer in animals, and have so generously furthered our investigations by placing material or specimens at our disposal. In particular, we desire to thank Professor Borrel, of the Pasteur Institute ; Professor C. O. Jensen, of Copenhagen; Mr. J. A. Gilruth, Chief of the Veterinary Department, New Zealand; Pro- fessor Landau and Dr. L. Pick, of Berlin; and Dr. Marianne Plehn, Munich. 1904. ] On the Zoological Distribution, ete., of Cancer. 73 Without the generous co-operation of these and many others it would not have been possible within so short a time to have covered the extensive ground indicated in this paper. Fie. 6.—Homotype amphiaster. Hpithelioma Mouse. Transversely arranged chromosomes, longitudinal splitting, reduced number. All the chromosomes are not figured. List of Specimens of Malignant New Growths examined by the Cancer Research Fund Committee during the year 1903. | | : : | Microscopic character. | Animal. A ge. | Primary site. 1 | | | | | Cow ............| Aged | Orbit ........| Carcinoma, large polygonal cells. | eer as. i Ramen 6) 5.44. | ‘ spheroidal cell. Sol ca a se | Re een a squamous cell. Oar mew oe » b Flee ai lath as ”? ” ” | eae Sip LENGE cco cone! “3 cubical cell. | | SoMeiar ciel sic, «16 sis 's« 6 | oc cette ee, 3 59 ap H y ee Pleura: Ee squamous cell. | | (secondary) | a eee. oe oe a's i ? Gastric gland | A is 2 | | (secondary) | | | 13 GG ae eres iE) ol Perineum-.:)..: | Melanotic, sarcoma. | | imme s sc. «| Need |) Ovary once... | Sarcoma. Hoss Be ¢ a Suprarenal ....| 02, | GS eae aie Aged | Ovary ........| Carcinoma. Dae ves Be fi a : | daw.........-.| Osteo-sarcoma. 55 ene ate bee Adrenal ......| Sarcoma. De eee || Need!) | River .......-| Carcinoma, cubical: cell. DR se sccter cee | ns Bowel ........| Sarcoma, spindle cell. | Bees) SOU E et ee By iINeck PEGE Se 3 ss si | 5) se eeeeees--.| Aged | Face and neck .| Carcinoma, squamous cell. Betleifer . sce. 2 Side of thigh .. Melanotic sarcoma. | OP fai.) IE, Mamme......' Carcinoma (scirrhus). ae ii5 : 10-11 | Upper lip.....| Epithelioma. 93 ev eNUpper lipiis. | Fibro-sarcoma. 10-11 Carcinoma, squamous cell Abdominal in Sarcoma, round cell. gland | Sympathetic Fibro-sarcoma. glands | - il eR sai Be .. | Testis ........! Sarcoma, mixed eell. | ” / 74 Dr. E. F. Bashford and Mr. J. A. Murray. [Jan. 12, List of Specimens of Malignant New Growths—continued. Wild mouse .... | Mouse (Pick) WEG eas Stet 535 ca hy tl aban 3 Even (Pick): wks 2: Horse, gelding ... as ks ay i stallion! Mirek. Seimc rere s Horse .. Sine View ea seen tite MFT ONBG? ies ew ec Sheep nein ”) IPAL’ <5 Pores agterema Ble White mouse..... Yellow mouse... Mouse (Jensen) .. Pe ai orreliy iy Indian parakeet .. Giant salamander | EH) | iuebyas COdM i eos. a pale arta Giuenard : 5% 2%%ecl Trout (Gilruth) .. eenebdelam i 5.2 ”” 39 Age. 10 12? | Mammary | Pectoral muscle | Floor of mouth | | Primary site. Microscopic character. VER vedas ae whe Moth iret. < AMNUIG Ys ths )e. See's Leg (subeu- taneous) Sup. max., or- bit, gld., 1. jaw Anai tumour .. Spleen, liver, stomuch Ibis BEA Palate, cervical Mandible Liver Sub-max glands Inguinal mamme Axillary tumour glands Leg (subeu- taneous) Asxaillay 3 Groin # bb) 3 Jaw 3 Baek e A OMe Wes oe Floor of mouth Vestis ... Air-bladder . Abdominal tumour | os » ” »? Carcinoma, columnar cell. a squamous cell. 29 9> Sarcoma, round cell. 5 spindle cell. Carcinoma (sebaceous adenoma). Sarcoma, spindle cell. - small round cell. Carcinoma, squamous cell. Osteo-sarcoma. Carcinoma, squamous cell. Epithelioma. Carcinoma, squamous cell. 99 39 39 bP) ) 4s columnar cell. s squamous cell. ” 9) 9 Medullary carcinoma. Osteo-sarcoma. Carcinoma, cubical cell. Sarcoma, mixed cell. Adeno-earcinoma (localised kera- tinisation). Adeno-carcinoma. Carcinoma, spheroidal cell. Hpithelioma. Adeno-carcinoma. bP) ” Epithelioma. Adeno-carcinoma (sweat glands). Carcinoma, squamous cell. Squamous cell epithelioma. Myxo-sarcoma. Cystic adenoma, malignant. Sarcoma, spindle cell. Adenoma, malignant. Carcinoma, columnar cell. | 1904.] On the Zoological Distribution, ete., of Cancer. . 75 Fie. 7.—Malignant adenoma of Gurnard. Primary tumour. x 2250 (reduced). Peritoneal cavity. 76 On the Zoological Distribution, ete., of Cancer. [Jan. 12, a 4 is 7 Fie. 8.—Spindle cell sarcoma. Codfish. Secondary nodule, wall of swim bladder. x 1500 (reduced). DESCRIPTION OF PLATE. ADENO-CARCINOMA OF TROUT. (1) The chromosomes in the stroma mitoses are long slender V-shaped locps, which split longitudinally, 24 in number. They are somatic in type ; compare fig. 2 from the tumour. (2) Somatic equatorial plate in margin of tumour, seen from the pole, slender V-shaped chromosomes 24 in number arranged transversely on the spindle, and showing longitudinal splitting. All the chromosomes are not reproduced in the figure. (3) Microphotograph (untouched) of somatic equatorial plate. (4) Microphotograph (untouched) of heterotype amphiaster seen from pole. (5) Heterotype amphiaster (polar view). Small ring chromosomes present in reduced number (12). All the chromosomes are not reproduced in the figure. This is a drawing of the mitosis photographed in fig. 4. (6) Heterotype amphiaster, lateral view (one centrosome only in this section). Ring chromosomes in reduced number arranged longitudinally on the spindle. All the chromosomes are not figured. is. yj. Soc. Proc., vol. 13, mg) Bashford and Murray. 1904.] Conjugation of Resting Nuclei in Epithelioma of Mouse. 77 “ Conjugation of Resting Nuclei in an Epithelioma of the Mouse.” By E. F. Basurorp, M.D., and J. A. Murray, M.B., BSc. . Communicated by Professor J. Rosk Braprorp, F-.R.S. Received January 28,—Read February 4, 1904. In a previous communication we have drawn attention to the fact that the power of cell proliferation, which has been proved to occur in an epithelioma of the mouse (Jensen), is a phenomenon unparalleled in the mammalia. A mass of tumour, 16 lbs. in weight, has been produced by artificially transplanting portions of the original growth and its descendants. In seeking to throw light on this fact, we have studied carefully the phenomena which follow the transplantations of portions of the tissue to new sites, and have found that the tumours which arise are the genealogical descendants of the cells introduced. We have studied the growth of the tumours which arise at successive stages of 24 hours. In a tumour removed on the eighth day, and less than half a split pea in size, conjugation of resting nuclei has been observed. To take a specific case, the nuclei of two adjacent cells are continuous through the cell wall by a tube-like bridge, in the middle of which a strand of nucleolar substance with fusiform swellings in either cell is visible. The cells of this particular case are adjacent to. the stroma, and close to the outer surface of the young tumour. Oo Dr. R. Staehelin. On the Part played by — [Dee. 11, “On the Part played by Benzene in Poisoning by Coal Gas.” By R. STAEHELIN, M.D., Senior Assistant in the Medical Clinie at Basle. Communicated by Professor E. H. STARLING, F-.R.S. Reeeived December 11, 1903,—Read January 28, 1904. (From the Physiological Laboratory, University College.) In a recent paper Vahlen* has maintained that a difference exists between the poisonous action of coal gas and of carbon monoxide, and Kunkelt has also drawn attention to a similar difference in the case of frogs. In the course of a research which I was undertaking in University College for other purposes, at the suggestion of Professor Starling, I have come across facts which may serve to explain the difference noted by these observers. My first object was to investigate the effect of deprivation of oxygen on the fatigue curve of muscles. To this end, a muscle was hung up in a closed chamber and the atmospheric air driven out by a stream of some other gas. When coal gas was used for this purpose, it was noticed that the muscle rapidly went into rigor mortis, whereas, in nitrogen, it remained excitable for many hours. I set myself, there- fore, to find out which constituent of the coal gas was responsible for this poisonous effect. The experiments were carried out on the frog’s sartorius, since the relatively large surface of this muscle renders it particularly accessible to gaseous poisons. The muscle was hung up in a glass tube, closed above and below by rubber corks. To the upper cork a hook was fastened on which was hung the bony insertion of the sartorius. Through the lower cork a glass tube passed. The movements of the muscle were transmitted to a recording lever by means of a steel needle hooked into the lower end of the muscle and passing through the glass tube. The lever was weighted near its axle, and was after- loaded, and its movements were recorded on a slowly rotating drum, on which the time was also marked by means of a signal. ‘The closed tube was provided with two small lateral openings by which any gas desired could be passed through it, the opening through which the needle passed to the lever being made sufficiently air-tight by means of a small plug of vaseline. In every experiment two similar * Ferchland und Vahlen, “ Ueber Verschiedenheit von Leuchtgas- und Kohlen- oxydvergiftung,” ‘ Archiv fiir exper. Pathologie und Pharmakologie,’ 48, p. 106; Vahlen, “Ueber Leuchtgasvergiftung,’ ‘Archiv fiir experimentelle Pathologie und Pharmakologie,’ 49, p. 245. + Kunkel, ‘‘ Ueber Verschiedenheit,von Leuchtgas- und Kohlenoxydvergittung,” ‘Sitzungsberichte der Physikalisch-medizinischen Gesellschaft zu Wiurzburg,’ 1902, Nr. 4, 5, p. 61. ft 1903.] Benzene in Poisoning by Coal Gas. eo apparatus were employed, so that the influence of two gases could be compared at one and the same time on the two sartorii of the same frog. When coal gas was led through the tube the muscle began to contract 1—7 minutes after the beginning of the passage of the gas, and in 4—1 hour the muscle was maximally contracted and had an opaque appearance. ‘The rapidity with which the contraction came on was proportional to the rapidity with which the gas was passed through the chamber, and varied with the temperature, warmth quickening and cold slowing the process. Once the muscle was fully contracted no recovery took place. This phenomenon could not be due to the CO in the coal gas, since, on passing pure CO through the muscle chamber, the muscle remained excitable as long as in nitrogen, and the fatigue curve took the same course as in that gas. Among the remaining more important constituents of coal gas which might produce this phenomenon, benzene merited first attention, since all aromatic bodies are more or less potent poisons. I therefore tried the effect of passing air through benzene into the muscle chamber. I found that in less than a minute the muscle began to contract, and the contraction reached its maximum height within a short time, the muscle becoming opaque and dead. The other sartorius of the same frog was placed at the same time in coal gas. It began to contract only after several minutes, and the contraction took half an hour to reach its maximum. Benzene vapour therefore showed itself much more poisonous for a muscle than coal gas, evidently on account of its greater concentration. According to Letheby, London coal gas contains only 3:8 per cent. of condensible hydrocarbons, of which benzene forms only a small proportion. I therefore, in another experiment, passed air into the muscle chamber, not through pure benzene, but through water which had been shaken up with benzene. In this case the contraction began 6 minutes after the beginning of the experiment and took 38 minutes to reach its maximum, while the control muscle in coal gas began to contract in 7 minutes and reached its maximum contraction in 40 minutes. It seemed, therefore, highly probable that benzene was really the constituent of coal gas which was responsible for its toxic properties on muscle. If that were the case the passage of coal gas through oil, which absorbs benzene, ought to deprive it of its deleterious effects. This was found to be the case. Coal gas, passed through oil, showed no difference in its effects from pure CO or nitrogen. In the same manner as benzene I investigated the effects of xylol and toluol. If the muscle was supplied with air blown through either of these fluids no poisonous effect was observed even when the fluids were warmed. The absence of effect in these two cases is perhaps to be ascribed to the lower vapour tension of these two substances. The 80 Dr. R. Stachelin. On the Part played by [Dee. 11, muscle was equally unaffected when supplied with air which had been passed over heated naphthaline. I also investigated the effect of certain hydrocarbons of the fatty series, namely, methane, prepared by heating sodium acetate and sodium hydrate, acetylene produced by the action of water on calcium carbide, as well as the mixture of substances obtained by blowing air through petroleum ether. In none of these cases was any effect produced on the muscle. If, however, a trace of benzene was added to petroleum ether, r2gor mortis of the muscle was almost immediately produced. In this case the volatility of the petroleum ether apparently aids the evolution of the benzene. On the other hand, the addition of small quantities of benzene to xylol does not impart poisonous qualities to the air bubbled through the mixture, the benzene being apparently held fast by the less volatile xylol. It seemed, therefore, most probable that the specific poisonous effect of coal gas on frog’s muscle was due to benzene only, and that it was to the presence of this substance in coal gas that the differences observed by Kunkel and Vahlen between the action of coal gas and carbon monoxide were to be referred. Vahlen states that warm- blooded animals and frogs die more rapidly in coal gas than would be expected from its percentage of CO, and also that frogs in coal gas present excitatory phenomena which are absent in pure CO. Although Kunkel denies the presence of any difference between the action of these two gases on warm-blooded animals, he also draws attention to the peculiar effects on frogs of coal gas, which are not produced by other gases free from oxygen, and describes them as “choreiform twitchings and spasms in the neck and legs.” To decide this question the following experiments were carried out .:— I. Three frogs were placed in air-tight bell jars, through each of which coal gas was conducted at constant rate. The coal gas had to pass in each case through a wash bottle, which in A was water, which, of course, left the composition of the gas unchanged, in B oil, which would absorb the benzene. In front of C were two wash bottles, the first one containing oil, the second one benzene. The passage of the gas through the three bell jars was begun at 11.20. At 11.25 all three frogs were restless. Frog C remained then still for a short time, and the breathing became irregular, and twitching occurred in the extremities and back. Movements were chiefly co-ordinated, though there were some twitchings of isolated muscles. After a little time the movements became shorter and less co-ordinated, the legs remained stretched out, and breathing ceased. Frog A betrayed phenomena similar to C, but the spasms were less evident, and came on more slowly. Frog B became quite quiet, the respiration becoming irregular and shallower. At 11.40 B was sitting up in normal position, though the breathing was some- what irregular, while A and C were lying on their bellies, with legs stretched out. At 11.45 all three frogs were taken out. C gave no signs of life, and in ten minutes was quite rigid; B still reacted slightly to stimulation, showed shallow respiratory movements, and graduaily recovered, so that at 1.30 it was apparently normal. Frog A at first showed no response to stimulus, and no respiratory =] 1903.] - Benzene in Poisoning by Coal Gas. 81 movements. By 11.55 it had recovered sufficiently to show both these phenomena, and at 1.30 it was so far recovered that it could recover its position when tnrned on its back, and in a couple of hours later was apparently normal. II. In a second experiment the arrangements were the same as in the first, except that in C the gas which had passed through the oil was allowed to pass through water saturated with benzene instead of through pure benzene. In this case the phenomena in A and C were practically identical, and were similar to those observed in A in the first experiment. We need not, therefore, give fuller details of this experiment. We thus see that, when frogs are exposed to coal gas, motor phenomena are produced, which are absent if the coal gas be pre- viously purified by passage through oil, and that these phenomena can be reproduced if the purified gas be made to take up benzene vapour. The poisonous properties of the gas can be increased by increasing the tension of the benzene vapour. We are, therefore, justified in concluding that the differences between the effects of CO and coal gas observed by Kunkel and Vahlen, depend on the presence in the latter of benzene. The slight motor excitation observed in frogs in coal gas, which had been purified by passage through oil, is exactly similar to that described by Kunkel, as the result of deprivation of oxygen, produced by placing the frogs in nitrogen or CO, as is shown by the following experiment :— III. Two frogs were placed, one in a bell jar. through which CO gas was led, the other in a similar jar through which coal gas was led after passing through oil. Both animals in a short time showed slight twitchings of isolated groups of muscles in the extremities and back, and occasional extensor movements of the hind limbs, which gradually diminished. No difference was observable between the two frogs. In three quarters of an hour they were taken out of the jars, and both recovered within a short time. In order to be certain of the part played by benzene in coal gas poisoning, we must have some idea of the effect of pure benzene on the frogs. I have been unable to find any published experiments over the effects of inhalation of benzene on the frog. Beyer* states that xylol acts as a narcotic poison, like the other odorous substances investigated by him. I have, therefore, made some experiments on the influence of benzene vapour in the presence of oxygen on frogs. TV. A frog was placed in a bell jar, in which a beaker of benzene was hung up. Eight minutes after the beginning of the experiment spasmodic movements and twitchings began in various parts of the body, accompanied by a considerable secretion of mucus. After a few minutes the spasms ceased, the frog lay still with extended limbs, and respiration, which at first was irregular, became gradually shallower and less frequent. Twenty minutes after the beginning of the experiment all respiratory and other movements had ceased. The frog was taken out of the jar, and recovered in a few hours. * Beyer, “ Narkotische Wirkungen von Riechstoffen und ihr Einfluss auf die motorischen Nerven des Frosches.” ‘Archiv fiir Anatomie und Physiologie, physiologische Abteilung,’ 1902, p. 201. VOL. LXXIII. G 82 Dr. R. Staehelin. On the Part played by [Dee. 11, In other experiments in which the frog was left longer exposed to the action of benzene vapour, 7igor mortis came on first in the hinder and then in the fore extremities before the heart had ceased to beat. One peculiarity was observed with regard to the reflex irritability of the animals, under these conditions, which is worthy of notice. Very early in the experiment, at the very beginning of the spasmodic move- ments, the frog reacted very slightly and incompletely to changes in its position, that is, such as would be produced by holding the vessel in which they were placed in an oblique position or turning them on their sides or backs. On the other hand, the reactions to tactile stimuli of the skin were much more pronounced than usual, so that a tap on the one foot might evoke muscular contractions throughout the whole body. When the animals began to become rigid, stimula- tion of a toe of the rigid limb could evoke contractions of the limbs which had not yet become stiff. Thus, in all these experiments the first result of the poisoning was motor excitation, which showed itself at first by co-ordinated movements affecting large portions of the body, and later by twitchings of isolated groups of muscles. A little later the respiration became irregular and finally ceased. The higher reflexes, ¢.g., the reaction to changes in position, were abolished, while the lower were increased. Finally, however, the paralysis became universal, so that also the spinal reflexes were abolished. This stage was followed by a rigidity of the muscles, and last of all the heart ceased to beat. The rapidity of onset of these phenomena is naturally dependent on temperature, being quicker the higher the external temperature. It is evident that we have, therefore, to deal with an action of the poison on the central nervous system. The general spasmodic movements are abolished by previous destruction of the brain and spinal cord ; the twitchings of the muscles and the rigor of the extremities persist, however, in the complete absence of the central nervous system, and occur in a hind limb, the nerve of which has been divided, as rapidly as on the opposite side. The onset is not pre- vented by curarisation, and must, therefore, be ascribed to a direct action of the benzene on the muscles themselves, as has been described at the beginning of this paper. In the poisoning by coal gas this rigor of the muscles was not observed either by myself or Vahlen and Kunkel, probably because the animals die of asphyxia before the small amounts of benzene present in the coal gas have time to bring about their direct effect upon the muscular tissue. As Experiment I shows, the increase in the percentage of benzene in the coal gas is followed by the onset of rigidity in the muscles. Rather more difficult is the explanation of the increased reflex excitability in the later stages of intoxication. It may be that the poisonous effects are first confined to the higher centres. On the “J 1903.] Benzene in Poisoning by Coal Gas. 83 other hand, the increased irritability of the muscles themselves is the chief factor in the spinal reflex hyper-excitability. In coal gas poison, when the lower centres are also paralysed by asphyxia, the reflex excitability disappears. It is thus possible to refer the difference between intoxication by coal gas and that by CO entirely to the influence of benzene, which determines in its first stage vigorous excitatory motor phenomena. In warm-blooded animals the conditions are quite different. Lorraine Smith* found that an addition of 0°65 per cent. of benzene to air had no effects on a guinea-pig, and Santessont did not succeed in producing either acute or chronic poisonous effects by administering benzene to a rabbit by inhalation. In man, too, cases of poisoning by benzene are few and far between. It is possible that the minute quantities, which are absorbed by the lungs, are rapidly oxidised and excreted as an aromatic sulphate. At any rate, in man, benzene plays no part in the poisonous effects of coal gas. Summary of Results. (1) Coal gas produces first excitation and then rigor of the isolated frog’s muscle. : (2) Frogs exposed to coal gas show excitatory phenomena which are absent when the animal is placed in an atmosphere of CO or nitrogen. (3) The specific effects of coal gas on frogs are determined by the presence of benzene in the gas, and can be produced by air containing the same percentage of benzene. (4) There is no reason to suppose that the poisonous effect of coal gas on mammals is determined by anything except its content in CO. * Lorrain Smith, ‘‘ The Poisonous Action of Coal-Gas and Carburetted Water- Gas.’ ‘Report of the Water-Gas Committee,’ presented to both Houses of Parlia- ment by command of Her Majesty, 1899, Appendix VII, p. 127. + Santesson, ‘‘ Ueber chronische Vergiftungen mit Steinkohlenbenzin,” ‘ Archiv fir Hygiene,’ 31, p. 336. 84 Mr. H. H. Dale. [Dees ak, “The ‘Islets of Langerhans’ of the Pancreas.” By H. H. Dats, B.Ch., George Henry Lewes Student. Communicated by Professor STARLING, M.D., F.R.S. Received December 11, 1903,—Read January 28, 1904. (From the Physiological Laboratory, University College.) (Abstract.) These structures were first described by Langerhans in 1869. They have since been found by many observers in the pancreas of every species of mammal, bird, reptile, and amphibian in which they have been looked for. Kiihne and Lea first recognised the peculiarly rich plexus of wide blood-capillaries in the islets. As regards their function, they have been regarded as connected with the nervous system, as lymphatic structures, as embryonic remnants, as patches of exhausted or degenerate pancreatic tissue, as furnishing a particular constituent of the pancreatic juice, and as internally secreting ductless glandular tissue, furnishing a substance necessary for normal carbohydrate metabolism, and quite unconnected with the externally secreting function of the pancreas. This last view has received support from many observations of the degeneration or absence of the islets in diabetes, and from the statement of several observers that, after occlusion of the pancreatic duct, the islets remain intact when the ordinary secretory tissue has disappeared. Lewaschew, in 1885, first stated that activity of the pancreas led to an increase in the number of the islets, and that intermediate forms between the ordinary secretory tissue and the islets could be found, and were more abundant after activity. This statement has been confirmed by Pischinger, Maximow, and Tschassownikow, and has also been repeatedly contradicted. Laguesse describes a perpetual change of secretory tissue into islets and wee versa, the islets being, in his view, pancreatic tissue in an internally secreting stage, and representing also the stage during which growth takes place. My observations have been made on the pancreas of the dog, cat, rabbit, and toad. The pancreas was hardened in a mixture of corrosive sublimate and formaldehyde, sections cut in paraffin and stained with toluidine-blue and eosine. ‘The islets appear, with a low magnification, as relatively unstained areas. The pancreas was examined in conditions of “rest” (normal activity), of exhaustion produced by prolonged administration of secretin, and of starvation. Exhaustion was produced in the mammals (cat and dog) by repeated injections of secretin into the jugular vein during 6—12 hours, accompanied by bleeding towards the end of the 1903.] The “Islets of Langerhans” of the Pancreas. 85 experiment, until the pancreas ceased or almost ceased to secrete. The animals were anzesthetised with morphia and A.C.E. mixture. In the toad, secretin was injected into the dorsal lymph-sac by a hypodermic needle during 2—4 days. The effect of starvation was observed in a stray cat, picked up in an emaciated condition and killed immediately, and in toads which had been for several months in the laboratory tank. In the resting glands of all the species the intermediate forms described by Lewaschew were observed, the islets being formed by an assimilation of the secreting epithelium to the centro-acinary cells and the epithelium of the ductules, with later rearrangement of the cells attended by formation of the wide tortuous blood- capularies. In the toad evidence was also found of reconstruction of secreting alveoli from islets and of cell-multiplication in the islet stage. The effect of exhaustion was in all cases the same—a very extensive conversion of the secretory tissue of the gland into large islets, of irregular outline, retaining obvious traces of their former alveolar structure, and containing numerous intermediate forms. Specimens have been obtained from a dog with the greater part of a lobule, and from a toad with the greater part of the whole pancreas thus converted. 7 The effect of prolonged starvation was, on the whole, very similar to that of exhaustion, but slighter in degree. Experiments on the dog* and rabbit were also made to observe the effect of occluding the pancreatic duct. There resulted in all cases an interstitial fibrosis. The areas of pancreas not destroyed assumed the islet condition, but the preformed islets showed no special immunity from destruction. The experiments leave the question of the function of the islets undecided, but the results of occlusion of the duct are in favour of Laguesse’s view that they represent an internally secreting stage in the life of pancreatic tissue. * On the dog only one experiment was made, in which the operation was performed for other purposes by Professor Stariing. 86 | Mr. R. P. Gregory. [Jan. 5 “ The Reduction Division in Ferns.” By R. P. Grecory, St. John’s College, Cambridge, University Demonstrator in Botany. Communicated by Professor H. MARSHALL Warp, E.R. . Received January 5,—Read February 5, 1904. The earlier work upon the spore-formation in Ferns having led to divergent results, a new investigation was begun in connection with other cytological work bearing upon the Mendelian hypothesis. During the progress of this work, Professor Farmer and J. E. 8S. Moore, in their preliminary communication to the Royal Society upon the reduction-phenomena of plants and animals,* indicated the occurrence of a true reduction at the heterotype division in both plants and animals. The examination of the early stages of spore-formation in Ferns leaves no doubt that the essential features of the phenomena described by the above-mentioned authors are present also in this group of plants. The species which I have examined are Pteris tremula, Scolopendrium vulgare, Asplenm marinum, the so-called hybrid between Scolopendriwm vulgare and Asplenwim ceterach, Onoclea sensibilis, Davallia capensis and Fadyema prolifera. All these are included among the Polypodiacee. The processes of spore-formation are identical in all these types. The reduced number of chromosomes, which appears at the heterotype division, is thirty-two.t Owing to the smallness of the nuclei it is not easy to determine exactly the number of the chromosomes in the case of the vegetative cells, but it is about sixty, and I think there is a strong presumption that the number is, as stated by Stevens,{ sixty- four. An examination of Dicksonia davalliowdes and of Alsophila excelsa (Cyatheacece) was sufficient to show that it is extremely probable that the following description applies equally to these plants, but as the quantity of material hitherto available,was small, only the early stages have been examined as yet. After the vegetative divisions of the archesporium are complete, the spore mother-cells undergo a period of rest during which the nuclei increase in size. At the end of this period, that is, in the earliest prophase of the reduction division, the spireme thread under- goes a longitudinal fission. The ensuing contraction of the spireme towards one end of the nucleus results in the formation of a series of loops. The double nature of the loops, which is a consequence of the longitudinal fission of the thread, is clear. * © Roy. Soc. Proc.,’ 1903, June 18. + In Pteris tremula no exact counts were made, but the number is apparently the same as in the other species described. f£ “‘ Ueber Soa omer ned ae bei der Sporenbildung der Farne,” ‘Ber. d. Dastseh) Bot. Gesellsch.,’ 1898, p. 263. 1904. ] The Reduction Division in Ferns. 87 As the polarity of the spireme becomes more pronounced the limbs of each loop approach each other, and segmentation into the chromo- somes takes place. Each chromosome has its origin in one of the loops of the spireme and thus forms a double U-shaped body, the hmbs of the U being twisted upon one another to varying degrees in the different chromosomes of the same nucleus. The approximation towards one another of the distal ends of the limbs of each U, often resulting in the appearance of a “ring” chromosome, is a common feature of the heterotype division in Ferns. As prophase leads up to metaphase the original longitudinal fission becomes more obscure, and as the chromosomes begin to group them- selves in the equatorial plate, each appears to consist of two parallel rods, which represent the approximated limbs of each loop of the spireme, and are joined at one end. The increasing difficulty during these successive stages of recognising the original longitudinal fission in the limbs of the chromosomes has led to an incorrect interpretation of their structure. The two limbs, of which each chromosome consists, were interpreted as being the result ot the original longitudinal fission in the now shortened and thickened ST Isa A similar conception led to the interpretation of the “ring” form of chromosome as being due to the divergence of the halves into which each chromosome was separated by that fission. The examination of numerous preparations of the stages intermediate between that of the looped spireme and that of early metaphase reveals the incorrectness of this interpretation, inasmuch as_ the original longitudinal fission can be clearly recognised in each limb of the chromosome. In the same way favourable preparations of the “ring” type of chromosome reveal the double nature of the ring, while in many cases the distal ends of the limbs either do not quite meet, or on the other hand may overlap, thus providing forms transitional to the U and X types of chromosome. The spindle fibres are attached to the limbs of the chromosomes near the distal ends of the latter; as, therefore, the two daughter- ane? 2 EES chromosomes are drawn apart, the familiar shaped figures are obtained, and the final separation takes place at a point corecoeou ine with the apex of the original loop. The exact time when the transverse fission, which separates the two limbs of each loop, takes place is not easily determined and appears to be variable. In some cases it appears to have been com- pleted before metaphase is reached, so that the chromosomes as they move to the equatorial plate consist of two separate rods. In others, on the contrary, the separation is synchronous with the commence- ment of the contraction of the spindle fibres, and consequent divergence of the limbs of the chromosomes. 88 Mr. RB. P. Gregory. [Jan. 5, As the chromosomes move toward the equatorial plate the longitudinal fission of each limb once more becomes clearly apparent, so that, seen in face, the diverging daughter-chromosomes form a ()-shaped body. A splaying of the ends of the rods at this or a slightly earlier stage often gives rise to those figures, not unlike tetrads, which were assumed by Calkins* to have an origin similar to that of the tetrads characteristic of the heterotype division of, for instance, Gryllotalpa, as described by Vom Rath. In the small chromosomes of the Ferns it is impossible, in the majority of cases, to trace the presence of the original longitudinal fission through the late prophase condition up to the beginning of metaphase. Nevertheless, a study of the successive forms assumed by the chromosomes indicates that the gradual obliteration is apparent rather than real ; for it can still be recognised by means of the slightly bifid ends of the limbs of the chromosomes. These appearances are sufficiently convincing as to the correctness of the interpretation of the so-called second longitudinal fission, as nothing more than a reappearance of the original fission undergone by the spireme in the early stages of prophase. | The second (homotype) division follows very rapidly upon the com- pletion of the heterotype division, and is provided for by the longi- tudinal fission already noticed in the diverging chromosomes of the heterotype division. The result is, therefore, a transverse true reduction division of the bivalent chromosomes which characterise the heterotype division. This work, therefore, provides an extension to another group of plants of the results obtained by Farmer and Moore?7 in certain plants and animals. It is not within the scope of the present paper to discuss the con- siderations tending to support the belief in the universal occurrence of a reduction division leading to the formation of the gametes. I shall confine myself to a consideration of the significance of the reduction division in connection with Mendelian segregation. Viewed from this standpoint the occurrence of a qualitative reduction in plants as well as in animals is extremely important as affording a possible provision for that purity of the gametes, in respect of allelomorphic characters, which is demanded by Mendel’s hypo- thesis. The work of Boverit upon the qualitative differentiation of chromo- * “Chromatin Reduction and Tetrad-formation in Pteridophytes,” ‘ Bull. Torrey Bot. Club,’ vol. xxiv, 1897, p. 101. + Farmer and Moore, loc. cit. ft “‘Mehrpolige Mitose als Mittel zur Analyse des Zellkerns,”’ ‘ Verh. d. Phys. Med. Ges. Wiirzburg,’ 1902, vol. 35. 1904. | The Reduction Division in Ferns, 89 somes, supported by that of Sutton,* McClung,t and others, affords strong evidence in favour of the theory that the development of certain characters in the zygote corresponds with the presence of certain chromosomes or groups of chromosomes in the nuclei. Cannoni has suggested a “cytological basis for the Mendelian laws” founded upon the occurrence of a qualitative reduction division, and, at a time when the general concensus of opinion among botanists was adverse to such a conception, went so far as to predict the discovery of a qualitative reduction in plants. A somewhat similar suggestion, based upon work on Jbrachystola (Orthoptera), was inde- pendently made by Sutton.§ Cannon’s hypothesis consisted in the assumption that in fertile hybrids, as well as in pure races, “the chromosomes derived from the father and the mother unite in synapsis and separate in the metaphase of one of the maturation divisions. . . so that the end is attained that the chromatin is distributed in such a way that two of the cells receive pure paternal, and two cells pure maternal chromosomes, and no cells receive chromosomes from both the father and the mother.”’|| Thus enunciated the hypothesis is applicable only to “monohybrids” (de Vries) ; it is insufficient to explain the phenomena observed in the offspring of Mendelian hybrids whose parent races differ from one another in respect of more than one pair of allelomorphic characters. This was recognised by Sutton,{] who was thus led to make a more careful study of the whole division process in brachystola, paying particular attention to the positions assumed by the chromosomes. He says (p. 233), “the results gave no evidence in favour of the parental purity of the gametic chromatin as a whole. On the contrary, many points were discovered which strongly indicate that the position of the bivalent chromosomes in the equatorial plate of the reducing division is purely a matter of chance, that is, that any chromosome pair may be with maternal or paternal chromatid indifferently towards either pole, irrespective of the positions of the other pairs, and hence that a large number of different combinations of maternal and paternal chromosomes are possible in the mature germ-products of an individual.” The view that the gametes may contain both chromosomes of paternal and of maternal origin is strongly supported by the recent results obtained by Valentin Hacker in his study of certain Copepoda. * “On the Morphology of the Chromosome Group in Brachystola magna,” ‘Biol. Bull.,’ 1904, vol. 4. * a Spormiatoerte Divisions of the Locustide,”’ ‘Kansas Univ. Quart.,’ 1902, vol. 11, No. 8 (contains other references). tT ‘ Bull. Torrey Bot. Club,’ December, 1902. § “The Chromosomes in Heredity,” ‘ Biol. Bull.,’ 1903, vol. 4, No. 5, p. 251. || Cannon, loc. cit., p. 660. {| “The Chromosomes in Heredity,” ‘ Biol. Bull.,’ 1903, vol. 4, No. 5. 90 Mr. R. P. Gregory. | [Jan. 5, He showed in 1892* that in Cyclops, as in Ascaris and other forms, the two germ nuclei do not fuse in fertilisation, but give rise to two separate groups of chromosomes which lie side by side in the spindles of the dividing nuclei. Riickertt was able to trace the paternal and maternal groups into the later stages of cleavage. Both observers have shown that the two distinct groups of chromosomes appear also in the germinal vesicle. Hacker§ has since traced the autonomy of the paternal and maternal chromatin in Cyclops from fertilisation up to the formation of the mother- cells of the gametes; and the double structure of the nuclei at the maturation divisions is such that he is able to show “das die Vierer- gruppen auf der einen Seite der Scheidewand vaterlichen, auf der anderen miitterlichen Ursprungs sind” (p. 341). In the “ secundiaren Keimbliaschen ” the paternal and maternal chromosomes pass between one another ‘‘in einer ganz gesetzmassigen Quadrillen-ahnlichen Ordnung ” (p. 342), with the final resuit that ‘‘ Wahrend der Hireitung von Cyclops findet eine Umordnung der Chromatinelemente in der Weise statt, dass die Eizelle in gleichmassiger Mischung g grossviiterliche und grossmiitterliche Elemente erhalt” (p. 374). _ The regularity observed by Hacker in the movements of the chromosomes at the reduction division, leading as it does to a sym- metrical distribution of the chromosomes, may be our first indication of a more comprehensive symmetry which probably underlies the pro- duction of the different types of gametes in Mendelian hybrids. “It is impossible to be presented with the fact that in Mendelian cases the cross-bred produces on an average equal numbers of gametes of each kind, that is to say, a symmetrical result, without suspecting that this fact must correspond with some symmetrical figure of distribution of the gametes in the cell divisions by which they are produced” (Bateson).|| On the hypothesis that the segregation of characters occurs at the reduction division, we shall expect that the mitoses in a Mendelian hybrid will be perfectly regular, and in our present condition of inability to recognise qualitative differences between chromosomes alike in form, we should further expect that the mitoses will differ in no visible way from those of the pure paternal and maternal races. Cannon{] has shown this to be the case in race-hybrids of Pisum * “Die Libildung bei Cyclops und Canthocampus,” ‘ Zool. Jahr.,’ 1892, vol. 5 + “‘ Ueber des Selbstandigbleiben der vaterlichen u. miitterlichen Kernsubstanz, etc.,” ‘Arch. f. Mikr. Anat.,’ 1895, vol. 45. £ See Wilson, ‘ The Cell,’ 2nd edit., 1902, p. 299 § “Ueber das Schicksal der elterlichen u. grosselterlichen Kernanteile,” ‘Jena Zeitschr.,’ 1903, vol. 37, p. 297; a review in ‘Zool. Zentralbl.,’ 1903, Jahrg. 10, No. 11, p. 365. | ENeraelis Principles of Heredity,’ Cambridge, 1902, p. 30. §| “The Spermatogenesis of Hybrid Peas,” ‘Bull. Torrey Bot. Club,’ 1903, vol. 30, p. 519. 1904. ] The Reduction Dwision in Ferns. Bal sativum, a result which is confirmed by my own observations upon race-hybrids of Lathyrus odoratus, for the material of which I am indebted to Mr. Bateson. Further light upon this question may be expected from the study _of hybrids between races which differ from one another either in the morphology or number of the chromosomes. The only observations upon this point are those of Rosenberg* on the hybrid Drosera longifolia x Drosera rotundifolia. The number of chromosomes characteristic of the vegetative and of the reduction divisions respectively in the former is 20 and 10; i the latter, 40 and 20. In the vegetative cells of the hybrid 30 (i.c.. 10+20) chromosomes appear. In the formation of the pollen of the hybrid there occur three types of nuclei, all of which may occur in the same pollen-sac. Commonly 15 bivalent chromosomes appear, but in many cases there are 20 (as in D. rotwndifolia), and in two cases Rosenberg observed 10 (as in VD. longifolia). He was, however, unable to determine whether dissimilar numbers of chromosomes appear in the daughter nuclei of the same pollen mother-cell. The investigation is, therefore, not sufficiently complete at present to permit a useful discussion of the results. The sterility which characterises many hybrids follows upon the abortive development of the sex cells, and the suggestion has been madej that this may be due to the inability of the hybrid to separate, in the formation of the gametes, the characters which were united in the hybrid zygote. It is well known that sterile plant-hybrids are particularly characterised by abortive development of the pollen, or (in the case of the hybrid Fern described by Farmerf{) of the spores. Among the offspring of a race-hybrid of Lathyrus odoratus fertilised with its own pollen, Mr. Bateson obtained a number of individuals which failed to form good pollen. In the plants with coloured flowers the sterility was, with a few exceptions, correlated with the develop- ment of a somatic character—the sterile plants generally possessing a green leaf axil, while the fertile coloured plants with rare exceptions had red axils. In these plants the divisions of the vegetative cells are quite normal, as are also those of the archesporium up to the forma- tion of the pollen mother-cells. The irregularity makes its appedianve only in the heterotype division. The longitudinal fission of the spireme takes place quite normally, but the segmentation into chromosomes is, if carried out at. all, irregular, and the pollen mother-cells degenerate. Since the equation divisions are quite normal, this would seem to indicate that the union * “Das Verhalten der Chromosomen in einer hybriden Pflanze,” ‘ Ber. d. Deutsch. Bot. Gesellsch.,’ 1903, vol. 21, p. 110. + Bateson and Saunders, Report to Evolution Committee, I, p. 148. f ‘ Annals of Botany,’ 1897, vol. 11, p. 533. 92 Dr. A. D. Waller. The Secreto-motor Effects in [Jan. 16, of the chromosomes in synapsis is such as to prevent any subse- quent separation, the result being that no sex-cells can be organised, since the essential condition of a qualitative separation of the chromatin is not fulfilled. “The Secreto-motor Effects in the Cat’s Foot Studied by the Electrometer.”. By Aucustus D. WaLiErR, M.D., F.RS. Recelved November 17,—Read November 19, 1903. Re- ceived in revised form January 16, 1904. In a previous communication* it was stated that the electrical signs of secreto-motor action by tetanisation of the sciatic nerve are demonstrable in the pads of a cat’s foot after death, best so during the second half-hour after death, when the action of the nerve upon muscles of the limb has ceased. The subsequent study of these effects, by means of electrometer records, has brought out with great distinctness the chief classical events with which we are familiar in the case of the contraction of voluntary muscle, viz., the latency and course of a single response to a single stimulus, the super-position of two or more responses and the composition of tetanus, summation of stimuli, fatigue and recovery, and the staircase phenomenon. The difference between the muscular and the secreto-motor series of phenomena is principally a difference of time, the former being about 100 times more rapid than the latter. I may preface the description by stating that I have experimentally satisfied myself that the electrical effects are in reality of glandular origin. The response is completely abolished by atropine, and it is restricted to the pads (glandular) of the skin, being completely absent from the hairy (non-glandular) skin, 7.¢., it is not a pilo-motor concomitant. The description itself will be best given by means of the following electrometer records of :— 1. A single response to show the latent period and duration of the response. . 2. A series of single responses to show staircase phenomenon. 3. A series of four responses to show composition of tetanus. 4, A series exhibiting post-mortem decline. * “Proc. Roy. Soc.,’ November, 1901, “On Skin-currents. Part IIT.—Observa- tions on Cats.” The electrical effect of indirect excitation is always ingoing _ through the skin. This direction has been conventionally indicated throughout this paper by a downward movement of the mercury column. 1904. ] the Cat’s Foot studied by the Hlectrometer. OB. 5. A series to exhibit the relation between magnitude of stimulation and magnitude of response. 6. A single response before and after tetanus to illustrate ‘ facilita- tion ” (Bahnung). 7. A single response before and after tetanus to illustrate fatigue. 8. A series to illustrate summation of stimuli. : 9, A series to show the difference between infrequent and frequent. stimuli. No special comment upon the records appears to be necessary, beyond, perhaps, a remark to the effect that “summation of stimuli,” as distinguished from “summation of effects,” is, by reason of the great length of the latent period, a particularly evident phenomenon. The latent period itself has its seat at the organ of intermediation between nerve and secreting cell, as is chou by the absence of demonstrable lost time to direct excitation and along the nerve itself. The declining excitability of the secreto-motor nerve fibres is very evidently in the centrifugal direction, stimulation of the nerve nearer to the periphery being effective after stimulation further from the periphery has ceased to be effective. Similar effects are obtainable on nerve-skin preparations of the frog, “summation of stimuli” and “staircase effect ” being, as in the case of the cat, particularly evident. O S /O IS SEC Fie. 1.—Cat. Nerve-skin response to a single induction shock 40 minutes post-mortem. 4 94 Dr. A. D. Waller. Zhe Secreto-motor Effects in [Jan. 16, 6) 10 20 30 0 50 60 secs. 70 Fie. 2.—Cat. Six nerve-skin responses to single induction shocks of uniform strength. Staircase increase from 0°0165 to 0°0195 volt. (The initial deflection is that of a standard ,1,th volt.) 100 O /0 ZO 30 40 SECS. Fic. 3.—Cat. 35 minutes post-mortem. Imperfect tetanus by four instantaneous make-break induction shocks at intervals of about 5 seconds. 1904.] the Cat’s Foot studied by the Hlectrometer. a) O 5 10.S€C8; Fic. 4.—Cat’s pad. 30 to 65 minutes post-mortem. Five single responses to exci- tation of the sciatic nerve by instantaneous make-break induction shocks at intervals of approximately 10 minutes. The dotted line shows the curve given by =i,th volt through the preparation and electrometer. Time Magnitude of post-mortem. Latency. response. 30 mins. 1 °4 sec. 0 -0128 volt. 40 ,, 16..,; G-O0L00) >; 48 ,, Bee 55 4700080) ;,, dD Cy, MS) ye 0°0048 _,, Gor, Bes Oe eos 00018 7, S Q g ie) iS) 9 ie) Q 9 9 ME oS ta BS os: BS QJ SS) st no) S S) ro) r ny we KA VOLEA 0 10 20 30 40 50 60 sSéCs. 70 Oe ee EEE Fig. 5.—Responses to single shocks of increasing and diminishing strengths. 1 hour post-mortem. Strength Voltage Strength Voltage of stimulation. of response. of stimulation. of response. 2,000 0 0005 10,000 0 -0090 3,000 0 -0015 5,000 0 °0015 4,000 0 0020 4,000 0 0015 5,000 0 -0030 3,000 0 -0010 10,000 0 0085 2,000 Nii N.B.—This electrometer record is not very satisfactory, as the magnification was taken too low. The following series of numbers observed without record with higher magnification 45 minutes post-mortem is a better one :— Strength Voltage Strength Voltage of stimulation. of response. of stimulation. of response. 1,000 Nil 4,000 0 -0065 2,000 0 -0015 5,000 0 -0125 3,000 0 -0035 10,000 Off scale 96 Dr. A. D. Waller. The Secreto-motor Effects in [Jan: 16, Telar. | Fig. 6.—Effects of a single shock, S.8., before and after tetanisation for 4 minute. Before .6.2s-< cece = 00115 volt. (Daring tet. <6 -_ 2 070800 oe) FATVOT so Sie aint acateres = O-0155. ;, Fie. 7.—Effects of a single shock, 8.S., before and after strong tetanisation for 1 minute. Before: sh. 55 es eetee = 00183 volt. (During tet. ...... 0°0283 ,, ) ATELY. (oT conns sie 0 COLT «4. A second tetanisation at T gives only 0 ‘0067 volt. 1904. ] the Cat’s Foot studied by the Electrometer. on / Zz 3 FZ a yi. O /O ZO 30 FO 50 60 S€CS. 70 Fie. 8.—To illustrate summation of stimuli. Effects of 1, 2, 3, 3, 2, 1 instantaneous make-break shocks. Response to 1 shock........ = 00035 volt. O 10 20 30 40 50 60 Secs. 70 Fic. 9.—To show that the effect of frequent is greater than that of infrequent stimuli. The first response, aroused by five shocks at an interval of 2 seconds, is an incomplete tetanus with a maximum value of 0°0040 volt. The second and third responses are each to five shocks at an interval of about jth second, and reach maximum values of 0'0123 and 0:0120 volt. VOL. LXXIII. H 98 The Secreto-motor Hffects in the Cat’s Foot. [Jan. 16, Addendum, December 10, 1903. The latent period and the duration of the response are smaller at high temperature, greater at low temperature. And under similar conditions the voltage of the response is greater at high than at low temperature. ‘These three points are illustrated by fig. 10 of a single response from the left foot enclosed in a cool chamber at +9°, and from the right foot enclosed in a warm chamber at + 35°. 30 Secs. Warm (35°). Cold ¢9°. Fie. 10.—Cat. Nerve-skin response to a single induction shock at 9° and at 35° of the surrounding air. Cold | Warm. Latent period..........| 4 secs. | 1 °5 secs. Duration of response....| 30 ,, 10 G5 Maximum voltage...... | 0 -0065 volt. 0 -0122 volt. The progressive alterations exhibited by fig. 10 are no doubt influenced by the falling temperature that normally occurs in a limb after arrest of the circulation. G 1904.] Joining Cervical Sympathetic with Chorda Tympani. 99 “On the Effects of Joining the Cervical Sympathetic Nerve with the Chorda Tympani.” By J. N. Lanetey, F.RS., and H. K. Anperson, M.D. Received January 26,—Read February 4, 1904. It is well known that the cervical sympathetic nerve and the chorda tympani have opposite actions upon the blood-vessels of the sub- maxillary gland, the former causing contraction of the vessels, and the latter, dilatation. Evidence has been given by one of us* that the chorda tympani if united with the cervical sympathetic, can in time make connection with the nerve cells of the superior cervical ganglion and become in part vaso-constrictor fibres. Our experiments have been directed to determine whether the cervical sympathetic if allowed an opportunity of becoming connected with the peripheral nerve cells in the course of the chorda tympani will in part change their function from vaso-constrictor to vaso-dilator. Two experiments were made on anesthetised cats, both give similar results, but one was much more conclusive on the point at issue than the other, and here we shall speak of that only. The superior cervical ganglion was excised and the central end of the cervical sympathetic nerve was joined to the peripheral end of the lingual, which contains the chorda tympani fibres. After allowing time for union and regeneration of the nerves, the cervical sympathetic was stimulated ; it caused prompt flushing of the sub-maxillary glands, and the effect was repeatedly obtained. The experiment shows we think (1) that vaso-constrictor nerve fibres are capable of making connection with peripheral vaso-dilator nerve cells, and becoming vaso-dilator fibres, and (2) that whether contraction or inhibition of the unstriated muscle of the arteries occurs on nerve stimulation, depends upon the mode of nerve-ending of the post-ganglionic nerve fibre. The cervical sympathetic gave a less scanty and more prolonged secretion than normal, so that some of its nerve fibres had become connected with the peripheral secretory nerve cells of the chorda _ tympani. A full account will be published later in the ‘ Journal of Physiology.’ * Langley, ‘ Journal of Physiology,’ 1898, vol. 23, p. 267. Pech atk 100 Dr. G. H. Bryan and Mr. W. E. Williams. [Jane z,- “The Longitudinal Stability of Aerial Gliders.” By G. H. Bryan, Se.D., F.R.S., and W. E. WitiiAMs, B.Sec., University College of North Wales. Received June 18—Read June 18, 1903,— Received in revised form January 7, 1904. 1. Introduction. The main difficulty connected with the attempt to fly by means of a machine heavier than air is that of longitudinal stability. It is not difficult to construct an aeroplane system which shall be transversely stable. But for this difficulty the problem of artificial flight would probably have been solved already. Experiments in gliding under gravity have been always made with machines not too large to be kept balanced by the skill of the experimenter, and the glides, though undoubtedly successful, have been of short duration. Experiments have invariably stopped short of the performance of contintious flight by a mechanically propelled machine. The problem of artificial flight is hardly likely to be solved until the conditions of longitudinal stability of an aeroplane system have been reduced to a matter of pure mathematical calculation. A theoretical investigation, even if calculated under conditions slightly different to those occurring in nature, will serve as a basis of comparison by which experimental results can be co-ordinated and interpreted in their true light. The object of these investigations is (1) to show that the longitudinal stability of aeroplane systems can be made the subject of mathematical calculation; (2) to draw the attention of those interested in the problem of artificial flight to the necessity of acquiring further experimental knowledge concerning the quantities on which this stability is shown to depend. 2. General Investigation of the Longitudinal Stability of any Symmetrical Aeroplane System. Consider any system of aeroplanes, having a plane of symmetry, descending in a vertical plane, in air or in any resisting medium whatever. To specify the motion (which .we suppose to be two- dimensional) choose two axes at right angles fixed in the body, having the centre of gravity as origin. The motion will be completely determined if at every instant we know— (1) the angle 6 which the axis of makes with the vertical.* (2) the velocity components w, v, of the body along the two axes. * This angle is supposed measured from the downward drawn vertical in the positive direction. The axis of y must be drawn adove the horizontal at an inclination of @. 1904. | The Longitudinal Stability of Aerial Gliders. 101 The angular velocity of rotation of the body will be d6/d¢ or 6. We shall use m to denote the mass of the body, mk? its moment of inertia about the centre of gravity. Whatever be the law of resistance, the resistances of the air on the several parts of the system will in general be functions of wu, v and 0. These resistances are always equivalent to two forces, which we shall call mX and mY, along the axes, and a couple mG about the origin, so that X, Y, G, denote the forces and couple, divided by the mass of the body. The equations of motion of the body are— du dei 4 (OU _ Ue _ fa m ( aan } mg cos 6 -mX, | {dv dé : Keen — = — yn 0 —/ GES) WO Done ale e m \di a uF) mg sin 0 — mY, t (1) mk? = —mG. | In steady motion u, v and 6 are constant, and equations (1) give WG gcose =x, OO -gsmn0—Y, 0. —G ...:.- (2). Knowing the forms of the aeroplanes and other parts of the system and the law of resistance, X, Y, G are known functions of , v and 6. Moreover, in steady motion, 6 = 0. ’ Equations (2) thus determine the values of wu, v, 6, for steady motion. Fluctuations about Steady Motion—We must now examine what happens when the system is slightly disturbed from its state of steady motion. Let the disturbance be represented at time 7 by small increases du, dv, 60 in the values of w, v, 0, and let wo, v%, O be their values in steady motion, so that in the disturbed motion, U = Uy+ ou, v= I+, 6 = 6,460. | Also let X,, Yo, Go be the forces and couple in the steady state ; then in the disturbed state we have, neglecting small quantities of the second order, Shan Sey a Ee (3), du dv d6 and two similar equations for Y and G. _ We shall denote differential coefficients such as dX/du... dG/dé by > 36.5 Go ° Substituting in the equations of motion, we obtain, to the first order, 102 Dr. G. H. Bryan and Mr. W. E. Williams. [Jan. 7, d | * al fbu~ vy i 56. = asin 6:80 — X,,bu — Kody | ene # 39 419 8.86 = = 9 eos Oy80-— Vidu— V 80 Sees dt dt : ; a § I? ap (88) = —G,,dou—G,6v — G00. To solve these equations, put Ou) = Pee du = Qe, OO — nee Substituting, dividing by «**, and re-arranging the terms, we have P (A+ Xz) + QX,+ R(AXs — Avo+g sin A) 0, PY, +Q (A+ Y,)+ RAY + Au, + 9 cos %) 0, PGy + QGy+ R (2k? +AG6 ) = 0. Eliminating P, Q, R, Na KG ee Xy , —Av +AXe +gsin A Vagpoihgl pases ie duy +rAY4 +9 cos A = OU a): GG ee 2k? + AGe If the determinant be expanded in powers of we get an equation of the form I AM+ BAP 4 CA24 DALE = 0 22 (4a), . where pe ; | A B= 2, etre C D = Kk? (XyYy— X,Yu) + %Gu — WoGy -— Xe Gu— Ye Gi, = Up (XyGu — XuGv) + %0 (YoGu — YuGr), Loo) —g sin Gy —g cos Gy+ Yo (X,Gu — XuGy) — X% (YyGu — YuGy) + Go (XuYr — X,Yu), | E = g cos &% (X,Gu — XuGv) -— 9 sin 4 (Y,Gu — YuG), ; In many cases it is possible to take the axes of co-ordinates, so that 6) shall be zero, and also that Y and its differentials shall vanish. In such cases the coefficients take the following simple forms :— Ae ) B = F*X,4+G6, | C = Gy —UoGy — X6 Gu, ech ae (6). D = u(X,Gu- XuGv) — gGr, E = 9(X,Gu - XuGo), J 1904. | The Longitudinal Stabslity of Aerial Gliders. 103 Equation (4) or (4a) is the period equation for small fluctuations about steady motion. In order that the steady motion may be stable, the roots of this biquadratic must either be real and negative, or complex with their real parts negative, and Routh* finds that this will be the case if the six quantities : A, B, C, D, E, and BCD — AD? — EB?, are all of the same sign. Since A is essentially positive the remaining five quantities must all be positive. We shall denote the last quantity by H. 3. General Theorems. t (1) The general transformation formule, connecting the nine co- efficients X,,... G6 referred to any given system of axes with those referred to any other system of axes, may be easily written down and need not be discussed here. (2) The work of calculating these coefficients for a given system may be reduced if the resistance is proportional to the square of the velocity, for X, Y, G will then be homogeneous quadratic functions of u, v, 8, and Euler’s theorem of homogeneous functions gives, remember- ing that 6) =0 and applying (2), uXy+vX, = 29 cos 9, UYy,+vYy = —2gsin 6, uGy,+vG, = 0. (3) If V is the velocity of gliding, the coefficients mX, .. . mGé are all linear functions of V for a given angle of gliding. But mg being, in steady motion, equal to the vertical resistance, is proportional to V?. Hence if m be eliminated, the values of X,,...Gg are inversely proportional to V. In this case A is independent of V, B is of dimensions V—!, C is of form P+ QV~, D is of dimensions V—!, E is of dimensions V~?, and the expression H or BCD-—AD?-—EB? assumes the form PV-2 + QV-+ Thus the conditions of stability C > O and H > O impose limits on the value of V2, the remaining conditions only * Routh, ‘ Advanced Rigid Dynamics,’ p. 167. + This section and also the subsequent parts enclosed in [...... | have been rewritten October 27, 1903.4 104 Dr. G. H. Bryan and Mr. W. E. Williams. | [Jan. 7, depend on the form and dimensions of the machine and the angle of gliding. We shall now show how the nine coefficients X,,...G g can be calculated for a system of aeroplanes if the laws of variation of the resultant pressures and of the positions of the centres of pressure are known. We shall assume that ,the resultant pressure on a plane lamina varies as the square of the velocity, and acts in a direction normal to the lamina. We may therefore write R = KSV2f(a), where S is the area of the lamina, V the velocity, and « the angle between the plane of the lamina and the direction of motion, K a constant depending on the units employed. - The function f (a) has been determined by lanes for certain rectangular planes. Also, let the distance of the centre of pressure from the centre of figure be a@ (a), 2a being the breadth of the lamina. Experiments to determine ¢(«) for square planes have been made by Joesel,t Kummer,{ and Langley,§ and Kummer has also experi- mented on oblong planes, with their longer side in the direction of motion. - Their results show a certain amount of discrepancy, and there appears to be considerable difficulty in | obtaining consistent results at small inclinations. Further experiments are very necessary, and it is to be hoped that more attention will be given to determinations of ¢ (z), when their importance, as affecting the stability, has been recognised. _ Experiments show that f(a), ¢(a) are, to a first approximation, independent of the translational velocity of the lamina, but the effects of a rotational angular velocity, 6, have never been considered. It would not be difficuit to determine these effects by experiments with a whirling table, making 6 the angular velocity of the table. Failing such experiments, these effects must be neglected, and, as 6 is zero in the steady motion, and only small oscillations are considered, they are probably small. When the system consists of a number of planes rigidly connected together, we may now write the component forces and couple in the form mx = YKSi Vif (a1 = 3) COS jes TN) — =KS, Vi2/ (a = Bi) sin py Sastce (7), mG = TKS, V2 (a1 — Pi) {pi + ud (41 — Pr) J * * Experiments in Aerodynamics,’ p. 62. + Joesel, ‘Mémorial du Genie Maritime,’ 1870. ft ‘ Berlin Akad. Abhandlungen,’ 1875, 1876. § Loc. cit., p. 90. 190+. | The Longitudinal Stability of Aerial Gliders. 105 where S, is the area, 2a, the breadth of the lamina, &1,9 the co- ordinates of its centre, V, the resultant velocity of its centre, a, = tan—! w/v, the angle which the direction of motion makes with the axis of y, 2; the angle which the plane of the lamina makes with the axis of y, so that (a,—/;) is the angle between the plane and the direction of motion of the centre of mass, and __ finally, pi = m Cos By + & sin fy. . The summation is to be extezded over all the planes of the system. If, now, u, v be the component velocities of the centre of gravity the velocities 7%, v; of the centre of the lamina will be given by Ww = u—mm6, y= TES, and for steady motion, since 6 = 0, 1% = u, % = 7. We may therefore write the expressions for the forces in the form m& = TKS, (ui? + 0)”) f (tan /v — Br) cos Pi = TKS; (w+v?- 2um0 + 288) 4 tan—1 ( ead ae py } cos 4, (7) + &6 neglecting terms in 6°. Similarly, mY = KS, (w+? - Qu 0 + 2&0) f { tan? (2) = ps } sin 61, \U+ + & 4, mG = = —~ DKS, (wu? + 0? — 2um6 + 2uk.6) f { tan=! (“= ai) - By } 10 {pi +a (tan i FF. a) )- 61) } 0+ &6 Differentiating these expressions with respect to wu, %, 6 we have :— mX, = =[KS; (Qu — m9) f(a — B1) cos By + KSy01 f’ (a1 — B1) cos Bi] ; or, since 6 = 0 in the steady motion, mXy = =[2KS,uf (a1 — 1) cos Bi + KS" (41 — Pi) cos Bi]... (8), Similarly, mX, = >[2KSyzf (ce — B1) cos Bi — KSyuf’ (a4 — Pi) cos Pi], mY, = > [2KS,uf (a — Br) sin By + KS,ef’ ( — Pi) sin Pil, mY, = > [2KSyef (a: — Bi) sin B; — KSyuf’ (@, — Fi) sin Pi], mXg = ~[2KS, (-— wnt vé1) f (a1 — Pi) cos Pr — KS, (v n+ U ff (a= P1) cos Pi]; 106 Dr. G. H. Bryan and Mr. W. E. Williams. _—_[Jan. 7, mYa = >[2KS, (— um + v&1) f(a — Bi) sin By — KS, (ony a wE1) Wi : (a “a 1) sin Bx}; mGy = =[-2KS8; (u- m9) f (a — 2) {pi + a1 (a — Bi)} —~ KS f’ (a1 - Bi) {pi + 1d (aa — Bi)} — KS0 f (a1 — Bi) a1’ (a — 1). Since in the steady motion G = 0, 1.0., 2 (u? +0) ZKSi f(a — Bi) (pi + 1g (4 — Bi)} = 0, mGy, = v=[ — KS) f’ (a1 — Pi) {pi +14 (a1 — Pr)} — KSi f(a — Pi) ard'(aa — Bi), mG, = w=[KS, f’ (a1 — Bi) {p1 +14 (a — Pr)} + KS) f (a1 — 81) a1’ (a1 — fi)], mG = =[- KS, f’ (a1 — Pi) (- om + U8) {pp + ah (a — B1)} — KS) f (1 — Bi) (— vm + ux) af (4 — 1). 4, Numerical Calculations for Particular Cases. Single Lanune. We now consider the stability of certain particular systems, beginning with a single plane lamina. _ Since the lamina is falling steadily under gravity, its plane must necessarily be horizontal. Take the axis of y parallel to this plane, then 6 = 0, 6 = 0, and, therefore, Y = 0; also, for equilibrium, 7 +«a¢ (a) = 0. The coefficients, therefore, become mXy = 2KSuf (a) + KSe/'(a), mX, = 2KSof (a) — KSuf’ (a), mX4 = —2KS (wm - 2) f(a) + KSf ‘(a) (on + U8), mGy = —KSof(a)a¢@ (a), mG, = KSuf (a) a¢’(2), mGg = — KSaf(a) (vn + u€) af’ (a). fz. 1.—Consider a square plane, balanced so as to fly at an angle of 10°, with the centre of gravity in the plane, and suppose its radius of gyration given by k? = 4a’. Professor Langley’s results give for this particular angle Ga) 103; ik) ciliata and from Joesel’s formula b(a) = 0°49, g~' (a) = —0°59, 9 = 10-490; 1904. ] The Longitudinal Stability of Aerial Gliders. 107 also w= Vsine = O17V, v= 0:98V. Substituting these values in the above expressions for X,..., we have mX, = 1:66KSV, mGy = 0 17KSVa, mX, = 0°32KSV, mG, = —0:°03KSVa, mx = 0'73KSVa, mGsg = 0:09KSVa?. Substituting these values in the expressions for A, B, C, D, E, given in § (2), and remembering that by the conditions for steady motion mg = KSV2f(«), we obtain AY es Oe Br 292 \02/N7, C = 18a— 1340 a?/V?, D = 295 a/V, E = 36400 a/V?, and therefore H = BCD - AD?- EB?= 48.105“ - 3-6. 107 We aoe e at e a® - 43. 1045 - 3-08. 10°F. This expression will be positive if V2 > 774a, and this condition will also make C positive. The glider will therefore be stable if its velocity is greater than ,/(774a), the units being feet and seconds. He. 2.—Let us now take the angle of gliding to be 35°, and assume as before k? = 4a. At this angle we have f(a) = 0°84, f’ (2) =06, (a) = 0°26, ap ((e))) == (0%), With these values, we have mXy = 1:34KSV, mG, = 0°35KSVa, ROS — AIRS. mG, = 0:°21KSVa, mxXxg = —01KSVa, mGs = 0:08KSVa. Substituting in the expressions for A, B, ete., we have A = 4a’, B = Dil a?/V, © = 15-20 +50 a2/V?2, D = 380 a/V+240a/V = 620a/V, E = 25600 a2/V2, H = 5600 a3/V2— 17400000 a3/V4. The latter is positive when V2 > 3100a, which is therefore the con- dition of stability. Ex. 3.—Let us now take an oblong plane lamina ; the values of f (a) 108 Dr. G. H. Bryan and Mr. W. E. Williams. _—‘[ Jan. 7, are given by Langley for a lamina of 30” by 4:8” the shorter side being in the plane of motion. In the absence of any information as to ¢@(«), we shall assume (whether correctly or incorrectly) the same values as for square planes. Let the moment of inertia be given, as before, by 4? = sa’. For an angle of 10° we have f(a) = 0°44, J @).— ts p(a) = 0-49, g(a) — 0:52) Substituting these values in the expressions for the coefficients, we have mX, = 1:6KSV, mGy, = 0:26KSVa, mX, = 0°6KSV, mG, = —0-4KSVa, mX4 = 0'56KSVa, mG6 =0:13KSVa?. Substituting again in the expressions for A, B, C, D, and E, we have, assuming k? = $a?, Ay i Sa Bex ee 1 (16 $ +0130 )KSV = 6002/V, C = 0G, —uG,— X6 G, = 19°4a — 760a2/V2, D = u(X,Gy — XuGv) — gGy = 280a/V, E = g(X,Gu—- XuGy) = 35000a/V2. Therefore a* 4 5 = =< g 0 os 12 10s This is positive if V? > 470a, which is therefore the condition of stability. 5. Gliders Formed of two Planes. Proceeding now to consider the stability of systems made up of several planes rigidly connected together, we shall first consider the — stability of a gliding system supported on two slats 8; S2, which are so narrow that displacements of their centre of pressures due to varia- tions of the angles of incidence of the wind may be neglected. We shall first suppose that the two slats are in the same plane, and that the centre of gravity is also in this plane. 1904. ] The Longitudinal Stability of Aerial Gliders. 109 Putting (a) = 0 and f; = fy = O in the expressions of § (3), we have all the Y’s equal to zero and ONG — KO(Sale 2) (20 (a) OF (aN) eck en eae (9) mX&, = K (Si +82) (2ef (a) + uf’ (@)), mX6 = —K (Sim +827) (2Qufat of’), my = — Kof’(«) (Sim +822), mG, = Kuf' (a) (Sim + Son), mGg = Kof’ (a) (Simi? + S277). Now for equilibrium we must have Sim t+ Sen = 0. In this case, G,, G,, and X¢ vanish, and therefore, from (6), we see that the coefficients C, D, E, will be zero, and the equilibrum is critical or neutral. [This result, which is also evident from first principles, holds good equally when the centre of gravity is not in the plane of the lamine. In order to make the system really stable, the lamine must be inclined at small angles to the line joining their centres. Calling these angles 6, and f2, neglecting ¢ (a), and writing a =a — P,,a’ =a — fo, we get in the equations (8), mXy = KS; Quf(@’) + of («’)) cos fi, m&y, = K2S; (Qef (a) — wf’ (z’)) cos Bi, mX— = — KES (2uf (2) + vf’ (@)) cos fy, mYy, = KS; (2uf («’) + vf’ (&)) sin Bi, mY, = KS, (ef (2’) — uf’ (z’)) sin fi, mY = —K2Sym (2uf (2’) + of’ (’)) sin fi, mGy = —Kvzsipis («), mG, = Ku2S8; pif’ («), mG6 = KerSin pif’ (@’). | | The conditions for steady motion give O = kS)f (@’) sin By + kSof (2) sin Bo, mg[V? = kS11f (a@') cos 8, + kS8of (2) cos Bo, O = Si f(a’) 7 cos B+kS,f (a) n2 Cos Bo. With these substitutions, those of the nine coefficients X,,... Gg which vanish in the limiting case of coplanar lamin are given by 110 Dr. G. H. Bryan and Mr. W. E. Williams. = [Jan. 7 9 ee Gary Colo ie) nem a ee) Tee t, X6 a Te u V?a2—m LF (@) f(a") Vu Yo gis brsim Ba fie) ae ow Wsm(B,-By LF @) Fes’ and an expression for Yg, which is not required. The remaining coefficients may for a first approximation be taken the same as in (9). We notice the following points. (1) The expressions for X, Yu, Y., Gu and G, contain the factor Ft’ (@) [fi (e’) — f' (@’)/f(e’). In the laws of resistance commonly assumed ff’ (a)/f(a) decreases as « increases, and this coefficient, therefore, becomes negative if a’ O appears at first sight difficult to reconcile with the smallness of the coefficients C, D, E for small inclinations of the planes, as BE is of a lower order of small quantities than BCD. To satisfy this fifth condition, however, the important thing is to make E small compared with C. Now C will be found to consist of two parts, one independent of V and the other negative and proportional to 1/V?, while E is proportional to 1/V?, hence stability can best be secured by making V? sufficiently large. ] (4) If the planes are not infinitely narrow, it will be found that Xu, Xv, Yu, Yy are the same as before, and —mG,/v, and mG,/wu are both inereased by 2KSia; (f(a) (a) + f(@) ¢ («)). This will have the effect of altering C, HE, D in the same proportion, and will, there- fore, not alter the stability, except when « = 0, when the additional term will prevent E,D from becoming zero. Gé is increased by the term v>KSiaym (f'(e)o(@) +f(v)¢(«)). This term will, in general, be negative, and therefore Gg will be diminished, which will have the effect of diminishing B. 6. Examples of Two-Plane Gliders. We shall now assume certain particular values for the ditmencreme of gliders of this form, and proceed to calculate the conditions of stability. Ex, 4.—Let us first consider the case of the two slats of equal area set at angles such that ay = 1b a, = Die 1904. | The Longitudinal Stability of Aerial Gliders. ih gi) Then we have from Langley’s results, f(@)=056, f@y=1, f(a”) = 0°28, f(a") = 2°5; ata") = —ef(m) therefore, if we put 7; = a, we have n2 = 2a. Let 4? be assumed = 222. Substituting in the expressions for X,, etc., we have for equilibrium mX, = 3°6KSV, mG, = 4KSVa, Moke = WISN: mG, = 0'7KSVa, mX4 = 4KSVa, mGg = 11KSVa?, and therefore, I 4 | A = 202, B= 680<, C V 156a — 22800a?/V?, —— 5 oe —— & y gat $ 7 ou D 2121 > E 297000 =, 5 H 7.10° =, yall | é a a. a3 —2 9.108 —,-1 ST rie This is positive if V? > 590a. | Hx. 5.—Let us now take the case of two equal square planes inclined at a small angle to each other. Let a= Or By = — 5°, Bo = Di and let 2a be the breadth of either plane, For these angles we have f(a — Bi) = 0°44, Ff (@1— Bi) = 14, Ff (a — Be) = 0°15, Ff’ (4 — B2) = 18, CN BY 0.45% MeO) ans hy See 0155. B' (a = fs) = 0°58. For equilibrium we must have f (41 — Bi) (m+ O$ (a1 — B1)} = f(e2 — B2) (n2 + OP (42 — Br2)}, whence 371 + yz = 0°28a ; therefore, if we put n2 = — 4a, we have 7; = 1:4a. Let /? be assumed 112 Dr. G. H. Bryan and Mr. W. E. Williams. [Jan. 7,. to be = 4a?. Substituting these values in the expressions-for the coefficients, we have MX, = 3'6KSV, ©) iN, — O021KS ve mX, = 0:58KSV, mY, = —O:1KSVe mX4 = 1:8KSVa, mY6 = O:00KSV, mG, = 2°61KSVa, mG, =0°44KSVa, mG, = 30KSVa?. Substituting again in the expressions for A, B, C, D, K, we obtain, putting kh? = 4a’, 2 2 A = 402, iB =2380— C= 42013100 a v? a 700 ye’ D = one E = 2840002 . 5) y 840 V2 H is positive if V2 > 2000a, which is the condition of stability for this form of glider. By putting / = 7-4a, 1 represents approximately the extreme length of the glider, and the condition of stability reduces to V2 > 2701. Hix, 6.—Consider next the case of two unequal square planes, inclined at an angle, and in the first case suppose that the smaller plane is in front. Let S and S’ be the respective areas, and suppose S = 108’. Let 2a be the breadth of the large plane, and 2a/3:1 that of the smaller, also let the distance between the centres be equal to 3a; and leugees— a2. Let the angle « between the large plane and the direction of motion be 10°, and let the angle between the two planes be also 10°, so that the small plane is inclined at an angle of 20° to the direction of motion. Then we have f(a) =03, f(a) =16, p(a) = 0°49, f(a) = 0°59, ° f(a+B)=05, f’(a+f) = 157, d(a+f) = 04, ¢(a+B) = 056. For equilibrium we have, if 7 be the distance between c.g. and centre of large plane, Sf («) (9 — ah(a)) = SP(a +P) O+a' $ (a+ B) —). Therefore 7 = 0°94a, so that y, = b—7 = 2:06a. Substituting in the expressions for X,, etc., i OTS, IX, 0 ANOS mX4 = 1:4 KSVa, mGy = 0°57 KSVa, . mG, = 0:097 KSVa, mG6 = 1:13 KSVe?. 1904] The Longitudinal Stability of Aerial Gliders. 113 Substituting these values in the expressions for A, B, etc., assuming that 4? = a2, we have y J, eR Bh 213 Cr )035 = 15050 475 a a a4 “? = — } = YD —— i fu S peuaae ee Oe 4 Diss Dre OL. v? 1B, =I 0000 <5, Jal == I, 10 Ve 123 lO vi This will be positive if V? > 10400. This condition will also make C positive, and is therefore the condition of stability. _ Ex. 7,—Let us now suppose the small plane to be placed behind the other § ‘ye being inclined to it at an angle of 5°, the direction of motion again making an angle of 10° with the large plane and /? being = «?, as TDi iag, 6. In this case we have Ihe) == (rs) Gy = We ae) ON aye) =e OI), aa Om fag) WB. aia) = 0:DD,)'s bi (a) 056. If the distance between the centres of the planes be 3a, the condi- tions of equilibrium give n = —0°35¢, m = — 2°6da. Substituting, we obtain Wi xa LDA V a, mGy, = 0°33 KSVa, fi == OBIS Gi. Tr OOD ORS a, mX = 0°65 KSVa, mGe = 1:32 KSVa?. i. 2.65. B 936 7 C = 34:80 — 2400 - earrney e! peeisoggol® "> Be 18108 3" 199 @ D= 17805, E = 80000, 12. 10° 7, - 3.10 H will be positive if V? > 250u. The condition of stability is therefore that V > ./(250a). (7. Lffect of Moment of Inertia on Stability. It will be seen from (5) that the radius of gyration, /?, occurs only in the expressions for A, b, the first two coefficients of the deter- minantal equation. We have A = /?, B = i? (X,+Y,)+Ge. In all the cases considered X,,+ Y, and Gg¢ are positive, and therefore these two cceflicients are positive for all values of 4°. VOL. LXXIII. I 114. Dr. G. H. Bryan and Mr. W. E. Williams, — [Jan. 7, Taking the expression | H = BCD - AD? - EB», it may be written CDG¢ — EGe? +? {(X. + Y,) CD — D? - 2 (Xu + Y,)GgH} — /4(X,, 4+ Y,)?E. H is therefore increased with increase of i” if (X.+ Y,) (CD — 2GgE) — D? — 2k? (X,,+ Y,) E > 0. At the critical velocity, if H = 0, this becomes CDG6 + (Xz, + Y,)2Ekt< EG,?. Unless this condition is fulfilled, the critical velocity given by H = 0 increases with increase of /?. In all the numerical examples considered above, CDGg > HG@? and E is positive, so that the critical velocity is increased by increasing £?. Thus in Ex. (1) 4? was taken equal to a7, and the critical velocity obtained was ,/(774a); if, instead, we had taken /? = a, the critical velocity would have been ,/(1240a).—Jan., 1904. | 8. Character of the Fluctuations about Steady Motion. Mode in which the System Overturns. The character of the fluctuations about steady motion depends on the nature of the roots of the biquadratic (4a), the expressions for the displacements du, 6v, 66 being evidently of the following forms :— Rorerootsallirealt .cs eee eee CEM + Co@Aat + cgerst +. Cet, For roots two real, two imaginary ... ce + yest + yet cos (ft — €). For two pairs of imaginary roots ...... yen" cos (Pit — 1) f + ye! cos (Pat — €2). The last form indicates two different sets of undulations of different lengths. Photographs of the paths of gliders taken by magnesium light distinctly show these two undulations, thus confirming our theory. An important further consequence is that a glider may perform ~ undulations decreasing in amplitude, corresponding to a pair of complex roots of the biquadratic with their real part negative, but the motion may be unstable through the other roots having their real part positive, or one or both of them being real and positive. This indicates a real danger in experimenting with gliders. Stability may be broken either if a real root of the equation (4a) changes from negative to positive, or if the real part of a pair of imaginary roots changes from negative to positive. The condition for the latter is that H = 0, whereas a real root 1904. | The Longitudinal Stability of Aerral Gliders. iLiL5 changes sign if HK = 0. If stability is broken by a fall of velocity the only quantities which can vanish are H and ©, of these H is more difficult to make positive than C, and hence it appears that the most likely way for a glider to overturn in general is by commencing with a series of oscillations of increasing amplitude. This again agrees with experiments. 9. Conclusions. 1. For a glider or other body moving in a vertical plane in a resisting medium of any kind whatever, the small oscillations about a state of uniform rectilinear motion are determined by an equation of the fourth degree, so that the conditions for stable steady motion are those obtained by Routh. 2. The coefficients in the period equation involve, in addition to the ordinary dynamical constants, nine quantities X,...G¢, which, when referred to rectangular axes fixed in the body, represent the differential coefficients of the forces and couple due to the aerial resistances with respect to its translatory and rotatory velocity components. 3. In the case of a system of aeroplanes these nine quantities can be expressed for the separate planes in terms of /’(a) and ¢'(a), where f(a) and (a) are functions determining the resultant thrust, and the position of the centre of pressure when the direction of the relative motion of the air makes an angle a with the plane. These functions have been tabulated for certain different forms of surfaces, but further data are greatly needed. _ 4. The longitudinal stability of the gliders is thus seen to be capable of mathematical investigation, and it is of paramount importance that the present methods should be practically applied to any aerial machines that may be designed or constructed before any actual glides are attempted. 5. The methods of calculation are exemplified by numerical deter- minations of the criterion of stability in the cases of a single plane lamina and a pair of planes one behind the other. Most of the calcula- tions have been performed for an angle of gliding of 10° with the horizon, and it has been necessary to assume arbitrary values for the moment of inertia of the lamina. 6. The condition that any steady linear motion may be stable in all these cases assumes the form V2 > ka, where a is a constant depending on the linear dimensions of the glider, and /: is a constant depending on its shape, the angle of gliding and the law of aerial resistance. 7. For a pair of narrow slats, in which the variations in the positions of the centres of pressure of each are neglected, certain coefficients ot stability vanish. If the planes are square so that the displacements of the centres of pressure are not neglected, the system is less stable than a single plane of breadth equal to one of the squares. 8. By inclining the planes at a small angle to each other the stability AYR ts it Pe eno 116 Mr..E. 8. Salmon. Ovultwral Experiments urth [Dee: 2, : is much increased. On the other hand, if they are made to slope away from each other the ghder becomes unstable. 9. Two square planes of equal size placed one behind the other at a small angle are less stable in the examples considered than a square equal in size to one of the two, but more stable than a single square whose side is equal to the total length of the ghder formed by the pair. 10. A pair of unequal squares of which the smaller forms a rudder are more stable, in the examples considered, when gliding with the rudder behind than with the rudder in front. 11. In general, the stability is increased by making the moment of inertia as small as possible. “ Cultural Experiments with ‘ Biologic Forms’ of the Hrysuphacea.” By ERNEST 8S. SALMON, F.L.S. Communicated by Professor H. MArsHaLL Warp, F.R.S. Received December 2, 1903,— Read February 4, 1904. (Abstract. ) In the introductory remarks the author poimts out that through specialisation of parasitism “biologic forms” have been evolved in the Erysiphacee which, both in their conidial (asexual) stage and ascigerous (sexual) stage, show specialised and restricted powers of infection. -The powers of infection, characteristic of each “biologic form,” are under normal conditions sharply defined and fixed, and hitherto the result of the experiments of numerous investigators—both in regard to the present group of fungi and to the Uredineee, where the same specialisation of parasitism occurs—has been the accumulation of evidence tending to emphasise the immutability of ‘“ biologic forms.” The second part of the paper gives the result of cultural experi- ments with “biologic forms” of Lrysiphe Graminis DC., carried out during the past summer in the Cambridge University Botanical Laboratory. It has been found that wnder certain methods of culture, in which the vitality of the host-leaf is interfered with, the restricted powers of infection, characteristic of “ biologic forms,” break down. In the first method of culture adopted, the leaf, which was Sie attached to a growing plant, or removed and placed in a damp chamber, was injured by the removal of a minute piece of leaf-tissue. In this operation the epidermal cells on one surface, and all or most of the mesophyll tissue, were removed at the cut place, but the epidermal cells on the other surface (opposite the cut) were left un- injured. Conidia were sown on the cuticular surface of the uninjured epidermal cells over the cut. In a few experiments the conidia were 1903.] Biologie Forms” of the Erysiphacee. ity sown on the internal tissues of the leaf exposed by the cut, and these gave the same results. Using this method of culture, over fifty successful experiments, of which details are given, were made. In these the conidia of certain ‘biologic forms” were induced to infect “ cut” leaves of host-species which are normally immune against their attacks. | The experiments proved that the range of infection of a “ biologic form” becomes increased when the vitality of a leaf is affected by injury, and also that species of plants “immune” in nature can be artificially rendered susceptible. Further experiments showed that the conidia of the fungus produced on a “cut” leaf are able at once to infect fully uninjured leaves of the same host-species. In other experiments, a method suggested by Professor H. Marshall Ward, with the object of avoiding lesion of the leaf, was adopted. Leaves were injured by touching the upper epidermis for a few seconds with a red-hot knife, and conidia were sown on the injured place. It was found that the cells immediately surrounding the place of injury were rendered susceptible to the attacks of a “biologic form” which is unable to attack uninjured leaves of the plant in question. In the third part of the paper, dealing with general considerations, the following hypothesis is advanced as to the actual manner in which the injury to a leaf causes it to become susceptible to a “hiologic form” otherwise unable to infect it. It is supposed that the leaf-cells of each species of host-plant contain a substance or substances— possibly an enzyme—peculiar to each species which, when the leaf is uninjured and the cells are vigorous, are able to prevent the successful attack of any mildew except the one “biologic form” which has become specialised to overcome the resistance. When the vitality of the leaf, however, becomes affected by injury, this substance is destroyed, or becomes weakened, in the leaf-cells in the neighbourhood of the injury, so that the conidia of ofher “biologic forms” are now able to infect them. The author suggests that injuries to leaves, caused in nature by hail, storms of wind, attacks of animals, etc., may produce the same effect as the artificial injuries described above in rendering the injured leaf susceptible to a fungus otherwise unable to infect it. Conidia produced on these injured places would be able to infect uninjured leaves, and would spread indefinitely. Such may be the explanation of a common phenomenon—the sudden appearance of disease caused by parasitic fungi on plants hitherto immune. A case is described which, it is believed, gives evidence that the injuries produced by Aphides caused leaves previously “ immune ” to become susceptible. In the concluding remarks, reference is made to the antagonistic 7 Ped kn 118 ‘Mr. G. Massee. > Seas forces concerned in the evolution of a “biologic form,” viz., “specialising factors” and “ generalising factors.” Attention is also drawn to the close parallel between (1) the behaviour of the fungus in the experiments in which the conidia were sown on the tissues of the leaf exposed by the cut; and (2) the biological facts obtaining in the class of parasitic fungi known as “wound parasites ” (Nectria, Peziza willkommit, ete.), which are able to infect their hosts only through a wound. “On the Origin of Parasitism in Fungi.” By GEORGE MASSER, Principal Assistant. Herbarium, Royal Gardens, Kew. Com- municated by Sir WiiuiaAmM T. THiseLron-Dyer, K.C.M.G., C.LE., F.RS. Received January 11,—Read February 4, 1904. (Abstract.) Up to the present no definite explanation has been offered as to why a given parasitic fungus is often only capable of infecting one particular species of plant. This, however, is well known to be the case, for although the spores of fungus parasites germinate freely on the surface of any plant when moist, infection only takes place when the spores germinate on the particular species of plant on which the fungus is known to be parasitic. ‘This apparently selective power on the part of the fungus I consider to be due to chemotaxis. An extensive series of experiments were conducted with various species of fungi, including Saprophytes, facultative parasites, and obligate parasites, and the results are given in tabulated form i the full paper. The chemotactic properties of substances occurring normally in cell-sap were alone tested ; among such may be enumerated saccharose, glucose, asparagin, malic acid, oxalic acid, and pectase. In those instances where the specific substance, or combination of substances, in the cell-sap assumed to be chemotactic could not be procured, the expressed juice of the plant was used. These experiments proved that saprophytes and facultative parasites are positively chemotactic to saccharose, and this substance alone is sufficient in most instances to enable the germ-tubes of facultative parasites to penetrate the tissues of a plant, unless prevented by the presence of a more potent negatively chemotactic or repellent substance in the cell-sap. As an illustration, Botrytis cinerea, which attacks a greater number of different plants than any other known parasite, cannot infect apples, although saccharose is present, on account of the presence of malic acid, which is negatively chemotactic to the germ-tubes of Botrytis. 1904. | On the Origin of Parasitism in Fungi. 119 In the case of obligate parasites the cell-sap of the host-plant proved to be the most marked positive chemotatic agent. Malic acid is the specific substance that attracts the germ-tubes of Monilia fructigena into the tissues of young apples ; whereas the enzyme pectase performs the same function for the germ-tubes of Cercospora cucumis, an obligate parasite on the cucumber. Immune specimens of plants belonging to species that are attacked by some obligate parasite owe their immunity to the absence of the substance chemotactic to the parasite. Purely saprophytic fungi can be educated to become parasitic, by sowing the spores on living leaves that have been injected with a substance positively chemotactic to the germ-tubes of the fungus experimented with. By a similar method of procedure, a parasitic fungus can be induced to attack a different species of host-plant. These experiments prove what has previously only been assumed, namely, that parasitism in fungi is an acquired habit. A series of experiments prove that infection of plants by fungi occurs more especially during the night, or in dull, damp weather. This is due to the greater turgidity of the cells, and also to the presence of a larger amount of sugar and other chemotactic substances present in the cell-sap under those conditions. VOL LXXIIL. K 120 Dr. J. A. Ewing and Mr. L. H. Walter. — [Jan. 11, ‘“A New Method of Detecting Electrical Oscillations.” By J. A. Ewine, LLD., F.RS., and L. H. Watrer, M.A. Received January 11,—Read February 11, 1904. The magnetic detector of Rutherford,* though now well known, appears to have created little interest until attention was directed to the subject by Marconi’s adaptation of the method to his telephonic detector. Marconi’s apparatus employs the change of hysteresis, which is produced in iron by the influence of electric oscillations, when these are caused to pass through a coil surrounding the iron, the change being made manifest by means of a telephone. In his view, the electric oscillations act by reducing the hysteresis. T It occurred to us to exhibit the alteration in hysteresis by a different method, namely, by applying the principle which is used in an instrument invented some years ago by one of us for the mechanical measurement of hysteresis. In that instrument} the hysteresis is measured by the mechanical couple between a magnetic field and the iron, when either the iron or the magnet providing the field is caused to revolve. Thus, if the field revolves, the iron tends to be dragged after it, as a consequence of hysteresis in the reversals of its magnetism, and if the motion is prevented by a spring or other control, it assumes a deflected position. Suppose, now, the electric oscillations to act on it, any change of the hysteresis caused by them will be exhibited by a corresponding change in the deflection. We anticipated, in accord- ance with the generally accepted view that hysteresis is reduced by the oscillations, that their presence would be detected by a fall in the deflection. With this expectation an experimental apparatus was arranged, consisting of an electro-magnet, capable of being rotated on a vertical axis by an electric motor. The magnet poles were bored out circular, and between them was suspended, by a phosphor-bronze strip, a ring made up of three thin, flat annuli of soft iron, clamped together, and provided at the foot with an axial pivot. The ring was free to turn inside of two bobbins wound with fine copper wire, the windings being at right angles to the plane of the rmg. ‘Through these copper windings, electrical oscillations, produced in the usual manner by means of a distinct spark-gap, were passed. The first experiments resulted in a very small deflection from the position due to normal hysteresis, indicating, as was expected, a decrease of hysteresis when the oscillations arrived. The apparatus * ‘Phil. Trans., A, vol. 189, p. 1, 1897. + ‘ Roy. Soc. Proc.,’ vol. 70, pp. 8341—-344, 1902. + ‘Journ. Inst. Electr. Engineers,’ vol. 24, pp. 398—430, 1895. 1904.] A New Method of Detecting Electrical Oscillations. 121 was also tested with an alternating current of about 100 periods per second in place of the oscillations, with the effect that the normal hysteresis deflection was almost entirely wiped out. Various other forms were also tried with indifferent results, when it occurred to us that there would be advantages in passing the oscilla- tions through the magnetic material itself, making it of magnetic wire. A small bobbin was therefore wound with insulated soft iron wire, and the ends soldered to the upper and lower halves of the spindle, which was itself divided at the centre, the upper half bearing the controlling spring, and the lower dipping into mercury, from which a connection led to the other terminal. On passing oscillations through this winding, a remarkable and unexpected result was obtained. The change of deflection was much more marked than in the former experiments, and was in the opposite sense, indicating an increase of hysteresis while oscillations were present. Afterwards, hard steel wire was substituted for the soft iron, and a very great increase in the effect was observed, still in the same direction—that of increase of hysteresis. | Owing to these encouraging results, it was decided to continue the experiments in this direction, abandoning the older form, in which a decrease of hysteresis was dealt with. The first bobbin constructed was about 3°; inch in external diameter, and had a vertical wire space of 4+ inch. The winding was a single No. 32-gauge iron wire, double cotton-covered, wound straight round from beginning to end. Later, No. 40 and No. 46 steel wires were employed, of which the latter gave the best results. It was soon noticed that any method of increasing the oscillatory current in the wires, as by winding the bobbin with two wires having a slightly unequal number of turns, was of advantage in giving a larger deflection. Later a fine copper wire secondary, wound on the bobbin parallel to the magnetic wire, was tried, first with the ends insulated, and then with the ends soldered together. A marked increase in deflection was observed when the secondary was closed, showing that the magnetic nature of the wire itself was influential. Accordingly, a bobbin was then wound with insulated steel wire, doubled back on itself. This non-inductive winding gave by far the best results hitherto attained, and is now used, except when special results are required. The instrument, though described as a detector of electrical oscilla- tions, may be said to measure rather than detect, giving quantitative as well as qualitative results, and being capable of regulation from a sensibility of the same order as that of an average coherer down to practical insensibility to powerful sparks in the same room. In the instrument, as shown in the figure, the electro-magnet takes the form of a ring capable of moving round a vertical axis, and is K 2 122 Dr. J. A. Ewing and Mr. L. H. Walter. — [Jan. 11, provided on the interior with two long wedge-shaped pole-pieces, M, M, the current to the winding being supplied through brushes | rm, = 77 i 10 HT Tt bearing against insulated rings below. The magnet is made to revolve by an electro-motor, the best speed being about five to eight revolu- tions per second, but the electro-magnet may be replaced by a 1904.] A New Method of Detecting Hilectrical Oscillations. 123 permanent magnet system giving a similar field. A structure is built up, external to the magnet, to support the vessel containing the pivoted bobbin and its centreing arrangements. The bobbin itself is made of bone, and is about 2 inches long. It is provided with a steel spindle at each end bearing in a jewel hole, the two halves of the spindle being insulated from one another. ‘The winding, which is, as far as possible, non-inductive, consists of about 500 turns of No. 46-gauge hard-drawn steel wire, insulated with silk. The bobbin is immersed in petroleum, or a mixture of petroleum with thicker mineral oil, which serves the double purpose of fortifying the insula- tion, and giving the damping effect necessary to steady the deflection due to the drag of the revolving magnet. Neadings are taken by means of a spot of light, as with speaking mirror galvanometers, but a siphon-recording attachment has’ been fitted, and any form of contact for working a relay could be employed. The detector, as before mentioned, gives quantitative readings, and, in some cases, the deflection may be too large to be easily read by the scale. For this purpose a variable shunt is provided, by which the deflection can be regulated. For the purpose of wireless telegraphy, the instrument has the advantage of giving metrical effects. The benefit of this in facilitating tuning, and in other respects, need not be insisted upon. From the physical point of view, the augmentation of hysteresis is interesting and unlooked for. It is probably to be ascribed to this, that the oscillatory circular magnetisation facilitates the longitudinal magnetising process, enabling the steel to take up a much larger magnetisation at each reversal than it would otherwise take, and thus indirectly augmenting the hysteresis to such an extent that the direct influence of the oscillations in reducing it is overpowered. The net result appears to be dependent on two antagonistic influences, and, in fine steel wire, under the conditions of our experiments, the influence making for increased hysteresis, as a result of the increased range of magnetic induction, is much the more powerful. 124 Mr. E. Matthey. [Jan. 11, “ Constant-Standard Silver Trial-Plates.”. By Epwarp MArTruHey, C.B., F.S.A., F.C.S., Assoc. Roy. Sch. Mines. Communicated by Sir WILLIAM CROOKES, F.R.S. Received January 11,— Read February 11, 1904. Referring to my paper communicated to the Royal Society, February 16, 1894, and read March 15, 1894,* in which it was shown that moderately sized plates of sterling silver of a uniform standard could be obtained by casting from thin castings, my attention since then was drawn to the difficulty of casting larger quantities than those described in that paper, which were only of an average weight of 4to5 kilogrammes per plate, and to the desirability of obtaining a large plate, say of some 8 or 10 kilogrammes in weight, without difficulty, and I have therefore resumed my “attention towards effecting this. It appears that considerable difficulties have been experienced with regard to obtaining large plates of constant standard. In the Royal Mint Report of 1873 is a memorandum appended by Professor W. Chandler Roberts, which refers to a series of well-known experiments with regard to obtaining a constant alloy of 0-900 silver by Levol in the Paris Mint, he himself being at the time engaged in the preparation of a standard silver trial-plate. Professor Chandler Roberts states :—‘“‘ From the foregoing remarks it will be evident that it as mmpossible to cast a standard silver plate or bar of wniform composition, and it was necessary therefore to resort to an artifice in order to obtain a standard trial-plate of the required dimensions.”7 The means adopted by him were to cast 1000 ozs. of standard silver into a skillet-mould 30 cm. long, 25 cm. wide and 5 cm. broad, to plane off 4 mm. from its surface and to roll the planed skillet to 1:8 mm. thickness, a sheet being produced 1:5 m. long, 45 cm. wide. From this the portion was cut which showed constant results about 925, but which varied from 924°6—925:°1; and the rest of the plate which varied from 924 (lowest) to 928-4 (highest) was abandoned. He shows all these results by diagrams accompanying his memoranda. The portion of available constant standard so cut out from the 1000 oz. sheet weighed 104 oz., about one-tenth of the whole plate. And this 104 oz. (= 3230 kilogrammes) formed the mint trial plate from a mass of 1000 oz. (= 31°103 kilogrammes) specially cast for | the purpose. Notwithstanding that many experiments were subsequently made by Professor Roberts Austen to find a means of obtaining a constant standard trial-plate, in 1899 he was compelled to resort to what he calls the ‘‘cumbrous expedient” of 1873. His statement is :—‘“ None * “Roy. Soc. Proc.,’ vol. 55, 1894, p. 265. + Fourth Annual Report, Deputy Master of Mint, 1873, pp. 44—46. 1904. | Constant-Standard Silver Trval-Plates. 125 of the results were satisfactory, and eventually, after no less than 31 plates had been cast without success, recourse was had to the method adopted in 1873, of casting a large mass of metal and detaching a particular portion which proved by assay to be of approximately uniform standard.”* ) The casting of the standard silver into thin instead of into the thick moulds ordinarily employed having been attended with such excellent results,} I was induced to believe that it must be due to the more rapid cooling of the metal, by which liquation was arrested; so that if the standard silver were cast into moulds sufficiently cooled, liquation might be induced to disappear altogether. In order then to overcome the difficulty of obtaining the standard trial-plates in larger sizes, I commenced by casting quantities of not less than 8 kilogrammes into a mould cooled externally by ice, and also by freezing mixtures as low as 10° C., and by these means I obtained most encouraging results— results which confirmed me in the supposition that by cooling with rapidity there is less time for liquation, which appears to be the direct converse of what has been supposed hitherto, viz.: “that a uniformity of standard was best attained by slow and uniform cooling.”{ But although I thus obtained exceedingly good results (two of these are subjoined, see p. 126), I was not satisfied that this was the best way of producing a constant-standard plate. I therefore adopted a different method of casting the standard silver. Instead of pouring the melted alloy into a mould from the top, I poured it into a mould by which the skillet was produced from the bottom, thus :— A Rough Section of Cast-iron Mould employed. By pouring the alloy into the gate A the metal passes by B into the space C. And instead of cooling the mould by ice or freezing mixture I used the mould simply cold. I have obtained excellent results by * Thirtieth Annual Report, Deputy Master of Mint, 1899, pp. 69, 70. + Vide my paper of February 16, 1894, before referred to. t Vide memorandum already referred to in Mint Report, 1873. 126 Mr. E. Matthey [Jan. 11 ‘hing 6°624 kilogrammes, 10) fo) i We Both 75 em. by 75 em. and 1 mm. in thickness. Weighing 6°655 kilogrammes. this method. Subjoined is one of the plates produced rolled to 1 mm, thick, measuring 90 cm. x 75 cm. and weighing 8-700 kilogrammes. Trimming the rough edges from the plate about 1 cm., the results came out as under, showing a constancy of 925 nearly all over the plate. I have drawn a line where the only wave of lower variation occurs. 1904. | Constant-Standard Silver Trial-Plates. 127 Weighing 7°870 kilogrammes, 12) 925 925 925 925 925 75 cm. by 90 cm. and 1 mm. in thickness. Another cast simply into a cold mould, and rolled to a thickness ot 1 mm., gave— This only weighed 3°640 kilogrammes, is without any wave of variation, and is absolutely constant. The hitherto accepted theory “‘that the molecular rearrangement is comparatively slight if the mass of metal is slowly and uniformly solidified ”* is contradicted by the results I have obtained; and the results all bear out the fact that there is little or no difficulty in obtaining an 8 or 10-kilogramme plate of constant-standard silver, or even more if necessary. * Fourth Annual Report, Deputy Master of Mint, 1873, p. 46. 128 Dr. A. E. Wright and Capt. 8. R. Douglas. [Jan, 11, “ Further Observations on the &éle of the Blood Fluids in con- nection with Phagocytosis.” By A. E. Wricut, M.D., late Professor of Pathology, Army Medical School, Netley, Pathologist to St. Mary’s Hospital, W., and STEwarT R. Dovetas, M.R.C.S8., Captain, Indian Medical Service. Com- municated by Sir J. BURDON SANDERSON, F.R.S. Received January 11,—Read February 25, 1904. (From the Pathological Laboratory of St. Mary’s Hospital, London, W.) [Pate 3.] In a previous communication we showed that the phagocytosis which occurs when cultures of the Staphylococcus pyogenes are added to human blood, is directly dependent upon the presence of certain substances in the blood which exert a specific effect upon the bacteria. We suggested that the bacteriotropic substances here in question might appropriately be denoted by the term “ opsonins.” In the present paper we propose to bring out certain further points in connection with the “ opsonic power” of the blood. RELATION OF THE Opsonic POWER oF HUMAN BLOOD TO THE CAPACITY OF RESISTING INVASION BY THE STAPHYLOCOCCUS PYOGENES. It has already been shown* by one of us that patients who are the subjects of acne, sycosis, or boils are characterised by a defective phago- cytic power for the Staphylococcus pyogenes. We have recently been able to satisfy ourselves that this defective phagocytosis is dependent upon a defect of opsonic power. It has also been shown by one of us that the cure of these bacterial infections, which can in almost every instance be achieved by the inoculation of appropriate quantities of sterilised staphylococcus cultures, is associated with the acquirement of an increased phagocytic power. We have now succeeded in establishing the fact—already adumbrated in our previous paper—that the increased phagocytosis which is associated with the achievement of the condition of immuni- sation here in question is dependent, not upon a modification of the white corpuscles, but upon a development of opsonins in the blood fluids. The results of the subjoined experiment bring out this fact into clear relief. Details of the Hxperiment. Immunised Patients Blood.—The patient, F. F., who had long been the subject of aggravated staphylococcic sycosis, had, after prolonged and * ‘Lancet,’ March 29, 1902. 1904.] Role of Blood Fluids in connection with Phagocytosis. 129 ineffectual treatment with antiseptics, been subjected to three successive inoculations of a sterilised staphylococcus culture. Under these inocula- tions his clinical condition had ameliorated itself in an astonishing manner, and his phagocytic power, which had previous to the date of inoculation been less by half than that of the normal man who served as a control, had increased in a progressive manner after each inoculation. A sample of blood was now (by the technique elsewhere described)* drawn off and mixed with ;,th of its volume of 10 per cent. citrate of soda. A second sample of blood was drawn off and allowed to clot in the ordinary way. In the case of the first sample of blood the corpuscles were isolated from the plasma by repeated washing with physiological salt solution, and centrifugalisation. The corpuscles thus isolated are referred to below as ‘‘ washed corpuscles.” : In the case of the second sample of blood the serum was simply separated from the corpuscles in the ordinary way by centrifugalisation. Control Blood from a Normal Man.—The blood which served as a control was obtained from a normal healthy man. It was drawn off in exactly the same manner end was treated in each case by exactly the same procedures as the blood obtained from the patient. Bacterial Culture.—The bacterial culture employed in the experiments set forth below was obtained by suspending in physiological salt solution a portion of a 24 hours’ growth of Staphylococcus albus on agar. The quantities of serum, washed corpuscles, and staphylococcus culture which are specified below were then in each case taken up into a capillary tube, mixed on a glass slide, re-aspirated into the tube, and digested together at blood heat for 15 minutes. Films were then made and stained by Leishman’s stain. Finally the number of ingested bacteria were enumerated in a series of polynuclear W.B.C. taken in order as they came. The phagocytic index given below—and the same applies through- out this paper—represents in each case the average number of bacteria ingested by the individual P.W.B.C. The number of polynuclear white blood corpuscles which have furnished the index is in each case inserted in brackets :— Experiment. AG Immunised patient’s washed corpuscles... ....... 3 vols. immunised patients serum: <1... cot ea Bs Suspension of staphylococcus culture........-... 1 vol. Phagocytic index (20 P.W.B.C.), 25-7. * © Lancet, January 23, 1904. 130 Dr. A. E. Wright and Capt. S. R. Douglas. B: Washed corpuscles from normal man..... 3 vols. Serum from normal man ...... By Suspension of staphylococcus niliaker 1 vol. Phagocytic index (15 P.W. B. C. . 13. C. Immunised patient’s washed corpuscles .......... 3 vols Serum from normal man ...... seer Ses Suspension of staphylococcus aitbaree 1 vol. Phagocytic index (15 P.W.B.C. _ 13. ive Washed corpuscles from normal man .. 3 vols. Serum from immunised patient . sn otis Suspension of staphylococcus culture ............ 1 vol. Phagocytic index (15 P.W.B.C.), 28-2. EXPERIMENTS ON THE Opsonic POWER OF HUMAN BLOOD IN ITS RELATION TO THE BACILLUS OF PLAGUE. In these and all subsequent experiments, unless where otherwise specified, the technique employed was exactly the same as that employed in the experiments set forth above. It may further be premised that the bacterial suspensions employed were in each case suspensions of very young agar cultures—in most cases 24-hour cultures—in physiological salt solution. By the term “heated serum” is in each case to be understood serum which has been subjected toa temperature of 60° C. for 10 minutes or more. Experiment 1. A. S. R. D.’s unheated serum.......--.--....-- S. R. D.’s washed corpuscles. . Suspension of plague bacillus ........ Phagocytic index (20 P.W.B.C. ©): 3°0. B. S. he D.’s heated Serum eet epee erie S. R. D.’s washed per ee ie ee Suspension of plague bacillus. Phagocytic index (25 P.W.B.C.), 0°7. 3 vols. Briss 1 ¥ol; ~~ oo 4 “a [Jan. 11, 1904.] Role of Blood Fluids in connection with Phagocytosis. 131 Experiment 2. x: SetveDs UMheated SOVUMscc 02.5 <-.- es ceeisiaciee & VOLS. S. R. D.’s washed corpuscles.. She iter ckavehene owiehe eee 5 Suspension of plague bacillus......... beeies vevell. Ehaeouyie index (201 Wes. ae Pook. B. S. R. We SehesbeG SeKUlit e000. ede od oo bees oe VOLS. Seer DASuwasWed COLPUSCIES Geen ccs vce e ssn ce 8) 55 Suspension of plague bacillus.......... es ele vous Phagocytic index ee ie W- B.C: a. 8 Experiment 3. A. BED V AS mUMMeCAGEOL SELUMN went cis- 2c ec ee cose ease & VOlS: Sees wasted COrpuUseles:... <0. j..)¢ena.s sislelarscslem (8 05s Suspension of plague bacillus ......... : 1 vol. Phagocytic index (21 P.W.B.C. 4), USO) 134, AeebewNV. s heated Seri ¢..c..uss. 0. semsaceees 8 VOLS, Sele Di swasmed CORPUSCLES: ahi.c- cielo es oa wiceis ale Ol igy Suspension of plague bacillus........ nels ll Ol: Phagocytic index (54 P.W.B.C. 0), 8:4. Experiment 4. TX eee TS sy uheated serMii” 2). 15,9 ccpeceesnisin le wiey sia Po VOlss PAUSE. Via Wasted) COMPUSELES itd crnoctscm 2h Mom (yi cays Suspension of plague bacillus. plague Phagocytic index is P.W.B.C. ©),5 ee Eb as -silledheds SELUNMe st oo seis Sia sera visjosecdinyiouate a a VOLS: A. E. W.’s washed corpuscles. . EE en, ROE ae ie gee Suspension of plague bacillus. ce a eit Lvl: Phagocytic index es PW BC), 1-4. It may incidentally be noted in connection with these experiments that while the plague bacilli which lay free in the films were in each case quite unaltered, many of those which had been ingested showed 132 Dr. A. E. Wright and Capt. S. R. Douglas. [Jan. 11, extremely characteristic involution forms* such as we have not seen since we worked with freshly isolated plague cultures in Bombay in connection with the Indian Plague Commission. So typical were the involution forms of the ingested plague bacilli, that we should not hesitate to employ the method of phagocytosis as an aid to diagnosis in the case of a doubtful plague culture. EXPERIMENTS ON THE Opsonic POWER OF HuMAN BLOOD IN RELATION TO Micrococcus MELITENSIS. Experiment 1. A. S. RD 2s unheated serum. < os...) ose moe - 3vols. S. R. D.’s diego aenaneaet 3) Suspension of Micrococcus Melitensis............ 1 ,, Phagocytic index (10 P.W.BO.) 26°9 B. S..R. D.’s heated serumici.? e555 0S... 1 2S S.R: Di’s washed corpuseles;.. 3205.2). << ee Suspension of Micrococcus Melitensis............ 1 ,, Phagocytic index (10 P.W.B.C.), 9°2 Experiment 2. A. A. H.W 's unheatediserum yo. ses sce. cee sen ORM A.B. Ws washed uconpusclesc.. .. <<)... sek sec eee Suspension of Micrococcus Melitensis..........:- 1 ,, Phagocytic index (21 P.W.B.C.), 10-0. B. A; H. W.'s heated serum... oc -cicsciss se casanie=«) | Cumaiee AH). Ws washed corpuscles.” wer so. ss.) See Suspension of Micrococcus Melitensis...... 1 vol. Phagocytic index (21 P.W.B.C. \, 9 2°4, Hxperiment 3. A. S.'R. D's heated serum:.7) een oe es ts ee es A. i, W 28 washed ‘corpuscles seme. se feces el eee Suspension of Micrococcus Melitensis ..... 1 vol. Phagocytic index (21 P.W.B.C. \ DAD. * It may be observed that our plague culture—like other plague cultures which have been cultivated on artificial nutrient media for a number of genera- tions—has altogether lost the property of developing in a spontaneous manner tite involution forms which are characteristic of freshly isolated plague cultures. 1904.] Role of Blood Fluids in connection with Phagocytosis, 133 B. Serinerbe s heated. SOLUM. 2 cess cs ctoceseeteeeecee. | SVOls, Pee bPaw Ss Washed COrpuUuccles”.c vc. ve sate cee ss Oh 5g Suspension of Micrococcus Melitensis ......,..... l-vol. Phagocytic index (21 P.W.B.C.), 0:9. EXPERIMENTS ON THE Opsonic PowrER oF HuMAN BLOOD IN RELATION TO THE BACILLUS DYSENTERICUS (SHIGA). Experiment 1. Ia Seehie Ds univeated: SOGUMVEL. 10 sicis «/eie(- 4 eealaisawics'e . oO VOLS, S. R. D.’s washed corpuscles.. Iitwic wacker aieeslairearny Ten Mas Suspension of Shiga’s Paeillaa Sereereyetss: : 1 vol. Phagocytic index (20 W.P.B.C. ©), 4-2. B, SpleP he SuNGALOONSCRUM. |. si calcu wens oe ec ssesue es 8 VOIS, S. R. D.’s washed corpuscles ..... Riis ecEMle erase She Suspension of Shiga’s bacillus ........ rr 1 vol. Phagocytic index (20 P.W.B.C. ©), ¢ 0-0. Tixperiment 2 A. eer oWes unbenbed senume. © holes lua kcaet ce So vols: 8S. R. D.’s washed oa Se es Suspension on Sntasis DACwllUss.. 4. ss esses eee en Livolk Phagocytic index (20 P.W.B.C.), 5:4. B, BAN cs) heated serumn sites vias coon Ob VOB. Baek es wasmeds COCPUSELOSHs fie~ ae ciniae © cle hata IN yg Suspension of Shiga’s bacillus......... 1 yol. Phagocytic index (33 P.W.B.C. 0), ¢ Ork Experiment 3. A. Shh Baambeahemmserumy. sto eb. soe awe | 2UVOLS, Shae sh was hedkeGrpusGler. <0 solce tele 62-48 sc, 8 Sh ¥y Suspension of Shiga’s bacillus ........ 1 vol. Phagocytic index (20 P.W.B.C. ©): 3°6 Igt . Dr, A. E. Wright and Capt. 8. R. Douglas. [Jan 11, B. S. Ra D.’s heated serum (2.6 03.0.0 2c). he Wieck oe) Pome Se he 4s washed corpuscles! sce es isis . 2 Suspension of Shiga’s bacillus). ..2.).4)0s.. 4-16) leone Phagocytic index (20 P. W. B. C ) Opes A certain number of the bacilli (and these bacilli were found indifferently in the interior of the cells and free in the preparation) had, in the case of the experiments undertaken with unheated serum, undergone spherulation. EXPERIMENTS ON THE Opsonic PoWER OF HuMAN BLOOD IN ITS RELATION TO THE BACILLUS COLI. Experiment 1. A. IB. Hes 2s auniieate dis ertuanerme feo iels)>t-aeloi* <2 oie Senor 3 vols. iB; He S2sewashed tcovpusclesters 1.06 «1s 1s a) ee Suspension of the Bacillus Poth. Keg evens 1 vol. Phagocytic index (20 P.W.B. © 7 3°8. B. HTL, Ses Heated Sex Ui flere lieve rs) e/ieiecnel «ore cls a eke een B. H. S"siwachedieompneclea stallaiede alee isnloiels eke antes kOe Suspension of the Bacillus alin Be caste tae, as I Wolk. Phagocytic index (20 P.W. B ©), 0: 0-75, Experiment 2 A. K. F.’s unheated serum <5) lye. tes) oid sie cee acto ah, oO amOlee iy. Hes washed icompusclesien astelecr veters aijclele occ) | Ome Suspension of the Bacillus coli.......... 1 vol. ‘ Phagocytic index (20 P W. B. ©. SU: D. B. FE, Fs heated germs fo or. Miles «ie ralrsite ios id sialeteidel | tO Olee KV ERs washed corpse leswieeustieitscietetere ele! +11 Ommnre Suspension of the Bacillus coli......+..+. 1 vol. Phagocytic index (21 P W. iD: 0. , 0-76. —1904.] Role of Blood Pluids in connection with Phagocytosis, 135 EXPERIMENTS ON THE Opsonic PowER oF HUMAN BLOOD IN ITS _ RELATION TO THE PNEUMOCOCCUS OF FRAENKEL. Experiment 1. A. S. R. D.’s unheated serum..........+. nein. 8n Bn sel) 2 vols, | 8. BR. Ds washed corpuscles.........., diamioo- 274). op Suspension of the pneumococcus of Treniell, ee avo Phagocytic index (15 PWBO), 1 16. | B, S. R. D.’s heated serum o..... 2.000 see eieo cciee aa te MOlS: S. R. D.’s washed: corpuscles..,........ oy cidiar eel iel Lc ah ay Suspension of Fraenkel’s pneumococcus........-. 1 vol. Phagocytic index (40 P.W.B.C.), 1-1 Experiment 2, A, eel NNers: WolMeated SenUMAn rr oles <6 cies eliele Sor a OMlise Sp Ike ID. TS SABEMIEC, GoyONSOIES cong doooo8 sncoGoInnno Al ay Suspension of Fraenkel’s pneumococcus.......... 1 vol. Phagocytic index (23 P.W.B.C.), 6 B. A, KE. W.’s heated serum.. coe 2 vols. Seva Mamas Ne CNCOnMUSCleS atten cle) ying isin ciene sis LL) ayy Suspension of Fraenkel’s pneumococcus........- 4 vol. Phagocytic index (40 P.W.B.C.), 0:2 EXPERIMENTS ON THE Opsonic PowrER oF HuMAN BLooD IN ITS RELATION TO THE BACILLUS OF ANTHRAX. Experiment 1, Bs Shida, Diss minlseneel Sema BAAS Sooo mcanogcD sn oumo) aKOKE S. R. D.’s washed corpuscles........ coongaqaneo 8! By SuUsvensiomokOacvllus aNLNTACUS). 1.02.0 --42 6. | L VO). Enumeration was here impossible, but there was everywhere evidence of phagocytosis. In the few cases where the leucocytes had not ingested bacteria, they were found to have extended themselves in a characteristic grasping manner along the bacterial threads (fig. 5). VOL, LXXIL, Iu 136 Dr. A. E. Wright and Capt. S. R Douglas. [Jan. 11, B. S. K. Ds heated serum! — oo... 16 oes peer nasil nee S. BR. Dis washed corpuseles 2.8 e200. ee +s ee Suspension of the Bacillus anthracis..........--- 1 vol. Here there were practically no signs of phagocytosis. The cells were everywhere empty, and they had not drawn themselves into intimate contact with the anthrax threads (fig. 6). Experiment 2. A. A. H. Ws unheated serum =: «: ..:sc. cet be «cl sieeeeeavolss S. R. D2s washed’ corpuscles. .- .<.. . oelweieleleet te Cee Broth culture of anthrax ....... « + it lel alee tl earole Phagocytic index (36 P.W.B.C. ), 5 2°4 4 (approeneee only). B. A, EH. W.’s: heated serum . 24.2 s.cc0s «6 00. ax onc oe OISe S. R. D. kiran ae cease Pa Broth culture of anthrax ....... «0 0 occa aerate Phagocytic index (100 P. W. B C ), 0: Opsonic PowER OF HUMAN BLOOD IN ITS RELATION TO THE BACILLUS TYPHOSUS AND THE CHOLERA VIBRIO. It is well known that human blood exerts a very considerable bactericidal power upon cultures of the bacillus typhosus and of the cholera vibrio. The destructive effect in question manifests itself to microscopical observation in the form of very profound morphological changes which come under observation in cultures which have been digested with unheated serum. The bacteria in such cultures, after undergoing agglutination and spherulation, swell up and lose their chemical affinity for anilin dyes. Finally they are completely dissolved. It is manifest that where disintegrative changes of this kind are occurring under the influence of the serum, opsonic effects will be more or less thrust into the background. ‘These last will, in the case of phagocytic experiments conducted with unheated serum, be masked, on the one hand, by the fact that there will be fewer bacteria available for phagocytosis, and on the other hand by the fact that intracellular disintegration will, it may be presumed, be more rapid in the case where the serum has already exerted a disintegrative effect on the bacteria anterior to their ingestion. Lastly, ingested bacteria which have lost their characteristic chemical affinity for their stain may readily escape enumeration. Ma 1904.|] Role of Blood Muids in connection with Phagocytosis. 137 All these points must be taken into consideration in connection with the subjoined experiments :— Experiment 1. A Slt) amunineated users wai niseee.ce ce es dee cen a VOlS av Ms washledsCORDUSCLeS Gos..c sess eececass . 8 55 2 ” Suspension of the cholera vibrio .. Hiverywhere considerable phagocytosis. Complete spherulation of almost all the micro-organisms within and all the micro-organisms outside the cells. No indication of vacuolation round the ingested bacteria (fig. 3). Phagocytic index (14 P.W.B.C.), 24 (cire.). B. Serle) siMeateG STUN 4x6 csi cers alscsesecasciaees & VOLS. Seek Osi washed yeCOnpuscles 0.006 5.2 0.+0ce sce) | O45 Suspension of the cholera vibrio .........0+-.046 2 4, HKverywhere considerable phagocytosis. No spherulation of the micro-organisms either within or without the leucocytes. Very marked vacuolation of the leucocytes round the ingested bacteria (fig. 4). Phagocytic index (11 P.W.B.C.), 26:2 (circ.). Experiment 2. A. (ACMA VES unineated: SOLU! ces «ccs ace «es ceee os 3 vols S. R. D.’s washed corpuscles .........-. at Bot Adah 1 vol Suspension of the cholera yibrio Complete spherulation of all the bacteria, whether within or without the cells. Phagocytic index (21 P.W.B.C.), 8-1 (circ.). B. A. H. W.’s heated serum 2... .csees cece cess nene 3 vols Se ik. Dus washed corpuscles. « .isjcacceiss isce cso 88 45 Suspension of the cholera Vibrio ........ +e ee0. it vol: No spherulation of the micro-organisms, either within or without the leucocytes. Phagocytic index (13 P.W.B.C.), 0°8. 138 -- Dr. A. E. Wright and Capt. 8. R. Douglas. [Jan 11, Experiment 3. A. S. RB. D.’s unheated serum <...2)..¢0 cenbindsscns)s Somamuee 8. BR. D.’s washed corpuscles > sissctmcn ees «2 2 hoe Broth culture of the typhoid bacillus ........... 2 33 Much phagocytosis. Complete spherulation of all the extracellular micro-organisms. Many of the bacilli in the interior of the leucocytes have completely preserved their original contours, others—probably the later ingested ones—are spherulated (fig. 1). P B. S, Ry DMs heated seruniy sca. « . ss. ects el cers =) S, R. D/s washed corpuscles 2.) ../.2!. <<. ens = see Broth culture of the typhoid bacillus ............ 2. ,, Much phagocytosis. All the micro-organisms, whether within or without the leucocytes, are morphologically unaltered and have pre- served their staining properties unimpaired (fig. 2.). Experiment 4. A. A. E, W.’s unheated serum << ,.2...+50c sees cs =) emOlse S, R. D.’s washed corpuscles ,..... Sistemi cone OE Broth cultivation of the typhoid ee aie ahaha yehapiaees 29 Complete spherulation of all the extracellular bacteria which have escaped solution. In interior of leucocytes most of the bacteria have undergone spherulation, but in the centre of the corpuscles some— probably those which were soonest ingested—are morphologically unaltered and preserve their staining properties unaltered. Phagocytic index, 100 (estimated). A> Ei. W..'s heated sexim oe icyjectes ales u ae a) ste an ee S. R. D.’s washed corpuscles ...... sais siete | On Broth cultivation of the typhoid banalins «3 aa lave sieut eagle No spherulation, either within or without the cells. Phagocytic index (20 P.W.B.C.), 31°8 (circ.). Experiment 5. A. 8. BR: Di's-unheated ‘sermmix.72 rite ac os «a bialhale vec © cee oles Nv hh. D.'s-weshedseorpuselese tis... -' ee soos nomaes Suspension of the typhoid bacillus: 2... Seem vue) ude male All the bacilli both within and without the cells have undergone spherulation. Phagocytic index (11 P.W.B.C.), 13:6. 1904.] Réle of Blood Fluids in connection with Phagocytosis. 139 Be Seas neated: SOLU =, wees nadyeriersieciedaeesace & VOLS: S. R. ee ee one hated Sins Suspension of the typhoid bacillus. . ore outed. yi laVOls No spherulation either within or without the leucocytes. Phagocytic index (23 P.W.B.C.), 7:2. Of incidental interest in connection with the above experiments is the demonstration whichpthey afford, that the spherulation of the intra- cellular ingested micro- organisms, which has been often ascribed to the agency of the leucocytes, is in ey due to agency of the blood fluids. Opsonic PowER OF HUMAN BLOOD IN ITS RELATION TO THE DIPHTHERIA BACILLUS AND THE XEROSIS BACILLUS. Experiment 1. A. Peeters UMMeALCH SLUM. and basic constants k,= 1:2 x 107-4, 1:2 x 107, 1:2x 107, and 1:2 x 107! respectively, the corresponding ratios kp/K at 25° being 1, 10, 100, and 1000. The calculation was made for the most part by approximation, it being assumed in the first instance that the total concentration was equal to the concentration of the unionised substance. A second approximation in which the value of ~ obtained from the first calculation was adopted usually sufficed. If a great many dilutions have to be calculated, the employment of graphical methods may effect a saving of time. In the following table all values of a and d have been multiplied by 10°. For comparison, the values for /,/K = 0, z.¢., for a simple acid have been added. lip = 107°. | | k/K=0. | k/K=1. k|K = 10. ko| K = 100. i 1000. a. | d. a. | as a. d. a. | d, a. d. ihe | . gs waa J i Bip tes 1 | 316 0 | 224 | 223 95°31 943 |31°5 |3030 | 9-99 | 9091 10 | 100 Ou 405 8°53 | 70°53 | 69°53 | 30-1 | 291 | 9°94} 904 MOO |? 31-241) 0 31-0 | °0°3| 29°7| 2-9'| 2-1 | 211-0) Bape eae 1000| 97] Oo 95| O00] 94] O11} 9:06} O-8 | 6-79 ——— The values of a for a given dilution fall off as %, increases, and that * Loe. cit., p. 587. ™ 4 1904. | Theory of Amphoterre Electrolytes. 161 the more rapidly as the dilution is small. It will be noticed that / varies with the dilution much more rapidly than a, which for high values of i, becomes nearly independent of the dilution. What is most deserving of attention is that although for /,/AK = 1000 the value of the acid constant /, is still nearly a million times greater than the value of the basic constant /, the acidity of the amphoteric substance is greatly diminished at small dilutions, being for example, at v = 10 only one-tenth of that of a simple acid with the same constant. Although the concentration of the chief conducting ion of acids is thus greatly diminished, this diminution may be more than compensated by the comparatively great concentrations of the slower ion HX* which appear at the same low dilutions. With regard to the negative ions OH~ and XOH, it may be seen from (5) that in the cases above considered } cannot exceed the value 10°10, and may thus be neglected in comparison with the other ions. It follows that the concentration of the remaining negative ion ¢ is equal to the sum of the concentrations of the positive ions, viz., 7+d. All calculations made from the conductivities of solutions of amphoteric electrolytes have hitherto proceeded on the assumption that the same method of treatment might be adopted as that applicable to simple electrolytes. This is, as we see, far from being the case, and we may now consider what manner of results amphoteric electrolytes with the above constants would yieid if their conductivities were treated in this erroneous fashion. Perhaps this is rendered most clearly apparent by calculating what values the Ostwald dissociation ‘constant’ would assume at different dilutions when deduced by the ordinary process from the conductivities. _ The molecular conductivity p... corresponding to Ht, XOH~ at infinite dilution may be taken ag from 350—370 at 25° when referred to reciprocal Siemens units. The molecular conductivity of HX*, XOH™ under the same conditions would be 60—70. From every concentration of HX*, then, we may obtain a concentration of H* having equal conducting power, by dividing d by a number varying from 5 to 6 according to the substance considered. If we add this quotient to the real value of « we obtain a false value «, which is assumed as the value of a in the simple calculation of the dilution constant. The subjoined table contains the values of the apparent Ostwald dilution constant k, for the amphoteric electrolytes considered above when calculated in the customary way from the values ¢ = a+d/5 and a =a+d/6. In each case the constant has been multiplied by 10°. For fy/K = 1 it will be observed that a fairly good constant *, may he got, the values for the greater dilutions approximating within the limits of experimental error to the true value ky. For y/K = 10 the values of /, are no longer even approximately constant, increasing rapidly with the dilution to attain a value at v = 1000 approaching 162 Prot. Walker: [ Feb. 3, k, x 10° calculated from «=a + d/5 and a=a + d/6. key|K = 0. Ae = T. Ienf[K =10. | dn/K = 100: |, b;/K=1000. v | Za | 3 : at+d/d.) a+d/6) a+d/5.|a+d/6. atd/5. at+d/6. at+d/5.| a+ d/6| a+d/5. a+ aj6. 10 1:0 | 1:0 | 0-944 | 0-938 | 0-718 | 0-679 | 0-780 | 0-622 | 3-30 | 2-62 100, 1:0 | 1:0 | 0-992 | 0-991 | 0-947 | 0-940 | 0-710 | 0-672 | 0-73 | 0-58 1000 1:0 | 1-0 | 0-993 | 0-993 | 0-991 | 0-988 0-936 | 0-930 | 0°69 | 0-65 the true value. For k,/K = 100 we have in the ordinary range of dilutions a fall to a minimum for /,, which is apparent when a = a+d/5 is used, and also occurs between v = 10 and v = 100 when a = a+d/6 is employed. In this case /, at v= 1000 is about 7 per cent. beneath the true value of kg. Finally, with /;/K = 1000 the value of k, at » = 10 greatly exceeds the true value, and falls very rapidly with the dilution to reach a minimum at high dilutions which will generally appear in the usual range investigated. Here /, at v = 1000 is 30—35 per cent. short of the true value. For other values of /, than that used in the above calculations the change in the value of /, with the dilution is similar, because, as may be deduced from the formule on pp. 157—158, the relative values of « for two electrolytes with the same /, do not vary greatly with the dilution, being at all dilutions approximately proportional to the square roots of the acid constants as long as w does not differ sensibly from 1/r. Turning now to the experimental data, we find that Ostwald obtained for ortho-amino-benzoic acid, 1: 2-NH.2C,;H,COOH, values of /, which steadily increased with increasing dilution,* confirmation of this result being subsequently furnished by Winkelblech.t ‘Their numbers are given in the following table :-— v. k (Ostwald). k (Winkeiblech). 64 0:66) as the value of /, at v = 1000, we obtain 1:02 x 107° as the approximate value of ky, since the table on p. 162 shows that i, exceeds /, by 6—7 per cent. With regard to the speeds of the ions we may adopt Winkelblech’s value pw = 357 as the sum of the velocities of hydrion and the anion. There is less certainty as to the sum of the velocities of kation and anion. The velocities attributed by Winkelblech to the kation in this and in similar instances are, in my opinion, considerably overestimated. For the amino-benzoic acids he does not give directly the experimental data from. which he estimated the velocity of the kations, but presumably the values were obtained by the same method as that which he adopted for other substances, viz., by measurement of the conductivity of the hydrochloride in presence of excess of base.* Owing to the very considerable: hydrolysis of such substances in aqueous solution, the results obtained for » at the experimental dilutions * Compare Bredig, ‘ Zeit. fiir physikal. Chem.,’ vol. 13, p. 214 (1894). 164 Prof. J. Walker. [Feb. 3, must be somewhat too high, and in consequence the speed of the kation is estimated at too high a figure. In order to proceed in a systematic way, I have added 6 in each case to the anion values found by Winkelblech for the amino-benzoic acids, and adopted the figures so obtained as the kation velocities in the succeeding calculations. The value of us thus estimated is about 7 per cent. below that obtained by adopting Winkelblech’s velocity for the kation. For the ortho-acid we have po = 32 + 38 = 70. Calculating with these constants, we arrive at the values given in the following table :— o-Amino-benzoie Acid, 1 : 2-NH».CsH,.COOH. kp f KE = 112, kg=-1:02'x 10%, po Ht, XOH- = 35%) sea 0. a. d. ew cale. | w expt. | & cale.| k,(O.).| &o(W.). | G4. 23 29K 1052 | 401 497 226 7°21 6°6 6-6 {$5646 128 | 20°3 17-0 10°8 10°8 ye | 7 Al ee | 256 | 16°3 6°7 16-1 16-2 8°3 8:4 | 8-4 512 |.12°3 2°5 23°38 23°6 8-9 9-1 92 1024 | 9-0 0-88 33°5 33-7 9°5 9-7. 1.956 } | The experimental values of the molecular conductivity mw are the means of the concordant series of Ostwald and Winkelblech. It will be seen that the agreement between these and the calculated values is very close. For comparison the “constants” k, x 10° derived from the calculated and the experimental values of ». have been added. The conductivities of solutions of para-amino-benzoic acid have been measured by the same observers, and Winkelblech determined the ratio k,/K, which he found by the catalytic method to be 210. With this constant the value of kg appears to be nearly 10 per cent. above the value of k, at v= 1000. Adopting for the latter the mean of the numbers found by Ostwald and Winkelblech, viz. 1:11 x 10-°, we obtain kg—1:21 x 10-5. w-Amino-benzoic Acid, 1: 4-NH».CgHy.COOH. ky/K = 210, kg =1-21x 10-5, pe» H*,XOH- = 356, po HX*+,XOH-=68 d # B | ko k ko - a | (eale.) | (O.) | (W.) (cale.)) (O.) | (W.) 32 | 22-3x10-* |138-7x10-*| 5-55 | —_ | 543) 7-7| — | 74 | 64] 20-9 64°4 755) 7°53) 749) 72) 72) 71 | 128 | 18-7 28°7 | 12-00) | 10-86, |)1112,| 97 |) 7 Bae | 256 | 15°8 12-0 16°47 | 16°34| 1684 88| 87/ 911 | 612 | 12°5 4°64 24°37 | 24°24 | 26°29 9°8 | 9°8 | 10°6 1024} 9-4 1-70 | 85-40 | 35-01 | 36°86 10°7 | 105 | 11-7 of fp aa atte dl Ny A 1904. | Theory of Anmphoterre Electrolytes. 165 The values of » calculated from the theory agree very well with those found by Ostwald, somewhat less well with those of Winkelblech. The most interesting point about the “constant” &, is that there is a minimum in both the calculated and experimental values at v = 64. The experimental data for meta-amino-benzoic acid are not so satisfactory as those for the isomeric acids just considered. Ostwald and Winkelblech found widely divergent values of the conductivity. Since in the following calculations use is made of Winkelblech’s constants, the comparison of conductivities can only be effected with his numbers. The value of k,/K found by the hydrolysis method is in round numbers 1100. This involves an addition of over 30 per cent. to k, at v = 1000 in order to arrive at an approximate value of k,, which in this case comes to be 1°4 x 107°. m-Amino-benzoic Acid, 1 : 3-NH».CsH..COOH. ky/K =1100, hg = 14x 107, pp. H+,XOH- = 355, po HX*,XOH- = 66. ij Chey cs a. d. fe CBC His ENNIS) ellie oo CAle. linen GNVED) ain) | 64 | 10°9x10-> | 163:0x10-5| 9°36 9 -36 11°‘1 11-2 128 106 79 2 test 1 72 6:5 |) 8 oem | 2956 | 9:99 | 36:2 15°20 | 16:04 75 8-4 | eto) 9-07 16 °5 22-06 | 23-04 8-0 9-1 toed 7-73 6-88 32°77 | 35-24 9-2 10°7 Here it will be noted that the agreement is by no means so close as in the previous instances. This is probably connected with the fact that the meta-acid rapidly oxidises in contact with platinum electrodes, and assumes a dark brown colour. The effect of this oxidation would be most apparent in the dilute solutions. Notwithstanding the want of exact accordance, the run of the constants is similar, both the values calculated from the theory and the values given by Winkelblech exhibiting a minimum at v = 256. These instances exhaust the data in Winkelblech’s paper for which an exact comparison of the theory with experiment is possible. They cover a range for i,/K of 110—1100, and in each case the peculiarities of the “constants” are faithfully reproduced by the theory. An experimental investigation of some other substances to which the theory is applicable is at present in progress, and I hope in a future paper to communicate the results obtained. | VOL. LXXIII. N 166 Dr. N. H. Aleock. Zhe Hlectromotive [| Decris: “The Electromotive Phenomena in Mammalian Non-medullated Nerve.’ By N. H. Atcock, M.D. Communicated by A. D. WALLER, M.D., F.R.S. Received December 15, 1903,—Read February 25, 1904. (From the Physiological Laboratory of the University of London.) Up to the present time there would appear to have been no published researches on isolated mammalian non-medullated nerve, and indeed, except for the paper of Brodie and Halliburton,* all our knowledge of these nerves has been either by inference from similar nerves in cold- blooded animals or derived incidentally from experiments undertaken for a different object. When it appeared, therefore, that the technique for mammalian medullated nerve served equally well for the non-medullated,7 it became a matter of interest to examine the phenomena displayed by the latter, and the very evident advantages of dealing with nerves of considerable size and comparative longevity, and the possession of a ready standard of comparison in the medullated nerves of the same animal, greatly assisted in obtaining an exact result. Methods. The splenic nerves were found to be very suitable objects for this purpose. They consist almost entirely of non-medullated fibres,t and in the horse, which is the animal which has been used for these experiments, the various bundles form a plexus around the splenic artery, which can be separated by careful dissection into its component parts, giving isolated pieces of nerve 1—1°5 mm. in diameter and from 6—8 cm. in length. These pieces will commonly retain their irritability for several hours if kept in 1:05 per cent. salt solution at 18° C., into which they are placed from a quarter to half an hour after the death of the animal. Waller’s§ galvanographic method was employed as well as the capillary electrometer.| In the experiments on the negative variation, the exciting current was derived from an accumulator of large capacity, in order to secure the greatest possible constancy, and the excitation was maximal, except in Experiment 716. The temperature at which the experiments were * Brodie and Halliburton, ‘Journ. of Physiol.,’ vol. 28, p. 181. + ‘Roy. Soc. Proc.,’ February, 1902, p. 264. t The proportion of non-medullated to medullated fibres varies in different animals; sections made from the nerves actually used showed that medullated fibres formed less than 0°5 per cent. of the total number. § Waller, “Signs of Life,” 1903. || The analysis of the electrometer records will be considered at a future time- 1903.] Phenomena in Mammalian Non-medullated Nerve. 167 conducted was 17—19° C., except where otherwise stated. The current of injury was balanced against an equal fraction of a volt, and the value read off. This compensation was maintained throughout the experiment. No current, therefore, flowed through the nerve when at. rest. The experiments were ordinarily carried out in duplicate, with both medullated and non-medullated nerves from the same animal. The electromotive phenomena are considered in the present paper under two heads :— 1. Negative variation. 2. Hlectrotonic currents. 1. NEGATIVE VARIATION. (Eaperiments.) The electromotive phenomena in the nerves of the horse resemble those in the nerves of other mammalia in kind, but differ in degree, being of considerably less magnitude.* This is due, in part, to the large amount of connective tissue surrounding the nerves, and forming a derivation circuit. When this connective tissue is dissected off, the effect on the galvanometer is increased. Three experiments may be quoted on this point :— Table I. Uap 2; . Experiment. Nerve. Serene OF Dieeatiye Notes. injury. variation. millivolts. millivolts. a...| Median 2°7 0-172 720 eye i 41 0-309 pie. « as 8-2 0-242 og ‘ 8-5 0-359 749 13 o6obu So Splenic 9°7 0 °504 Whole nerve. | Beles er sisets 5, 11°3 1:910 Single nerve-bundle. The nerve was isolated in the usual way and the measurements made that are marked a, the connective tissue was then dissected off as far as possible, and the measurements 0) taken. Both injury current and negative variation are increased, the latter more than the former. But it will be noticed that even in the most favourable case the voltage of the negative variation is less in the medullated nerves of the horse * This has already been observed in the case of the current of injury, see Gotch, *Schifer’s Text-book,’ vol. 2, p. 520, and Biedermann, ‘ Elektrophysiologie,’ 1895, p. 638. N 2 168 Dr. N. H. Alcock. The Electromotive [Dec. 15, than is usual in the mammalia. It is not certain how far this is due to the greater amount of connective tissue present between the individual nerve-fibres, or to some other cause. Both current of injury and negative variation are considerably greater in the non-medullated nerves of the horse than in the medullated. The following table gives the result of several observations :— Table IT. Carrent of injury.* No. of | Negative variation.* No. | Nerve. la ae ae Be ent | of experi- Mean. Max. aneHtS. | Mean. Max. | seule ae | aN es eer OI seer 2 Sa newibooas Median....| 3°86 5°3 5 | 0°380 0°83 | “ Splenic ....| 5°81 15 °5 13 - | 0°860 2°3 | 11 ‘These are the values obtained with the galvanometer. They are probably too low, and for several reasons it is likely that the maximum values are more nearly correct than the mean. ‘They are, however, strictly comparative, and it is seen the voltages in the splenic nerves are approximately three times the median, a result agreeing with the observations on cold-blooded animals.t The non-medullated nerves of Ap | a TNA. | ea || f Fie. 1 (Exp. 714, V.).—Splenic of Horse. The vertical lines are the successive negative variations produced by tetanising currents (5000 Berne units, 2 volts in primary circuit) for 13 seconds, repeated once a minute. * Here and elsewhere in millivolts. -+ Kiihne u. Steiner, ‘ Unters. d. Physiol. Inst. d. Univ. Heidelberg,’ vol. 3, p. 149; Sowton, ‘Roy. Soc. Proc.,’ vol. 66, p. 379. rma 1903.) Phenomena in Mammalian Non-medullated Nerve. 169 Rest -0005 volt. 717. Vv. -0005 volt. Fic. 2 (Exp. 717, V.).—Median of same animal as Fig. 1. All details as before. the horse commonly retain their excitability for 8—10 hours post- mortem, the medullated for about half this time. I have been unable so far to trace any quantitative relation between the current of injury and the negative variation as determined above. Obtaining a non-medullated nerve, subjecting it to a series of repeated tetanisations, and recording the successive negative variations by the galvanographic method, it appears at once that the results are very different from the corresponding phenomena in medullated nerve. Figs. 1 and 2 are records from two nerves, the one splenic and the other median, from the same animal under identical conditions. The successive negative variations from the latter are approximately equal, those from the former fall off very rapidly, and this rapid decrease in the negative variation is characteristic of mammalian non-medullated nerve. The following experiments were made :— 170 Experiment. 702. Horse I MOA. es I 705.9) 55 EE (OTE POLL 1098 esp ALE C12 bas PIN. TiS a tet ALN, (Ways thee ee rela A Mean of eight |0°617 experiments Dr. N. H. Alcock. The Hlectromotive | Deerto; Table III.—Splenic Nerves. N og 78. Neg. var. see Diminu- =e seen tion Notes. Initial. Final. E b/a. ment | millivolts. | millivolts. | mins. | | 0°95 0°33 fl 0°35 | Exc. lasting 10 seconds, once a minute. 0 376 (1st) } 0:219 25 0°582 |Same nerve as 0 °397 (2nd) E xp. 702. Dee (a8) it 0-109 22 | 0-222 0-576 a 0°317 0°131 26 | 0-413 0-560 0 °396 30 0°710 0-788 (1st) | Doe (2nd) 0 672 23 0°853 | Same nerve as in fH GULy J (23rd) Exp. 710 below 0-792 (4th) ee al | staircase. 0 °859 0 °392 29 | 0°456 |Same nerve as Exp. 710. 0 593 (Ist) | rae ay 0-215 50 | 0-362 Submaximal exci- 0-607 (4th) . tation, 500 unis, staircase. 0 °307 Zi 0 498 The following experiments were made under identical conditions with medullated and non-medullated nerves from the same animal :— 171 CrVe. — / Mammalian Non-medullated N LENOMENA VN PI Lar se) Ss (oP) rl "A 9B, 9a8 syuouL -wedxe oy} Jo uorjszod 194j,VT OY JO sis fTVUuR oy] LOT "(% puv T ‘ssy) quiod uostredutoo oy puokeq yuouttodxe oy jo moenuryu0d oY} OF LOJA SANG pojoyovlq ot J, 66 , OSBITTBIG ,, 166: 0 Z88- O (82: 0) €1é- O 10. T “pq 66 66 6g 66 6G £E 6V €& 1g 1é ‘yuout1ed x9 jo uolRin¢g 066: O 649-0 (pugg) (228-0) (puzz) 00¢- 0 b6S- O GZI- 0 (pugpr) 86S. 0 (prge) 989 0 [2-0 { 009. O ee WLR "UBA "99 NT “Al IQ], 666: O 109-1 09-1 666: 0 VSL-O EG (puz) $82.0 (48ST) 612-0 (pig) 196-0 (puz) 020- T (48T) P88: 0 NE OAL “IVA “SON UIP TA, oruopdg otttoptdg ONIN INL UBIPOTAL orueldg URIPIT| oruaytdg "a ‘06L WL ‘TL sydxqy yo uve ¢ ¢ . ; Ted PIL ‘OLZ ‘3dxq jo uv; a Ce 66 “tay "tts TTA estoH “El “OBL eos ve oe oe «6 uy [Brtettetecrel ame reomnts LA ®StOFL “PTZ eee 60 © 8 oo ¢ ae Poiet eer ST osmom abl ‘quoturTodxy 172 Dr. N. H. Aleock. The Hlectromotive [ Dec. 15, These tables show very clearly the alteration in the negative varia- tion by successive excitations. The decrease is most rapid at first, and after a little proceeds quite slowly, especially if the excitation is sub-maximal, so that if the result of the first few minutes is dis- carded, there remains a considerable period during which the diminu- tion is small, and which can be used to test any desired procedure (vide figs. land 6). The effect of rest is that the subsequent responses are greater for a time, but soon fall off (fig. 1), and the difference in this respect from the medullated nerves is very marked (fig. 2); in the latter the subsequent negative variations are almost invariably less after a pause. This difference is very clearly seen in the electrometer photographs (Experiment 750, B, fig. 3, splenic of horse, and Experiment 751, fig. 4, + -O01 volt. Fia. 3 (Exp. 750, B,).— Non-medullated Nerve, Splenic of Horse. Excitation for 10 seconds (Tet.), interval:20 seconds, as shown by the lower interrupted line. Time—1 mm. = 2 seconds. “O01 Fie, 4 (Exp. 751).—Medullated Nerve, Ulnar of Cat. Details as in Fig. 3. 1903.] Phenomena in Mammalian Non-nedullated Nerve. 173: oS aD os een Fie. 5 (Exps. 750 B and 751).—Response of Non-medullated Nerve (upper) and Medullated (lower) to Single Shocks. Same nerves as figs. 3 and 4. ulnar of cat). The full analysis of these will be considered in a subse- quent paper, but these figures show (énter alia) that the electrical resistance plays an unimportant part in the production of the phenomena observed, and that the rate of transmission of the electrical effect is very much less in the non-medullated nerves. As, therefore, the progressive diminution in the negative variation is. characteristic of non-medullated nerve, both here and in cold-blooded animals,* further experiments were undertaken to ascertain whether this was due to events occurring along the whole length of the nerve,. or to changes localised at the place of excitation. Two pairs of platinum wire electrodes were used for stimulation, placed respectively further and nearer the leading-off electrodes; the nerve was excited first at the “far” pair and then at the “ near.” In Experiment 718 (fig. 6) the result of exciting through the near pair of electrodes was to increase the negative variation to very nearly the original amount (a). After 12 minutes of excitation at this point the current was again sent through the “far” electrodes, and the resultive negative variation (c) was much less than the original value. If the nerve had been simply resting, instead of being excited at a proximal place, the negative variation would have been increased (see fig. 1); the effect, therefore, of passing a recently excited spot is diminution. In Experiment 719 the excitation was through three pairs of * Sowton, loc. cit. 174 Dr. N. H. Alcock. he Electromotive — [Dee. 15, Table V. Far. | Near. , | Experiment. | Negative variation. | | Ne nuive var Notes tt 8 ’ ae ; (see below). Time. |= ees Initial. Final. Initial. | Final. WAG. Horse VI |a0°604(Ist)|1 > 419 | os | po -agn | @cahn| ee es GE bo 419 29° 50-800 0-872) 12 (| Fig. 6 c 0166 0-20 6. oF ee | 6 = near. d c = far again 719» VI /@0 T11(ist) | 1 9 7709] 34 | 60-400 0-138) aeip| mane 0-144. (2nd) pease : : e 0°322 /0°092| 144 | 6 = 2nd pair c*0°092 |0°115; 3 {| ec = 8rd pair | ae j c* = 3rd pair re- | 7210" sf Val. | 60 0°372) 52. | 61226 10 7e4) versed. | (0 -496)) (17). | (| Control with me- | | | || dullated nerve. 720,B.,, VII | @0°292 0 -294, 22 b 0°355 |0°378| 204 | a = far. (6 0-253 0-281 | 5 | | | d = near. | | | | \| ¢ = far again. it Pair. 2nd Pair. |15¢ Pair _ Far. “Near” ; Far’ -OO1 718. volt. VI. ie ’ , -OO1 olt. Fie. 6.—Splenic of Horse. Negative Variations. Result of exciting at (1) far, (2) near, and (3) far electrodes. electrodes, and then the excitation was reversed; the alteration produced by reversal was very small. 1903.] Phenomena in Mammalian Non-medullated Nerve. 175 Experiments 721 and 720 were with non-medullated and medullated nerves under the same conditions. _ Considering the result of all these experiments, it is clear that the diminution of the negative variation in non-medullated nerves is due to changes occurring at the point where the nerve is excited. Using a constant stimulus this spot becomes less and less excitable, in the sense that the response becomes progressively smaller. The control experiments on medullated nerve show but slight traces ‘of this effect,* either the medullary sheath prevents this loss of excita- bility in some way, or the two classes of nerves differ very widely in their reaction to stimuli. Speaking broadly, the evidence is in favour of the former hypothesis. While the work of Sowtont and Garten{ on cold-blooded nerves is In accord with my results on mammalian nerve, the paper of Brodie and Halliburton at first sight offers a contradiction. These experi- menters excited the splenic nerves in the dog for many hours, and blocking the impulse by cold, observed that when the block was removed the splenic contractions followed as at first, apparently un- altered in amount. I have no doubt as to the correctness of their observations, and the results they obtained, differing from those | | t,0 j “O01 -OO1 K volt. volt. ! 223. Fie. 7.—Splenic of Horse. Effect of Ether Vapour. * Dr. Waller has very kindly permitted me to measure a considerable number of his photographic records of frog’s nerve, of 1896—1897; in all a small regular diminution is detectable, amounting to from 2:0 to 5°6 per cent., in experiments lasting 40 minutes. + Sowton, loc. cit. t Garten. 176 - Dr. N. H. Alcock. The Hlectromotive [| Dec. 15, recorded above, may possibly be explained by supposing that the splenic contractions are not as delicate an index of the condition of the splenic nerves as the galvanometric response. Waller* has shown that with a gradually increasing stimulus the voltage of the negative variation reaches its maximum much later than the contraction of the attached muscle, and as in all the experiments here recorded a certain amount of the negative variation still persisted, it is possible that even this fraction—perhaps one-third of the initial value—indicated a sufficient intensity of nerve impulse to give a maximal splenic contraction. If the medullary sheath has any such action as the hypothesis suggested above necessitates, the inquiry may be extended to see if any light can be thrown on the manner in which this sheath acts. The following experiments on electrotonic currents were, therefore, undertaken. 2. ELECTROTONIC CURRENTS. Methods. The nerve rested upon two pairs of non-polarisable electrodes. The distal pair led off to the galvanometer ; the proximal were con- nected with the automatic reverser, used by Dr. Waller in 1897, which delivered in order— (1) A current in the proximal direction (giving anelectrotonic currents in the nerve). (2) Excitation by means of an induction coil. (3) A current in the distal direction (katelectrotonic currents). (4) Excitation as No. 2 The result is seen in the photographic plate as a cycle consisting of one upward mark, anelectrotonus, and three downward, the middle one katelectrotonus, the two side negative variations. Experiments. The following pair of experiments was made under identical conditions :— Table VI. | | | | Anelectro- | Experiment. Nerve. fous | hea gn BED. Inj. cur.| Notes. mean. mean. : mean. | 728 VIII..| Median 0-702 0 °356 0 595 0:20 | Fig. 8 (branch of) 729A VIII Splenic 0 | 0-540 0 0-80 | Fig. 9 * Waller, ‘ Brain,’ vol. 18, 1895, p. 210. 1903.] Phenomena in Mammalian Non-medullated Nerve. 177 A A=0 a | | | NV. | le -001 001 { volt. volt. be K=0 Median. Splenic. | Fie. 8.—Exp. 728, VIII. Anelectrotonic | Fre. 9.—Exp. 729, A, VIII. The same and Katelectrotonic Currents, and Nega- on Non-medullated Nerve. An. and tive Variation on Medullated Nerve. kat. are too small to be legible, and the negative variation only is seen. The voltage of the polarising current was 3 volt, the excitation 3000 units of the Berne coil, the distance between the centre of each electrode and the next was 11 mm. In the medullated nerve the electrotonic currents exceeded the negative variation ; in the non-medullated they were so small as to be imperceptible on the photographic plate. In order to see whether the electrotonic currents were completely absent, or merely very small, further experiments were then Ae using a higher voltage in the polarising circuit. These measurements are uncertain, both from the difficulty of measuring such very small currents, and from the fact that the cessation of anelectrotonus and the commencement of katelectrotonus excite the nerve, and the resulting negative variation in the latter case is added to the katelectrotonic current, so that except in special cases (as in Experiment 731) the readings of the latter are too high. It is also not absolutely certain that current escape has not some share in the result, though as all these effects are abolished by crushing the nerve, and as the an- and katelectrotonic currents are not equal in magnitude, and not exactly proportional to the polarising current, it is probable that this error, if it exists, is a small one. Bearing these reservations in mind, certain conclusions may be drawn from these experiments. The electrotonic currents in the non-medullated nerves are evidently very small, about one-fortieth of the same currents in medullated nerves. Further, while the an- and katelectrotonic currents in the latter are nearly equal (0°702 millivolt [ Dee: 1b; The EHlectromotive Dr. N. H. Alcock. I~ eet { GZ0-0 | $20-0 | 70-0 | 10-0 = H | (s#[o4tp[tuz) squoa G80: 0 $90. 0 £0: Orr GOEO Wo shoe OO FOR OCT degen ese. Soe “Cee “IVA ‘SON "8 ‘9 ‘p 74 nese ++ (gatoa) qUOLIND SUISILEOT (avout) | { $20. 0 ‘LOI|RITOXO WO WOTWwIAeA 860.0 = 3 730: O= ¥e0. 0 D2 oaTyesou OU SuTAIS MoU ‘Og/ BV Ve0- 0 ee SAMO OUIUS TIM JuOWTLOdxe LOzery (uvour) : TZO. 0 6 Vn ee OO OSI ie TA ee ‘9 'p "2 ooo oe + * (sqToA) qUOLING SUISIAGOT 909. 0 { d TFO-0 | $F0-0 | ThO-0 | ¢ | M (soatytrer) squoa g 9T- O GI: O 060: 0 ST0-0 d Fee ee ee elie id (GE tes ero TET AN SUIS OO == | ee IRA ‘SON 9 2. f “p "eg 74 versie ¢ (S914) qUoLINd SUISTIe[Og 1¢9.0 { 61-0 880- 0 080: 0 O80. O TT-0 TT- 0 880- 0 6 M (S}[OATTTFUL) suo LT. 0 IT-0 880: 0 990- 0 £90. O PPO: O cg0- 0 é Vig ROR OTC 701} 09M I Aerelin Gaz, “TARA a ‘ZS LG S ‘9-1 come T "99: 0 oe ey aes) ‘quoted xq BON | qUITAND SUISALTOT ; 2 JULaN Ed 1903.] Phenomena in Manmalian Non-medullated Nerve. 179 Polarising current. | 2, 3 4 | 48 6 volts. =. -OOLV. | . | | NV. A : A. -OO1 730. | volt. VII. K Fie. 10.—Exp. 730, VIII. A.and K. Currents only, with Increasing Strengths of Polarising Currents. in A and 0°595 millivolt in K in Experiment 728), in the splenic nerves the anelectrotonic currents are perhaps four times greater than the katelectrotonic. This favours the view that these currents are not due to the presence of a small percentage of medullated fibres in these nerves, but are the actual expression of the fine sheaths present around the axis-cylinder.* The result, therefore, of this series of experiments affords a very probable explanation of the diminution of the negative variation with successive excitations in the non-medullated nerve, and the relative absence of this diminution in the medullated. For in the latter the exciting current can diffuse up and down the nerve, in the former it is strictly limited to the spot where it is applied, and the current density at the excitable axis-cylinder must be many times greater in the non- medullated nerve. It is not surprising, therefore, that there should be a marked local effect, and whether this is termed “injury” or ‘“‘fatigue ” is rather a question of terminology than of fact. * Dr. W. M. Fletcher has recently examined the sheaths of medullated and non-medullated fibres between crossed nicols. He finds that in the former there is characteristically present an anisotropic cholesterin deposit ; in the latter this is absent. The sheath round the non-medullated axis-cylinder is therefore of a totally different character to that in the medullated fibre. (Note communicated to the writer.) See also authors quoted by Brodie and Halliburton, lec. cit. 180 Electromotive Phenomena in Non-medullated Nerve. As the electrotonic currents are so small, it was not to be expected that there should be any marked alteration of excitability im the neighbourhood of the kathode or anode of a constant current. Three experiments were, however, made to serve as a check on the preceding :— Table VIII. | Neg. var. | ca | Sa Neg. var. | Experiment. (mean). | Gurine |. duri (mean). Notes. sa uring = during : Initial. | Final. : anelec. | katelec. | | 743, X. Polarising cur-| 0-488 | 0°468 | 0-452 | 0-460 | rent = 0°5 volt. | | Excitation = 1000. | | | | 744, X. Polarising cur-| 0°492 | 0-453 0 444, 0°434 Same nerve rent = 2 volts. | as 743. 745,X. Polarisingcur-| 0-971 | 0°931 | 0°928 | 0-884 rent = 2 volts. | | | The exciting electrodes were at a distance of 3 mm. from the polarising, and on their distal side. From an inspection of the figures it is clear that even taking the most favourable case of the last experiment the effect is a minimal one. I have much pleasure in acknowledging the kind assistance and advice I have received from Dr. Waller in the prosecution of this research, and also in expressing my indebtedness to the Council of the Zoological Society and Dr. Chalmers Mitchell for permission to use the nerves of the horses which had been slaughtered for the carnivora. Conclusions. 1. Non-medullated nerves exhibit a negative variation and current of injury of about three times the magnitude of the similar phenomena in the medullated nerves of the same animal. 2. The negative variation of non-medullated nerves undergoes a progressive diminution with repeated stimuli. 3. The immediate cause of this diminution is a localised change at the place of excitation. 4. The electrotonic currents of non-medullated nerves are very small, about one-fortieth of those in medullated nerves. aval 5. This latter fact affords an explanation of 2 and 3, as the exciting current, being confined to the place of application, has a greater current density and therefore a greater local effect. On the Formation of Solids at Low Temperatures. 181 “Note on the Formation of Solids at Low Temperatures, particu- larly with Regard to Solid Hydrogen.” By Morris W. TRAVERS, D.Sc., Professor of Chemistry at University College, Bristol. Communicated by Sir W. Ramsay, K.C.B., F.RS. Received February 4,—Read February 18th, 1904. In the year 1902 Dr. Jaquerod and:[ carried out some experiments on liquid and solid hydrogen with a view to determining its vapour pressure on the scales of the constant-volume helium and hydrogen thermometers. We found that hydrogen remained liquid down to 14°-2 (He scale), the lowest temperature to which we could reduce a jarge mass of the liquid by means of the pump at our disposal. When, however, a small quantity of liquid hydrogen, cooled to 14°-2 in a glass tube immersed in the liquid contained in the large vacuum vessel, was allowed to evaporate under reduced pressure, it solidified when the pressure fell to 49 or 50 mm. of mercury. This pressure corresponds to a temperature of 14°:1 on the helium scale. The presence of the solid was determined by mechanical means, and it was not possible to observe its appearance.* Dewar gives the melting point of hydrogen at about 15° absolute, and the melting pressure at 55 mm. of mercury. He describes its appearance as that of ‘frozen foam,” or as ‘‘ clear transparent ice.” t It appeared to me worth while to carry out a few experiments to try to determine whether solid hydrogen formed definite crystal, or indeed whether the glassy substance was a true solid or merely a highly viscous fluid. My meaning will become clearer if I give an instance in which both such changes occur. If an organic liquid, such as ethyl aceto-acetate, is cooled slowly to the temperature of liquid air, it is converted into crystalline solid, the formation of the crystals commencing when the liquid is cooled to about — 150° C., usually at several points on the side of the vessel, and spreading rapidly throughout the mass. If, on the other hand, the liquid is cooled very rapidly, a hard glassy substance is formed, and though crystals may begin to appear, they will only do so locally, as the velocity of crystallisation decreases rapidly as the viscosity of the liquid increases. The glassy substance is really a liquid of high viscosity ; it is formed with perfect continuity from the normal liquid state, and should differ from the solid (crystalline) form in its physical properties. Such a substance might, for convenience, be called a pseudo-solid. * © Phil. Trans.,’ A, vol. 200, p. 170. + British Association, Presidential Address, 1902. See also paper on “Solid Hydrogen,” ‘ Brit. Assoc. Report,’ 1899, reprinted in ‘Nature’; also ‘Roy. Inst. Proc.,’ 1900. ) VOL. LXXIII, 0 182 On the Formation of Solids at Low Temperatures: [Feb. 4, In investigating solid hydrogen the apparatus shown in the accom- panying figure was employed. The liquid hydrogen was introduced into a small clear-glass vacuum-vessel 15 cm. long and 4 em. in internal diameter. ‘This vessel was placed inside a glass tube BB, which communicated with an exhaust pump through a tube DD sealed to it, and was closed by a rubber stopper C. A short glass tube E, 6 mm. in diameter, passed through the stopper, and through it passed the stirring rod FF. To allow of free rotating motion to the stirrer, and to make the apparatus gas-tight, a short piece of rubber tube G, was passed over the end of the tube E and was wired to F. The lower part of the apparatus was contained within the vacuum vessel H, which contained a small quantity of liquid air. When the liquid hydrogen was made to boil zn vacuo, its temperature fell, but the liquid did not appear to become more viscous. At length films of a colourless glassy substance formed at the surface, and broke away as the bubbles rose. After a short time the vessel became filled with these flakes, and while in this condition stirring, by. giving the top of the rod F a rotatory motion, did not appear to indicate that the portion which remained liquid had undergone any considerable increase in viscosity. After a time the mass con- tained so much solid that it became pasty, and finally the whole of it appeared fairly homo- geneous. The solid evaporated fairly rapidly, so that after about 10 minutes only a hollow cylinder of it, about 3 cm. long and 2:5 cm. in diameter, re- mained. This had the appearance of a film of ice which had partly thawed, consisting of clear granules connected by thinner and less transparent portions of solid. No crystals were. observed on either of the three occasions on which the experiments was carried out. An attempt was made to examine the solid in the field of a polariscope, but it was unsuccessful. Though there is no direct evidence of the formation of crystalline hydrogen, my experiments lead me to the belief that solid hydrogen is a crystalline substance and not a pseudo- solid. The sharpness with which the solid hydrogen is formed, and the constancy of the apparent melting pressure, are distinct evidence in favour of this conclusion, though it must be allowed that the rate of change in viscosity, when the temperatures are measured on the Centigrade scale, will probably. 1904. | Pharmacology of Indian Cobra Venom. 183 appear to be more rapid at low temperatures than at high tempera- tures. The whole question of the formation of solids at very low tempera- tures is of great interest both from a physical and from a biological standpoint. It is quite possible that if living organisms were cooled. only to temperatures at which physical changes such as crystallisation take place with measurable velocity, the process would be fatal, whereas if they once were cooled to the temperature of liquid air, no such change could take place within finite time, and the organism would survive.* These experiments were made in connection with some investigations which were being carried out at University College, London, with the assistance of a grant from the Royal Society. As Iam at the moment unable to continue the work, I have decided to publish this note. “ A Contribution to the Study of the Action of Indian Cobra Poison.” By Captain R. H. Evuiot, M.B., B.S. Lond., F.R.C.S. Eng,, D.P.H. Camb., etc., of the Indian Medical Service (Madras). On special duty for Snake Venom Research under the orders of the Secretary of State for India. Communicated by Professor Sir THomas R. Fraser, F.R.S. Received January 18, — Read February 25, 1904. (Abstract) Previous Work on the Subject. Brunton and Fayrert discussed the pharmacology of Cobra venom at some length ; they attributed the effects of the poison to its action on the cerebro-spinal nerve-centres, especially on the respiratory centre. They observed that Cobra venom had a direct action on cardiac muscle, and that it also affected the heart through the vagal system, but they did not lay much stress on circulatory failure. They surmised that the high and maintained blood pressure of a cobraised animal was due to arteriolar constriction, but did not attempt to explain how this was brought about. Amongst the many other points of interest they took up, was the influence of artificial respiration in cobraism. * Experimental results are given by Macfadyen, ‘Roy. Soc. Proc.,’ vol. 66, 1900, pp. 180, 339, 488; Swithinbank, ‘ Roy. Soc. Proc.,’ vol. 68, 1901, p: 502. t+ Owing to the kindness of Professor Sir Thomas Fraser and of Professor H. A. Schifer, the writer was enabled to carry out this research in the Pharmacological and Physiological Laboratories of the University of Edinburgh. Towards. the expense of this research, grants were received from the British Medical Association and from the Moray Fund for the Endowment of Research (Edinburgh). t ‘Roy. Soc. Proc.,’ vols. 21, 22, and 28. Oo 2 184 Capt. R. H. Elliot. 4 Contribution to the [Jan. 18, Cunningham in the ‘Scientific Memoirs by Medical Officers of the Army in India’* urged the opposing theory that Cobra venom acted on respiration, through the blood and not through the nervous system. Weir Mitchell, and Reichert? carried on Brunton and Fayrer’s views. Their paper was mainly concerned with the venoms of other snakes than the Cobra. They thought two factors were at work on the rate of the heart, viz., an increased activity of the accelerator centres, quickening the beat, and a direct action on the heart slowing it. They attributed the primary fall in blood pressure to depression of the vaso-motor centres, but thought it might be partly cardiac. The rise they considered “capillary ” and the final fall cardiac. Bagotzi{ laid great stress on the réle played by nerve-end paralyses (especially phrenic), and disputed Brunton’s views that respiration was attacked through the medullary centre. He did not find any action of the venom on the vagal mechanism. He surmised that death with a tightly contracted heart, the result of very large doses of venom, was due to a cardiac action. C. J. Martin in the article on snake venom in ‘ Allbutt’s System of Medicine,’ considers that, in Cobra poisoning, the circulatory mechanism is not easily affected, and contrasts this with the state of affairs in viperine poisoning. He found that vagal stimulations stopped the heart up to near the end of life in Cobra poisoning. Object of this Research. This was to accurately ascertain the precise part played by the various important centres, nerves and organs in the production of death from cobraism. Methods employed in the Research. 1. Perfusion of the frog vessels was carried out with solutions of Cobra venom of various strengths.—The central nervous system had been destroyed first in each case. The strength-limitation of the action of the venom on the arterioles was carefully studied. | 2. Perfusion of frog hearts was carried out with solutions of Cobra venom of various strengths.—The isolated hearts were perfused in Schiifer’s plethysmograph, and blood mixture was employed as the vehicle for the poison. The strength-limitation of the action of Cobra venom was again determined here. Certain drugs which resemble this poison in their action on heart muscle, were also experimented with, ¢.9., strophanthin and the sulphate of atropia. The risks apparently * 1895, Part IX, and 1898, Part XI. + ‘Smithsonian Contributions to Knowledge,’ 1890. ~ Virchow’s ‘ Archiv fiir Path.,’ vol. 122, p. 201. 1904. | Pharmacology of Indian Cobra Venom. 185 attendant on the use of the latter drug in Cobra poisoning are pointed out. 3. The study of the action of Cobra venom on the frog heart im situ was next taken up, by means of a number of devices, which included the direct application of the poison to the medulla oblongata, which was exposed for the purpose. 4, Perfusion of the mammalian heart was carried out with solutions of Cobra venom of various strengths.—The isolated heart was perfused through its coronary vessels with a nutrient fluid, in which the venom was dissolved. Cats’ and rabbits’ hearts were used. 5. By means of kymographic tracings, the blood pressure, respiratory movement, etc., of cobraised rabbits were recorded and studied.—The activity of the vaso-motor mechanism was studied, in various stages of cobraism, by stimulations of the depressor and sciatic nerves, the vagi were cut, likewise at various stages, and their ends were also stimulated, in order to ascertain the part played in cobraism by the vagal inhibitory mechanism ; injections of a solution of sulphate of atropine were also. made, and the effects were observed. The author received much help in this section from Drs. Sillar and Prentice. 6. A similar set of experiments to the last was carried out on dogs and cats, plethysmograpluc tracings of intestinal volume were also included here, im order to study the changes, if any, going on in the splanchnic area curculation. | 7. The movements of the auricle and ventricle were studied in cobraised cats and dogs by removing the front of the chest parietes, and attaching the auricular and ventricular walls (by means of hooks and silk threads) to levers recording on a kymographic apparatus.—The blood pressure in a large artery was recorded at the same time, and intestinal volume was also frequently taken by means of a plethysmograph. At various stages the vagi were divided or stimulated, and the results observed. The effect of giving further doses of Cobra venom with the vagi, intact or divided, was also studied. The condition of the vagal nerve-ends received close attention. 8. By kymographic experiments the influence of artificial respiration on the centres, nerve-ends, etc., of cobraised animals was carefully studied.—The experiments were " d (8RHwy GPL A ‘rad. Sol = iV 9-892 4 eG | | | aA | “TOF UIra ON at] | ; AIOA B PUB spnojo oY} © ‘purM | O'M "pur O° UL Wes OG PTNOD syvoaq ‘A'S TM cs ‘GS UMIRA @ qT “QsvotaAG Oe ‘G WIV B TIM “48vo.L0AG 0..6- WT solIy -4GBes«dNG: “GsuatOAG | 8..0+ T, + ‘68 ‘d “861 = V G-SSL a4 Z48'd *O6=-V 0-624 “A $6 d ‘88l=V 0-194 Oye = ‘OS POUIvUL “21 PUB 4STIIOA0 oUvIeEq O'M "MOUS 97491] PST GAN TOOS 41 4nq ‘spnojo yuoared O'M “MOUS OTFHIT BITIM YstotaaQ | 0..01— if | B YIM Qsvodoao Apravoyy | g.4T— ‘yg, | -sueay ySiq ATO qi IV | 8..88—- vod Pal — V. 0-692 ECO dl VALVE NG 8-952 a RG Lcdis = V.3) sav, O-FoL Ff 6 20d ‘Wd TT puv 6 woomyoq "Wd Og’) ‘Wd 9 put g W9d0M70q ‘Nd OST RV ZT pus 6 woe 4oq “HV 0&4 "C06 ‘(q) WUoIpead jeryueqgod pus 4B OP Vu ‘(q) JUoIptas jeyuogod pus 4@ opel ‘(q) qUoIpras jeryu0jod pus 4B apvul -oneq TOYyywom ‘(y) APAtpoV-orpBy | sMOTZVALoSsqG | oyyrom ‘(y) L7rATQ0R-oIpeyy SUOTFBATOEGG | LoyyVom ‘(VW ) AZTATJOV-OIpRyYy | STOTYALOSqCO | 1 rd eal ‘panunguor-—'T 91q¥T, 13 Atmospherieal Radio-activity mm High Latitudes. 1904.] ‘OILM UO 4SOAZ JO 4e0o ULYT, “puda Sparvsmog "Nog ur porvodde spnojo MOF V YN ‘sso~pnoro ATUTe AT ELOY. xd = 1G — avi ‘ayqissod uit 919M SUOTyvALosqo ‘oles G OF WOSIA SUIART, PULA OTL], ‘O[QISIA BLOANR ON ‘SUIMOAD §«ssoTpnopD ‘avaT—) ‘90T ‘d “OSB = V ‘O[ISTA OlBIAG qYSIT VIOINV 91971, B puv peivojo Ays pus sparen “07 4nq ‘4sBVo10AO 4SAIE FV ‘SOT “d “9tIT = V "OTQISTA OLOM SLVIS FIST SOUTy 4B 4eyy Os ‘spnoTo uIYyy AtoA WOAT SUTLMOTg ‘G96 ‘dq “GOT = VW O° M 6-08 — ib "eIOINB OU ‘SSOTPNLOTD ‘768 ‘d ‘sp=V "U998 aq UV s1ByIs OTM Ysnoayy ‘gpnoyo UIYY YALA “ysvo10AG "G86 ‘d ‘99 = VW *puo spaemoz uoty pure mou AXOAO OTQISTA SALAS FILSTTG ‘spnoypo UITG YILM “4svo.t9AGC ‘89d ‘V8s—- Vv "pues sparvmoy MOUS OTFFIT B UGLIM “GSBIIOAO ‘$8 '‘d ‘SST = VW “§SBI.LOAO ‘S&L d “P8= VW "O1IM WO 4SOAF FO 4ROO uly, “wtomme Jo udrs ON ‘poavoddesip woos Sphoypo yng “4sVo19AO YsAE FV ‘OSE “d “G4 = V ‘STM +s oy Ol Glee 6-874 4 'S TM. 0-01 — nh GLVL A OM 0-016 — WM 6-692 A O°M. 08 — iL LVVL A 0°M. GollL— WL 0-602 4 0 °M 0..6 — WL 0-692 a ‘puo ye MoUs OFN9ITT °9 pur e@ weemgoq AIvA Spno[O Sida OSGi We ‘pus pur suruuisoq qe spnop Ys uryy (Z—T) oz v ysnl ‘our Fo gavd 1oxvors ot} SsoTpPHoTAO 9IT'd “POC= VW "OTIM TO JSOTF JO 4ROd ull, ‘“mMoy pug ur somo uly} ‘ANoy JST UL Spnoyo ON ‘SUTUTOUT pTOO aRoTO ‘OI d ‘99= VW “SPHOTO UI} TTA “GSRdLZAO SLE dy Vel = WW. ‘gpnoyzo [SIT UY YIM surucoUL FLSA ‘OST ‘d ‘84=V "MOAMYOq WOES oq URO ONT oTAFIT B YL os ‘snonu1ya0oo os oytnb qOU ote spnozo ynq ‘48vI.10AG ‘ehS ‘'d *sOlL = W “ACN 6 OM $-6- WL Lv 8I PI 66 a3 ia (75 ia VOL. LXXIII. 214 Mr. G. C. Simpson. [Feb. 3, The first few measurements which I made having shown that the radio-activity is very much greater than in lower latitudes, I determined to make a thorough investigation of it. On November 23, I started a series of measurements of the radio-activity, taking three observations a day ; the first in the morning between the hours of 9 and 12, the second in the afternoon between 3 and 6, and the third in the evening between 9 and 11; the series continued uninterrupted for 4 weeks. The results are given in full in Table I, when the follow- ing symbols are used : B = height of barometer in centimetres. = temperature in Centigrade degrees. = estimated velocity of wind, the direction being added. potential gradient measured in volts per metre. Wah I It will at once be seen that the numbers are very much higher than the corresponding ones for Germany, the mean for the month being 102, which is nearly six times as great as the German mean (for the year); and the highest value, 432, being nearly seven times greater than the German highest. To turn now to a closer investi- gation of the numbers, Table II shows that there is a marked daily period. Table IJI.—Mean and Maximum Values of A for each Week. __——. - Morning | Afternoon Evening : A. A. Mean | | | | ay three. | Mean.| Max. | Mean.| Max. | Mean.| Max. : |S | | Nov. 23 to Nov. 28..| 62 | 114 | 86 | 150 | 146 | 294 | 98 | Nov. 80toDec.5...| 71 | 144 | 69 | 120 | 82 | 208 | 74 | Dec. 7 to Dec.12 ...| 105 | 168 68 | 90 | 129 | 198 | 101 Dec. 13 to Dee 19) 5, 110 204 | 131 | 384 166 | 432 186 Mean for month... 87 88 | Jame | 102 The maximum falls in the evening, there being little difference between the morning and afternoon means. Not only is the mean value greatest for the evening, but the absolute maximum falls in the evening each week, and on 13 out of 22 days on which three observations were taken the evening values were the greatest. It would be of interest to find the relation, if any, between the radio-activity and the other electrical conditions of the atmosphere ; but my observations of the dissipation and the ironisation of the air are not sufficiently complete for a rigid comparison, and so must be 1904.] Atmospherical Radio-actiity in High Latitudes. 215 left for future investigation. Table III gives the relation between the radio-activity and potential gradient. The mean potential gradient for the period of each exposure having been taken with a self-registering electrometer, the values of the radio-activity for each interval of 50 volts meter have been put together and the mean taken. Table III.—Radio-activity and Potential Gradient, measured in Volts per Metre. | Pees Ee 9504 || 50 100), 109-1150.) 150-200) pol. grad. No. of observations . 6 3 | 13 | iG 8 PATHE ME 6. ke we 338 150 96 | 140 150 BAS WMaxaM UI. . v.23 3 66 168 198 | 432 384 200—250. | 250—s00. 3800—350. 350—400.| 100—450. ee _ No. of observations. | 4, 6 8 2 3 | | PERSOFTNC ERY <1 510) 2 t< 0 3) eres 96 | 84 61 | 1il 83 A maximum ......; 144 _ 144 | 188 150 150 As can be seen, there is no direct relation between the radio-activity and the potential gradient. To come now to the etfect of the meteorological elements, Table IV shows that temperature and radio-activity are not closely related. Table [V.—Radio-activity and Temperature. ni O° — 40° to | a 30° to | — 20° to | {e) (o} | fo) ° | Temperature °C. | =30° IG kOe. | —10° to 0°.| O° to 10°. | | i ee ee EA irene AAA EU tr) vale dela!,|| Skea | | No. of observations. 12 10 iy 25 8 | PNGMCAT wie 2 ve wa 4's | see 166 80 82 110 ASmMaAxImMUumM 5... | 294 432 204 174 198 Nor does the barometric pressure appear to have any great influence. Table V.—Radio-activity and Barometer. | 740—750. | 750—760. | anometer ...-......:.|) 720-730. | 730—740. No. of observations... | 10 18 23 20 | AMATO: «2 «0: wie’ | 66 | 109 85 ei MMMTBVCAT @)5 elm <'aiies os v6) | 180 Iw 829% 204. 432 216 Atmospherical Radio-activity in High Latitudes. [| Feb. 3, The only meteorological element which appears to have a direct influence on the atmospherical radio-activity is the amount of cloud. To show this I have divided the observations into three classes: ist Sky clear, or nearly so; 2nd Detached clouds; 3rd Sky completely overcast, and taken the mean for each class. The results are given in Table VI. Table VI.—Radio-activity and Clouds. Detached | Completel Clouds. Clear sky. clouds. pe | | tit No. of observations... 18 | 26 27 Perici tkse is ee gee cae 130 107 | 76 | A Ma, ee oa | 432 384 ) 198 | | H | There does not appear to be any close connection between the aurora and the radio-activity, the greatest value of the radio-activity having been obtained when no aurora was visible. During the whole time these observations were being taken the sun did not rise above the horizon. The time used was mid-Huropean, local mean time being 42 minutes ahead. The place of observation is 140 metres above sea-level. The ground for a hundred miles round is hard frozen to a great depth and covered with a coating of snow the average depth of which is over 2 feet. Sudden changes in radio-activity are sometimes observed to take place, as, for example, on December 17, when with a rising barometer the activity rose for a few hours from the low value of 66 to the exceptionally high one of 384. There is some difficulty in reconciling these observations with Elster and Geitel’s view that the activity is entirely due to a diffusion of a radio-active emanation from the soil. 1904.) — On High Temperature Standards. 217 “On the High Temperature Standards of the National Physical Laboratory: an Account of a Comparison of Platinum Thermometers and Thermo-junctions with the Gas Thermo- meter.” By J. A. Harker, D.Sc., Fellow of Owens College, Manchester, Assistant at the National Physical Laboratory. Communicated by R. T. Gtazeproox, F.R.S. Received January 20,—Read February 11, 1904. (From the National Physical Laboratory.) (Abstract.) This paper contains an account of a continuation of the work of Dr. P. Chappuis and the author,* on a comparison of the scale of the gas thermometer with that of certain specially-constructed platinum thermometers, from temperatures below zero up to the boiling point of sulphur, and in one case to a point close to 600° C. The results of this work substantially confirm the experiments of Callendar and Griffiths, and show that the indications of the platinum thermometer may be reduced to the normal scale by the aid of Callendar’s difference formula : 1-36) i] 100 100 where pi is the platinum temperature, T the temperature on the normal scale, and 6 a constant which, for pure platinum, does not differ much from the value 1°5. The temperatures chosen for the determination of dare 0° C., 100° C., and the boiling point of sulphur. In the present paper the work is extended to a temperature of 1000° ©. Moreover, a number of standard thermo-junctions of platinum—platinum-rhodium were included in the comparisons. The gas-thermometer employed for this work was presented to the laboratory by Sir Andrew Noble; it was obtained, along with materials for the electric furnaces and thermo-junctions, through the kindness of Dr. Holborn, of the Reichsanstalt. The bulbs used were of porcelain, glazed inside and out, and the gas used was pure dry nitrogen. The thermo-junctions, which were carefully compared by Dr. Holborn with the standards of the Reichsanstalt, at a number of fixed points up to 960° C., were again tested and compared together before use, with concordant results. 8L— 7 GZ. IL — — pie a fens ne aed — se ereseeecssccneeoeses 209 qv ~ —') dura} poyenoyee : bs) seeecens Occ ecccccceeeces eoeee S LODL-¥ 9688. 01— 8629. ¥ 2188: F F00G- T 69188-0 89612. 0— ANT aA erteresanenneee HQ ~ ‘O F- 0o881 — ") SP 0381 — ‘O 1. o@8l— ‘D PE. 0881 — "DO G8. o&81-- "OD V+ 081 — "D F- o&8I— ‘0 9-0t8L— |" O “DIL FO padtasqo *y, ”n RQ 00. 198 — €8- 291 — 18. G83 — LL. 188-— €8. 8€3— 12+ 64% — 83+ LLZ— GZ. €93— RPO NE hai 8 A 112000. 0 T1900. 0 86P600- 0 129200. 0 181400: 0 848600. 0 109800. 0 661200. 0 ou iemaie ces a ay = ZRE- 0 ae ae = 612-0 PIL-0 hae eae aaa e 618. 9 162 0 6EF- 0 68: 0 020: 0 221-0 VES: 0 621-0 yoo Dy 5 (ad) LOV-9 stead That re —_ — == peas = eee p ‘Oo “bry 73 cc ; an) bbb 9 a Zb9: 1 = as = mm = rye” NOI ace eae 99F- 9 8TL-0 20+ % G8P- I 1&2. 0 68-0 621-0 ¥868. 1 Po Oe bi — GOL 9 — ae = a _ _— — vreeeress COD Qe QDURYSISOYT + 668.9 bP. 192. L GZ. € PPL 602-8 660-Z 082. ¢ ere “i 6F0- L 19G.¢ 108-6 801-F €29- I LbY. 968.2% 61GE. L TENSE SHEN SC SRP NAAR Ea Son — —— — —. — ences | —— ——$—$—$—— | — —ooe Py ‘(#S.1Y) ‘(FIN) ‘O'pd) ‘(*pa) “(np) “(?3y) “fny) ‘(d) S187 JOATIS UVULLE [O31OIN wHniIpelled WNIpLlyed reddop TVATIS plo» wUNULAe[ I ey. oH NI ‘usso1pATy prubrry Jo yulog Ssurplog ey} 4e ArjyouoULIEYT, oouRySIsoYy OLLOOT|—'T O[qeT, 1904. ] at the Temperature of Bowling Hydrogen. 247 are algebraically higher than in Centigrade degrees, a peculiarity which is shared with them by gold. It is also remarkable that in the cases of all the purest metals examined, their resistances calculated by either method of reduction vanish at temperatures above — 273° C. (Auss was not so pure as Aug, which was electrolytic gold.) As measurers of temperature gold and silver seem to be the best. One prominent characteristic associated with them is, that their 0's are the smallest. Clearly those metals (if there are any) are accurate temperature measurers for which 6 vanishes, so that we expect those to be best in which this constant is least. There is a further characteristic displayed by the best metals as shown in Table I, which may be explained thus: both methods of reduction rely on the parabola, and the farther away the representative arc of the parabola is from the vertex of the curve, the more nearly straight does this arc become and the smaller will 6 be. Now in both metals referred to, especially in the purer specimens, this characteristic is most marked compared with the other metals employed.* It is worthy of note that for these pure platinums the average value of 6 is very nearly 2°5, while Callendar’s platinums, also pure, gave 15—1°6. Is the parabola determined by the resistance at 444°°53 C., 100° C. and 0° C., different from that determined by the resistances at 100° C., 0° C., and —182°°5 C.? In the sequel I shall show that these must be different, and, in fact, that we must look for an entirely different hypothesis to correlate resistance and temperature. As a matter of interest, in the last line of Table I, I have noted the ratio in which the resistance of each metal at 0° C. is reduced on cooling it to the boiling point of hydrogen. This seems to be a quantity showing no connection with other properties of metals. So far we have looked at the results rather from the point of view of metals as thermometers. But a much more important question is, What is the relation between resistance and temperature in metals ? We are entitled to consider the temperatures at which liquid oxygen and hydrogen boil under atmospheric pressure as being known to within one- or two-tenths of a degree, namely, —182°°5 C. and —252°5 C. Further observations made with the constant-volume hydrogen gas thermometer lead to the conclusion that hydrogen * Tn the Callendar parabola (A—T)? = P(B—R), the values of A for Auy and Agi are —27,053° and 14,520° (the only other one greater than 10,000° being —18,087° for Au;3), and the corresponding values of B are —598° and 177° (the only other one greater than 100 being 135° for P.;). Similarly in the Dickson parabola (a7 +R)? = p(6+T) the values of a for Auy and Ag,; are —1080° and 339° (the only other two greater than 100 being 251° for P,; and —129° for Au,;), and the corresponding values of } are —11,841° and 7319° (the only others greater than 2000° being —8400° for Auzz and 2881 for Ag;,). This characteristic comes out equally strongly when the curves are reduced to a common resistance of (say) 1000 ohms at 0° C. 248 Prof. J. Dewar. LHlectric Resistance Thermometry |Feb. 25, freezes about 5° below its boiling point. In the present experiments I have been able to get eight observations in liquid hydrogen boiling under pressures varying from 5—50 mm., and it will not lead us appreciably astray to take the temperatures of these observations as (say) 4° below the boiling point. If the law connecting resistance with temperature be parabolic, the very gentle curvature at the boiling point of hydrogen will allow us to consider the rate of drop in resistance per degree of temperature for the 4° below the boiling. ‘point of hydrozen as roughly the same as that between the boiling points of oxygen and hydrogen (70°), so that the ratio of these two drops on this supposition should be about 4:70, or say one-eighteenth. These ratios are given in Table II. Table II. Tate ae ses Pt+ Rhy. | Atgs. | Arig. .| Aga meee 0-015 0-121 0-06 0-015 0-083 | 0-018 | 0-057 | 0-006 | | 1:269 6 -836 3-21 0-495 2-999 | 1-425 | 1°583 ) 1-512 ee pe ek ns ay | 84 57 53 33 36 |. 79 28 252 | Now these ratios are all much smaller than one-eighteenth, hence we infer that the curves have taken a more or less quick turn in the neighbourhood of the boiling point of hydrogen, or perhaps above it. On Plate 10 the observed resistances are displayed graphically. For convenience I make seven groups, namely platinum, gold, silver, copper, palladium, magnetic metals, and alloys; and in order to bring the characteristics of these groups into- comparison, each metal is supposed to have the resistance of 30 ohms at the freezing point. This number was chosen to suit the scale—the intention being that, roughly, the “plot” of resistance and temperature should be a line equally inclined to the two axes of resistance and temperature. This has been accomplished by taking a centimetre to represent 20° C. in temperature, and 2 ohms in resistance. The diagram for each group has the reading at 0° C. placed 24 cm. higher than that of the group below it, in order to obviate confusion among so many approximately coincident lines. Attention paid to this will enable each group of curves to be hey seen and compared with the others. The first noticeable peculiarity is the close coincidence of the two silver curves. For them, Callendar’s «’s, the ’s, and the ratios of the resistances at 0° to that at the boiling point of hydrogen are almost the same, although the 6s differ much. In like manner the Dickson constants for Agyg are all in the same ratio (about 5:2) with those for Ags4. 1904. | at the Temperature of boiling Hydrogen. 249 Next, the curves for P;, Auyo, the two silver, and Pdys, are very approximately parallel. Of the two palladiums, Pd3; would appear to have contained so much impurity as to have behaved almost like an alloy. The two alloys take up quite independent positions compared either with the purer metals, or with themselves, tending more to parallelism with the axis of temperature. In this connection I may mention that I constructed and used once or twice a carbon thermometer ; in its case the resistance diminished as the temperature rose (a result already known), and its “plot” departed still farther from the pure metals than the alloys do, its 2 being — ‘08048. The magnetic metals present the most striking curves, being at first sight quite unlike any of the others. But on closer inspection we shall find that this is not so, and in fact they give the clue to the general connection between resistance and temperature in metals. The magnetic metals and gold were found to have negatwe values of 6. Now, if we examine the curves of the other metals, they will all be found concave towards the axis of temperature, for the arcs extending from the boiling point of water, through the freezing point, down to the boiling point of oxygen; while below the boiling point of oxygen these curves are convex to this axis. On the other hand, gold and the magnetic metals are already convex to this axis from the boiling point of water to the lowest temperature reached. This leads me back to the research made by Professor Fleming and myself in 1896* on the electric resistance of mercury, in which we were able to observe the resistance of the metal from far below its melting point, and considerably above it when in the molten state. The curve connecting the resistance of mercury with temperature, throughout this range, including the change of state, was somewhat like the disused old English 4 the temperature being measured hori- zontally to the right, and the resistance vertically upwards. In the present instance, though in different circumstances, this same curve reappears. For platinum, silver, copper, palladium, and the alloys these experiments include a part of the curve (fig. 1) starting from (say) P, passing through I, the point of inflexion, and through Q down towards the absolute zero; whereas, in the case of: gold and the magnetic metals, the corresponding part of the curve begins below I, (say) at Q, and proceeds thence towards the absolute zero. Professor Callendar, from former experiments of Professor Fleming and myself, had noticed this behaviour in the case of platinum.7 The eight observations below the boiling point of hydrogen are shown in Plate 10. * “Qn the Electric Resistivity of Pure Mercury at the Temperature of Liquid Air,” ‘ Roy. Soc. Proc.,’ vol. 60, p. 76. + ‘ Phil. Mag..,’ vol. 47, pp. 218, 222. VOL, LXXIII. 48 250 | Hiectric Resistance Thermometry. - ~— [Feb. 25, BiG iE It is clear that in no case can anything parabolic connect resistances and temperatures ranging from the boiling point of water to that of hydrogen. Just as we seek for a circle of curvature at any point of a curve, so in the present case we may, at a point, or over a short range of the /, seek for an approximate parabola; but any such parabola will be of no, or little, use for extrapolation. I have mentioned that the value of 6 for my platinum (Callendar) parabola is about 2°5, whereas observers at high temperatures finds its value about 1°5 for pure platinum. Such differences have been found by others also, but they do not seem to have attracted attention. On looking at fig. 2, we shall find the discrepancy easily accounted for. The portion ABCD of the curve represents roughly (on a much reduced scale, in order to show the curvature more clearly) the curve of P2;, given in Plate 1. The points B, D, I, represent the boiling points of oxygen, water, and sulphur ; and C, F, represent the freezing points of water and silver. I iinet, Ys 7007 0 l000° DeuSoc, Proc., vol. 73, Plate 10. , 135 Gane | [Se Bee ae i ee es Es is ea pees SES ke See is Ee S Peery Ei an : Fal Se > epee ee 3 oe SN Pe eee IS A i 5 © le) Ur © [e) = 2) % | Roy. Soc. Proc., vol. 78, Plate 10. Dewar. 3 ATS, Pale aa 8 BERGER: Vee VAAAEYAG AG) | TT |] Ts, EGRRE BE) BBR S Gi. PT Tes Fs | BE Be ie ka = | a BE ae nea: 8 |_| | | - | | | | | : f a P ei o° YA e L | | Kael _| SeusGuasvsesns. - 7) Ke Ta DO baad Now, at the critical point we have found f, = 3°04 p,, hence combining these results we have finally, t, = 33°88", and thence p, = 11:145 atmospheres. Before leaving these results we may write the Rankine equation in Van der Waals’s form, namely, fs. 54° 34783 l—m ie Mm —log « where « and m are the reduced pressure and temperature, and Van der Waals’s f is 54°34783/¢,, or with the above result, 1-604, about half the usual value of f In lke manner Trouton’s constant, namely, the ratio of the molecular latent heat to the absolute temperature of the boiling point, is 125/20°5 or 6-096, again only about half the usual value. In this connection we may refer to Olzewski’s experimental observa- © tion of the temperature, 192° absolute, at which the Joule-Thomson effect vanishes when the expansion is from a great pressure, in this instance between 110 and 117 atmospheres. With the usual Van der Waals’s notation we may express the connection between this inversion temperature and the critical temperature in either of the forms, ii = : [ee ee ee eka Leones seers ce ee 4\ 9 ) ; (9); or the former equation shows that for small pressures and consequent values of v so great that b) may be neglected in comparison with 1, the critical temperature is 4/27 of the temperature of inversion. But 1904. ] Physical Constants at Low Temperatures. 261 when the pressure is great, the ratio }/v cannot be neglected and the critical temperature is a greater multiple of the temperature of inversion. The second of these equations shows that the initial pressure must not exceed nine times the critical pressure. Assuming a critical pressure of 16 atmospheres for hydrogen and taking p = 117 atmospheres and? = 192°, this equation gives the critical temperature as 42° absolute. Again, for pe = 15 atmospheres, ¢, = 46° and fOk%, — 32, p- = 41 atmospheres. Results derived from a discussion of similar equations depending on Clausius’s formula, Berthelot’s ‘“ modified” Van der Waals’s, or Reinganum’s formula, are still farther from the value we expect. In part explanation of this failure it is to be noted that these formule are but the best theoretical approximations we have at hand, and while they are useful within short ranges, we can hardly expect the same accuracy over a temperature range of five or six times the critical temperature. Again, for a very large number of bodies the product of «a, the co-efficient of expansion at the boiling point, and the critical tempera- ture is constant and about 0°6 to 0°7. Thus for oxygen from equation (2), we have ai, = 0°61. For 0°0750 0°8042 x 15 nitrogen we get at, = x 127 = 0-79, but ior hydrogen 0:0054 0:07 x 58 perature as high as 42, af, only reaches 0°56. 7. There are, therefore, as far as we can see at present, and as far as theoretical considerations can aid us, great departures shown by hydrogen from what are fairly general results. Van der Waals’s f and Trouton’s constant are each only about half the usual values; and we have now found, from the consideration of the point of Inversion of the Joule-Thomson effect, and of the product af, variations greater than the average from the values we should have expected. Further experiment will be necessary before these discrepancies can be cleared up. we have x 34 = 0°45 and even if we take the critical tem- VOL, LXXIII. U . UE 5 ae 262 Prof. K. Pearson. On a Criterion whieh may [Mar. 4, “On a Criterion which may serve to Test various Theories of Inheritance.” By Kart Pearson, F.R.S., University College, London. Received March 4—Read March 17, 1904. (1) One of the most difficult problems in the treatment of heredity is that of obtaining a satisfactory criterion which will enable us to dis- tinguish the truth or falsehood of various hypotheses. As a rule, all the criteria used have been based upon a determination of the type of offspring due to parents of selected types. Unfortunately such a method of approaching the problem of heredity fails wholly to reach some of the most important modern theories, for the reason that these theories start from the assumption that the type of the offspring is not any, or at least any precise and simple function of the parental types. The type is said to be a factor which at present can only be determined by direct observation or by experimental crossing. Mr. Galton in his ‘ Natural Inheritance,’ it 1s true, used the term “ midparent ” to denote an individual compounded, in a simple way, from the two parental types, and giving offspring of a definite type. In generalising, however, Mr. Galton’s conceptions in my ‘“ Law of Ancestral Heredity,”’* I purposely placed before myself the aim of reducing the theory to a purely statistical theory, and discarded entirely the conception that the type of offspring was settled by the parental types. The generalised midparent of any generation became a compound of the devations from type of the ancestry of that generation, and no assumption was made as to any inheritance of absolute type; the theory became purely a statistical theory of the distribution in various generations of the deviations from type. At a somewhat later date the Mendelians gave up the conception that the type of the offspring was known from the parental types. The actual effect of crossing two individuals was compared to the formation of a chemical compound, the character of which could not a prior be predicted from the known nature of the components. It was a matter to be determined by observation or experiment only. With this wider view the original Mendelian theory of “dominant” and “recessive ” characters has disappeared, and that theory has thus far approximated itself to the “ Ancestral Law.” In a second paper communicated to the Royal Societyt entitled the ‘Law of Reversion,” I endeavoured to work out a general theory of alternative inheritance, on the hypothesis that a certain number of the offspring were for any character like one or other parent or like some one or other ancestor, the proportions of offspring like ancestral types diminishing in number with the distance of descent. This — * ‘Roy. Soc. Proc.,’ vol. 62, pp. 887 and 388. 7 ‘ Roy. Soc. Proc.,’ vol. 66, p. 142 e¢ seq. 1904. ] serve to Test various Theories of Inheritance. 263 theory was developed with special reference to certain characters in man and hound, which were said to be alternative, 7.¢., the offspring, if the parental types were different, took after one or other parent. Quite recently Dr. Franz Boas* has published a very suggestive paper on “Heredity in Head Form.” He propounds a theory that the cephalic index in man is a case of alternative inheritance, and that the offspring take after one or other parent. His theory is iess general than my theory of 1899, because he excludes from consideration the reversion to grandparents or higher ancestry. It is more general than mine, in that he assumes imperfect and not perfect correlation between the groups of offspring and the individual parents whom they respectively follow. ‘This I consider a distinct gain. But the neglect of ancestry, other than the immediate parents, renders the application of his theory to so-called Mendelian phenomena absolutely impossible. Thus, when a white mouse is crossed with a grey mouse the hybrid generation can hardly be considered as made up of two groups taking respectively after white and grey parents. In the following, or segregating generation, it is possible to consider the groups as a result of reversion to grandparental or higher ancestral types; it is not possible to deal with them on Dr. Boas’s more limited theory. Hence, I think he errs in terming his theory a generalised form of Mendel’s Law. It is a theory of alternative inheritance, and no such theory which stops at resemblance to the paternal and maternal types can describe the fundamental phenemenon of segregation in the second generation. We must deal with reversion to higher ancestors, whether such reversion be physiologically brought about .by the purity of the gamete or by any other process. A general theory of alternative inheritance may cover Mendelian phenomena ; a theory of the individual dominance of either parent in one or another group of offspring, a theory of what I have defined as intermittent unit prepotency, cannot do so.t Still, Dr. Boas considers that he has evidence for his theory in the inheritance of cephalic index in man, and it seems to me that his paper indicates the manner in which it may be possible to still further generalise my results of 1899.- It is clear, however, that we badly need some criterion to distinguish between these competing theories in the case of measure- ments of the inheritance of any given character. Since none of the three theories referred to is essentially based on the determination of the type of the offspring from the parental types, we are thrown back * “The American Anthropologist’ (N.S.), vol. 5, pp. 5830—5A38. + ‘ Biometrika,’ vol. 2, p. 389. ft Dr. Boas, I think, has not fully understood my theory of the midparent. He repeatedly speaks of the mid parental value of a character and of the offspring clustering round this value on the theory of “ Galton and Pearson.’ There is abso- lutely no antagonism between my theory and the stature of Americo-European half-bloods exceeding both parental types. My midparent is based solely on U2 264 Prof. K. Pearson. On a Criterion which may [Mavr. 4, on a consideration of the variability of the offspring due to parents of given types. Luckily the three theories give us totally different values of the variability of an array of offspring due to parents of given types, and we have in this question of variability a crucial test of the applicability of one or other of the theories to the inheritance of a given character. In order to bring this point out I must briefly consider the variability of arrays of offspring under the three theories. (2) Variability of an array of offspring on the pure statistical theory developed as the “‘ Law of Ancestral Heredity.” If o, be the standard deviation of the offspring, say of one sex, py their correlation with father, py_ with mother and py, the coefficient of assortative mating, then | D =o a/ —- Pine ~ Pf’ + 2pfe Pe Po 1 — pfm? is, whatever be the nature of the frequency distribution provided the regression be linear, the mean of the standard deviations of all the arrays due to parents of given types. Ifthe characters be distributed according to the normal law of deviation, then = will not only be the mean of all the array standard deviations, but the actual standard deviation of each array. If, therefore, the character selected be in each generation distributed according to the normal law, we should expect to find that if we take all pairs of given types, the offspring due to such pairs will have a variability given with reasonable closeness by the above result. If we deal with all the offspring due to fathers, say, of a given type, the mean standard deviation of the arrays will be o; ,/(1 — px”), and in so far as the distributions are approximately normal the standard deviations of all arrays will be the same.* Hence arises the importance, when we use = as the variability of the offspring, of showing that the regression for the given character is linear, and that the frequency is not widely divergent from a normal distribution. These points were dealt with by Mr. Galton in his very first investigation of the subject. He actually considered in the case of stature whether some of the offspring followed the father and some the mother, and showed that > did not vary sensibly from array to array.t Subject, therefore, to a demon- stration for each character that the frequency is approximately normal and the regression linear, we see that the purely statistical theory of deviations from type (‘ Roy. Soc. Proc.,’ vol. 62, p. 387), and the offspring type itself may be wholly different from both parental types, exceeding or falling short of them. * The property that = is the mean of the standard deviation of all the arrays was first stated by Yule, ‘ Roy. Soc. Proc.,’ vol. 60, p. 477, for the case of linear regression. + ‘Natural Inheritance,’ pp. 89—90, and Table 10, p. 207. This investigation seems to have escaped Dr. Boas, see luc. cit., p. 530. 1904. ] serve to Test various Theories of Inheritance. 265 heredity summed up in the “ Law of Ancestral Heredity ” would assert that the offspring of all parents of a given type would have a constant variability, whatever that type might be. (3) Mendel’s Theory.—lt we take as a fair sample of this theory the generalised Mendelian theory, discussed by me in a recent communica- tion to the Society, and now published in the ‘ Phil. Trans.,’ we find that this constancy of the standard deviation of the array is no longer true. It only becomes true if the number of Mendelian couplets on which the character depends is indefinitely great. In other cases, while 7 /(1 — pe”) is still the mean of the standard deviations of the arrays, the actual value of the standard deviation alters sensibly and con- tinuously as we cross the correlation table, always tending to increase in one direction and decrease in the other. Clearly we have, as I have pointed out in the paper referred to, an excellent criterion here between the two theories.* (4) Lastly, let us turn to the theory of individual parental domi- nance. I will give the analysis for this case, extending and generalising Dr. Boas’s formulee. I suppose the total offspring n of a pair of parents to be divided into two groups 7, and n, in number. In the first groups with a mean 7; the fathers are considered as predominant without the mothers being supposed at present entirely without influence ; in the second group with mean mz, the mothers are supposed to have the predominating influence. We may speak of these two groups, for convenience only, as “father’s offspring” and “ mother’s offspring.” Let o,, and o,, be the standard deviations of ‘father’s offspring” and “mother’s offspring” for a given character z; let o7 om, o be the standard deviations of the fathers for the same character, of the mothers, and of the offspring as a whole. The mean m of the offspring as a whole will be given by m=(m m+ 2 m2)/n. Further let ry, T2¢ be the paternal offspring correlations for “father’s offspring ” and “mother’s offspring,” and 7, Tam the maternal offspring corre- lations for the same two groups respectively ; 77, shall be the coefficient of assortative mating between parents, #, y, 2 are the characters in father, mother, and child, x and y being measured from the parental means and 2 irom some other origin. Then, if 8; stands for a summa- ration of all the offspring of the first and S, of the second class, we have NO? = My {Fo,7 + (mM — m)?} +N {oo,? + (2 — m0)?! , or Oe? = (MO o,7 + NaGe,2)/M + MyNe (My — Mz)?/M «22... eee (1). Let py and pine be the total paternal and maternal correlations ; we have Pfe = [Si {a (2 -—m)} +8p {x (2 — m)} |/ (nope) = (Mopo ali + Noo pO o,1of)/ (NTO e)s ‘* ¢Phil, Trans.,’ A, vol, 203, p. 66. 266 Prof. K. Pearson. On a Criterion which may [{Mar. 4, or Ny Ce, 1D) Ce ee = 2 rip oe ii). an prada (il) Similarly, Nn; T No o ese Pie = bre a Vim + — ee T9 DI cece eecereeeccesee (ii1). Fp N Oe Here we have supposed that although the fathers of ‘father’s offspring” and of “mother’s offspring” will, when weighted with their offspring, be unequal in number (7.e. m; and ng), yet their variabilities are the same, and similarly for the mothers. This is equivalent to supposing no correlation between the dominant effect of a parent and his or her deviation from type. Otherwise we cannot put oy in the S; sum the same as o; in the Sp sum. We are now able to write down the general regression equation of bi-parental inheritance, 7.¢., Eh See 4. Pee ep —h = = 1 = Pmnf oF i! ae, pe “fin Om where 2» is the probable value of the character in offspring of parents of characters x and y. Hence, if we remember that ram = pyin, We have on substituting from (11) and (iii) :-— ay = m+(™ 2 Ry, an We “Ra) w+ ‘2 wh) peed fe 7+ Rom ) Yon (iv), nN OF OF it) ho nv m where Ry, Rim are the bi-parental co-efficients (71¢— mp T1m)/(1 — m4) and (tim — Tmf Ti)/(1 — Tm2s), of the “father’s offspring” and Re, Ry» similar quantities for the “ mother’s offspring.” Now, fixing our attention for a moment on “ father’s offspring,” we should expect parents of characters « and y to produce an array of father’s offspring with a mean: fl = = 1d, + Ray EG + Ria oat Ny) eco 0 6 oe 6 cele os eletetetnrte (v), OF On and with a standard deviation s; given by 9 aa JUS - 9 ms 2 * . re eee cal 51° = Tp," (1 — Ty? — Ty? — Vinf + 2rigTiml mf) | (1 —Tin a) tes (v i). Similarly the arrays of ‘‘ mother’s offspring” for parents of the same characters would have a mean: 2 — VIL) + Roy 29 + Ran 2 Yy €.0. 0 0: 010) 600,066) 0) u otutele (vil), Lip and a standard deviation s, given by Sq” > Gre (1 art Tog? ara Tom — Teng? + QropTomt mp) | (1 rar Tonf?)- ee (Vill). 1904. ] serve to Test various Theories of Inheritance. 267 Now, if v; and vz be the numbers of children in the array of a y parents belonging to either group, and v = 11+ v2, we shall have for the standard deviation of the total offspring of « y parents : Vig? = Vy {81 + (oa — Sp)"} + V2 {827 + (po — Sp)*} neo ee (ix), where 2, is given by (iv) and is the mean of the whole array. If the relative influences of mother and father depended upon their characters, we could go no further with (ix) until this had been determined. If, however, we suppose this influence on the average to be not sensibly dependent on the characters « and y, we may write m/n = v/v and n/n = vo/v. Substitute from (iv), (v) and (vii), we find after some reductions : 2 eS) ‘ 14817 + NeSo7 . NN» omn on a = —— -+ 2 { (iy = M2) 5= (Ry — Roy a 2 nu n° - . Of OF To verify this equation I transferred to the mean, summed for every possibl@ array, z.c., for all values of x and y, and divided by the total number of arrays. The left-hand side should be o,”, the right-hand side became, after some considerable reductions, identical with the right-hand side of (i), as it should be. We have in equation (x) accordingly, the variability of an array of all offspring on the hypothesis that the children may be divided into two groups, differently related to the two parents. We see at once that the variability of an array will depend on the actual values of the parental characters, unless : Riypoe,/oy 7a Ropo,/ay ’ and res Ora = Bone Con: But this is asserting that the bi-parental regression co-efficients for the two groups are the same, or, as we may put it, that there is no distinction between the parental influences in the two groups. ‘This is the case usually assumed under the “Law of Ancestral Heredity” with its constant variability within the limits of random sampling for the arrays. In every other case the arrays alter in variability with the magnitude of the parental character. Let us look at the matter from another light and sum 2?,, for every value of y only, or we can obtain the same result by direct investigation, namely, we find if =, be the standard deviation of an array of offspring due to fathers of character «, oe mo,” (1 — rif’) ee ex (1 — 19”) te v ) 2 Ny Ns com o . + 6 my — Mo + ( Typ — — Tog “es iD | he Cale nu Seep OE Of i ae , , 268 Prof. K. Pearson. On a Criterion which may [Mar. 4, Or, we see that unless Tif Te,/ OF as ia, Tc,/ OF; z.c., the paternal influence be the same in both groups, or there be no question of individual dominance of the parents in “father’s offspring ” and ‘‘ mother’s offspring” respectively, the arrays of offspring due to different classes of fathers will not be equally variable. It is clear that if the standard deviation of the array be plotted to the size of the father’s character, the resulting curve will be a hyperbola with real axis vertical, and in the two directions across the correlation table taken from the parental value Cras o “@ = (Mm — M2) / (rx 2 Ty v2) OF OF the variability of the arrays will rapidly increase from a minimum. As there can hardly be a sensible distinction between m, and ia, for it would mean bimodality in all the characters dealt with, which is contrary to experience, we may say that the variability °of the ofispring arrays increases in both directions with the deviation of the father (or mother) from the mean. If we put m = m, we have: N9 ay xu a mga (ry) teat Lara") 4 i (yyy), (al n nv o o which shows clearly how the variability of the array increases hyperbolically with the deviation of the father from the mean. (5) Now it is as well to take one or two numerical cases to appreciate the degree of curvature of this hyperbola, for if it were a very flat hyperbola, it possibly could not be readily distinguished from a horizontal straight line (Ancestral Law) or from a parabola (Mendel’s Law). I take the hypothesis suggested by Dr. Boas, ie., a negligible influence of the father on “mother’s offspring,” and a negligible influence of the mother on “father’s offspring.” We have ro7 = 71m = 0. Further, if we suppose the two groups of ofispring to be equal in number and equally variable, ny = m2 and =o, From (i) it follows that Therefore from (xi) it follows that 5 1 La" = C2 { (1 => 2.p7 40) + Die of } a 1904. | serve to Test various Theortes of Inheritance. 269 Accordingly, we have: Be = Y) oe ob Jone) b= oR Ly? = 62 (1 — py”), C= / 20-7 Di = Giese % = oy, ee Oe (Lp), whence we have the following table :— Table of 2z/o¢. Pie: x 0°3 0-4 0-5 | Oe. 0°91 0°82 0-71 Gp 0-95 0-96 0°87 er: 1-00 1-00 1-00 Qap... 1:09 1°15 1:22 Now parental correlations in my experience of extensive masses of good data are rarely as low as 0:3, generally over 0:4, or nearer even 0-5. But even with 0-3 we see that there ought to be about 20 per cent. increase in the standard deviation of an array as we pass from the mean father to a father with a deviation equal to twice the paternal deviation, 7.¢., to an array which will be given with at least a moderate number of cases In any parental correlation table. As we approach a parental correlation of O-4—0°5, this increase of the standard deviation amounts to 40—72 per cent., and should be still more conspicuous. We conclude, therefore, that if the parents are respectively dominant in two separate groups of offspring, then when we plot the standard deviations of the arrays of offspring to the deviations of the parent from the mean, we ought to get a very sensibly hyperbolic curve, Or, if we plot the squares of the one to those of the other, we ought to get a sloping straight line of very sensible slope, 0°1—0-25 about, instead of the horizontal line of the “¢ Ancestral Law.” If we suppose the same conditions to apply to a bi-parental array, 16., My = Mo, M1 = Na — £N, Go, = Oc, = Ge, Toe = Tim = O, we find that (x) reduces to SSE Tala) 1 Tif + em? Srp a 41am Y 5 2ny” = Fe Nas cron) A Oly Saleh ARPS VO) st | : t eH Tinf 1 a i OF 1 aa | mf On Or, using (ii) and (iil), which give pye = 4riy, Pine = $2! 2 2 2 Se ae { ( LNG) = )+( Pfe ss Cos eee | \ (xiii). Pf} \L—pmp oe 1 omy? om " | 270 Prof. K. Pearson. On a Criterion which may [Mar. 4, Let us apply this to the case of stature of parents and sons in man. Here I have deduced from my “ Family Data Records ”* py = 0°5140, ps =.0-4938, pnp = 2801 (xiv). We find: 2 sy ees — 0°1028 + (0°5579 © _ 9:5 359 +) pp (xv). OF On Hence we have the impossible value 320, or o; for the variability of the arrays of offspring due to mean parents. Generally we deduce— Values of 2,,/o, for Arrays. Father: 2z/o7 Mother: y/o. | Array =) 27/07 —————————— | | | 0 | 0 (0 324/(—1) ?) eA | 0 0-46 ii | zits 1-05 Si | = 1-60 | AS | BO 2°19 | | | We are dealing here with measurements on upwards of 1000 families ; the probable errors, therefore, of py. and pme hardly allow of our supposing the first term on the right in (xiv) to be zero. But if we do make this hazardous assumption, we see that as we pass from pairs of mean parents to fathers of 6’ 1” and mothers of 4’ 10” we should pass from an array of offspring of no variability to one twice as variable as the general population. Further, for any given male there would exist a female relatively only very slightly taller than he is, who would have offspring with him of sensibly no variability. In view of this result we may safely assert that the hypothesis of ‘father’s offspring” and “mother’s offspring” cannot apply to stature In man under any conditions in the least approximating to Dr. Boas’s assumptions in the case of cephalic index. (6) Our conclusions may be summed up as follows :-—The variability of the array of offspring due to a group of parents of definite character can be satisfactorily used as a criterion between various theories of inheritance. In particular if the variability of the array be plotted to the character of the parent— (a) On the hypothesis of the ‘“‘ Ancestral Law” a horizontal straight line is the resulting curve. (5) On the generalised Mendelian theory the result is a paraboia with horizontal axis. * « Biometrika,’ vol. 2, pp. 373 and 378. 1904.] serve to Test various Theories of Inheritance. 271 (c) On the generalised theory of alternative inheritance which divides the offspring into two groups more intimately associated with one or other parent, the resulting curve is a hyperbola with vertical real axis. (7) I have applied the criterion here developed to my measurements on father and son in more than 1000 families. The three characters stature, span and forearm are dealt with. The correlation tables for the paternal inheritance of these characters will be found im a recent memoir by Dr. Lee and myself on the “Inheritance of the Physical Characters in Man.”* We excluded all arrays with less than eight individuals in them, deeming it absolutely untrustworthy to find a mean and standard deviation from fewer than eight cases. The standard deviation of each array for a given paternal character was found by Dr. Alice Lee. These were then plotted to the paternal character by Mr. W. L. Atcherley, and results are shown in the accompanying Diagram 1. The zigzag polygons in each case give the plotted variabilities of the arrays, the vertical numbers being the total on which the variability is based. The horizontal line AA gives the mean value, - «/(1 — pyc”), of the standard deviations of the arrays according to the statistical theory. The broken lines ccc and ¢'c'c’ give o, /(1—p2x) + twice the probable error of the deviation of an array from a, /(1 — p%e). Thus if 2=o, ,/(1 — pe) and 2, = the standard deviation of an array of m individuals out of a total of n, we have plotted up and down from > the quantity SS! 2 x 0°65449 NAG hes z Now unless a difference is at least twice its probably error we certainly cannot assert it to be significant. Now, if the diagram be examined, it will be seen that almost without exception the zigzag polygons fall well within the non-significant areas bounded by ccc and ¢c'¢’ There are, indeed, three exceptions, but all three occur in arrays vith less than twenty individuals, or arrays where some eccentricity of individual or measurement might easily make itself felt. But there is really no need to appeal even to this explanation, we have thirty-nine observations on arrays, and three of these only exceed the double of their probable error; this is actually less than half the excesses we might have expected on the theory of probability. There is clearly absolutely nothing in the observed results opposed to that constancy of variability in the arrays which is suggested in the usual treatment of the “Law of Ancestral Heredity.” (8) I next look at the theory of alternative inheritance, or at the hypothesis that some children follow the one, some the other parent. An examination of the diagrams show that there is not the least * © Biometrika,’ vol. 2, pp. 415—417, ”'SS=SzS S'S O 272 Prof. K. Pearson. On a Criterion which may [Mar. 4, approach to a hyperbolic distribution of the array variabilities. The hyperbola would have its axis vertical and vertex downwards. In the case of stature I give the hyperbola that would result from a special FS (263) (62) (#42) (22-5) (46) L | G/ O62 63 Cf 65 66 67 68 62 70 Ww 22 73 TH T5 76 A ather’s SS pam. S72 of Farearw. SID ef & tature. es | a Kan I Bakes Waele . — a RS c) ‘SS z me Cot ha wiok Ce = ) Ou etO alas he Le Ae : Q K} a Bs 6) IR See CMY SSR ECUM els Ye ree alia SS aaenCOM e ! PoLbats trv SMITAYS ' Bey 2 a a EE ay 16 ; Gan 18 ie) o Bf Acather's €lozmeanse, Diagram I.—Arrays of sons. case, namely, Dr. Boas’s hypothesis of equal numbers of offspring with equal variability and of like mean following each parent. The equation to this hyperbola may be found from (xi), by putting 1904. | serve to Test various Theories of Inheritance. 2713 to I be = jal) =) oh; Teo, = Te, a Ce —— 2h LO} Law => 2 Pfe = 1:0280, Ny = n Tor = 0. We have (Zafoe)? = 04716 + 02642 (w/o) where of = 2'°568. It will be seen at once that the variability of the array is too small for the mean father and far too large for the exceptional father. I think we may safely conclude that for man in the case of the three characters under investigation no such theory of alternative inheritance applies. (9) I now turn to the third or Mendelian hypothesis. Are the variability distributions better represented by parabolas of horizontal axes than by horizontal straight lines? We have a series of points in each case and the question is: What is the best fitting parabola ? The equation to the Mendelian parabola, as given in my memoir already referred to,* is in the notation of the present paper Pee + Mics (xvi) Sra ocmasn rine Xvi), where gis the number of Mendelian couplets, and z is measured from the mean.t Writing 2,?/o.? = &, x/o¢ = », this is of the form, a and b being numerics : & = at+bn. Applying the method of least squares, since a is an absolute constant, we find for the best value of 0, pz being the total in the % array : peso S (Ma: (2a? — 57e") ay Ge" (uw) or, ¢ the number of Mendelian couplets is to be found from — Lo _ 9) of 8 oe (22? — Gre") a} V@ 4 C2 S (Mx az) Working this out for the case of stature ‘first, I found, after some rather laborious arithmetic, that g = 48. Thus the best fitting Mendelian parabola needs no less than 48 couplets. We may write the parabola in the form De Hy = -$ (1422), To OF * © Phil. Trans.,’ A, vol. 203, pp. 66-67. + To reduce to the result of the above memoir we have Zz = tos, oc = of = eV (3;)n w = e(s—4n), g =n, where « is an undetermined constant depending on the relation between actual scale and number of Mendelian couplets. 274 Prof. K. Pearson. On a Criterion which may [Mar. 4, or, since z does not exceed 2 to 3 af we have for the part of the parabola involved very nearly the straight line Soh, SNe (1 ear = erates :.:. (xviii). This, substituting the values of o, and oy, is a straight line of slope 0-021. Now the best fitting line to the observations, A’A’, has a slope of 0-022, or we conclude that the Mendelian parabola is when the number of couplets is as large as in the present case sensibly parallel to the line which best represents the variability of the arrays plotted to the parents’ character. The Mendelian line B’B’ of Diagram I is not as good a fit as the line AA’, because the theory constrains it to pass through the point given by o; = ,/(8/9)o., and not through the actual mean point o, ,/(1—7r,”). This is owing to the fact that the Mendelian theory gives 77. constant and equal to 1/3. Hence, we see that the Mendelian theory will not, as a rule, give as good a fit to the observations as the best fitting lne, when the number of couplets is large as in this case and the correlation differs from 1/3. The parabola thus sensibly coincides in direction with the best fitting straight line, but is raised above it in position. I give the best fitting straight lines for the three characters we have been considering For Span— . y — 2'°762 = 0:011 (x — 67-396). Probable error of the slope 0-011, equals 0:013. For Stature— y — 2°°344 = 0-022 (w — 67-686). Probable error of the slope 0:022, equals 0:013. For Forearm— y—V°773 = —0:008 (# — 18279). Probable error of the slope — 0-003, equals 0-009. Thus, of the three slopes all differ by less than twice and two of them by less than once their probable error from zero. We may accordingly conclude that neither in the best fitting straight lines, nor consequently in the Mendelian parabolas for the measurable range, is there any sensible deviation from horizontality. In fact we have 48 couplets in the case of stature, and roundly 150 for span* and 2000 for forearm. With such numbers the Mendelian theory cannot on the problem of variability of arrays give any other sensible answer for the range available for investigation than the constant variability * Tf the Mendelian theory discussed were correct, it would be difficult to grasp how the forearm-inheritance could be determined by far more couplets than the span is, or why one slope should be negative and the other positive. 1904. | serve to Test various Theorves of Inheritance. 275 of the “Ancestral Law,” but it does not equal the latter law in its description of the facts, because it shifts the mean variability of the array out of its proper place owing to its absolute rigidity in the value of parental correlation. I think we may say that for the three characters here considered the ‘“‘ Law of Ancestral Heredity ” is the only one of the three theories which clearly describes the facts, and the facts at no point differ significantly from its statements. (10) The theory of alternative inheritance, which differentiates the offspring into “father’s offspring” and “ mother’s offspring,” permits of being tested in another manner, namely, by considering the variability of the array of brothers who have a brother of given character x Let us first look at the problem generally. A man of extreme character value will have a group of brethren, say, ‘father’s offspring,” who allowing for regression are like himself, but the other group of his brethren will be far less like himself, and on the average of all mothers near the population mean. ‘Thus, the array of brethren of a man of extreme character will be made up of two components, one tending to be like himself, the other like the general population. Hence on the whole the array of brothers corresponding to men of a given type must become more and more variable, the more marked the deviation of the given type. I will now give the results of the analysis. Let p be the general correlation of brothers, m their mean, and o their standard deviation. Let 7); be the correlation of brothers who are “father’s offspring,” m, their mean, and o their standard deviation ; let 722, m2: and o be the corresponding constants for the brethren who are “mother’s offspring.” Let 1; and vz be the number of brethren of each class. Then if an individual have a deviation a from the mean of all brethren, he will if a “father’s offspring” have a co-fraternity of standard deviation o, ,/(1—71;7) and at distance from the mean of offspring given by Oy fay = Mm +7, = (@+m— Mm) — mM — px O7 (711 — p) & + (mM — mM) (1 — 713). He will have v; — 1 of such brethren, but he will also have vz brethren of mother’s offspring type with a standard deviation o, ,/(1 —7 2”) and at distance from mean of offspring given by 02 /: : fag = Mz4+ 2 = (@+m — my) — mM — px : O71 | @ aes p) @ + (m2 —m) (1 — 742). vj Similarly, if the brother of deviation « were a ‘“‘ mother’s offspring,” his brethren would be made up of .ve—1 of standard deviation a2 ,/(1 —729”) and mean pgs, and of v; brethren of standard deviation 7 /(1 —12?) and mean p21 where 276 Prof. K. Pearson, On a Criterion which may [Mar. 4, P22 (rap = p) 10 se (mg = my cl 7 12) p21 = (rw = = p) H+ (ny = nv) (1 = 112). 2 Hence, generally for the variability of the brethren of a given brother of deviation 7, we have SS) 9 e3 4 4 == “i i 2) oe ey 1) aa na) +o, = 1 — ¥99) + = mn rt lee ) > by SEE na vy,—-1 2 Vo—1 V9 a yaya Py MY Osu) | BIG sy) ee Vy + p21? ae 2 (v1 +2 -1) Bote (six): Clearly this gives a hyperbola for =, in terms of #, with its real axis perpendicular to #. ) Similarly we lave : v1 (v1 — 1) wes +2)(¥1) +2 —1) {ry1017 + (my - 1)?} v2(v2 — 1) ae + V2) (v4 + 2 — Ty (799027 + (Mz —m)?} a 2v1V2 (11 + V2) (4 + V9 — 1) and further {1490102 + (1 — M) (M2 — M)} ....-. (ax) : V V9 o? = 3 __ fo42 + (amy — m)?} + —*— {or? + (amg — m)} , Vy + V2 VAs 2 VyMN, + VoIMNe My + Me and m= In any actual case we may reasonably suppose our pairs of brothers to be a random sample from families with indefinitely great numbers,* or put v; and v2 infinite in the ratio of m to m2. Hence: m = (MmN1 + MgNn2)/N , a? = (moy? + Neoy”)/n + “ae — M2)? , 2 2 n Ny 2nyn Q ~ Tio" ar pt T9909" a = COR Oy Woanonbstioundeddonaodesec (xxi), N 1A () > q I * «Phil, Trans.,’ A, vol. 203, p. 77. 1904. ] serve to Test various Theorres of Inheritance. 270 B= 24 5 (1-8) + 2rd) b he ~ra°) +32 (=r) b 5 (at® + p21”) +5 = ~ (va2” =F (4127) onoanoodedod (excl To simplify still further, assume no distinction in the totals of ‘mothers’ offspring” and “ fathers’ offspring,” or take. my = m2 = m, ooo. 1) — i) — 47, lence P = (rir +722) + $712, Xg? = 07 {1-4 (rr? + P20") — 3 1 — 712?)} ied) top) t= 2'(Fip p)) 2. Hence we see that if the variability of the arrays of brethren is to be constant, it is absolutely necessary that 7; = 722 = Tig = p, or the degree of likeness between brothers whether they belong to “ mothers’ offspring ” or “ fathers’ offspring” must be identical. If this be not true, the variability of the array oi brethren of a brother of given character must obey a hyperbolic law, being least for a brother of mean character. If we adopt Dr. Boas’s theory of complete alternate inheritance we have i Ti, = T22 and Ty. =0. Whence PU = 2p, and Ge (W202) As OPM crak (seo ealstas elace (xxiii). This is a very easy result to test. On the whole, however, it is better to ask the general question: Are the variability of arrays of brethren hyperbolically distributed ? (11.) I propose to answer this by appeal to my data for nearly 2000 pairs of brothers measured for their cephalic index. Selecting the arrays of 20 brothers and upwards the results plotted in Diagram II. were reached. Here p = 0°4861 ando = 3°314 and the mean cephalic index = 78°92. Hence the “ Law of Ancestral Heredity ” gives for the mean value of Zz, 2 = 2°896. On Dr. Boas’s hypothesis Qe? = 5°7924 + 0°236327. These results are shown in the same manner as in the first diagram, the broken lines marking the limits of twice the probable error of = and the actually observed =, We may, I think, safely conclude from this’ result that the VOL, LXXIII, x BP ts il on! 278 Prof. K. Pearson. On a Criterion which may [Mar. 4, horizontal or mean line of the “ Ancestral Law” gives a better result than the hyperbola. Asa matter of fact in the small arrays at the extremes we get as before very erratic results; almost any value of the standard deviation may be reached, not only because the probable errors are so large, but because the appearance of any single abnormality or of any slip due to measurement or classification becomes so very disturbing in these small arrays. Even then we only find four cases in seventeen which deviate by more than twice the probable error from the mean line. In this case the mean error of the ‘‘ Ancestral Law ” for the seventeen arrays is 0°31, and of Dr. Boas’s hyperbola 0-49.* &.D.of Arrays of Praweltharen, 7 67f F213 Ye 175 16 FF 18 (9 BO SL SR BSS BH GS GE 87 Cephalic Index ot I"” Faxother: Diagram II. (12.) It may be asked how Dr. Boas has reached the conclusion that in the case of cephalic index the “ Ancestral Law” does not apply and that the children break up into “‘ mothers’ offspring ” and “ fathers’ offspring” ¢ He has used the measurements of Dr. Maurice Fishberg of New York, on forty-eight families of “East European Jews.” Now, I think it may be reasonably questioned whether a population defined as “‘ East European Jews” can be considered as homogeneous. I mean by this: Would any such category breed true to itself? Is it to be looked upon as a “race” in the sense used by met of a popula- * Tf the large values often found for the variability of the extreme arrays of a correlation table have real significance, which I doubt, I personally am inclined to think it is due to deviation from true linear regression, the regression curve being a cubic curve with its inflexional portion representing the regression line of the ordinary range. + ‘Biometrika,’ vol. 2, p. 511, and compare with the stable Mendelian popula- tion discussed in my paper, ‘ Phil. Trans.,’ A, vol. 203, pp. 58—60. ee 1904.) serve to Test various Theories of Inheritance. Ae tion which has been isolated and intra-bred for some generations ? If not, it would very possibly exhibit ‘‘segregation” in the second generation ; thus if there were any Aryan or other non-Jewish blood in the immediate ancestry, this might be induced by a segrega- tion tendency increasing the variability in the offspring of such pairs. Further apparent segregation would arise in the case of any unfaith- fulness, and this might be correlated with physical dissimilarity. Shortly, we want a good deal further information as to the nature of these particular American Hast European Jews. But quite apart from all this, the forty-eight families had only 158 children or an average of less than 3:3. Now, some families must have had four and five children and Dr. Boas gives the standard deviation of all the separate families, or he must have found a standard deviation here and there on the basis of at least two and possibly even fewer individuals!* Ido not think standard deviations so found can be of any valueat all. Further, and most important, he has not determined the standard deviation about the mean of the individual family, which he ought to have done, but about the theoretical average mean of all families having a father and mother of the given characters. Now the individual family has all its ancestors in common beside the father and mother and hence its mean is nof in our previous notation : UG m+ Rye Je Ree Y, OT the value used by Dr. Boas, but the much longer expression given by the general formula for “ Ancestral Heredity.” 7 It is only when we take the average results for all families of parents x and y, that the other ancestral terms will disappear, and we reach the above result His only method of reaching an approximation to the mean of an indurdual family, would be to actually find it from the family itself, and then investigate the standard deviation from this mean.{ This he has not done, and I think we must doubt, even if he had done it, the validity of means and variabilities based upon two to four individuals. Still his memoir is very suggestive, and it seems to me that the investi- gation of a large series of head measurements in parents, and, if possible, adult children, say, in 1000 families forming a homogeneous population, would be of great value. * TI have shown in ‘ Biometrika,’ vol. 1, p, 399, that the values of the standard deviation found from two individuals will, on the average, only be about 0°563 cf its true value. Similar large reductions occur if it be found from three to four individuals, and soon. This is quite apart from the large probable errors intro- duced by paucity of numbers. + ‘Roy. Soc. Proc.,’ vol. 62, p. 394. ~ There would, I take it, certainly be correlation between r—y and the omitted ancestral terms in the family mean. lp 280 Ona Criterion for Testing Theorres of Inheritance. [Mar. 4, (13.) The present paper has, I think, brought out the following points :— (i) A variability criterion between contending theories of inheritance is possible, and is easily applied. (ii) Variability plotted to parental character will give (a) a horizontal ’ straight line on the Theory of Ancestral Heredity, (b) a parabola with axis along the axis of parental character on Mendelian Principles, (c) a. hyperbola with its real axis perpendicular to the axis of parental character on the Theory of Alternate Inheritance which is summed up in the idea of parentally differentiated groups within the family. (iii) As far as my own measurements on stature, span, forearm, and cephalic index in man go, there is nothing to support the view that the variation curve of (11) is effectively represented by either a parabola — or a hyperbola. Within the limits of the probable errors of random sampling it appears to be a horizontal straight line. Further applications of the criterion will no doubt be soon forth- coming, but it is essential they should be made with a full under- standing of what the various theories amount to, and how they must be applied to observations. In particular it is very needful that we should distinguish between the mean of an individual family as deter- mined from all its ancestry, and the mean of an array of offspring of given parentage-type—but with all varieties of earlier ancestry— determined from that parentage only. 1904. ] Some Uses of Cylindrical Lens-Systems, 281 “Some Uses of Cylindrical Lens-Systems, including Rotation of Images.” By GerorGE J. Burcu, M.A., D.Sc. Oxon., F.RB.S., Lecturer in Physics, University College, Reading. Received February 29,—Read March 10, 1904. It was, I believe, first shown by the late Sir George Stokes that if two similar cylindrical lenses are placed in contact, with their axes of curvature at right angles, the combination acts as a spherical lens, and that if the axes are not at right angles the system is equivalent to a cylindrical and a spherical lens combined, the spherical element dis- appearing when the axes are parallel. I have recently had occasion to employ a modification of this arrangement which has proved of some practical utility as well as theoretical interest. When, as in Sir George Stokes’s experiment, the two lenses are placed in contact, the combination acts as a spherical lens in every Deel anes But when the two lenses are not in contact there is only one pair of conjugate points at which a real object will give a real image. Reference to the ordinary focometer formula will show that this must be the case. Let / be the distance between an object and a screen, and let a convex lens of focal length f < 4/ be moved along the line joining them. Then there must be two, and only two, positions of the lens at which it will form images on the screen, one at a distance wu from the object and v from the screen, and the other at a distance v from the object and wu from the screen where respect, that is to say, the formula aa holds in regard to it. U U0 U-v— 8, and ae In either of these two positions a cylindrical lens of equal power will, if its axis be vertical, form sharp images of vertical lines, or of horizontal lines if its axis be horizontal. If therefore we place one cylindrical lens with its axis vertical at a distance u from the object and a second of equal power with its axis of curvature horizontal at a distance wu from the screen, there will be produced on the screen a sharp image magnified vertically v/w times, and horizontally u/v times. This method may be employed for comparing by photography curves plotted to different scales, or for increasing or diminishing the ordinates of a curve or record the scale of which is unsuitable. Figs. 1 and 2 are an example of this latter use. Fig. 1 is a record taken with the capillary electrometer of the dis- charge of the electrical organ of Malapterurus, the two sides of the 282 Dr. G. J. Burch. Some Uses of Cylindrical [Feb. 29, organ being connected with the outer coatings of a pair of condensers, the inner coatings of which were connected with the electrometer. Hing: ale A full description of the experiment is given in the papers by Gotch and myself, from which this illustration is taken.* In order to find the E.M.F. indicated by such a record it is necessary to measure the subtangent—or, in this case, the polar subnormal to the curve at various points according to the method described in my papers on the “Capillary Electrometer.”+ The method is, however, inapplicable if the angles are too steep. It occurred to me that if the ordinates could be diminished and the abscissee magnified by photography the angles might be brought within the measurable range. Accordingly I fixed the original of fig. 1 at 56:25 cm. from a photographic plate, with a cylindrical lens of +12°5 cm. with its axis horizontal at a distance of 17°5 cm., and a second with its axis vertical at a distance of 43°75 cm. The result is shown in fig. 2. The definition is remarkably good in the negative, “Fie. 2. * ©Phil. Trans.,’ B, vol. 187, p. 347; ‘ Proc. Roy. Soc.,’ vol. 65, p. 434. + ‘Phil. Trans.,’ A, vol. 183, p. 81; ‘Proc. Roy. Soc.,’ vols. 48, 59, 60, p. 388, and vol. 79, p. 175. 1904. ] Lens-Systems, including Rotation of Images. 283: but the field of view is small. The abscisse are magnified two and a half times and the ordinates diminished in the same proportion, so that: the total effect is the same as if the original photograph had been taken on a plate moving 25/4 times as fast, and the resulting curve can be easily analysed. It should, however, be noted that fig. 1, being a curve with polar co-ordinates, is not properly adapted for this process, which requires, to be accurate, curves with rectangular co-ordinates. It serves, never- theless, to illustrate the method, and as the portion photographed occupies not more than an angle of 2° of are, the error involved is hardly noticeable although great enough to render it not worth while to spend time over a minute analysis of the modified curve. It is not necessary that both lenses should be of the same focal length. If they are, the image is magnified as much in one direction as it is diminished in the other. If it is desired to preserve the dimensions in one direction unaltered—as, for instance, in comparing the time relations of two electrical responses of different intensity— é.g., nerve under the influence of CO, and in a normal condition—one lens may be put midway between the object and the screen. If m be the magnification required in the other direction, the focal length necessary for the other cylindrical lens is f a, . land its distance from the object is — ! 1+m In practice it is more convenient to employ a pair of lenses of rather more than the required power when close together, and to adjust the distance between them until the right magnification is obtained. The image thus modified by the cylindrical system is projected on to the curve with which the comparison is to be made, and examined with a lens. The principal difficulty is the smallness of the field of view, which is elliptical in shape, the major axis being limited by the diameter of the lenses and the ratio between the axes being as 1:m?. Unfortunately the use of lenses of larger aperture introduces errors of chromatic and spherical aberration which spoil the definition. Owing to the fact that each cylindrical lens has to be separately focussed, and that no image is formed until both are correctly adjusted, not only as to distance but also as to the direction of the axes of curvature, focussing is a matter of some difficulty. I have found the following method answer very well :— | Two lines, one vertical and the other horizontal, are ruled, across the middle of a glass plate—a spare negative answers very well, and the lines can be ruled with a needle on the film. This is placed in the object holder. ‘The first cylindrical lens is inserted in its cell and the horizontal line focussed with it sharply on the screen. The first lens 284 Dr. G. J. Burch. Some Uses of Cylindrical [Feb. 29, being removed, the second is inserted in the other cell and adjusted to focus the vertical line. The first lens is then replaced in position and rotated till both lines are sharply focussed. When three lenses are used, the two employed to obtain magnification must be contained in a sliding tube furnished with a pin working in a slot to prevent rotation. In the preceding pages the axes of curvature of the two components of the cylindrical lens-systems are, in all cases, at right angles to each other. Equally interesting properties are possessed by those systems in which they are parallel to each other. CASE I.—Two equal convex cylindrical lenses set utth their axes of curvature parallel, and at such a distance apart that thew principal foc coincide as in a telescope. A clear and undistorted view of distant objects is obtained, but on rotating the object the image appears to rotate with equal angular velocity in the opposite direction, and on rotating the tube containing the lenses, the image appears to rotate with twice the angular velocity in the same direction. The reason is obvious. Suppose the axes of curvature are vertical, then there is no vertical deviation of the image-forming rays, and so far as its vertical components are concerned the virtual image coincides with the real object. Inthe horizontal plane there is deviation. The rays from the cylindrical objective cross as in a telescope before reaching the cylindrical eyepiece. If, therefore, we focus the eyepiece so that the final virtual image is at the same distance from the eye as the real object, a clearly defined image will be produced, erect, but enantiomorphic, as if reflected in a plane mirror. If the object is viewed through two cylindrical telescopes in tandem, then if either is rotated, the image rotates in the opposite direction, but if both are rotated together, the image remains stationary as in the case of two erecting prisms in tandem. Two causes may disturb the sharpness of the definition, namely, deviation from exact parallelism of the axis, and error in adjusting the distances between the lenses, so that the horizontal components of the virtual image are not focussed at the same distance as the vertical. If the two lenses are not of equal focal length the resulting images are clearly defined, but not symmetrical, being magnified or diminished in one direction and of natural size in the other. CASE II.—Two cylindrical lenses with ther axes of curvature parallel, the distance between them being greater than the sum of their focal lengths. Such a system acts as a compound microscope, giving well defined images of objects situated at a certain distance from the objective. —'? 1904. ] | Lens-Systems, vncluding Rotation of Images. 280 The same phenomena of rotation of the image on rotating the lens system are observable in this case as in the cylindrical telescope, but with this difference: the images, though sharply defined, are not symmetrical, the object being magnified at right angles to the axis of curvature, and not magnified parallel to it. The focal distance at which alone good definition is obtainable is that at which the virtual image is the same distance from the eye as the object. But to an astigmatic eye there are two distances at which the instrument gives perfect definition, the direction of the axis of curvature coinciding with the astigmatic axis of the eye in one case and being at right angles to it in the other. The cylindrical telescope cannot by altering the focussing be made to project an image on a screen like an ordinary telescope. To do that we must employ two cylindrical telescopes with their axes of curvature at right angles. They need not both be of the same power, and, curiously enough, one may be situated inside the other without interfering with its action. Polarised Light. Another peculiarity of the cylindrical telescope is rather remarkable and extremely valuable. As would be expected, rotation of the tube of a cylindrical telescope, though it rotates the image, is absolutely without effect as regards the polarisation of the ray. There is, so far as I know, no other means of rotating an image without altering it in this respect, the ordinary erecting prism, whether silvered or unsilvered, introducing a difference of phase. I have found the cylindrical telescope extremely useful in some experiments where this was 1mportant. Note added March 12. If the axis of a telescope of which the magnifying power is m is moved sideways through an angle 6, the displacement of the image is (1+m)6, taking the upper sign if the image is inverted, and the lower sign if it is erect. In the telescope formed of two equal cylindrical lenses, parallel to the axis of curvature m= -—1, and at right angles to it m= +1. Accordingly, if the axis of such a telescope is moved through an angle in the plane of the axis of curvature, the displacement of the image is zero. If it is moved in a plane at right angles to this the displace- ment is 26. And if the plane of movement makes an angle of 45° with the axis of curvature, the image appears to move at right angles to it. The effect of rotating the telescope about its axis while moving it in azimuth is very striking. A cylindrical telescope may be made with one lens as follows :—A 286 Prof. A. Gray and Mr. A. Wood. fect of a [Mar. 12, plane mirror is set at right angles to the optic axis in the principal focus of a cylindrical lens. Between the lens and the eye a plate of unsilvered glass is placed at such an angle as to reflect the rays from some distant object through the lens on to the mirror, whence they are reflected once more through the lens into the eye. The unsilvered reflection is not necessary if the eye is held at some considerable distance from the lens. In such an arrangement the image is not enantiomorphic, the effect of the cylindrical lens-system being neutralised by that of the mirror, but the image rotates when the lens is rotated exactly as in an ordinary cylindrical telescope, and the plane of polarisation is not affected by the rotation. “On the Effect of a Magnetic Field on the Rate of Subsidence of Torsional Oscillations in Wires of Nickel and Iron, and the Changes Produced by Drawing and Annealing.” By Professor ANDREW GRAY, F.R.S., and ALEXANDER Woop, B.Sc. Received March 12,—Read March 17, 1904. On May 15th, 1902, we communicated to the Royal Society a paper entitled “On the Effect of a Longitudinal Magnetic Field on the Internal Viscosity of Nickel and Iron, as shown by Change of the Rate of Subsidence of Torsional Oscillations.” We described in that paper the results of experiments on the rates of subsidence of torsional oscil- lations in nickel and iron wires in fields of different strengths, and showed that the effect of the field, or, more properly, of the magnetisa- tion of the wire, is to increase the rate of subsidence in nickel and to diminish it in iron. In nickel, it was pointed out, this effect rose to a maximum at a certain field, from 100—180 C.G.S., according to the initial amplitude, and thereafter diminished as the field was increased ; while in iron the effect was in the main all produced at a field of about 160 C.G.S. or rather less, and increased only slightly with further increase of field intensity. The experiments described in the present paper are referred to at the end of the former one as in progress, and some account of their results is given; and we propose now to describe them a little more in detail. Experiments on the Effect of Drawing Down and of Annealing a Nickel Wive-—A piece of the nickel wire formerly experimented on was tested for subsidence in the manner already described, and then drawn down, by being passed through a draw plate, from the diameter 1:4 mm. to 0°775 mm. The results are illustrated by fig. 1. Take first curve I of that figure. It has for ordinates the differences between i le al Ig 1904. | Magnetic Field on Torsional Oscillations, ete. 287 the amplitudes at a given stage in the course of the subsidence with various fields, and the amplitudes at the same stage in the subsidence with zero field. The stage chosen was in each case the instant after the completion of the fortieth vibration after the attainment of the initial amplitude indicated on the curve. The abscisse are galvano- Hyco ut field and no-field Be 40 vibrations (e) ie») eEaING as iss CN AO -10 : O 5 z0 Te 20 Galv. deflections. meter defiections and are proportional to the fields: each unit of deflection means a field of 15-54 C.G.S. units. Curve I shows the effects of the various fields on the wire as it was received from the makers. The wire was then hardened by being drawn down as described ; and it was then found that no effect of magnetisation on subsidence was perceptible. The nickel wire was next annealed by heating it toa bright red heat LESSEE 288 Prof. A. Gray and Mr. A. Wood. Hfect of a [Mar. 12, and then plunging it into cold water; when it was found to be softer than it was originally. For the original wire, for example, the modulus of rigidity was 7°97x10 C.G.8., for the wire as drawn down it was 8:06x10 C.G.S., and for the annealed wire it was 7:81 x 10 C.G.S. Curve II is the curve for the annealed wire corresponding to curve I, that is, showing the differences after forty oscillations of the amplitudes of oscillation of the wire under the various fields and the amplitude with zero field. The latter amplitude it is to be remembered was the greater. After annealing and re-drawing had been performed in succession, the wire gave curve III, and its rigidity modulus was 8-318 x 10 C.G.S. These three curves make it clear that the effect of the magnetic field on the internal viscosity of a nickel wire depends very largely on the hardness of the metal, as, of course, we should expect. The progressive modification of the groups of molecular magnets by the magnetic field is opposed by the greater resistance introduced by the hardening, and the changes due to the field are not produced. In the annealed condition the wire has its groups much more at liberty to take up a new arrangement, and this is shown also, of course, by the smaller magnetic susceptibility in the hardened condition. Curves Ig, Ila, give the results as in I and II, but after twenty oscillations. It will be observed that while II, is above II up to and beyond the field of maximum difference of amplitude, and is then below IT, curve I, is below II throughout its entire course. Effect of Drawing Down on the Iron Wire.—This is shown in fig. 2, in which the ordinates and abscisse represent the same quantities as do the ordinates and abscisse of the curves in fig. 1, after twenty oscil- lations in each case. The amplitude for iron, it is to be remembered, is greater with field than with no field. The steep rise of curve I, and the horizontality of the remaining part, show the fact noted in the former paper, that the effect of the magnetisation in iron is attained with low fields, and that higher then produce little further effect. After the results shown in curve I had been obtained the wire was drawn through two holes of the draw plate, and after this treatment gave curve II. This curve is lower and rises less sharply than I, showing that the internal viscosity of the wire was not so much affected by the magnetic field as previously. The drawn wire was then annealed, and gave curve III, which does not differ much from I. Further drawing down resulted in curve IV, which is practically a straight line, that is, the difference is now nearly proportional to the field. Re-annealing of the now much thinner wire gave curve V, which shows a very distinctly greater effect of magneti- sation on viscosity than ever before. The repeated drawing and annealing process thus seems to result in the finally annealed wire in 1904. ] Magnetic Field on Torsional Oscillations, ete. 289 a collocation of molecules much more easily affected by a magnetic field. Liffect of Permanent and Temporary Strain in Nickel Wire.—This is illustrated on fig. 3, which is constructed on the same principle as figs. 1 and 2. Curve I describes the behaviour of the wire in its original state. After this curve had been obtained for the different fields, a weight (that of a large vibrator) of 11,202 grammes was hung Ny qu e for field and no-field after 20 vibrations. fo) Et fe) Difference of Amplitud cn Galv. deflections. on the wire and left for two days. At the end of that time curve II was obtained, showing that the permanent strain had increased the differences of amplitude remaining after the forty oscillations. The weight was left on for another 2 days, and curve III was observed. Two repetitions of the process gave curves IV and V, which practically coincide with III, and are not drawn. It is remarkable that permanent strain should have in nickel practically the same effect as annealing in increasing the effect of field - 290 Prof. A. Gray and Mr. A. Wood. Hfectofa ([Mar. 12, on the rate of subsidence, which effect it is to be remembered is in nickel an increase.* The wire was now allowed to rest for a few days without any weight o-field after 40 vilrations. [S HEX f/ | Difference of ee for field and n O Galv. deflections. upon it ; but it was found that the curve then obtained coincided with curves III, IV, and V, so that the rest seemed to produce no effect. The heavy vibrator was now hung on instead of the light one, which had been used in all the previous experiments with both wires. * See the former paper, loc. cit. 1904. ] Magnetic Field on Torsional Oscillations, etc. 291 Curve VII was obtained. The light vibrator and the heavy one were then used again in succession and gave curves VIII and IX. Probably the difference between curves VIII and IX was due to the additional longitudinal strain (in this case temporary) given to the wire by the additional load. The subsidences were taken for all the curves after forty vibrations no matter which vibrator was used. The difference between curves VII and IX, it will be noticed, is a twofold one; a much greater maximum of effect and a shifting of the maximum to a greater field. It seems not impossible that the main differences between nickel and iron disclosed by the experiments described in the previous paper, and the effects now discussed, may be explained by supposing that the groups of magnetic molecules in nickel lie in layers across the wire, separated by a matrix of conducting material which becomes hardened by drawing, and prevents the progressive changes in the groups from the initial condition of closed chains, which is brought about by magnetic force. On the other hand, the supposition would be that in iron the molecular magnets are in longitudinal groups with non- magnetic material between. ‘Thus, in the vibrating wire in the first case, the conducting material moving in the field of the molecules would give rise to dissipation of energy, and there would be a greater rate of subsidence in the field than without it. In the other case it is conceivable that the changes of the longitudinal and initially more or less nearly closed chains might result in such a modification of field as to result in less dissipation of energy by induction currents due to relative motion of the conducting substance and the molecular magnets, for the conducting matter between the longitudinal rows of elementary magnets may move in a feebler field after the magnetisation than before, owing to the breaking-up of the closed chains. The effect of permanent and temporary longitudinal strain in nickel seems contrary to what we should have expected. It is known that longitudinal strain on nickel diminishes its longitudinal magnetisation, and judging from this we should have rather expected the contrary effect to that which we have observed. ‘There is however, no doubt as to the result, which is borne out by many sets of observations. The cause must be matter for further consideration and experiment. (et Ce ee ae 292 Mr. R. W. K. Edwards. (Jan. 28, “A Radial Area-Scale.” By R. W. K. Epwarps, M.A. Com- municated by Professor A. G. GREENHILL, F.R.S. Received January 28,—Read March 3, 1904. This is a contrivance for finding the area of a plane figure by means of a transparency. The design in the transparency consists of a number of radiating lines. Hach of these lines is graduated. There are various patterns of this design, and their respective claims to convenience and accuracy form a wide field for discussion. In the accompanying transparency (reproduced in the figure), which is fairly simple and effective, there are eleven straight lines radiating from a point at equal angles. The way in which the transparency is used is as follows :— 7 The figure whose area is to be found is placed under the transparency, in close contact with it, so that its contour lies just between the two outside lines of the transparency, 2.¢., so that each outside line touches the contour, or passes through a cusp or angular point, or contains some rectilineal portionsof the contour. Hach of the radiating lines thus becomes a tangent or transversal, or contains a side of the figure. The graduations of the right-hand points of intersection of these transversals are read and added together ; then the graduations of the left-hand points of intersection are read and added together. The second sum is subtracted from the first; and the result records the number of square inches in the figure. It will be seen that if each of the outside lines touches, or passes through an angular point of the figure, there will be eighteen gradua- tions to be read—those on the outside lines cancelling each other. It one of the outside lines contains a rectilineal portion of the contour, there will be twenty graduations to read; if both outside lines do so, twenty-two graduations must be recorded. If a quicker use of the area-scale, with less chance of accuracy, is desired, the figure to be quadrated may be placed so as to lie just between the first and ninth, or the first and seventh, or the first and fifth lines, in which cases fourteen, or ten, or six graduations will be read respectively. If the figure is too large to be included between the outside lines, it may be divided into two parts by a straight line drawn across it, or into three parts by a pair of straight lines inclined to one another at the same angle as the outside lines, which may be done by means of the cardboard slip accompanying the diagram. The second of these methods of dividing up the figure may be also employed when it is desired to obviate the inaccuracy that may result from the two outside lines otherwise being tangent to the figure. A Radial Wn sSeale. 1904.] UA =e, a EG Zo Lo : AZIS LVN gpa ° fat “x O.¢ 6: ‘ * TE hue Ly d: HH gh te eta SA Gif el a a gee E226 Ieoe ez g Me 12 deen C2 a 3 cea i / mae meee Woec! ei Al of Gi vi ci a om ONG ated E oo panne Se ig { wy Gi VOL. LXXIII. 294 Mr. R. W. K. Edwards. [ Jan. 28, Principle of the Area-Scale. Suppose the equation of a curve to be 2 12 = a +004 662+ d63, the area between the curve, the initial line, and the radius vector a 26 making angle 26 with the initial line,is$ | 17? d0; 22, 0 ab + 2b62 + $c63 + 4d64, 7 ¢., S, say. Now if 70, 72 be the outside radii of the sector so quadrated, and 7, be the radius vector bisecting the angle between them, it can be seen that Ry = Gh re? = 44+604¢6P?2+d6, 192 = 4+ 2604 46? + 8d@ ; whence, assuming (A79?+ Br\? + Cr2*) @=8, and solving three of the four simultaneous equations for A, b, C, we get values A=4, B=2, C=, which also satisfy the fourth. We have then S = 16 (7? + 47? +197). The design on this particular transparency is made on the assump- tion that each of the ten separate portions of the curve between the first and third, the third and fifth, . . . the ninth and eleventh radiating lines, approximates to some member of the above family of curves. Anyone acquainted with the discussion of areas, in Cartesian co- ordinates, in terms of a series of equidistant ordinates, and their common distance apart, given in Bertrand’s Calcul Intégral, Section 363, will see an analogy in the foregoing expression of the area of a sector in terms of the squares of three equally inclined radi and their common inclination. Other designs of the transparency can be made on the assumption of larger portions of the curve approximating to members of families of curves in which r? is equal to a rational algebraical function of 6 in ascending powers of 0, of degree higher than the third. On the whole it is easier, and possibly more effective, to use the method employed above, in the same way as ‘ Simpson’s Rule” uses the corresponding theorem in Cartesian co-ordinates. There is no reason to suppose that “Simpson’s Rule” gives a less accurate approximation in the generality of cases than ‘‘ Weddle’s Rule” * or the numerous rules deducible in terms of the co-efficients calculated by Cotes. An inyestigation of the family of curves r = a+b04c6? + de * See Boole’s ‘ Finite Differences.’ 1904.] A Radial Area-Scale. < 295 shows that they form a large variety of spirals. An important point to notice is that if, as usual, ¢ denotes the inclination of a radius to the tangent at its extremity, we have tan 4, 7.¢., rd6/dr, equal to 2 (a+b6 + cP? + dé?) 42043002” which, in general, cannot be zero unless at the origin, when the curve passes through the origin. This means that error is likely to be perceptible when the curve quadrated is such that any tangent to it passes through the point of radiation ; ¢.9., when the outside lines are tangents to the curve; or when the curve is re-entrant in such a way that any one of the radiating lines gives three or four readings. The design may be expected to give the best results, therefore, when arranged so that the outside lines past through cusps or angular points of the curve, and so that no radiating line crosses it more than twice. For curves in which there are no sharp points, it is best, therefore, to divide into two or three areas. . There is no reason why the presence of points of contrary flexure should be supposed to vitiate the results. For the equation for 0 giving the positions at which such points occur in the stated ea of curves, will be found to be, in general, of the sixth degree. Calculable deviations from strict accuracy may be expected in the case of nearly all the well-known regular curves quadrated by means of the area-scale. If, as in the present pattern, the angle between the outside lines is half a radian, these deviations will a found to be insignificant, except in the cases of oval curves touching the outside lines and not treated as suggested on p. 292. 296 Mr. J. Y. Buchanan. (Petra “On the Compressibility of Solids.” By J. Y. BUCHANAN, id Jeet Received February 11,—Read February 25, 1904. The solids dealt with in this research are the metals platinum, gold, copper, aluminium, and magnesium. Their absolute linear compressi- bilities were directly determined at pressures of from 200—300 atmospheres at temperatures between 7 and 11° C. The determina- tions were made by the same method, and with the same instrument which I used for the determination of the compressibility of glass in 1880.* As nearly a quarter of a century has passed since then it will Ibe expedient to recall the principal features of the instrument, and of ‘the method. The idea of it occurred to me on the evening of March 23, 1875, the ‘day on which the ‘“ Challenger” made her deepest sounding, namely, 4475 fathoms (8055 metres), and I was able to put it in practice ‘6 days later, on March 29, when, however, the depth was only 2450 fathoms (4410 metres). The observations which I was making during the voyage on the compressibility of water, sea-water, and mercury, were of little value without a knowledge of the compressi- bility of the envelope which contained them. It was a matter to which I had given much thought. I had studied all the methods which had been used up to that date, but they had all turned out to be faulty. The idea of utilising the linear compressibility of glass in order to arrive at its cubic compressibility had occurred to me, as it had, no doubt, occurred to many others, before. The difficulty lay im giving the idea experimental expression. It was clear that the instrument would fall to be classed as a piezometer, and would have to be a self- registering one, because what takes place in the depths of the sea is removed from observation. All my piezometers contained a liquid, and this I had recognised to be fatal to absolute measurements. The problem had, therefore, come to be: to design a piezometer which should contain no liquid; and it was the solution of this problem which eceurred to me on the evening of March 23, 1875. The form which the instrument took was very simple. In my laboratory outfit I had included some lengths of tubing suitable for the stems of piezometers, of which I had to make a number during the voyage. In order to be able to use the indices of broken deep-sea thermometers, the tubes had the same internal diameter as the stems of these instruments, about 1 mm. On the outside of the tubes a scale of millimetres was etched. I took the greatest available length of this tube, namely 60 cms. I then drew out a wire of the same glass and passed it into the tube until it appeared at the other end of * *Roy. Soc. Edin. Trans.,’ vol. 39, p. 589. 1904. | On the Compressibility of Solids. 2977 the tube, ‘This end of the tube was then sealed up, and the end of the glass wire was fused into it, so that, while free throughout its whole length, longitudinal motion was prevented. The length of the glass wire was 57 cm., so that there was an empty space in the tube of 3 cm. above it. The magnetic index of a broken deep-sea thermometer was re-haired and passed into the tube above the glass wire. The open end of the tube was then sealed up. The result was a piezometer consisting of nothing but glass. In principle it was precisely the same as any of the other piezometers. The indices of these give the difference between the compression produced by the pressure on the contents and on the envelope. In the case of the other piezometers, which contained liquids, the balance was on the side of the contents. In the all-glass piezometers the contents, besides being of the same material as the envelope, were completely protected from pressure, and the whole of the change of length measured fell to the envelope. It has, therefore, a feature which is possessed by no other instrument ; withit the absolute compressibility of a solid is determined by one measurement. Before the instrument was attached to the sounding line, the index was brought down by means of a magnet to rest on the end of the internal glass wire, exactly in the same way as if it had been the mercury column in a maximum and nunimunm thermometer. The instrument was then sent to the bottom, or to whatever depth might be decided on. During the descent the temperature of the glass, both inside and outside, fell with that of the water through which it passed, but as the contraction produced was the same on the wire and on the tube, there was no differential effect to be recorded by the index. On the other hand, the increasing pressure, as the instrument descended, affected only the outside tube, which it shortened. In contracting, it was obliged to pass the index, which was kept in its place by the internal wire. When the instrument was being hove up, the reverse process took place; the tube lengthened, and lifted the index clear of the internal wire by an amount equal to the lengthening of that portion of the tube. As the whole clearance produced by the expansion from the greatest depth did not exceed 1 mm., its amount had to be estimated by the eye with the assistance of a magnifying glass. The experiment made on March 29, 1875, was quite successful, and it gave 3°74 as the cubic compressibility per million per atmosphere, of the glass of which the tube was made. The exact figure found in 1880 for glass from the same source was 2°92. A number of observa- tions were made with the instrument, both on the sounding line and in the compression apparatus with which the ship was supplied, and figures from 3—5 per million were found. These were sufficient to give the order of the constant which was sought, but it was impossible = 298 Mr. J. Y. Buchanan. [ Feb.Abw, with the appliances at hand to measure such small distances with sufficient accuracy to enable a definite value to be determined. On the return of the ship I embodied the principle in an instrument of precision, which I had constructed in the early part of 1880, and I used it in the month of June of that year for the exact determination of the compressibility of the glass which had been used in the construction of my “ Challenger” piezometers. It is this instrument, and without any alteration, which I have used for the purpose of the present research. Bigs: With the assistance of fig. 1, its features, and the distribution of its parts, will be apparent without any lengthy description. It consists of three parts: the force pump on the left, the receiver for the reception of piezometers or other bodies on the right, and behind these, the block with tubes projecting on either side to receive the rod or wire of the solid, the compressibility of which is to be determined. Lvery part of the instrument is made of steel. The part which most concerns the present research is the steel block, in the rear, with tubular prolongations. When the rod or wire to be experimented with has been introduced, the ends of the tubes are closed with thick glass tubes, which are kept in their places by open steel caps. Each of these tubes is commanded bya microscope with micrometer eye-piece. In 1880, when the instrument was housed in a room with a stone floor, these microscopes stood on three-legged stools, as shown in the figure. As the room with the stone floor was no longer available, I had to instal the instrument close to the windows of the laboratory, which has a wooden floor, and fix metal brackets in the wall to carry the microscopes. In both cases the micrometers, which measure the expansion or contraction of the body under examination, are independent of the instrument which holds it. The manometer, which indicates the pressure in the instrument, is - or s 1904. | On the Compressibility of Solids. 299 seen under the steel block which carries the tubes. It is simply a mercurial thermometer with a very thick bulb. The scale on it is an arbitrary one, and its value as a measure of pressure is fixed by observing its reading when the principal piezometer which I used during the voyage of the “Challenger” was in the receiver. This piezometer, known as C No. 1, contained distilled water, and from very many carefully executed experiments at depths from 800 fathoms (1440 metres) up to 2500 fathoms (4500 metres), made in the South - Pacific where the oceanic conditions were most favourable, the apparent compression of distilled water in this particular instrument at the temperature ruling in these depths, which averages in round figures 2° C., and when exposed to measured columns of sea-water, of known quality as regards density, was accurately known. ‘The indications of the manometer are, therefore, equivalent to those of piezometer C No. 1, the standardisation of which was effected under an open-air water column. The observations made with C No. 1 on board the “Challenger,” which form the basis of the scale of pressures, are collected in Table I. They are expressed in terms of the apparent compressibility of distilled water deduced from them. In the table the vertical lines represent apparent compressibility in volumes per million per atmosphere, rising by steps of 1 per million from 45—55, so that all the values of the compressibility falling between say 45 and 46, or 49 and 50 are arranged in one column. Above each entry of apparent compressibility will be found the depth in metres to which the instrument was sent, and the temperature (° C.) of the sea-water at that depth. The depth is expressed in metres because it so happens that the average density of the water in this part of the South Pacific, allowance being made for the vertical distribution of temperature, compression, and salinity, is such that a vertical column of it 10 metres high exercises very exactly the same pressure as 760 mm. of mercury. So that the depth in metres, divided by 10, gives the pressure in ordinary atmospheres. At great depths a very slight correction has to be made; the nature of this will be apparent from the following table, in which, for different depths in metres, D, the pressure P in atmospheres is given :— 1 ae 1400 2000 3000 4000 5000 6000 Bacco 139°96 200°14 300-82 401-98 503-62 605°75 Owing to the preponderance of water of low temperature and of very uniform salinity in a vertical column of water in any part of the open ocean, the pressure exercised by it per thousand metres does not differ appreciably from 100 atmospheres. Inspection of Table I shows at once in which column the true value of the apparent compressibility is most likely to be found. It is the one which includes values between 49 and 50. Outside of this 300 ° Mr. J. Y. Buchanan. [Feb. 11, column it is only the adjacent one containing values between 48 and 49 which enters into competition. The mean of all the values in these two columns is 49-16, and this figure forms the basis of the measurement of pressure in this investigation, and it is used in interpreting the pressure-value of the readings of Piezometer C. No. 1, when being compared with those of the manometer used for the ordinary measure- ments of the pressure in the apparatus. The change of apparent compressibility of water with change of temperature for the small range of temperatures with which we are concerned was found in 1880 to be at the rate of 0°33 per degree (Celsius), and this figure is used in the present research. Micrometers.—The same microscopes and micrometers, which served in 1880, were again used in this research. ‘Their value was determined by reference to a stage micrometer, ruled into hundredths and thousandths of an inch. This was then verified at the National Physical Laboratory. The changes of length measured by the micrometers are therefore given in terms of the standard inch ; and, it may be added, the values attached to the readings of the micro- meters in 1880 were exactly the same as those now found by reference to the standard of the National Physical Laboratory. In the microscope which was always placed on the left hand, one division was equivalent, on the stage, to 0°0004219 inch. In the one on the right hand one division was equivalent to 0°0004167 inch on the stage. As the contractions or expansions are given directly in terms of the inch, the total length of the rod is given in inches also. In order to bring the ends into a suitable position for observation with the microscopes the length of the rod or wire had to be not less than 75 or greater than 75°5 inches. The actual lengths were measured exactly in each case. The average was 75°32 inches (1:915 metres). To facilitate the observation of the ends through the thick glass tube a piece of microscopic covering glass was moistened with a drop of water and laid horizontally on the tube, producing the same effect as if a flat surface had been ground and polished on it. The effect observed and measured is the lengthening of the rod when the pressure is relieved. As the compressibility of solids is very small, the highest pressures have been used which were found to be compatible with the reasonable persistence of the glass terminals ; the usual pressure was in the neighbourhood of 200 atmospheres. Very few of the glass terminals stood over 300 atmospheres. The pressures actually chosen were as nearly as possible those at which the manometer had been compared with the “‘ Challenger ” piezometer. The body under observation is in the form either of a rod ora wire. Ifitis in the form of a rod then it is fitted with wire ends ot sufficiently small calibre to enable them to enter the glass terminals. | Table I G3 C ord the rig] | { i j | | Table [.—The Apparent Compressibility of Water in Glass, as observed in Piezometer C. No. 1, on board H.M.S. “Challenger,” in the South Pacific, in October, November, and December, 1875. The observations are arranged in order of apparent compressibility, which is expressed in volumes per million per atmosphere. Above each entry of the apparent compressibility is, on the left, the depth, in metres, of the water at the particular station, and, on the right, its temperature, in Celsius’s degrees. Apparent compressibility per million per atmosphere. | | ; | = oY, Apparent als 46 47 4lg 4l9 5\0 51 52 5/3 54 compressibility. | | et a | | ena a 2475 1°°9| 1800 2°6| 1890 2°8 | 4140 1°6 | 1620 2%4 | 1800 2°°4| 2700 1°°9 | 1800 31 | 2880 1°8 45 0 46°8 AT 1 | 48-7 49 °6 50 °6 51°0 53 “4 55°5 2385 2°21/3195 17%9 | 3330 1°%6 | 1620 2% | 1800 2%4| 2700 273 46 °8 47 3 48-1 49 °5 50°8 51°7 005 270 | 4050 1° 2880 1°8 | 1800 2°3 4 4050 1°4 47-8 48:0 49 °8 50-1 2610 2°09 | 2700 270 | 2520 272 47-0 48 °9 49 +4; 1620 370 | 3645 13 48 °8 49 24700 2°0 | 1440 372 48 °4. 49 -65 | 2700 2°70 | 4085 14 48 -15 49 *2 24700 178 49°75 3285 1°°8 | 49-1 3195 1°95 49 « 1800 2 °7 49 °8 °. 3240 2 °2 49-7 1800 28 49-0 Tae ae aS 800 2%] 2700 271 | 1800 3°1 | 2880 1°8 z o. 224 | 2925 27°18 | 3035 1795 | 2540 2 17)! Boek ; “s i Mean values Hy ee i 9 Pars 8 an Hoe 48° 49°48 | 50 °5 51°35 | 53 °4 55°5 ee 49°16 2°1 | a est ay Fa earn y 3; 3 2 Olea a 1 4 7 13 | Number of ob- " 2 -——_-— S| | servations 20 | 1904. | On the Compressibility of Solids. 301 During an experiment with a rod it contracts while the pressure is being raised, and expands again when the pressure is relieved. The steel tube which holds it, however, acts in the opposite sense, it expands while the pressure rises and contracts while it falls. If the two surfaces were perfectly smooth, one half of the change of length would be measured at the one end and the other half at the other end. As the surfaces are- not perfectly smooth, this does not usually occur. Moreover the steel tubes are prolongations of the central steel block which holds them. The block is bored with holes at right angles to each other in the three principal directions. Consequently for a distance of about an inch and a half in passing through the block the rod is not supported at all. With the exception of this small portion, however, the rod is supported throughout the whole of its length by the steel tube. Now, although it is thus nominally supported equally throughout the whole of its length, we know that in reality this is pretty certain not to be the case. At some place, either in the right arm or in the left arm of the apparatus, the rod is sure to bear more heavily than in any other part. The contraction under pressure and the expansion under relief of pressure will then apparently take place as trom this point as origin. Supposing this point itself to be motionless, it is evident that the change of length measured at the two ends will be in the same proportion to each other as would be the arcs which they would describe if the rod were a lever oscillating on the point as a fulerum. As there is no support at all at the centre, this point must lie on one side or on the other of it.and the motions of the ends must be unequal. But the fixed point of the tubular receiver is the central block; therefore any point in, let us say, the right-hand tube will, when pressure is being raised, move to the right, and, on relief of pressure, retreat by an equal amount to the left. Consequently when we observe and measure the change of position of, for instance, the right-hand extremity of the rod, when the pressure is relieved, that change of position is composed of two motions, the expansion of the part of the rod which lies between the right-hand extremity and the point in it whose motion with respect to the steel carrying tube is nul, along with the proper motion of that point. Similarly, when we measure the change of position of the left-hand end, it also is composed of two parts, the expansion of the part of the rod which lies between the left-hand extremity and the same point in the length of the rod where its motion with respect to the steel tube is nal, along with the proper motion of that point. But at the left-hand end the motion of expansion is to the left, and at the right-hand end it is to the right, while the proper motion of the position of the common point on the rod and on the: tube is always in one direction, and in this case, to the left. Therefore the distance measured in the right-hand microscope is the expansion of the portion of the rod which lies to the right ot te uke eek 302 Mr. J. Y. Buchanan. [Feb. 11, the point on it which is motionless relatively to the tube minus the proper motion of this point: and the distance measured at the left- hand end is the expansion of the remainder of the rod plus the proper motion of the common point. Consequently the algebraic sum of the two motions measured is the expansion of the rod under the relief of pressure. When the substance is used in the form of a rod, as, for instance, in the case of glass, its ends are drawn out into wires, such that they can enter and be visible in the glass terminals. What we really measure then is the change of length under change of pressure of the axial glass wire in the rod, which may be looked on as a fascine of a very large number of similar but somewhat shorter wires. The sole function of these other wires is to maintain the wire that falls under observation in an axial position. It is obvious that this function can be performed with equal efficiency by wires of any other material, and that the conditions are in no way altered if these are fused into a tube of which the wire to be measured may be regarded as the core. Consequently by my method the linear compressibility of a solid can be determined as well on a wire as on a rod; and there is no limit to the thinness of the wire, so long as it can be handled, and be perceived in the microscope. These two conditions are, in a way, antagonistic, because for the microscope the finest possible point is desirable, while for the handling of the wire a sensible thickness is essential. Only in the case of glass can a good working compromise be effected, because the wire which enters the glass terminal can be drawn out at the end to the finest possible hair, and the end of the hair can be fused into the minutest possible sphere, which can then be observed in the microscope with the sharpness with which a barometer can be read with a good telescope. When the substance under observation is in the form of a wire, it lies in a glass tube which fits the bore of the steel tube as closely as possible. Its bore is a very little larger than that of the glass terminals, or about 1 mm. ‘This tube acts as a bearer, and its length is as nearly as possible equal to the distance which separates the inside ends of the glass terminals when in position. When the pressure in the apparatus is raised, both the wire and the glass tube which carries it are shortened, while the steel tube which carries both of them is lengthened, and when the pressure is relieved the reverse takes place. The glass tube behaves exactly like the glass rod, that is, it is able to a slight motion of translation. Similarly, the wire, which is carried by the glass tube, generally expands and contracts under pressure at a less rate than does the glass, producing again a slight apparent motion of translation. But again, as in the case of the rod, the algebraic sum of the observed motions is the expansion or contraction of the wire. 1904. | On the Compressibility of Solids. 303 There is an advantage in having a very slight leak in the apparatus. The routine of an observation 3s then that the observer in charge of the pump and the manometer gets the pressure up somewhat higher than that desired; he then settles himself with the relieving lever in his hand and calls out as the mercury in the manometer in falling passes each division. The observers at the microscopes read their micrometers at the same moment. When the pressure has fallen a little below the desired pressure, the pressure is very carefully relieved, and the readings of the micrometers and of the manometer are taken at atmospheric pressure. The algebraic sum of the movements of the two ends on the micrometers gives the linear expansion of the body which has taken place, and the difference of the two readings of the manometer gives, when interpreted by the help of Piezometer C. No. 1, the difference of pressure which has caused the expansion. The micro- meter measurements are then reduced separately to their absolute values in terms of the inch. The algebraic sum then gives the linear expansion in terms of the inch. It is then divided by the length oi the rod or wire in inches and by the pressure in atmospheres; the resulting quotient is the linear compressibility of the metal or other substance. Multiplying this by three, we obtain the cubic com- pressibility of the substance, if truly isotropic. Tt will be evident that, to work with this instrument, three observers are necessary, namely, one for each microscope, and one to raise and relieve the pressure and observe the manometer. I was fortunate in being assisted during this investigation by Mr. Andrew King, who was formerly my regular assistant, and is now of the Heriot-Watt College, Edinburgh, and by Mr. J. Reid, Demonstrator in the chemical laboratory of that institution. These gentlemen gave up their Christmas vacation for this work, and I owe them a deep debt ot gratitude for the willingness and the efficiency of their help. The metals experimented with have been used in the form of wire, and the size chosen was No. 22 of the standard wire gauge (8.W.G.). In the case of aluminium, however, the size was No. 20. The dimensions corresponding to these numbers are given in the following table :— 25 Sectional | Leneth of | No. of wire. Diameter of wire. area of | ies i i wire. , a | SW Ge inch. mm. : sq. mm, | metre. 29 0 036 0-914. | U *656 1-824 | | 4 | a fh | 22 | 0.028 | 0-711: | O:39%6 4 2-519 The degree in which the actual wires corresponded with the tabular specification was checked by weighing measured lengths of them. 304 Mr. J. Y. Buchanan. [Petri The- weight of 1 metre of each wire was as follows :—Platinum, 8:156 grammes; gold, 7320 grammes; copper, 3°375 grammes ; aluminium (No. 20), 1642 grammes; and magnesium, 0°552 gramme. : Neglecting the magnesium which, being pressed and not drawn, is very uneven in its calibre, these figures show that the actual wires were very slightly smaller than they should be by the gauge. Thus, in the case of the platinum wire | c.c. occupies 2°636 metres (lineal) instead of 2°519 metres as by the table. The platinum and gold wires were pure specimens obtained from Messrs. Johnson and Matthey in the year 1880. The copper was “high conductivity” copper, and it as well as the aluminium and magnesium wires were of the best quality obtainable at the present day. The platinum and gold wires were heated to redness over a Bunsen lamp before use, so that they were thoroughly annealed. The aluminium wire was also heated, though to a much lower temperature, so as to soften it. The other metals were used in the state in which they were supplied. All the wires were straightened, but not stretched, before use. The temperature of the wires during the operations was always that of the laboratory, and every care was used to keep it as uniform as _ possible, and it was as nearly as possible that of the air outside. Working in the middle of winter and in a comparatively high latitude, I hoped to be able to do so in conditions which, as regards temperature, would be similar to those which obtain in the depths of the sea, but the extraordinary mildness of the weather this year made it impossible, and the temperatures fell, mostly between 9° and 11° C. The results of the investigation are set forth in detail in Tables IL to VI, and they are summarised in Table VII. Table Il.—Platinum. Date, January 9, 1904. Temperature, 7 C. Wire No. 228.W.G. Length, 75°35 inches. Gen ges Oe Longe. Compression Linear ar. | Pressure, per million, | compressi- “i 8 bilit Rightarm,| Left arm, Sum, =) | 0h ae g/P ae r. l. r+l=8. 75°35 ; atm. in. in. in. il 204 | 0:008750 | —0°000844 | 0 002906 38 °57 0°188 2 204: 0 -003750 | —0-000970 | 0002780 36°88 0 °180 3 204 | 0:003750 | —0-000928 | 0002822 37°45 O -184: 4 300 0 005292 | —0:001181 | 0004111 54°56 0 182 1904. | | On the Compressibility of Solids. 305 Table III.—Gold. Date, January 10, 1904. Temperature, 10°°6 C. Wire No. 22 8.W.G. Length, 75:4 inches. Ohamzearobylens Compression Linear No. Pressure, = ? .per million, | compressi- | | No. ai oe Boy oop hl bility, | ightarm,| Leftarm, Sum, 10s S: gp — r. Me Nett = s 75-4 SS ds | i atm. | in. in. in. | 1 231°5 | 0:905208 | —0-000591 | 0-004617 | 61°21 | 0-264 2 230°0 | 0:005208 | —0-000970 | 0 004288 | 56 *20 0 °24:4 3 230°0 | 0:005417 | —0-000844 | 0°004573 | 60 °64: 0 °264 4, 204:0 | 0:005000 | —0-000590 | 0 -004410 | 58 °48 0-287 5 | 247-0 0 -005834 | —0-001181 | 0 °004653 | 61°70 O 254 6 230°0 | 0:005208 | —0-001094 | 0°004114 | 54°55 0-237 7 288°5 | 0:005000 | —0-000422 | 0 -004575 | 60 “70 0 255 8 238 "3 0 005000 | —0-000422 | 0 °004578 | 60 °70 0 °255 9} 238°5 0 :005208 | —0-°0006383 | 0 -004575 | 60°66 0 °254 10 | 273°5 0 006042 | —0:000548 | 0 -:005494. | 72 °85 0 °264 11 273°5 | 0:0058384 | —0-000211 | 0 -005623 74-56 0°273 12 | 273°5 0 006042 | —0 000591 | 0°005451 | 72°28 0 °264: 18 273 °5 0 006042 | —0 000464 | 0 °005578 | 73°96 0:270 14 273 °5 0 -006042 | —0-000717 | 0 -005325 70 61 0 258 15 | 273°5 0:°006250 | —0°000717 | 0°005533 | 73°37 0 °261 16 . 273°5 | 0:006042 | —0-°000506 | 0 005536 73 AL 0-268 | 17 269 °O 0 ‘005834 | —0 000422 | 0°005412 | 71-76 0-267 | 18 273°5 | 0:°006042 | —0-000548 | 0:°005494 | 12°85 O -266 | | 0-260 | Table 1V.—Copper. Date, January 9, 1904. Temperature 10° C. Wire No. 228.W.G. Length, 75:3 inches. Chace oe lenecn: | Compression Linear No. Pressure, ~| per million, compressi- FE. Rightarm,| Left arm, Sum, 1@3 2. = Gi, pe P 1. fn 2S 753 S/P = 2. | atm. in. in. | in. i 195 °5 0 005664 | —0 -001687 | 0:008980 52°85 0-270 2 230 °U 0 :006334 | —0:001687 | 0 -004647 61°70 0-268 3 195 °5 0005875 | —0:001814 | 0 -004061 53°93 0°276 A 195 °5 0 :006125 | —0°001772 | 0-004353 57 ‘81 0 °296 5 247 -O 0 :007417 | —0:001856 | 0-005561 73°85 0 :299 6 230 :0 0 °006750 | —0-001434 | 0:005316 70 °60 0 °307 7 282 ‘5 0 :007751 | —0-°001519 | 0 006232 82°76 Q 293 0°288 306 Table V.—Aluminium. Wire No. 20 S.W.G. My. J. Y. Buchanan. Date, January 11, 1904. [Feb. 11, Temperature, 9° C. Length, 75°35 inches. | Ubonpee engi: Compression Linear Pressure, | per million. | compressi- AEA aaa? | s bilit ; Right arm,| Left arm, Sum, 12 a= g/P ey) ie i, | pet Sel) aes | (Pa | atm. in. in. in. Fan 195 °5 0 005542 0 002616 0 008158 114 °22 0 584 4 161 °5 0 -604900 0 °002152 0007052 93 58 0-579 3 230 °0 0 006667 0 -002742 0 -009409 124-86 0 °543 4. 178°5 0 °005334: 0 -002109 0 007443 reer 0 °553, 5 256°0 | 0:007251 0 002995 0:9010246 135 -96 | 0°5381 | | | 0 °558 Table V1I.—Magnesium. Wire No. 228.W.G. Length, 75-2 inches. i i Date, January 17, 1904. Temperature, 9° C- i Cpevecesol length. | Compression Linear _. | Pressure, per million, | compressi- Eo eee | s bilit ; Right arm,| Left arm, Sum, 106 _§_ = 9, Jy r. 1. rt+l=S8. 75:2 S/P =a. | atm. in. in. in. u 204: 0-009167 | 0:°C07120 | 0°016288 216 61 1-062 2 204: 0010001 0 006202 | 0:016203 215 48 1056 3 204: 0 009917 0-006829 | 0°016246 216 07 1-059 4 204: 0:010418 | 0 005991 | 0°016409 215 °57 1-057 5 204 0010543 | 0005442 | 0 °015985 212-60 1 042 6 204: 0 °011251 0 004852 0 -0161038 214 16 1 -050 1-054 | a: | | Table VII.—Summary. Compressibility. Substance. Year. ane Density. at Linear. Cubic. Plata hy. 2. = 1904: 194, 21°5 0 °1835 0 °5505 Glolil 5 a3 a soe oe 55 197 i) a3 0 260 0-780 Copper . 2 63 8 °9 0 °288 0 864 Aluminium .. mi 27 2°6 0 °558 0 °1674 Magnesium..... i DA Ws 1 054 3 :162 Mercury -....... 1875 200 13°6 1°33 3-99 Glass, AMIN for) ariavte oe 1880 ole ahs r@) 973 2, -92 - Oe scat 1904 2 968 1°02 3°06 » German.. S 2 494 0 846 2°54 1904. On the Compressibility of Solids. 307 In the summary, Table VII, the compressibilities of English flint glass and of the glass of which ordinary German tubing is made as well as that of mercury have been included for purposes of comparison. The compressibility of mercury rests upon a large number of observa- tions made in the “Challenger,’* by which its apparent cubic com- pressibility was found to be 1°5 per million per atmosphere. The piezometers which were used for this purpose were made by myself on board. The divided stems were of lead glass, because I had no other, and the bulbs or reservoirs, which had a capacity of about 20 c.c., were made of German glass, for the same reason. I have, therefore, applied to the values then found for the apparent compressibility of mercury, the value of the absolute compressibility of German glass found in January of this year, and the result is that the absolute cubic compressibility of - mercury at temperatures between 1° and 3° C. is 3°99. With regard to the metals quoted in the tables, the figures speak for themselves. The number of different metals is very small and, until the investigation has been extended so as to include at least the greater number of the metals which can be easily procured in the form of rod or wire, it is not likely that any very general features or laws will be apparent. It will, however, be observed that in the case of the five metals used as wire, their compressibility increases as their density and atomic weight diminish, yet there is no reason to suppose that the compressibility is a continuous function of the atomic weight, like the specific heat. Mercury, although in the fused state, shows this clearly. But besides this, it happens that two pairs out of the five metals, namely, platinum-gold and aluminium-magnesium, are contiguous in the atomic weight: series, yet the compressibility of magnesium is, roughly, double that of aluminium, and the compressibility of gold is half as much again as that of platinum. If, however, we compare gold and copper, which occupy parallel positions in Mendeléiefi’s scheme, we see that they are very much alike, and the same holds with regard to magnesium and mercury which occupy a homologous position. It these facts indicate anything more general, we should expect the metals of the palladium and the iron groups to have a low compressi- bility like platinum, zinc and cadmium to havea very high compressibility like magnesium, and thallium an intermediate but still considerable compressibility like aluminium. It will be observed that the two kinds of glass mentioned in Table VII are more compressible the greater their density. This may, however, be due to a specific feature of the oxide of lead which enters largely into the composition of the flint glass. It is obvious that there is here a great field for interesting research, and fortunately the method is capable of great refinement; only, the successful application of it requires considerable manipulative skill, * “Chem. Soc. Jour.’ (1878), vol. 33, p. 458. 308 Mr. J. Y. Buchanan. | [Feb. 11, as well as great patience. The necessity to have, as part of the apparatus, two glass tubes which are exposed to the high pressure on the inside only, introduces an element of chance into the work which is sometimes annoying and sometimes exciting. It is im- possible to say beforehand whether a particular glass terminal will stand or not. It is necessary to be provided with a large reserve of them before beginning work, and when one fails another is put in its place without loss of time. Hitherto I have taken no particular care of my glass terminals, because I can always depend on finding plenty of them which will stand from 200—300 atmospheres, and there is abundance of work to be done at these pressures. When, however, it is desired to use higher pressures, it will be prudent to take some measures for preventing the points of the wires scratching the internal _ surfaces of the terminals. When some precaution of this kind has been taken, casualties will be less frequent, and the attainment of higher pressures will be merely a question of how many glass ends the observer is prepared to sacrifice in the service. In the work connected with this paper, which extended over the greater part of 4 weeks, fifteen glass terminals gave way ; and oddly enough, the failures were as nearly as possible equally distributed between the two ends ; eight of them fell to the left arm and seven of them to the right arm. ‘The bursting of a terminal causes no incon- venience beyond the trouble of replacing it, because the construction of the instrument enables air to be completely excluded from it, and the quantity of water in it to be kept within such limits that its resilience is of no account. When a tube bursts it usually splits longitudinally up the middle into two slabs. One of these almost always remains entire, the other is sometimes broken into fragments, but there is never any projection of material unless the instrument has been carelessly put together and air admitted. Microseismic Effects—In a research like the present where the primary object is the numerical determination of a physical constant, the secondary phenomena which reveal themselves are often of equal and sometimes of greater interest, because they generally affect pre- ferentially the natural history side of physics. To this class belong the phenomena observed in connection with the behaviour of ice under the relief of high pressure in my earlier investi- gation.* In the present case the frequent bursting of the glass terminals afforded the opportunity of observing another and very interesting phenomenon. It is illustrated in fig. 2. It was first noticed when copper wire was being experimented with. The pressure had been raised to 300 atmospheres, and had begun to fall when the tube gave way. On proceeding to replace the broken tube with another I was astonished to find the copper wire twisted into a regular spiral * *Roy. Soc. Edin. Trans.,’ vol. 29, p. 598. 1904.] On the Compressibility of Solids. 309 in the tube. It made three complete turns in | Fre. 2. the length of an inch, and the undulatory form was visible throughout one-half of the length of the wire. Instead of fixing new glass terminals, I cut off the end of the copper wire, which showed this curious seismic effect, and put another wire in its place. An exactly similar effect was produced on the magnesium wire, when a glass terminal burst; only the effect was even more marked. The spiral produced in the glass end was closer, and, indeed, the wire had been shoved over itself and broken, for magnesium wire is very brittle. The undula- tions of greater amplitude extended through the whole length of the wire, and there were maxima at distances of about 35 cm. and 85 cm. from the seat of the explosion. The bursting pressure in this case was no more than 150 atmospheres, yet the effect produced was very much greater than it was in the case of copper. The experiments with gold and aluminium were carried out without the loss of a terminal. In the case of platinum, a terminal burst at about 250 atmospheres, but it produced no apparent seismic effect. On the last day of my experiments I proposed to determine the com- pressibility of a wire of mild steel, but, owing to hurry in putting the apparatus together, it was impossible to get any satisfactory observa- tions, but one of the terminals burst, and ata pressure over 250 atmospheres. Here again there was no seismic effect. The platinum wire had been thoroughly annealed before being used, and the mild steel wire was as soft and ductile as copper, yet, though copper showed the seismic effect beautifully, it was imper- ceptible in both platinum and steel. Before the experiment with the steel, I supposed that the high density of platinum caused the shock to be opposed by more inertia than it could overcome, but the density of steel is less than that of copper, therefore its immunity to shock must be due to something other than its density. The open ends of the glass terminals which are inside of the water- tight collars are cut sharply off and the edges are not rounded in the VOL. LXXIII. Zi 310 Dr. W. M. Bayliss and Prof. E. H. Starling. [Mar. 21, flame. Special directions were given to the glass blower about this, because the effect of it would be the production of considerable tension in that part of the glass. Notwithstanding my directions, some of the tubes were rounded off in the lamp and the effect was as I had foreseen. The only one of these ends which I used burst. In the case of ends which have been cut off and not heated, the fracture is confined to the part of the tube outside the apparatus. In the case of the end with rounded edges the outside part was fractured in the ordinary way, and in addition the rounded portion, which was exposed to no difference of pressure, exploded out of sympathy, much after the fashion of a Prince Rupert’s drop. I am continuing this investigation, and I hope shortly to be in a position to be able to communicate further results to the Society. Croontan Lecture.—< The Chemical Reeulation of the Secretory Process.” By W. M. Baytiss, F.BR.S., and E. H. STARLING, F.R.S. Received March 21,—Read March 24, 1904. In the complex reactions which make up the life of an individual, and the evolution of which has been the determining factor in the individual’s existence, we may distinguish two main types; though, as in all attempts at classification of biological processes, the line of division between the two types must be more or less indistinct. In the first place we have those reactions which depend for their production on some special structural arrangement, and are therefore determinant factors in the evolution of form. In some cases the adaptive production of organs or protective mechanisms may be associated with a direct chemical reaction, as in the formation of protective tissues. Im most cases, in higher animals at any rate, such an adaptation will be intimately associated with the development of the central nervous system, of which the peripheral parts of the body must be regarded as the executive mechanisms. In general, however, we may say that this type of adaptation is dependent on the adaptive growth of cells. The second type, the more primitive of the two, involves, in the first place, not so much a change in the growth or arrangement of cells, as a change in the metabolism of pre-formed cells or struc- tures. It may, perhaps, be looked upon as a preliminary to the first type, namely, structural change. It is, however, of special interest, since its mechanism is subject to analysis by physiological methods. Instances of chemical adaptation may again be divided into two groups. In the first place we have the chemical adaptation to the 1904.) The Chemical Regulation of the Secretory Process. d1l environment, which is found in all living organisms from the lowest to the highest. This adaptation may be conditioned by changes in the food, or may arise as a reaction to the presence of harmful substances in the surrounding medium. As an example may be mentioned the mould Penicilliwm glaucum, which, as shown by Duclaux,* when grown on calcium lactate forms invertase only; on starch, however, it produces amylase in addition, while on milk it produces a proteolytic ferment and rennet. In the higher animals we have all the complex processes by which an animal reacts to the introduction of living or dead poisons, and which result in’ the production of an acquired immunity. As part of the same process, if we accept Khrlich’s views, we must include the process of assimilation of food, and the adaptation of an animal to profound changes in its diet. But in all the higher animals the reaction of any part of the body to external changes involves alterations in its relations to other parts, and there is evolved a complex system of internal correlation of the activities of organs, effected partly by the action of the central nervous system, and therefore determinant of changes in form, partly through means of the internal medium—the blood or similar fluid. This latter mechanism of internal correlation has only recently entered the domain of exact investigation. Thus the profound influence exercised by the thyroid gland on the nutrition of the whole body, specially of the central nervous system, and the production of a substance by the suprarenal bodies which maintains the tone of all contractile tissues in the body, have been disclosed to us during the last 15 years. When chemical adaptation occurs in response to changes either in the environment or in definite organs of the body, the adaptive reaction may affect all cells of the body, or may be specific in the case of certain cells. Only in the latter case will the results be apparent to the morphologist as determinant of form. The researches which we wish to bring briefly before the Royal Society have reference entirely to the last two groups we have mentioned, and deal with the mechanism of adaptation to changes in the food and the chemical correlation of the activities of different organs engaged in the digestion and assimilation of the food. As we proceed down the alimentary tract, we find that each cavity has its own set of reactive mechanisms arranged so as to pour on the ingested food a juice which shall dissolve one or more of the constituents of the food. In the mouth, as has been shown by the researches of Ludwig, Heidenhain, Langley, and Pawlow, the mechanism for the secretion of saliva is entirely nervous. The mucous membrane is endowed with distinct sensibilities for different classes of food, and the activity of the salivary glands is excited reflexly according to the nature of the substance present in the mouth. In * © Microbiologie,’ vol. 2, p. 86. 312 Dr. W. M. Bayliss and Prof. E. H. Starling. [Mar. 21, the stomach, the researches of Heidenhain, and especially of Pawlow,* have shown that the secretion of the gastric juice is, in the first place, controlled by the nervous system, and is excited by appetite, or by reflex impulses arising in the mouth. Only later on, in gastric digestion, does a secretion come on, determined, in some way or other, by the presence and nature of the food in the stomach. This secondary secre- tion is independent of the central nervous system ; but whether it is to be looked upon as a local reflex, or as a chemical excitation, directly or indirectly from the gastric contents, has not yet been determined. As the strongly acid fluid containing the products of gastric digestion leaves the stomach to enter the duodenum, it comes in contact with two other secretions, the bile and the pancreatic juice, which are secreted in such an amount that the duodenal contents become prac- tically neutral. According to Pawlow, the secretion of the pancreatic juice is exactly comparable to the secretion of saliva, and is effected by a nervous reflex. The starting point of this reflex is the stimu- lation of the duodenal mucous membrane by the chyme, or by substances such as oil, ether, or oil of mustard. Not only is the pancreatic juice turned out into the intestine just at the time when it is required, but, according to Pawlow, the composition of the juice varies according to the food, the proteolytic ferment being increased by a diet of meat, while the amylolytic ferment is increased by a starchy diet. ‘This adaptation of the glandular activity was ascribed by him to a species of “taste” in the mucous membrane. It was imagined that the different constituents of the food excited different: nerve endings, which, in their turn, caused reflex activity of different mechanisms in the pancreas itself. ‘The field of these assumed reflexes was considerably narrowed by the researches of Popielskit and Wertheimer,t who showed that the introduction of acid into the duodenum was productive of secretion even after destruction of all nerve connections of the pancreas and alimentary canal with the central nervous system, and even after extirpation of the sympathetic ganglia of the solar plexus. It was with a view to determine the mechanisms of this reflex secretion of the pancreas, as well as of the adaptation of the pancreatic secretion to variations in the food of the animal, that we began our researches. The last-named authors had also shown that the secretion occurred, but in smaller quantities, if the acid was inserted in any part of the small intestine, with the exception of the lower end of the ileum. It was thus easy to examine the effects of the introduction of acid into a loop of ileum in which all nerve connections with the pancreas, or with the rest of the body, had been destroyed. This crucial experi- ‘Le Travail des Glandes Digestives,’ Paris, 1901. ‘Gazette Clinique de Botkin,’ 1900. ‘Journal de Physiologie,’ vol. 3, p. 335, 1901. Ver 5 Ss 1904.) The Chemical Regulation of the Secretory Process. 313 ment had, curiously, not been performed by previous workers in the subject. On carrying it out, we found that destruction of all nerve connections made no difference to the result of introducing the acid. The pancreatic secretion occurred as in a normal animal. It was therefore evident that we had to do here with a chemical rather than a nervous mechanism. Previous work had narrowed the question down to such a degree that the further steps were obvious. We knew already that the introduction of acid into the blood-stream had no influence on the pancreas; hence the acid introduced into the intestine must be changed in its passage to the blood-vessels through the epithelial cells, or must produce in these cells some substance which, on access to the blood stream, evoked in the pancreas a secretion. This was found to be the case. On rubbing up the mucous membrane with acid, and injecting the mixture into the blood-stream, a copious secretion of pancreatic juice was produced. It was then found that the active substance, which we call secretin, was produced by the action of acid from a precursor in the mucous membrane, probably in the epithelial cells themselves. Once formed by the action of acid, it could be boiled, neutralised, or made alkaline, without undergoing destruction. The precursor of the substance ( pro-secretin) cannot be extracted by any means that we have tried from the mucous membrane. Even after coagulation of the mucous membrane by heat or alcohol, however, secretin can still be extracted from the coagulated mass by the action of warm dilute acid. The question then arose whether this chemical mechanism repre- sented the normal mode in which secretion of the pancreatic juice was excited by the presence of food in the gut. It had already been shown by Wertheimer that the secretion evoked by the presence of acid diminished as the acid was placed further down in the small intestine, and was absent when the acid was placed in the lowermost section of the ileum or in the large intestine. We found a correspond- ing distribution of pro-secretin. The most active extracts of secretin were to be obtained from the duodenum. ‘The extracts from the jejunum were less powerful, while those from the lower 6 inches of ileum or from the large intestine were practically inert. The proof that secretin is really carried by the blood to the gland has been furnished by Wertheimer,* who has shown that the blood coming from a loop of intestine into which acid has been introduced, when injected into another dog, evokes in the latter a secretion of pancreatic juice. All authors who have investigated the matter since our first publication on the subject have confirmed our results; but many of them are still loth to give up the idea of a nervous connection between the gut and the pancreas. Pawlow had obtained evidence of the existence of secretory nerves to the pancreas in the vagus as well * “C.R. Soc. de Biologie,’ 1902, p. 475. 314 Dr. W. M. Bayliss and Prof. E. H. Starling. [Mar. 21, as in the splanchnics. In all his experiments, however, it was difficult to exclude the possibility of the secretion having been excited by the contraction of the stomach or relaxation of the pylorus, causing the passage of some acid contents of the stomach into the duodenum, since both these results may occur on stimulation of the vagus. We have been unable to obtain secretion from stimulation of any nerves in any case where this possibility was excluded, and we are inclined to believe that the chemical mechanism we have described is the only method by which the pancreas is awaken to secrete. The inhibition of secretion obtained by some authors in an unanesthetised animal on stimulation of the vagus is, we believe, a secondary phenomenon due to interference with the blood supply or more probably with the flow of acid chyme from the stomach, or perhaps to the rapid emptying of the upper part of the gut of its acid contents. Secretin can be split off from its precursor in the mucous mem- brane by the action of acids or boiling water. Many acids are able to effect this conversion, their power being roughly proportional to their ionic concentration. We have, therefore, concluded that the process is one of hydrolysis. According to Fleig,* a secretin can also be prepared from mucous membrane by the action of soaps, and secretin has been detected in the blood flowing from the loop of intestine into which oil of mustard had been introduced. Fleig regards the secretin produced by the action of soaps as different from that produced by the action of acids ; but it is difficult to see on what grounds he makes this distinction, since the action of the secretin prepared in the two ways is identical. The production of secretin by the action of oil of mustard as well as the well-known secretion of pancreatic juice evoked by the introduction of ether into the duodenum, suggests that the hydrolytic dissociation which gives rise to secretin may occur in the living cells as a result of stimulation or severe lesion, since neither of these two substances will produce secretin from an excised and dead mucous membrane. , We have not yet succeeded in determining the chemical nature of secretin, though we have obtained chemical evidence which will serve to exclude certain classes of substances. Thus the fact that it will stand boiling shows that it is neither a coagulable proteid nor a ferment. It is soluble in 90 per cent. alcohol in the presence of ether, but it is insoluble in absolute alcohol and ether. It is slightly diffusible through animal membranes. It can be filtered through a gelatinised Chamberland filter. It is not precipitated by tannic acid, thus excluding bodies of alkaloid nature as well as di-amido compounds. This evidence, slight though it is, points to secretin being a body of relatively small molecular weight and not a colloid. It may be compared to the active principle of the suprarenal glands, adrenalin, * “CO. R. Soc. de Biologie,’ 1903. 1904.] The Chemical Regulation of the Secretory Process. 315 which has been obtained in a crystalline form and the chemical constitution of which has been approximately determined. This is, indeed, what one would expect of a substance which has to be turned out into the blood at repeated intervals in order to produce in some distant organ or organs a physiological response proportional to the dose. The bodies of higher molecular weight, such as the toxins, which owe their activity, according to Ehrlich, to the fact that they can be directly assimilated by the cells of the body, and built up into the protoplasmic molecule, always give rise to the production of anti-bodies, a process which, while not preventing necessarily their utilisation in the body, would prevent their acting as a physiological stimulus to certain definite cells. Adrenalin and secretin on the other hand belong to the class of drugs which act by their physico- chemical properties, and whose physiological effect is determined by the total configuration of their molecule. It was suggested to us early in our experiments that the secretion of pancreatic juice, evoked by secretin, was essentially a sudden production of an anti-body ; such a sudden production is unknown in the animal body, and the anti-character of the secretion is at once negatived by the fact that secretin can be mixed with a freshly secreted juice without in any way destroying its efficiency. Like adrenalin, secretin is extremely easily oxidised, and it is probable that it is got rid of in this way from the body, since, even after repeated injections of secretin, it is impossible to find this substance or any precursor of it either in the pancreas, the urine, or other tissues of the body. Just as in the case of adrenalin, so we find that secretin is not specific for the individual or species. An extract of the mucous membrane of the dog will evoke secretion in the pancreas of the frog, the bird, rabbit, cat, or monkey. In the same way the pancreatic secretion of the dog can be excited by injection of secretin prepared from the intestine of man, cat, monkey, rabbit, fowl, salmon, skate, frog, or tortoise. ‘The evolution of this mechanism is, therefore, to be sought at some time anterior to the development of vertebrates. The action of secretin is not confined to the pancreas. It has long been known that the pancreatic juice, in order to exert its full activity on the food stuffs, needs the simultaneous presence of bile, and the fact that in many cases the two fluids are poured into the duodenum by a common orifice shows the close connection which must exist between them. Digestion of fats is impossible unless both fluids have access to the gut, and even in the digestion of carbo- hydrates, as was shown by S. Martin and Dawson Williams many years ago, the presence of bile greatly hastens the digestive powers of the pancreatic juice. Whenever, therefore, a secretion of pancreatic juice is required, a simultaneous secretion of bile is also necessary. It is * i ee 316 Dr. W. M. Bayliss and Prof. E. H. Starling. [Mar. 21, interesting to note that this simultaneous secretion is provided for by the same mechanism by which the secretion of pancreatic juice is evoked. If the flow of bile be determined by measuring the outflow from a cannula placed in the bile duct, it will be found that introduction of acid into the duodenum causes a quickened secre- tion of this fluid. The same increase in the secretion of bile can be produced by injecting solutions of secretin into the blood stream. This influence of secretin on the liver has been fully confirmed by Falloise. This observer has shown that acid extracts of the intestinal mucous membrane cause an increase in the bile secretion most marked when the extract is made from the duodenum and diminishing as the extract is taken from the lower parts of the gut, that from the lower section of the ileum being quite ineffective. In some cases the injection of secretin is followed by a secretion of glairy saliva, but this is at once abolished on section of the nerves going to the salivary glands, and is simply a result of the lowering of blood pressure which occurs when any extract of the intestinal mucous membrane is injected into the blood stream. On no other glands of the body has secretin the slightest influence. We must, therefore, regard secretin as a drug-like body having a specific excitatory effect on the secreting cells of the liver and pancreas. The discovery of secretin has placed in the hands of physiologists the power of controlling the activity of a gland by purely physio logical means, and we have taken opportunity of the control thus acquired to investigate the exact character of the changes induced in the pancreas under this physiological stimulus. So far as we can tell secretin has no specific ‘influence on any one constituent of the pancreatic juice. When injected it causes secretion of a juice which is normal in that it resembles the juice secreted on entry of food into the duodenum, and contains a precursor of trypsin, amylopsin, and steapsin. Secretin, in fact, appears to cause the pancreatic cells to turn out the whole of the mesostates which they have accumulated during rest in preparation for the act of secretion. If secretin be injected at repeated intervals until the gland will no longer respond to the injection, it is found on microscopic examination that the cells have discharged the whole of their granules. In sections stained with toluidine blue and eosin the whole of the cells stain blue in marked contrast to the normal resting gland, where one-half or two-thirds of the inner margin of the cells is taken up with brilliantly stained red granules. This effect is not produced in all cases. In some animals we have injected secretin at frequent intervals over a period of 8 hours, and obtained at the end of the experiment a secretion as vigorous as after the first injection. The pancreas in this case was evidently not fatigued, and on killing the animal and examining this organ microscopically, it was found to give the typical picture of a 1904.] . The Chemical Regulation of the Secretory Process. 317 resting pancreas. One may say, therefore, that under healthy con- ‘ditions the activity of the pancreas is two-fold in character, and that the normal stimulus of secretin excites not only a breaking down of the protoplasm and a discharge of granules, but also a building up of the protoplasm and a new formation of granules. So marked, in fact, is this power of self-restitution that it is often advisable to diminish the resistance of the animal by bleeding or other means if it is desired to obtain a specimen of exhausted gland. A study by Mr. Dale of the stages of exhaustion carried out in this way has brought to light a remarkable behaviour in the cells of the pancreas, to which we have no analogies in other secreting glands of the body. After the discharge of the granules the cells seem to undergo a still further involution, losing the whole of their chromo- phile substance, diminishing im size or undergoing vacuolation, and finally being transformed into cells undistinguishable from those which have long been known as forming the so-called “islets of Langerhans.” Mr. Dale has, in fact, shown that in all probability these “ islets,” which are generally regarded as pre-formed structures, really repre- sent stages in the functional activity of the secreting cells of the gland, and he is of opinion that the activity of the gland is always associated with a cycle of changes in which the islets are formed, to be afterwards regenerated into secreting tissue. Other observers have noted in the embryo a development of secreting tubules from tissue undistinguishable from the ‘islets of Langerhans,” and it is interesting to note that the depletion of the gland caused by long starvation has a similar effect to that caused by over-excitation, namely, the conversion of a large proportion of the gland tissue into “islet” tissue. Although secretin acts in this apparently coarse manner in turning out all the pre-formed secretory products present at the time in the pancreatic cells, the conditions of its formation determine a close adaptation of the pancreatic activity to the needs of the animal. Formation of secretin depends on the presence of acid chyme in the duodenum. This acid chyme is squirted in small quantities into the stomach at varying intervals after the taking of food. As soon as it enters the gut, secretin is formed in the mucous membrane, absorbed by the blood-vessels and carried to the pancreas, and it will continue to be formed until the secreted pancreatic juice exactly neutralises the acid of the intestinal contents. The presence of an excessive amount of acid in the duodenum is prevented by the reflex pyloric mechanism revealed by the researches of Von Mering and of Serdjunow.* These observers have shown that so long as the contents of the duodenum are acid the pylorus remains firmly closed. As soon, however, as they become neutral or alkaline the pylorus opens and allows a further * Pawlow, ‘Das Experiment,’ 1900, p. 17. 318 Dr. W. M. Bayliss and Prof. E. H. Starling. [Mar. 21, quantity of acid gastric contents to enter the duodenum. By this double mechanism, which is partly nervous, partly chemical, it is pro- vided that the acid contents of the stomach shall pass on into the gut only insuch quantities as can be dealt with by the secretory mechanisms of the intestine. One more chain in the link of adaptive reactions may be briefly mentioned. ‘The pancreatic juice, as secreted, contains only a weak proteolytic ferment. But it contains also trypsinogen. As soon as this juice enters the gut it causes a profuse secretion of intestinal juice. This latter contains another ferment, enterokinase, which acts on the trypsinogen, converting it into a body trypsin, one of the most active proteolytic ferments with which we are acquainted. So far we have dealt only with the correlation of the activities of the cells lining the intestinal tube with those forming the masses of the pancreas and liver, and have seen that a very large part in this correla- tion is played by a chemical substance which acts, so to speak, as a chemical messenger between these various organs. A striking feature, however, of the pancreas is its alleged power of adapting its secretion to the nature of the food taken in by the animal. It has been stated by Pawlow that according as the food consists chiefly of proteids, carbohydrates, or fats, so do we find a relative preponderance of the ferments acting respectively on each of these three classes of foods. The evidence on which this statement is based, although lending to it considerable support, is not absolutely convincing. Vasilieff* examined the pancreatic juice of dogs which were fed on meat, or bread and milk alternately for periods extending over several weeks for each kind of diet. This observer found that the transition from bread and milk diet to a meat diet caused a rapid rise in the proteolytic power of the juice, which reached its maximum after several days of meat feeding. A return to a diet of bread and milk caused a slower fall in the proteolytic power of the juice, but a rise in the amylolytic power. Similar results were obtained by another pupil of Pawlow— Jablonskyt—who also extended his observations to the fat-splitting ferment. At the time that these observations were made the function of enterokinase was unknown, and it is therefore impossible to say what proportion of the trypsinogen of the juice secreted in these experiments had been converted into trypsin by the small amount of intestinal mucous membrane at the mouth of the duct. While, there- fore, we are unable to ascribe much importance to the results as regards the proteolytic power of the juice, there seems no reason to doubt the results obtained by these workers as regards the starch-digesting power of the juice. In 1899 Walther} made a series of observations * © Archives des Sciences Biologiques,’ St. Petersburg, 1893. + Ibid., 1896. + Ibid., 1899, vol. 7, p. 1. 1904.) The Chemical Regulation of the Secretory Process. 31 on a dog with pancreatic fistula in order to determine whether the amounts of ferments secreted were determined by the nature of the food at any given meal. He was satisfied that his results showed that, even without prolonged adherence to one diet, the composition of pancreatic juice was adapted to the nature of the meal taken. His results do not entirely bear out his contentions, as is seen by the following table, in which it will be noticed that although milk contains no starch it evokes the secretion of a large amount of amylopsin, and that meat causes a secretion of more steapsin than does milk, although this latter contains much more fat than the meat diet. Table I.—Results of Walther’s Experiments. Total amount of enzyme secreted. Diet. Proteolytic. Amylolytic. Fat-splitting. GOOKeves WM. ies Ls aire 1044; 2310 4125 | 250 zrammes bread ...... 2360 6343 1218 ! 100 grammes meat...... 1720 2498 44.10 | Of course Walther, as well as the other observers mentioned, regard the adaptation as determined by the stimulation of special nerve endings in the mucous membrane by each constituent of the food, a conclusion: hardly borne out by the results just quoted. Another disturbing factor in these experiments is the large variation in total quantity of juice secreted with different food stuffs. Table I1.—Amount of Pancreatic Juice Secreted for different Food- stuffs (Walther). Hours of secretion. amount. 1. 2. | 3. 4, 5. 6. FMW Se et Oy 600 c.c. milk | 8°2| 6:°0/ 23:0! 6-2/1°75| .. |.. |.. |.. | 45 ec. 250 grammes 35°5 | 47:0 | 20°5 | 16°5 |10°0 12°0 |6°5|3:°0|.. | 151 ,, bread 100 grammes | 45°0 | 52°0 | 35°0 9°75 142 ,, meat | | The quantity of juice secreted will depend on the amount of secretin turned into the circulation, and this, in its turn, on the amount of acid entering the duodenum from the stomach. ‘The mt “ling 320 Dr. W. M. Bayliss and Prof. E. H. Starling. ([Mar. 21, amount of juice will, therefore, be measured by the stay and resistance to digestion of the substance in the stomach rather than to any direct nervous or other influence of the duodenal contents on the pancreas. A repetition of Walther’s experiments by Popielski,* working independently, has in fact led the latter to deny altogether the adaptation of the pancreatic juice to the nature of the food. Popielski concludes from his experiments that variations in the juice depend only on the intensity and duration of the stimulus, the intensity of the stimulus determining the amount of enzymes, whilst its duration determines the total quantity of juice. In the meantime the question had been attacked from another side. It had been shown by Fischer and Niebelt as well as by Portier{ that watery extracts of the pancreas of the cow, horse, and dog had no influence on lactose. Weinland in 1899 confirmed these results so far as concerns the pancreas of dogs on an ordinary diet free from milk. On the other hand, he found that extracts of the pancreas of dogs, which had been fed for several days on milk, sometimes with the addition of lactose, invariably contained lactase in considerable amount, and these results have been confirmed lately by Bainbridge working in our laboratory. Here then we have a definite instance of adaptation of the pancreas, the pancreatic juice or pancreatic extracts of dogs on normal diet containing no lactase, while the administration of lactose to these animals causes the appearance of lactase in both cases. Since in this case we have to determine, not simply an increase or diminution in the amounts of ferments always present in the juice, but the presence or absence of a definite substance, this was evidently the best starting point for an investigation of the mechanism by which the pancreas can adapt itself to the nature of the food, an investigation which has been carried out and completed by Dr. Bainbridge. What are the limiting conditions? In the first place the reaction is absolutely specific. Unless the animal is taking lactose in its diet no lactase is ever found in the pancreas or in its secretion. The pancreas of new-born animals, for instance, is quite free from lactase, which, however, makes its appearance 2 or 3 days after birth as the result of the milk diet. The production of lactase is not a direct reaction of the pancreas to the presence of lactose in the blood, since subcutaneous or intravenous injection of lactose does not cause the appearance of lactase in the pancreas. The intestinal mucous mem- brane of all animals, whether on a milk diet or not, contains lactase and has an inverting action on lactose. It might be thought therefore that the production of lactase by the pancreas was a reaction to the % * “Centralblatt f. Physiologie,’ vol. 17, 1903. + ‘Sitzungsberichte der K. Preuss. Akad. d. Wiss.,’ 1895, p. 73. ft ‘C. R. Soc. de Biologie,’ 1898, p. 387. 1904.] The Chemical Regulation of the Secretory Process. 321 presence of the products of inversion of lactose in the blood. This was found not to be the case. Subcutaneous injection of galactose for several days was not followed by any appearance of lactase in the pancreas or its juice. Nor was the appearance of lactase due to the increased production of this ferment in the mucous membrane, and its escape into the blood. Injection of an extract of mucous membrane rich in lactase, repeated several days in succession, was not followed by any appearance of lactase in the pancreas. Injection of lactose into the duodenum, and the subsequent injection of secretin after an interval of 1 hour, was inefficacious in causing the appearance of lactase in the pancreatic juice. For the production of lactase in the pancreas, or its juice, it is therefore necessary that lactose should act on the intestinal mucous membrane for some time. ‘The reaction is a slow one, like the adaptation in Vasilieff’s experiments, and is certainly not due to the stimulation of certain nerve endings in the mucous membrane by the lactose. The problem was somewhat similar to that presented by the action of acid in the duodenum, since this introduced into the. duodenum produces secretion of juice, whereas, when introduced into the blood stream, it has no effect whatever on the pancreas. The question suggested itself whether, under the influence of lactose, a special secretin was formed in the intestinal mucous. membrane which, on access to general circulation, evoked the forma- tion and secretion of lactase by the pancreas. Secretin was there- fore made in the usual way (z.¢., acidification, boiling, neutralisation,. and filtering) from the mucous membrane of milk-fed dogs. The secretion evoked by the injection of this liquid resembled that obtained from the injection of ordinary secretin, and contained no. lactase. Yet it was evident from the results already obtained that lactose: must act on the pancreas through the mucous membrane of the intestine. An extract was therefore made from the mucous membrane: of the whole small intestine of a milk-fed dog. This was filtered through muslin, and about 10 ¢.c. injected subcutaneously into a biscuit-fed dog once a day for three days. The dog was then aneesthetised, a cannula placed in its pancreatic duct, and ordinary ‘secretin injected. A flow of pancreatic juice was obtained, and this juice was found to contain lactase. This experiment was performed eight times, and in each case the juice obtained from a biscuit-fed dog which had been injected with an extract of the mucous membrane of a milk-fed dog contained lactase. Here then at last we have some glimpse into the mechanism of the adaptation of the pancreas to the nature of the food. As the result of injection of lactose some substance which we may call x is produced in the mucous membrane of the small intestine. This substance is carried. 322 The Chemical Regulation of the Secretory Process. |Mar. 21, by the blood to the pancreas, and there slowly gives rise to the formation of lactase, which is turned out in the juice when secretion is excited by the entry of acid chyme into the duodenum. We have no knowledge as yet as to the nature of this substance v. All we can say is that it is destroyed at a boiling temperature, since boiled extracts of the mucous membrane of milk-fed dogs do not, when subcutaneously injected, cause the appearance of any lactase in the pancreatic juice of biscuit-fed dogs. ‘Table II].—Effect on Milk Sugar of Pancreatic Juice from Bisewt-fed dogs, which had received Subcutaneous Injections during 3 days of Extracts of the Mucous Membrane of Mzlk-fed dogs. The figures represent c.c. of lactose solution which reduced 50 c.c. Pavy’s solution. Controls. | apes Lactose + pancreatic en Solution of | Lactose + pancreatic ee inversion. | lactose. juice (boiled). | 1 7°4 i 6°8 18 ‘1 2 8 2 8:2 7°6 16 °5 3 8:2 8°15 7°85 9°7 4 7°95 7-9 7°65 8°5 Hs) 7°8 oe 7°5 88 6 7°0 7°05 6°75 Sal 7 4°] a 3°75 20°8 8 9°25 a 8°2 259 Whether the qualitative adaptation of the juice in respect of its trypsin, amylopsin, and steapsin is carried out in a similar fashion we cannot as yet say. We hope that an investigation of the mechanism of this adaptation, which is now proceeding, may throw light, not only on the factors involved, but also on the nature of the substance which is formed in the mucous membrane, and has this marked effect on the activity of the pancreatic cells. Involving, as it does, two distinct sets of cells, this chemical adaptation is more complex than any yet investigated, and shows the intimate relation which must exist between the chemical activities of very different organs of the body. 1904.] Hasy Method of preventing Death from Snake Bite. 323 “Hxperiments on a Method of Preventing Death from Snake Bite, capable of Common and Easy Practical Application.” By Sir Lauper Brunton, M.D., F.B.S., Sir JOSEPH FAYRER, Bart., K.C.S.1., F.R.S., and Leonarp Rogers, M.D., B.S., etc., Indian Medical Service. Received February 22,—Read May 5, 1904. Although this paper is a joint one, the authors wish to mention that each has had a different part in its production. The whole research may be fairly regarded as the natural outcome of the work begun in India nearly forty years ago by one of us (Fayrer), and this is the only ground on which his name can be associated with this paper. The instrument employed was designed by another of us (Brunton), and the actual experimental work was entirely carried out by a third (Rogers). The first experiments on the use of permanganate of potash as an antidote to snake poison was made by one of us (Fayrer), in 1869, both by the local application of a solution and by injection into the veins,* on the ground of its being a chemical antidote. The animals experimented upon were dogs, but the permanganate of potash did not seem to have any power to avert the lethal action of the poison. It was shown also by Wynter Blytht that Cobra venom when mixed in vitro with permanganate of potash becomes innocuous. His results were confirmed by two of us, who showed that some other substances had a similar power.t They tried by the injection of strong solution of permanganate of potash, and also by its local application to an incision made over the bite, to destroy the lethal action of Cobra poison previously injected, but their experiments were unsuccessful, the permanganate appearing to be unable to overtake the poison which had got the start of it. In 1881 Messrs. Couty and Lacerda § made a number of experiments upon the effect of permanganate of potash on serpents’ venom, and Lacerda found that permanganate of potash not only destroyed the lethal action of the venom when mixed with it in wiro, but also preserved life when a 1l-per-cent. solution of permanganate was injected into the tissues close to the place where the venom had been * “The Thanatophidia of India,’ 1872, p. 95, by J. Fayrer, M.D., London, J. and A, Churchill. + “The Poison of the Cobra,” by A, Wynter Blyth, M.R.C.S., ‘The Analyst,’ February 28, 1877, p. 204. t “Note on the Effect of Various Substances in Destroying the Activity of Cobra Poison,’ Brunton and Fayrer, ‘Roy. Soc. Proc.,’ June 20, 1878, vol. 27, p. 465. § Couty and Lacerda, ‘Comptes Rendus,’ vol. 92, p. 465. 324 Sir L. Brunton, Sir J. Fayrer, and Dr. i Rogers. [Feb. 22, previously injected, and also when both venom and antidote were injected directly into the vein. At the time of presenting his note to the Academy of Science in Paris, M. Lacerda was apparently unaware: of the previous experiments by Blyth, Brunton and Fayrer. In a later publication* he discusses their experiments, but claims for himself to: have scientifically demonstrated permanganate of potash to be a. precious antidote to serpent venom, and to have brought it into common use, and thinks, therefore, that the priority belongs to him ;. but he was apparently unaware that instructions for its use with the ligature had many years before been promulgated by Fayrer in India. In the winter of 1881 a number of experiments were made by Dr. Vincent Richards, who found, like the previous experimenters, that Cobra poison was compietely destroyed by permanganate of potash when mixed with it in vitro, so that death did not follow the injection of the mixture either hypodermically or into a vein. He found also that when Cobra poison was injected into a dog, and the injection followed either immediately or after an interval of 4 minutes by a hypodermic injection into the same part of a solution of permanganate of potash no symptoms of Cobra poisoning resulted, but after the development of symptoms of Cobra poisoning permanganate of potash failed to have any effect whether injected locally or into a vein or both. These results obtained both by Lacerda} and Richards seemed to give good hope that permanganate of potash might be used to lessen the appalling fatalities from snake bite in India, but it is evident that the hypodermic injection of a solution can never be widely employed because the hypodermic syringe is expensive, it is liable to get out of order just at the times that it is wanted and the solution may become dried or spilt or may not be available. It is evident that the first requisite for any antidote to snake poisoning is that it shall be always at hand; second, that it shall be easily applied; and thirdly, that it shall be cheap. About two years ago one of us (Brunton) was asked on behalf of a young officer going out to India, to design an instrument which might. be used in case of snake bite. He did so accordingly, and he has since had a similar one made for him by Messrs. Arnold and Sons which seems to combine the three requisites just noted. It consists of a lancet-shaped blade about half an inch long, long enough in fact to reach the deepest point of a bite by the largest snake. He has had some instruments made with a double edge like an ordinary lancet, and others with one edge sharp and the other edge blunt, so as to * Lacerda, ‘Comptes Rendus,’ vol. 93, p. 466. + “© YVeneno ophidico e seus antidotos,’ Dr. J. B. de Lacerda, Rio de Janeiro, Lombaerts, &c., 1881, p. 64. 1904.) Hasy Method of preventing Death from Snake Bite. 325 press in the permanganate. The lancet is set in a wooden handle about an inch and a half long, which is hollowed at the other end so as to form a receptacle to hold the permanganate. Two wooden caps are fitted over the ends of the instrument, one to keep in the permanganate, and the other to protect the lancet. Such an instru- ment, if turned out in large numbers, could be sold at such a small price as to be within reach of even the Indian labourer, and might be sold everywhere in the same way as packets of quinine are at present. Wie. 1.—Lancet for use in snake bite, showing the steel blade, the cap which covers it, the hollow wooden handle for holding crystals of permanganate of potash, and the cover which retains them. The plan now proposed is to make a free opening into the site of - the bite, and to rub in crystals of permanganate. For this purpose the limb should be surrounded by a tight bandage above the bite, the puncture of the tooth or teeth should be freely cut into by the lance- shaped blade and the crystals of permanganate introduced and rubbed round. A few drops of saliva may be added. To test the efficacy of the proposed plan several lethal doses of venom dissolved in a few drops of water, so as to resemble, as far as possible, the natural poison, are to be injected into the limb of an animal, a ligature placed round the limb above the seat of injection, an incision made, and crystals of permanganate placed in the wound, moistened and rubbed in. Experimental Investigation, by Leonard Rogers. In order to test in as practical a manner as possible the value of the suggestion of the two first-named authors of this communication, the following experiments were carried out at the Physiological Laboratory of the London University by the third-named author. In the first place it was necessary to ascertain if crystals of permanganate destroy the activity of other venoms besides that of the Cobra, for we are not aware that its action in this direction has been tested against any extensive series of snake venoms. As the value of the suggested treatment would evidently be greatly enhanced if the per- manganate could be shown to act efficiently against every class of snake venom, a series of experiments were carried out to test this point. The venoms in solution were mixed with small quantities of VOL. LXXIII. 2A 326 Sir L. Brunton, Sir J. Fayrer, and Dr. Th: Rogers. [Feb. 22, a 10-per-cent. solution of pure crystalline permanganate of potash in 0-9 per cent. NaCl, and after given times the mixtures were injected into pigeons, several times a lethal dose of each venom being used, so that if recovery took place it would be evident that the permanganate had destroyed the activity of the poisons. The following table (p. 327) summarises the results of these: experiments. It will be seen that the table includes venoms of each main sub- division of snakes, namely, the two true vipers, the Daboia Russellu . of India and the Puff Adder of Africa, the Pit Viper, the Crotalus: horridus, the Colubrine snake the Bungarus fasciatus, and one of the Hydrophide or Sea-snakes, namely, the Hnhydrina bengalensis. In the case of each ten or more lethal doses were neutralised by very small quantities of permanganate in solution, and in most of them twenty lethal doses were readily thus rendered harmless. The only failure was in Experiment 7, in which 32:2 milligrammes of Bungarus fasciatus venom was added to 25 milligrammes of permanganate of potash in solution, and in this case by far the greater part of the poison must. have been neutralised, for in previous experiments one-eighteenth part of the venom per kilogramme, used in Experiment 7, killed a pigeon in 1 hour. Further experiments showed that 25 miligrammes of the permanganate of potash did entirely neutralise 16:1 milligrammes of Bungarus fasciatus venom. It is evident then that the salt will neutralise about its own weight of this venom, but that its power in this direction has a definite limit as might have been expected. It is clear, then, that this agent does act on every class of snake venom and renders them inert. Owing to the limited time available and the small number of animals ~ for which a license had been obtained, the actual experiments on the treatment after injection of the venoms have been so far limited to those of the Cobra as a typical representative of the Colubrine class, and of the Daboia Russell as a common and deadly viper. Rabbits and cats were used in the investigation, the latter on account of their mixed diet and firmer tissues resembling more closely the human subject. The venoms were dissolved in as small a quantity of sterile normal saline solution (0°9 per cent. NaCl) as possible, so as to: resemble in concentration the natural venom. The portion of the limb to be operated on was cleaned of hair by scissors beforehand (as the human subject is free from this obstacle to treatment). The strong solution of venom was then injected into the subcutaneous tissues of the cleaned part of a hind limb a little above the paw, as: most snake bites in the human subject occur on the distal parts of the extremities. After a given measured time a ligature consisting of a piece of bandage was tied loosely round the thigh and twisted up: tightly by means of a piece of stick or a pencil so as to temporarily stop the circulation through the distal part of the limb in order to 327 eq | ‘0d | OT | eZ. 0 | | 0z | CZ 0gz rete ee ee nce ‘sisuappbuag Duiuphyugy ° A A 2950 312 2360 300 1770 200 240 950 > 1480 ie 6 202 1180 ui if 152 880 950 50 150 121 800 860 My eH 5, 700 106 590 650 >670 Ae ie 600 91 530 650 > 029 “ 5 "6 440 550 a‘ A 61 420 440 “ ie 3800 50 410 3307 2L re ba 38 290 210 ie 26 if 140 ie é 20 ct 100 ce i 13°5 an 95 re i 6 i 50 Me 2 3 ae 25 po is 1h 100,000, for which g > 2. 2. “ Periode des Dezimalbruches fiir 1/p wo p eine Primzahl,” by H. Hertzer,. printed in Grunert’s ‘ Archiv der Math. und Phys.,’ vol. 2, 1902, p. 249. The Table (pp. 249—251) is a continuation of the preceding Table for primes up to: p } 112,400, and is arranged in the same manner. Shanks’s MS. Tables give the period-length (say &) of 1/p for all primes from 30,000 to 120,000. The collation was effected by simply multiplying Shanks’s value of € by Kessler’s or Hertzer’s value of q;. the product of €¢ should = (p—1) in every case. The collation was, of course, only possible for such primes as have g > 2 (being the only ones shown by Kessler and Hertzer), thus excluding about two-thirds of the total number of primes ; the €, g of these excluded primes are, however, easily computed when required. By this collation a number of discrepancies (102 in all) were discovered between Shanks’s MS. and the printed German tables. The values of € have in all these cases been re-computed,{ with the result of detecting errata, as follows :— ROIS es son ese p missing 5; € wrong 66; total 71 Wessler oi... Total 102. Ber AEN 22200. . s 6 p missing 1; g wrong 2; total 3 * Part of the MS., viz., for primes from 30,000 to 60,000, is bound up with a small volume marked Constants and Primes (with the Press-mark 103 d 15), and the rest, viz., for primes from 60,000 to 120,000, is bound up with the ‘ Royal Society Archives,’ vols. 60,61. Mr. Shanks’s MSS. bear dates as follows :— For primes 30,000 to 60,000, dated 1875. For primes 60,000 to 75,000, dated 1876. For primes 75,000 to 110,000, dated 1877. For primes 110,000 to 120,000, dated 1880. + By the writer of this paper, with the help of an assistant (Miss HE. Cooper), by permission of the Council of the Royal Society. { By the writer himself, and verified in part by Miss E. Cooper. 360 Corrigenda in Tables of Reciprocals of Prime. The table following gives the Corrigenda on Shanks’s MS. This list is probably far from complete; there is reason to suspect a good many errors in Shanks’s MS. among those primes for which q + 2, but, as the table has not been published, it did not seem worth while examining these. Corrigenda on Shanks’s MS. Table III | primes from 30,000 to 112,400]. Insert five primes (p) missing in MS.: 33797, 59369, 94111, 95089, 104383. Correct the period-lengths (&) opposite the primes (p) as below. Hee S IN Dp. p- . Dp. &. emer | | | | | | a 33,797 | 8,449 || 65,011 | 21,670 | 86,143) 1,758 || 103,813 | 17,302 34,871 | 1,585 || 70,001 | 35,000 || 86,323 | 14,387 || 104,381 | 104,380 42,773 289 || 70,867 | 3,937 || 87,121 | 4,840 || 104,888 3,866 43,758 | 14,584 | 70,921 | 3,546 || 87,151 | 2,075 || 104,707 | 17,451 44,893 | 22.446 || 71,821 | 4,788 || 87,517 | 21,879 || 105,367 | 35,122 46,153 | 15,384 || 72,559 87 || 87,697 | 29,232 || 105,613 | 26,403 46,649 686 || 72,661 | 24,220 | 88,003 | 14,667 || 105,929 | 26.482 47,093 | 23,546 || 72,871 | 1,735 || 89,689 | 22.422 || 106,031 2 305 47,711 | 1,835 || 72,901 | 8,100 |) 92,107| 6,579 |) 106,921 2,430 53,857 | 17,952 | 73,851 | 7,335 | 93,319 | 15,553 || 107,647 | 15,378 55,021 | 7,860 || 74,413 | 37,205 | 94,111 | 9,411 || 107,887 | 26,959 55,681 | 9,280 | 74,687 214 || 95,089 | 23,772 || 109,441 9,120 55,933 | 27,966 | 78,079 | 3,003| 95,791 | 15,965 || 110,051] 22,010 57,457 | 2,128 || 79,111 | 7,911 |. 96,601] 4,830 | 110,917| 18,486 57,493 | 14,373 || 80,347) 13,391 || 96,911 | 4,405 | 110,969] 27,742 58,031 | 4,145 || 82,021 98,898 | 24,723 || 111,149 148 59,369 | 14,842 || 85,411 | 17,082 || 101,051 | 2,150 | 112,249 | 14,081 61,007 | 1,034 || 85,447 | 1,818 || 102,793 | 34,264 Ker) IN © ise On Heat Regulation and Death Temperatures. 361. “A Research into the Heat Regulation of the Body by an Investigation of Death Temperatures.” By Eprep M. Corner, M.A., M.B., B.C. (Cantab.), F.R.C.S., B.Sc. (Lond.), Surgeon to Out-patients, St. Thomas’s Hospital, and Assistant-Surgeon to the Hospital for Sick Children, Great Ormond Street, Erasmus Wilson Lecturer, Royal College of Surgeons, and JAMES E. H. Sawyer, M.A., M.D. (Oxon.), M.R.C.P., Honorary Anesthetist to the Ear and Throat Hospital, Birmingham, lately House Physician to St. Thomas’s Hospital. Communi- cated by Professor J. N. LANGLEY, F.R.S. Received April 4, —Read May 5, 1904. The fact has long been known that, in many forms of disease, variations of the bodily temperature occur as death approaches. Our knowledge of such variations is limited to certain apparently sporadic cases.. Hitherto, no attempt has been made to ascertain the special class of disease in which such deviations occur most frequently, nor has the subject been examined in a scientific light, so as to bring these death temperatures in line with the present knowledge of the mode of production of pyrexia. It is towards this latter object that this inquiry has been directed. Further encouragement to bring forward this new source of knowledge is given by Sir John Burdon Sanderson, who says “‘ the subject (of pyrexia) is one in respect of which results as valuable can be obtained by clinical investigation as by experi- ments on animals.’”* This communication has been divided into two parts. In the first part, a general account is given of the bodily temperature immediately preceding death, and of the variations which are seen in these thermo- metric records. A series of problems are shown, in order to ascertain, where possible, the different factors which may influence the variations in the death temperatures. The figures quoted are, in all cases, the absolute minimum, as numbers of these cases have to be rejected for many reasons ; for example, those in which tepid sponging has been employed, and others where the records are incomplete. In the second part, an attempt is made to ascertain the possible influences which may cause such variations in the bodily temperature, and to explain, from a study of these changes, the mechanism by which the temperature of the body is regulated in health, and the reasons for which deviations from the normal occur in disease. ParT I. As the fatal result approaches, the curve on the temperature chart may exhibit several changes. In 49 per cent. of surgical, and in * Clifford Allbutt, ‘System of Medicine,’ vol. 1, p. 152. 362 Mr. E. M. Corner and Dr. J. E. H. Sawyer. =‘ [ Apr. 4, 19 per cent. of medical cases, a definite change of over 1°5 F. occurs during the 12 hours immediately preceding death. These figures are the absolute minimum for about 2500 consecutive cases which have been collected from the records of St. Thomas’s Hospital. In instances where no such sudden change occurs, the temperature may remain about the same level, but the more usual procedure is for it to slowly and steadily fall. On the other hand, it is not common for the bodily temperature to rise slowly, for, when an elevation occurs at all, the change is generally a sudden one. The character of the temperature charts as death approaches is naturally affected by the types of fever from which the patients may be suffering, such as remittent, intermittent, etc. The influence of these fevers cannot be eliminated, and it is an interesting point that such cases frequently show a very decided variation in the character of the temperature near the fatal termination. The performance of tepid sponging, or the exhibition of antipyretic drugs, just before the death of the patient, ‘ renders the chart valueless for this research. For these, and similar reasons, many variations of temperature have been necessarily neglected. There are other changes in the temperature charts which have to be considered in these investigations. Sometimes, about 24 hours before death, the temperature falls suddenly, anything from 1—5° F., and then rises rapidly, and at this point death may occur. On the other hand, the fatal result may be postponed, and, after the temperature has reached its maximum, defervescence may begin again. Such fluctuations as these are more commonly found among surgical than among medical cases. The following is a good example of these changes :—a baby girl, aged 7 weeks, was severely burnt, and died on the fifth day ; the temperature during the last 36 hours of life repre- sented these variations; it fell 4°°3, rose 6°°4, falling again 4°-2. Again, in a woman, aged 28 years, who suffered from general peritonitis and appendicitis, there was a fall of 4° in the temperature, followed by a rise of 8°, when death occurred. These early deviations of the temperature begin about 24 hours before the death of the patient, and may exaggerate either the death rise or the death fall, or may mask them altogether. Of the 2500 cases which were examined, 1305 were medical cases, the remainder, 1195, being surgical. Changes in the bodily tem- perature of over 1°°5, occurring during the last 12 hours of life, were found in 34 per cent. of all the records examined; in 49 per cent. of surgical, and in only 19 per cent. of medical cases. On account of this great difference of 30 per cent., the two lists of cases have been kept separate, although in many instances they overlap. The explana- tion of the difference which is found between the medical and surgical cases has to be sought for in the other particulars. 1904. ] On Heat Regulation and Death Temperatures. 363 Of the variations found among the fatal medical cases, 75 per cent. were rises of temperature and 25 per cent. falls. For the surgical cases, the corresponding percentages were 73 and 27; the two classes agreeing very closely. Hlevations of temperature occurred in 26 per cent., and falls in 8 per cent., of all the cases examined. The relative proportions of male to female deaths were also investi- gated, and the following figures were obtained :— Medical Cases. Proportion of male to female deaths during MOMS MMe POLIO: 2 to wu sjsanmalbinan ceeds. sack 1°75 male to 1 female. Ditto, showing death rise of temperature ...... Oi ae ae Leaeee Ditto, showing death fall of temperature ...... wey Line Surgical Cases. Proportion of male to female deaths during MOMS AMMO MOTION: eve sv reace deeds 8 eS Jou 1:63 male to 1 female. Ditto, showing death rise of temperature ...... 1°75 AH 1a hi Ditto, showing death fall of temperature ...... 1°39 Fr 1S The following deductions may be made from these figures, indicating lines along which an explanation of the phenomenon may be sought. In surgical cases males are more apt to show a death rise; females, on the other hand, show a marked tendency to death falls. In medical cases the difference is not so marked, but female patients are relatively more liable to death variations of temperature than are male. A fall of temperature in both medical and surgical cases is more common in females than in males. : There is one point on which medical and surgical cases differ from each other, and that is, that in the latter deaths due to injury are included. As to the actual cause of death in disease and in injury the difference is not so great, but, as will be seen in the following table, the death temperatures vary considerably in one important point in the two classes. Only the surgical cases are here considered :— Proportion of deaths due to disease and AN UU eee a BN aaa Ue MRR c cto se. sales 1-88 disease to | injury. Wicro, showing death Tiseaciecs.. sala... <0 1:84 i bea Divto, showing death fall .c.0.00).522..00........ 3°29 i 1a These figures show that rises in the bodily temperature, just before death occurs, are found to be present in fairly equal proportion in patients dying from disease and from injury. Falls of temperature, however, are very much more common in patients suffering from disease. The following table shows the same proportions in_per- centages :— 364 Mr. E. M. Corner and Dr. J. E. H. Sawyer. _—‘[Apr. 4, Disease. * Tnjury: Proportion of death changes to total... 51:7 per cent. 46-0 per cent. e wt. ifises i Cnupat oORe i Sion 33 ») : dalls SF d cee ys ees io Sith -inge From this table it would appear that changes of temperature, as death approaches, occur less frequently in injuries than in disease, and that this difference is due to the comparative rarity of death falls in the former. The rises of temperature are present in equal proportion in the two classes of cases. In the following charts an attempt is made to show the amount and frequency of the various deviations of temperature :— CHART 1. CHART 2. 0.6f Medical & Sur gical | Rises of Temperature! No.of Medical § Surgical Falls of Temp oo EP ER LAE CET ES EF SiS ihre CASes. oe Be Pe LOle EP TIE [scrap ati OR 120 The following conclusions may be drawn from Charts 1 and 2 :— 1. That death rises of temperature are naturally larger than death falls. 2. That small variations of temperature are more common than large. 3. That death rises are comparatively rare over 5° ; more so over 6°. 4, That death falls are comparatively rare over 4° ; more so over 5’. 5. That the variations of temperature in the medical and surgical cases agree fairly well, although death rises in medical cases are relatively more common from 3—4’. 6. That death rises are more frequently of greater magnitude in surgical than in medical cases. Charts 3 and 4 show the actual temperature at the time of death (or rather the last recorded while life was still present) in cases in which there had been a previous thermometric variation of over 1°°5. 1904. | On Heat Regulation and Death Temperatures. 365 CHART 3. CHART 4. a Temperatures at time of death, after rise of 5 or more| oh , After fall of 5 ormore. | 70 [S28 29° 108 10lt 102° 105° 104" 105" 106" 107" 108" oF NO” Ii} PS >|_85° 96° 97° 96° 39° 100° 101" 102° 105° 104" lof 60 |—+- 60 50 | | 50 40 | 40 30 30 20 20 10 10 0 0 From these charts the following deductions are made :— 1. That there is very little difference between medical and surgical cases as regards the final temperature, whether it be preceded by a rise or by a fall. 2. That, when death rises occur, the final temperature is most frequently between 101° and 103°, but quite commonly ranges from 100—106°. 3. That, after death falls, the most common final temperature is between 97° and 100°, but falls as low as 95° are fairly frequent. The high temperatures which are recorded under the death falls are due to the previous rises of temperature, just as the low temperatures under the death rises are due to previous falls. The variations of the temperature of the preagonistic stage are considered in the following table with respect to the duration of the illness as estimated roughly by the length of the period between admission to the hospital and the death of the patient. This method by no means tells correctly the duration of the illness; but by no other way can it be estimated on account of the histories being so unreliable. It was decided in consequence to accept the above method only in surgical cases, as indicating fairly well in this class the state of afiairs :— Duration of stay in hospital. Rises. Falls. MOLAR e Ri Sore ea al (OE anes 2 Ia Dealenee Me eats. an See sean 12 | Splatter Ye Wey ica dey: USAGE sea) 2s 8 AN ey ReCSAN Dike | Se 21 bist week, 6 r ae wens By ebhasincv i Pe, nla a iets: dino 2360) | con ie pe epee te phere e's Kamen AnUe TAR I, 7 TN eek hee ACO ae 2 | COs ae uO MLO Geis epee. 29) 2nd week, 12) 2nd week, DOR vcas.) + pee at Ab a 34 63 10 22 A Deka vhs DIAS 80158 Shs ee 18 iliay cas) a OO Meter Ae. MOR gcc ot ake Ae 9 Leah ea LUTON PGi ago” bou3k eee 16 2,4 GR SS Aba PEW), ot ce 0 ane 0 Oren your! Ae wis. eed cami A pi ot LO 4 Over 4 __,, Pe NON ris saat od Get 2 366 Mr. E. M. Corner and Dr. J. E. H. Sawyer. [Apr. 4, Owing to the obvious errors that enter into the composition of such a table, it was decided not to continue it into details. The nature of the disease would be expected to affect the death changes of temperature far more than the mere duration of the illness would. But one point stands out in the above table, namely, that the shorter the illness the greater the number of cases in which variations of temperature are found. How far this is relative or absolute has not been calculated, for the reasons which have just been given. The greatest number of rises are found in short illnesses; but the corresponding statement, as founded on the above table, which is composed only of surgical cases, does not hold good for the falls of temperature, but, as will be seen later, there is distinct evidence that sudden depressions of temperature as death approaches occur frequently in diseases of long duration. The disease which brings about the fatal termination may be regarded as the most important factor in causing the changes of temperature which occur as death approaches. The influence of the disease upon the character of the pyrexia, when death is associated with a rise of the bodily temperature, has already been slightly indicated. The following diseases are those in which the greatest variations were found :— Medical Cases. Rise of 8°.—Acute yellow atrophy of the liver. 7°.—Septic meningitis, tubercular meningitis, typhoid fever (hemorrhage, preceded by a fall of temperature). 6°.—General tuberculosis, cerebral hemorrhage, — septic meningitis, peritonitis, marasmus and diarrheea. 5°.—Peritonitis; intussusception, preceded by a fall (two cases) ; general tuberculosis (three cases), pneumonia (two cases), diphtheria, chronic renal disease, carcinoma of the liver, intestinal obstruction, carcinoma of the small gut, hemiplegia, ulcerative endocarditis, acute yellow atrophy. The list of diseases for the smaller elevations of temperature is too long for reproduction. Two points are clearly seen in this table—that in severe toxic diseases a large rise of temperature is common, and that in heart diseases it is a rare affection. Surgical Cases. Rise of 10°.—Fractured base of skull. : 9°.—Fractured spine, fractured skull, burn. 8°.—Strangulated hernia and peritonitis, tetanus, pyzemia. 7°.—Septicemia, intestinal obstruction and _ peritonitis, fractured vault of skull, fractured base, burn (two cases), pyzmia, suppurative nephritis, septic broncho-pneumonia. (aa) F 1904. | On Heat Regulation and Death Temperatures. 367 Rise of 6°.—Peritonitis, various causes (five cases); meningitis (two cases), burn, scald, spina bifida, imperforate anus, gluteal abscess. From this list the following conclusions may be drawn :— 1. That injuries to the head and spine generally give rise to high death changes of temperature. This may be emphasised by the low temperature which precede the final rise and which are the result of the shock caused by the injuries. Such a condition is frequently found in those patients who succumb within 24 hours after admission to the hospital.* 2. That the following diseases are the most frequent in taking the first places among the death rises of temperature :— lesdenmries 10), 1s 918%. 1:75, 25°6°,.1; 5°, 5—L11 cases. Spinal injuries: 9°, 1; 6°, 1; 5°, 1—3 cases. ounep Oe i 2): 6 4) 3’ 610 cases, meade 6/1: 5; 1-2 cases. Meningitis: 6°, 2—2 cases. 3. Besides the above, death in those cases associated with a rise of temperatures is almost always due to some form of poisoning by septic -organisms. ‘Thus the remainder of the list of cases with a death rise of 5° or more, can be summed up as follows :—Peritonitis, septiczemia, pyzemia, septic meningitis, septic broncho-pneumonia, suppurative nephritis, cellulitis. It appears from this that, besides the injuries already mentioned, a septic process causes the death rise in almost all cases. The predominance of this process in causing death in surgical cases probably accounts for the difference in the numbers of the death variations in medical and surgical cases. Medical Cases. Fall of 5°.—Phthisis. 4°—Typhoid fever, hemorrhage (two cases); intestinal obstruction, broncho-pneumonia, cardiac failure, phthisis (two cases). 3°.—Diphtheria, mitral disease, cirrhosis of the liver, heemor- rhage, cardiac failure, bronchitis, myelitis, pneumonia, chronic renal disease (two cases), tubercular menin- gitis (two cases), phthisis (two cases). This list shows that large falls of temperature occur in diseases of long duration, such as phthisis, the absence of diseases of the nervous system and the comparative frequency with which cardiac affections is found are also striking facts. * Sawyer, “The Temperature of Coma,” ‘ Brit. Med. Journ.,’ Dec. 26, 1903. 368 Mr. E. M. Corner and Dr. J. E. H. Sawyer. [Apr. 4, Surgical Cases. Fall of 6°.—Intestinal obstruction, cyst of ovary, erysipelas. 5°.—Sarcoma of pelvis, carcinoma of rectum and obstruction, compound depressed fracture of vault, scald, septi- cemia, pyzemia, abscess. 4°,.—General peritonitis (two cases), pyemia (two cases), tubercular laryngitis and phthisis, epithelioma of tongue, sarcoma of tibia. 3°.—Injury to chest and bronchitis, fractured ribs, sarcoma of face and scalp, suppurative nephritis, erysipelas, pyeemia, sarcoma of neck, cellulitis, imperforate anus, burn, cut throat, ruptured gut. The general absence of injury and the great preponderance of sepsis are striking features in the above list. Some of the falls of temperature are dependent upon the shock following a surgical operation. The distribution as regards the age of the patients with variations of temperature is shown in the following two charts. CHART 5. CHART 6. ‘ Ages of Patients with Death Rises. ee of eae with Death Falis. feck So 5 ee eee es if aes al 20 aoe = ales i acca 00 me @ © (ey (2) Beeler AL —— at + ma | aN a Be. He | 3 os \g hast | A a a [| z “| = _ te 7 — © (ey (e) > al Ges \ ‘s [i PS \ \ Ce A 7 aca t.ho i | / The deductions, which may be made from Charts 5 and 6, are :— 1. That the variations in the temperature as death approaches are more frequent in children under the age of 5 years than at any other vee of life. 2. That between the ages of 5 and 10 years there is a marked decrease in the frequency of the variations of temperature. 3. That between the ages of 20 and 50, that is, during the most active period of life, variations of temperature more commonly occur than at any other period, except in children under 5 years of age. \ 1904.] On Heat Regulation and Death Temperatures. 369 4, That rises and falls of temperature are relatively of the same frequency to each other at all periods of life. IPaNiea JUD As in so many physiological controversies the various theories fall under two headings, firstly the inanimate, in which the question is examined from the point of view of chemistry, physics, etc., and secondly the animate, which deals with the phenomena of living tissues, so also in the subject of pyrexia there is found a similar condition of aifairs. The inanimate theories deal with the ratio of the production and the loss of heat to each other, the disturbances of which must lead to change of temperature. ‘The first question to be discussed is how the approach of death affects the interchange. (a) It seems hardly rational to expect that the capacity for the production of heat will increase as the body approaches death. The physiological processes of the body become less and less active until they cease. In diseases such as tetanus, In which tremendous con- vulsions take place, there must be an enormous production of heat, and yet in this condition there is usually no pyrexia. Towards death the frequency of the convulsive seizures may somewhat subside, and after this change there is often an elevation of temperature. Under such conditions the increase in temperature may be brought about by an excess of heat production over heat loss. The majority of the other cases of death rises of temperature cannot be accounted for in this manner. Again, curari causes muscular paralysis by suspending the activity of the neuro-muscular system and, as a result, a fall in the temperature of the body occurs; but during the convulsions which are first caused by the drug the temperature rises.* With regard to the falls of temperature at the approach of death, an explanation which is very tempting to urge, is that with the loss of the activity of the bodily processes a diminution also occurs in heat production. In support of this view may be put forward the fact that the greater the duration of the illness the more frequent is there a sudden depression of temperature. For, in diseases which are of long duration, the capacity of the organism for heat production is much more likely to become diminished than in illnesses lasting only a short time. (b) The heat loss as death'approaches must be diminished in almost all cases by the slowness of the circulation, shallowness of respiration, suppression of urine, etc. These changes would tend to cause a rise of temperature. Such an event;would naturally be expected to be more frequent when the jillness is short and the organs of heat production * Pembrey, ‘Text-Book of Phsiology,’ 1898, edit. by Schafer, vol. 1, p. 841; Bernard, ‘ Lecons sur la Chaleur Animale,’ 1876, p. 157. 370 Mr. E. M. Corner and Dr. J. EK. H. Sawyer. [Apr. 4, are not worn out by work under a prolonged strain. And, as has been pointed out above, if the disease should be of long duration, a fall of temperature rather than a rise would be expected to occur. The reason for this supposition is that as death approaches in long illnesses, there must presumedly be a considerable diminution of heat production, which is not counter-balanced by the small reduction of the heat lost at the surface. The investigations in the first part of this paper support these ideas, since they seem to show that rises in the bodily temperature occur more frequently in short diseases, and falls of temperature in those of longer duration. The animate factor in pyrexia is chiefly that of the action of the central nervous system. ‘The nervous system exercises a control upon the loss of heat by means of the vaso-motor system, which regulates the amount of blood in the deep and the superficial parts of the body, and by the respiratory centre which controls the frequency and depth of respiration; upon the production of heat through the nerves which control the activity of the tissues, chiefly the muscles.”* The heat production in a tissue is probably not under the control of that tissue itself, but its thermogenetic function is governed by its proper segment of the spinal cord. The nervous centres cannot of themselves produce heat ; they can only govern the manufactories. As the brain centres exercise a tonic control over the spinal centres which are in connection with the reflexes, sphincters, etc., so, in a similar manner, the higher centre or centres of the brain may hold in check the lower centres in the spinal cord, which give the tissues their power of heat production. The brain centres are the last to be evolved in the history of animal life, and it may be urged that they are the most complex, and therefore the most easily thrown out of gear. For this reason a generally acting death agent, such as a toxic condition, will affect and inhibit the action of the higher centres before or toa sreater degree than it will damage the lower. Whenever the higher centre is cut off from the lower, the latter becomes exaggerated in its action. This is well seen in the spastic condition when the reflexes etc., are exaggerated. In a similar way, with the commencement of dissolution of the higher centres, a death elevation of temperature may be expected to result. The fallacy in reasoning from the above analogy is that, in the one case, the facts deal with a reflex centre, and in the other, as far as is understood, with an automatic. How far these differ in regard to their relation to the higher centres, it is impossible to say, but one can instance the acceleration of the automatic and rhythmic centre of the heart when the inhibitory control of the vagus is cut off. An important fact which this investigation shows is that, as death approaches, there is a tendency for a sudden rise in the bodily * Pembrey, Schifer’s ‘ Text-Book of Physiology,’ 1898, vol. 1, p. 854. 1904. ] On Heat Reyulation and Death Tenvperatures. 371 temperature. In about 2,500 medical and surgical cases, an elevation of temperature of over 1°°5 was found in 26 per cent., or just over a quarter of all the cases, while a fall occurred in 8 per cent. only. The percentage of rises of temperature in surgical cases 1s 37, and in medical 15. It is a most remarkable fact that this sudden elevation of temperature is observed to take place so frequently in surgical cases (in over a third of all the cases examined), and that this change should be found to occur more than twice as often in surgical as in medical cases. As the surgical diseases in these records are of shorter duration, as a rule, than the medical, the tissues In consequence have not been so long exposed to abnormal conditions, and so their heat producing functions are less likely to be impaired. Again, falls of temperature are proportionally much more common in patients dying from disease than from injury. It may be presumed, therefore, that the loss of control of the nerve centres in an exhausted organism only occasionally results in an elevation of the bodily temperature, whereas with less exhausted tissues an increase in the animal heat occurs. From these considerations it would appear that heat production is constantly in excess, and that in consequence of this the organism must exercise some tonic control over the process, in order to keep its temperature at a constant level. The idea of the thermogenetic control of the higher centre over the lower gives point to the modified and accepted view of Liebermeister that, ‘‘in consequence of the injurious action of the fever-producing cause, the organism loses its power of keeping itself at the normal temperature.”* The poison will, unless it has special affinities, affect the higher and the more complex centres before the lower. Hence, the “spastic” over-production of heat which may result in fever. The great frequency of death rises of temperature in cases of head injuries, some spinal injuries, meningitis, brain diseases, etc., emphasises the possible cutting off of the controlling function of the higher centres. On the other hand, it should be remembered that many of the patients had pyrexial temperature charts, it may be for some days before the preagonistic variations of temperature occurred. The regulation of the bodily heat in these cases was already partially out of the control of the nervous system, and the further elevation of temperature during the last 12 hours of life may be thought to be due to an increasing loss of this control on the part of the organism. In small animals, after section of the spinal cord in the cervical region, the temperature falls rapidly ; in larger animals, such as dogs, if kept in an envelope of non-conducting material in an ordinary room, the temperature of the body rises to above that of fever, but without clothing the temperature rapidly falls until the animal dies. * Burdon Sanderson, Allbutt’s ‘System of Medicine,’ vol. 1. ue Mr. E. M. Corner and Dr. J. E. H. Sawyer. =‘ [Apr. 4, in collapse.* In man, according to Sir Benjamin Brodie, section of the spinal cord in the cervical region causes pyrexia, but very dis- cordant results have been obtained by other observers on cases with similar injury to the nervous system. Dr. Pembrey’s explanation of _these contradicting results seems to be the correct one. He says: “The section of the spinal cord high up in the cervical region abolishes the power of regulating temperature. When the patient is exposed even to moderate cold, his temperature falls owing to the increased loss of heat and to the diminished production of heat. On the other hand, if the weather be hot and the patient be too well covered with bed clothes, his temperature rises and may reach a dangerous height, owing to the diminished loss and the increased production of heat in the body. In the paralysed man the production of heat rises and falls with the external temperature.”+ And Sir John Burdon Sander- son writes: ‘“‘Section only shows the abnormal facility with which the body yields to the influence of outside conditions.” { It might be urged that the variations of temperature, just before death, are due to a similar condition, but obviously this cannot be the case, for after section of the spinal cord many other factors arise which are not present in those patients who suffer from no such lesion of the nervous system. After section, the respiratory movements are altered in character, and respiration is entirely performed by the action of the diaphragm, and in consequence there is less loss of heat through this channel. Again, the muscles are paralysed, and, therefore, cannot produce the normal amount of heat, while the sweat glands are no longer active, and thus less heat is lost by the evaporation of moisture from the external surface. For these reasons alone the death varia- tions of temperature found in the patients who are considered in the first part of this paper are in no way analogous to the changes which occur after section of the spinal cord in the cervical region. Arguing from the supposition that the higher centres have a tonic control over the lower, it 1s to be expected that stimulation of the higher centres should lead to a still further control and diminution of the action of the lower. In this way falls of temperature may be caused. How far a toxic agent will stimulate it is difficult to say. It is possible that some antipyretic drugs may act in this way, such as quinine and salicylic acid. Other substances seem too powerful to stimulate and would appear to paralyse the higher centres. Smaller doses of the poison, however, may act as a stimulant, first to the higher and then to the lower centres, the former being affected before * Burdon Sanderson, ‘‘On the Process of Fever,’ ‘The Practitioner,’ vol. 16, 1876, p. 426. + Pembrey, Schiafer’s ‘ Text-Book of Physiology,’ 1898, vol. 1, p. 862. + Burdon Sanderson, ‘‘ On the Process of Fever,” ‘The Practitioner,’ vol. 16, 1876, p. 428. 1904. | Heat Regulation and Death Temperatures. 373 the latter. It may be that for this reason, there is a slow rise of temperature in some diseases; while in others of more severe onset, there is a sudden elevation of temperature, which may be due to paresis of the higher centres. Septic conditions show frequently changes of death temperature, which may be the result of :— 1. Stimulation of the higher centres over the lower, producing fall of temperatures. ; 2. Paralysis of the higher centres over the lower, producing rise of temperatures. | 3. Simultaneous paralysis of higher and lower centres producing no change. 4. Special poisons may affect higher and lower centres differently. There seems to be some evidence, therefore, in favour of the view that a higher centre in the brain controls the thermogenetic centres in the spinal cord, but it does not follow, when there is an increased heat production through the removal of this control, that there must necessarily be pyrexia. An increase in the loss of heat may keep the temperature at the normal level. In patients, however, who are dying, there is a tendency for the amount of heat lost to be diminished, the evidence for which has been pointed out above. Besides this natural tendency, means are constantly employed to prevent falls of tempera- ture in the failing, by increasing the amount of bed-clothes, by using hot bottles, by bandaging the limbs before operation, etc. This slight diminution in the amount of heat lost may be able to prevent the temperature of the body falling, but it would hardly be sufficient to raise it many degrees without some increase in heat production. From these considerations it would appear that the variations in bodily temperature as death approaches must be dependent in many instances upon the increased amount of heat production and not upon the dimi- nution of heat lost. As these variations are so constant, there is an indication that, although the centre which controls thermolysis, heat- loss, may be the chief factor in keeping the body at a normal tempera- ture, yet the centre which controls thermogenesis plays a more important part than has lately been attributed to it. The thermolytic centre probably has an inhibitory effect upon the vasomotor centre, 7.¢., over vaso-constriction, in a similar manner, we think, to the action of the upper thermogenetic centre upon the lower in the spinal cord. As the lower: thermogenetic centres tend to exaggerate the amount of heat produced in the tissues, so the vaso- motor centre by constricting the vessels tends to diminish the amount of heat being lost; the latter effect being counterbalanced by the inhibitory action of the thermolytic centre. The action both of the lower thermogenetic centre and also of the vasomotor centre, when not controlled by the higher centres, seems to be that of raising the bodily temperature. VOL. LXXIII. 2D 374 Heat Regulation and Death Temperatures. [ Apr. 4, Whatever effect the fever-producing agent has upon the controlling power of the higher thermogenetic centre, it will probably act in a similar manner upon the thermolytic centre. If this should be the case, paralysis or weakening of the functions of these two centres would result in a much higher elevation of temperature than if the thermogenetic controlling centre in the brain alone were affected. _ The following diagram shows the hypothetical relationship of the different nerve centres which take part in the regulation of the bodily temperature. The afferent path from the surface to the vaso-motor centre has been left out for the sake of clearness. A B Thermolytic Centre (inhibitory to D). Vaso-motor Centre. Thermogenetic Centre atte Secret h alo Thermogenetic @ Centre (automatic) If A or B be paralysed or weakened, there is a tendency to pyrexia; if A and B be paralysed or weakened, there is a tendency to hyper- pyrexia ; if C or D be paralysed or weakened, there is a tendency to a fall of temperature. If A or B be stimulated, there is a tendency to a fall of temperature ; if A and B be stimulated, there is likely to be an extreme fall of tem- perature ; if C or D be stimulated, there is a tendency to pyrexia. From these considerations, and from the facts which have been elicited by a careful study of the death variations of temperature, the following theory for the causation of pyrexia seems to present itself. Pyrexia is due to two factors, to an augmented production of heat owing to the activity of the thermogenetic centres in the spinal cord being no longer perfectly controlled by the higher centre in the brain, and to a diminished loss of heat owing to the weakening of the functions of the thermolytic centre; the power of the two higher centres being weakened or paralysed by the morbid products or 1904. ] On the Action of Radium on Micro-organisms. 375 toxines of the affection from which the organism is suffering. In other words, normal temperature is preserved by a mutual see-saw action of these centres—the thermogenetic and the thermolytie. We recognise fully that, for a more perfect understanding of death temperatures, it is necessary for the observers to examine the patients for themselves, and not to trust to records, however many or accurate they may be, so that they can note in each case the changes in the skin, the circulation, the respirations, etc., concurring with the varia- tions of the bodily temperature. Nevertheless, we venture to put forward our investigation and views, not as physiologists, but as clinical observers, with the hope of pointing out new lines of research, by which may be increased the knowledge of the regulation of animal heat. “ A Note on the Action of Radium on Micro-organisms.” By ALAN B. GREEN, M.A., M.D. (Cantab.). Communicated by Sir Micuarn Foster, K.C.B., F.R.S. Received April 11,— Read May 5, 1904. [Puate 11.] The radium salt used in these experiments was 1 centigramme of radium bromide, bought of Messrs. Buchler and Co., of Brunswick, in June, 1903. It was. contained in a vulcanite and brass capsule, fronted with thin tale. The radium was enclosed immediately behind the tale, and the circular area over which it was spread was about 3 mm. in diameter. The radium emanations which were applied to micro-organisms were such as passed through the talc, 7.¢., the @ and y rays. Dr. E. F. Bashford, to whose kindness I am greatly indebted for the use of the radium, has informed me that Sir Wiliam Ramsay tested the preparation for the intensity of its combined f and y rays, the latter being a practically negligible quantity. The results showed that, on comparison with samples of radium bromide giving a virtually pure spectrum of radium, these rays were practically 100 per cent. This radium salt was, in fact, a pure preparation of radium bromide. Dr. Bashford also informs me that this radium bromide caused pig- ment to disappear after 18 days from a mole with 15 minutes’ exposure, the tale being in contact with the surface of the mole. Five minutes such exposure produced a marked skin reaction, while 20 minutes’ exposure caused a reaction proceeding almost to ulceration. I found that the radium was itself luminous, and that it could 2 De 376 Dr ADB. Green. [Apa tt, illuminate a screen of zinc sulphide through a sheet of lead over 1 cm. thick. It discharged a gold leaf electroscope, highly charged with +or-— electricity, at a distance of over 6 feet. It caused a brown colouration of glass or tale when applied at a distance of 1 mm. for 12—24 hours. The experiments which have been made are of two kinds. In the first set investigation was made of the germicidal action of radium emanations, and in the second, endeavours were made to ascertain whether micro-organisms exposed to the emanations became thereby themselves radio-active. | I.—The Germicidal Action of Radiwn Emanations. The following have been subjected to the emanations of radium :— (a) Calf vaccine, together with its contained extraneous bacteria, which in these experiments consisted of S. pyogenes aureus, S. pyogenes wlbus, S. cereus flavus, S. cereus albus. Both freshly collected and stored calf-vaccine pulp were exposed to radium emanations in the following way :—A layer of pulp not exceeding 0°5 mm. in thickness was spread in the centre of the depression of a hollow-ground glass slide, around the circumference of which depression a metal ring had been cemented. The capsule containing the radium bromide was placed upon the metal ring in such a way that the salt was brought within 1—2 mm. of the lymph pulp, nothing separating them but the tale of the capsule and the intervening air. The pulp was thus used in the thinnest practicable layer, in order that the emanations might act as uniformly as possible on all the component parts. The vitality of the vaccine and of its extraneous bacteria was tested before, and at varying intervals of time after the exposure to radium ; the former by inoculations on calves, the latter by cultivations on nutrient media. For each such test a small portion of vaccine was removed from the preparation and was mixed with enough sterile water to form a semi-fluid emulsion. A loopful of emulsion was used to inoculate a liquefied tube of nutrient agar-agar, and a plate was established in the usual way. The remainder of the emulsion was used for inoculating a calf. (b) The following species of micro-organisms have also been separately subjected to the action of radium:—sS. pyogenes aureus, S. pyogenes albus, S. cereus flavus, S. cereus albus, Streptococcus pyogenes, B. prodigiosus, B. proteus vulgaris, B. pyocyaneus, B. typhosus, B. colt communs, B. maller, B. pestis, B. tuberculosis, the bacillus of Malta fever, Spirillum cholere Asiatice, and sporing cultures of B. mesentericus vulgatus, Bb. mesentericus ruber, B. subtilis, B. anthracis, B. tetani, Gartner’s bacillus, the bacillus of malignant cedema, and the bacillus of Rauschbrand. 1904.] On the Action of Radium on Micro-organisms. 377 In the case of these micro-organisms, growth was removed from solid media and was exposed to the radium emanations as a film in the depression of a hollow-ground slide, the radium being applied as in the case of vaccine. The vitality of these micro-organisms was tested both before and after exposure to radium by cultivations on media by a similar method to that used in the case of vaccine. In some instances metal rings were placed round colonies of bacteria m situ on the surface of nutrient medium, and exposure of bacteria under these conditions made by placing the capsule containing the radium on the rings. In the case of each experiment with vaccine and with micro- organisms in pure culture, a control was carried out by making a similar preparation and subjecting it to similar conditions as the experimental preparation with the exception of exposure to radium ; it was also subjected to tests for vitality similar to those used for the experimental preparation, at corresponding intervals of time. All experiments and their controls were made at room temperature. The exposure of micro-organisms in liquid media to radium was found unsatisfactory owing to the presence of material of a complex nature between the radium and the micro-organisms, and owing also to the constant variation in the distance between the radium and the micro-organisms suspended in the liquid. It was found from the foregoing experiments and their controls that a marked germicidal action was exerted on the specific and extraneous micro-organisms of vaccine and on the other above-mentioned micro- organisms as a result of their exposure to the radium at a distance of 1—2 mm. for varying lengths of time. The following is a summary of the results of these experiments and their controls :— fesults of Huperuments with Vaccine. The specific germ in no case survived a longer exposure to radium than 22 hours, at the end of which time it had completely lost its ability to cause vesiculation or any visible irritation at the site of inoculation on a calf. In seventeen out of a total of twenty-five experiments its potency was destroyed after 10 hours’ exposure to radium and in four cases after 2 hours. The controls remained fully potent after the experimental vaccines had been rendered inert. The extraneous micro-organisms of these vaccines, as has been previously mentioned, consisted of S. pyogenes aureus, S. pyogenes albus, S. cereus flavus, S. cereus albus. In each experiment these bacteria were destroyed after exposure to radium in rather less time than was the potency of the specific germ. In no case did they survive a longer exposure to radium than 15 hours. The extraneous micro-organisms of the control vaccines were alive 378 Dr. A. B. Green. [Apr. 11, after those of the experimental vaccines had been killed. The following experiment may be related to explain in greater detail this action of radium. Experiment. On November 12, 1903, vaccine pulp was exposed to radium at 10 a.m. Portions of this pulp were removed from the influence of the radium at the end of 2, 6, and 10 hours’ exposure. Liquefied nutrient agar-agar tubes were inoculated with one platinum loopful of each portion immediately,after its removal from the radium, and plates were established in the usual way. The remainder of each portion was inoculated on a calf on the following day, November 13. Control. Plates were similarly established from the control at 10 a.m. and 8 P.M. on November 12, and the remainder of the vaccine was used for inoculating the calf on the following day, November 13. On November 18, the experimental portion of vaccine exposed to radium for 2 hours had caused good vesiculation on the calf; the portion exposed for 6 hours caused very poor vesiculation, and the remaining portion exposed for 10 hours caused no trace of vesiculation. The number of extraneous bacteria originally present in the vaccine were 1200 per platinum loopful of emulsion (the mixture of this emulsion has been previously described). In the portion of vaccine exposed to radium for 2 hours the number left alive was 1050 per platinum loopful; in the portion exposed for 6 hours there were fifty bacteria, and in the portion exposed for 10 hours there was no evidence of living bacteria at all. On November 18 the control portion of the vaccine had caused good vesicula- tion; and the agar-agar plate poured at 8 P.M. on November 12 contained practi- cally the same number of colonies of extraneous bacteria as were present in the plate poured from the same vaccine at 10 A.m. on the same date. Results of Experiment with Non-sporebearing Bacteria. All the non-sporebearing bacteria previously mentioned were killed after exposure to radium for 2—14 hours. A description in detail of the results of some of these experiments may be of use in illustrating this germicidal action. One of a Series of Experiments with 8. pyogenes aureus. Before exposure to radium, plate cultivations showed 84,000 bacteria present per platinum loopful of emulsion. After exposure to radium for 6 hours, this number had decreased to 31,000; at the end of 10 hours’ exposure to 260, and at the end of 14 hours’ exposure no bacteria were left alive. At the end of 14 hours the control preparation showed bacteria alive in practi- cally undiminished numbers. One of a Series of Experiments with B. coli communis. Before exposure to radium, 75,000 bacteria were present per platinum loopful of emulsion. After 3 hours’ exposure this number was reduced to 3000, and after 6 hours’ exposure all the bacteria were killed. 1904. ] On the Action of Radium on Micro-organisms. 379 The control at this time showed practically the same number of bacteria as were present originally. One of a Series of Experiments with Spirillum cholerx Asiatics. Before exposure to radium, 47,000 bacteria were present in a platinum loopful of emulsion. After 3 hours’ exposure to radium, 2100 were left alive per platinum loopful, while no bacteria survived an exposure of 6 hours. The control preparation showed at this time practically no decrease in the number of bacteria originally present. Results of Experiments with Bacteria containing Spores. Bacteria containing spores were by far the most resistant to the germicidal action of radium of any micro-organisms used in these experiments, for they were not killed by less than 72 hours’ exposure. This corresponds with the time given by R. Pfeiffer and HE. Friedberger* as necessary for the killing of spores by the emanations of the radium used by them. The following are examples of experiments with these micro- organisms :— One of a Series of Experiments with B. mesentericus vulgatus (Sporing). Immediately before exposure to radium the preparation showed 170,000 micro- organisms per platinum loopful of emulsion. After 48 hours’ exposure this number had decreased to 260, and at the end of 72 hours all micro-organisms per platinum loopful had been killed. The control preparation showed practically no decrease in the number of micro- organisms at the end of 72 hours. One of a Series of Experiments with B. anthracis (Sporing). There were originally present 11,000 micro-organisms per platinum loopful. After 48 hours’ exposure to radium, 120 only were left alive, and at the end of 72 hours all micro-organisms per platinum loopful were killed. The control showed micro-organisms present in undiminished numbers at the end of 72 hours. One of a Series of Experiments with B. tetani (Sporing). In the experiments with tetanus spores the actual numbers of micro-organisms were not investigated, but only the presence or absence of living germs by means of cultivations in the depth of sugar agar. After 48 hours’ exposure to radium, the presence of living micro-organisms was still evidenced, but after 72 hours’ exposure no growth followed the inoculation of a sugar agar tube. Thus in these experiments the non-sporebearing bacteria exhibited the least resistance to the germicidal action of radium emanations, withstanding exposure for 2—15 hours only. The resistance of the specific germ of vaccine was slightly in excess * “Berl. Klin. Woch.,’ July 13, 1903. 380 IDrvA. Green: [Apr. 14, of this; while by far the greatest resistance was shown by spores, these not being killed by less than 72 hours’ exposure. Experiments have also been made from which the following points have been noted. 1. As the distance between the radium and the micro-organisms subjected to its emanations was increased, the germicidal action which was marked at the nearest distance became less evident and finally ceased to be exerted. In these experiments Staphylococcus pyogenes aureus was used, a separate strain being used for each series of experiments. Portions of growth were subjected to the radium emanations for the same time and under the same conditions, except that the distance between the radium and the bacteria was varied. After 30 hours’ exposure it was found that— At 1 mm. bacteria were killed. At 1 cm. bacteria were usually lessened in numbers, but all were not killed. At 10 cm. no definite germicidal action was apparent. The following is an example :— } Experiment. Staphylococcus pyogenes aureus exposed to radium emana- tions for 30 hours at different distances. Number of bacteria present per Number of bacteria platinum loopful of emulsion after 9 5 originally present per 30 hours’ exposure to radium. erles. ane platinum loopful of emulsion. At 1 mm. At 1 cm. At 10 em. 1(a@) 1060 0 =: I UD) Res ia. 1700 50 160 oe IN(@)) 505k 1200 Se a 987 | 2. As extra thicknesses of mica or glass were interposed between the radium and the micro-organisms exposed to their influence, the time of germicidal action was delayed. Finely woven copper gauze also caused slight delay of germicidal action. A sheet of lead 0-1 mm. thick, placed between the radium and the micro-organisms, caused weakening of germicidal action, and as extra thicknesses of lead were interposed and the B-rays were cut off, germicidal action became less and less evident. Il.—Induced Radio-Activity of Bacteria. It has been found that after exposure at a distance of 1 mm. to the radium emanations for 24—120 hours, micro-organisms themselves may | A. B. Green. Roy. Soc. Proc., vol. 13, iate 1M Fic. 1.—Photograph of a mass of B. mesentericus vulgatus (sporing), made radio- active by exposure to radium bromide for 72 hours and killed by the exposure. Fig. 2.—Photograph, taken through a double layer of lead-foil, of a mass of B. mesentericus vulgatus (sporing), made radio-active by exposure to radium bromide for 72 hours and killed by the exposure. 1904. | On the Action of Radiwm on Micro-organisms. 381 show signs of radio-activity. It has not yet been ascertained whether living micro-organisms can exhibit induced radio-activity, but micro- organisms which have been killed by exposure to the radium emanations can do so. . | In these experiments no radio-activity has been found in bacteria not exposed to the action of radium. | Induced radio-activity of micro-organisms has been shown in the following manner :—A small mass of bacteria, removed from the surface of nutrient medium, after subjection to the radium emanations at a distance of | mm., for, as a rule, 72 hours, was removed from the depression of the hollow-ground glass slide in which it had been exposed to the emanations, and was placed between two thin sheets of glass, generally coverslips, which were not themselves radio-active. These sheets of glass, with the small mass of bacteria pressed between them, were next, in a dark room, brought into contact with the film of an Ilford “ special rapid” photographic plate. A cotton-wool pad was placed on the glass sheets to keep them in position, and the whole was wrapped up in a lght-proof package. Twenty-four hours later the photographic plate was developed and a photograph was obtained of the bacterial mass. An image has been developed after only 1 hour’s exposure of a sensitised plate to the radio-active bacteria, and in some instances after a fortnight’s exposure. Faint images have been thus produced on sensitised plates by S. p. aureus and albus which had been subjected to radium emanations, but, so far, the best photographs have been obtained from bacterial masses containing a.number of spores, after their subjection to the emanations for 24—120 hours (Plate 11, fig. 1). Radio-active micro-organisms have continued to give off photo-actinic emanations after 3 months have elapsed since their exposure to radium. Photographs of masses of micro-organisms, possessing induced radio- activity, have been obtained,through a double layer of lead foil (Plate 11, fig. 2). A sheet of lead 3 mm. in thickness interposed between a radio-active bacterial mass and the sensitised plate has prevented the _ passage of photo-actinic emanations from the glass to the plate. These photographs would seem, therefore, to be caused by P-rays emitted by the micro-organisms. My best thanks are due to Mr. Power, Medical Officer of the Local Government Board, for the facilities he has afforded me in the research, and for the kind and valuable advice he has given me concerning it. 382 Prof. B. Moore and Mr. H. E. Roaf. [Apr. 12, “On certain Physical and Chemical Properties of Solutions of Chloroform in Water, Saline, Serum, and Hemoglobin. A Contribution to the Chemistry of Anesthesia.—(Preliminary Communication.)” By Brengamin Moors, M.A., D.Sc., John- ston Professor of Bio-chemistry, University of Liverpool, and Herpert E. Roar, M.B., Toronto, Johnston Colonial Fellow, University of Liverpool. Communicated by Professor C. 8. SHERRINGTON, F.R.S. Received April 12,—Read May 5, 1904. The number of substances which have been shown to possess more or less well-marked anesthetismg properties reaches some hundreds, and hence it is obvious that the action cannot have a different explanation in each case, but rather depends upon some general type of interaction between the anesthetic and the active part of the cell, which is the cell-protoplasm. Further, the action occurs not only with nerve-cells, but with ciliated and other epithelial cells, with muscle-cells of all types, with bacteria, amoebee, and other unicellular organisms, and with all types of vegetable cells in which activity is suited to experimental demon- stration. In all these varied types of living cell, activity decreases alike with increasing dose of the anesthetic, and, with sufficient concentration, all sign of life becomes obliterated. Hence the action of the anesthetic must be due to some change brought about in the only material which is uniformly present in all these types of cell, that is, the cell-protoplasm. Accordingly, in briefly reviewing, as an introduction to our experi- ments, the previous theories of anzsthesia which have.been advanced by various observers, we believe we may justly cast aside those which attribute it fundamentally to anything peculiar in the structure or chemical composition of the nerve-cell, or to any alteration in the nutrition of the nervous system, brought about by variations in its blood supply or otherwise. It is true that cells differ in the degree of their reaction to anesthetics, but not in kind, and ultimately the metabolic processes of bacteria are stilled as effectually as are those of the mammalian nerve-cell. Any such effects as anzmia or hyperemia of the brain, which have been alternately described by various observers, must accordingly be only set down as secondary effects, and not as primary causes of anesthesia. Similarly, theories which are based on the peculiarly high content in cholestearin, lecithin, and fatty derivatives soluble in ether, of the nerve-cell and its processes, cannot furnish an explanation of anzesthesia, 1904.] Properties of Solutions of Chloroform mm Water, etc. 383 for these substances are not present in demonstrable quantity in by far the greater number of animal and vegetable cells. Turning to the views of anesthesia which rest upon an interaction between the anesthetic and the cell-protoplasm, we find the speculation first thrown out by Claude Bernard* in 1875, that anesthesia consists in a semi-coagulation of the substance of the (nerve) cell, a coagulation which may not be definite, that is to say, in which the substance can return structurally to its primitive state after elimination of the toxic agent. Bernard supports his view chiefly by analogy, and instances the stiffening and opacity of skeletal muscle when exposed to chloroform vapour. A similar view was expressed by binz,t who stated that sections of cerebral cortex placed in 1-per-cent. solution of hydrochlorate of morphia soon showed a cloudy appearance, and fine granules appeared in the nuclei; the protoplasm also became granular. The stage at which the cell-protoplasm was merely cloudy, and not discretely granular, could be recovered from by washing away the morphia, but, when once the granules appeared, they could not be made to dis- appear again. Similar results were obtained by exposing cortical ner've-cells to vapour of chloroform, or to solution of chloral hydrate. Neither Binz nor Bernard showed, however, that there was any precipitation or senii-coagulation at or near the concentrations which correspond tu anesthesia, nor were the optical methods used capable of demonstrating effects upon the protoplasm short of precipitation. Similar speculations of a general nature regarding the action of toxic agents, as being due to the formation of a loose, easily-dissociated compound between the toxic agent and the cell-protoplasm, have been thrown out by various writers, as, for example, Buchheim (1856) and Schmiedeberg (1883), Demoor{ has shown that, subsequent to prolonged and deep anesthesia, the dendrites of nerve-cells acquire moniliform swellings, and has founded on this a mechanical theory, which rests on the view that the swellings observed are due to a retraction of the protoplasm of the dendrites, so that the communication of cell with cell is inter- rupted. The swellings described by Demoor have also been observed by Hamilton Wright,§ who also found that they became larger, more numerous, and encroach more and more on the dendritic stems the longer the anesthesia is kept up. These effects are of importance as evidence of an interaction between ' protoplasm and anesthetic, but the retraction theory of Demoor based * “Lecgons sur les Anesthésiques,’ etc., 1875, p. 153. + ‘ Vorlesungen tiber Pharmakologie,’ p. 175. ft ‘ Arch. de Biol.,’ 1396, vol. 14. § ‘Journ. of Physiology,’ vol. 26, 1900, p. 30; vol. 26, 1901, p. 362. SATE SEDER EEE SEES ENE SU nase BSS OL PE gD SE ETL SRNODE ee ee a é 384 Prof. B. Moore and Mr. H. E. Roat. [Apr. 12, on them will not hold in view of more recent work on the fibrillar mode of communication between cell and cell. Further, as pointed out above, any adequate view as to how anesthetics produce their effect must be applicable also to the unicellular organism, and not merely to the nerve-cell, or any colony of cells. Wright obtained further effects as a result of prolonged anesthesia, which led him to adopt the view that the action is bio-chemical in character. He found, for example, that the Nissl’s bodies lost their affinity for basic dyes, such as methylene blue, but that this effect was only temporary, and disappeared as soon as the anesthesia had passed off. These changes, however, do not begin to appear immediately the aneesthetic commences to produce its effect, and are rather a signal of the changes produced in the protoplasm in marked degree by the prolonged action of the anesthetic than an indication of the first reaction between cell-protoplasm and anesthetic. Our own attention was first attracted in this direction by witnessing the experiments of Sherrington and Sowton* upon the effects of chloroform on the excised mammalian heart, fed by a current of Ringer’s or Locke’s solution, and through which later a similar current, but containing in addition small amounts of chloroform, could be perfused. These authors observed that a concentration of chloroform in the Locke solution, amounting to only 1 in 100,000, produced a marked and unfailing action in diminishing the extent of the cardiac contractions, and further that this effect appeared rapidly after the dilute chloroform solution reached the heart, lasted just as long as the chloroform in this excessively low but yet adequate concentration was passed through, and ceased almost immediately as soon as the normal Locke’s solution was recurred to, the heart attaining again its normal force. This effect could be repeated as often as was desired, and there was no cumulative action whatever, that is, no matter how prolonged the passage of the chloroform solution on passing back to the normal Locke’s solution, the chloroform effect rapidly disappeared, and again recovered when the chloroform was once more turned on. It was this latter effect which suggested the experiments on the chemistry of anesthesia recorded in this paper. It was quite obvious that the effect of the chloroform upon the cardiac muscle fibres depended solely upon the concentration (solution tension, or osmotic pressure) of the chloroform in the cell for the time being, and not at all upon the total amount of chloroform which had been fed to the heart up to the moment of observation. This experimental fact suggested the view that the effect upon the * Thompson Yatee, ‘ Johnston Laboratories Reports,’ vol. 5, Part 1, p. 69. j . 1904. ] Properties of Solutions of Chloroform in Water, ete. 385 cell was due to some combination being formed between the proto- plasm and the chloroform, and further that this combination was not a stable fixed one, leading to permanent removal of the protoplasmic activity, but an unstable one, which existed only so long as the pressure of the anesthetic was kept up to a definite level, and gradually dissociated as the level of chloroform pressure was allowed to fall, and as a result left the protoplasm free for a renewal of its metabolic processes—to choose a familiar analogy, that the protoplasm of a cell undergoing anesthetisation entered into a combination with the anesthetic, similar to that between hemoglobin and oxygen, unstable in character, and only lasting so long as the pressure of the anesthetic was kept up. t occurred to us, as protoplasm is built up chemically of proteid, that a certain amount of evidence as to the formation of such an unstable compound might be obtained, in the first instance, by experimenting with proteids. We accordingly experimented with the proteids of the blood, and have obtained a number of results which together point to the formation of such compounds as are indicated above. It is our intention to proceed further and study in a similar fashion the effects of chloroform upon various types of living cell, but we here present the work done upon proteids, which appears to us to prove that an easily dissociable compound is formed between proteid and chloroform. Our experiments may be described under the following headings :— 1. On the obvious physical and chemical changes produced in serum and in hemoglobin solution by the addition of chloroform. 2. On the relative solubility of chloroform in water, normal saline solution, serum, and hemoglobin solution. 3. On the relative vapour pressures of chloroform when dissolved in water, saline, serum, and hemoglobin solutions respectively, and on the variations in the coefficient of distribution in these solutions. 4. On the solubilities of gases in serum and hemoglobin solution in presence of chloroform. I.— Effects of Chloroform on Serum and on Hemoglobin Solution. On adding chloroform* to either serum or hemoglobin solution, and allowing the mixture to stand, changes occur which are obvious to the eye, and were to us previously unknown, but on consulting the literature we found that they had been observed by HE. Salkowskif in * The chloroform used for all the experiments described in this communication was presented to us by Messrs. Duncan and Flockhart, of Edinburgh. t+ ‘Deutsche Med. Wochensch.,’ 1888, No. 16; ‘ Zeitsch. f. Physiol. Chem..,’ vol. 31, 1900, p. 329. The fact that the red blood corpuscles combine with chloro- form is also mentioned by Schmiedeberg, ‘ Arch. f. Heilkunde,’ 1867, p. 278. a 386 Prof. B. Moore and Mr. H. E. Roaf. [Apr TZ: using chloroform as a preservative for these fluids, and were also described by Formanek,* the bearing of such phenomena upon the question of anzesthesia was not, however, appreciated by these previous observers, who had approached the matter from a different standpoint, and as we have in some respects amplified their observations, and in others have obtained results not quite in accord with theirs, we feel justified in here recording our experiments. It was observed by EH. Salkowski in 1888 that blood could not be preserved by adding chloroform, because it gradually became converted into a thick mass. In 1891 it was observed by Horbaczewskit that hemoglobin was precipitated from a solution containing it, and kept at a temperature of 40—50° C., to which chloroform was added as a preservative. The subject was investigated more minutely by Formanekj in 1900, and this observer found that a solution of hemoglobin kept at 50—55° C. for some time with chloroform was completely precipitated, the filtrate being entirely free from hemoglobin. Formanek dried and analysed the precipitate, and from the absence of chlorine after fusion with sodium carbonate and potassium nitrate came to the conclusion that the precipitate is not a chloroform com- pound of hemoglobin. In our opinion this is not a valid proof, as the chloroform need not be so stably combined with the chloroform as to stand drying at 130° C., and subsequent fusion as employed by Formanek. The precipitate was dissolved by Formanek after thorough’ washing to remove the chloroform by the addition of a few drops of sodium carbonate solution, and the solution gave the bands of oxy-hemoglobin, and on treatment with ammonium sulphide reduced hemoglobin. Formanek also found that blood serum and egg- albumin were precipitated (when the reaction of the fluid was acid or neutral) on keeping at a temperature of 50—55° C. in presence of chloroform. From these experiments this observer came to the conclusion that the precipitate with which he was dealing was a mixture of hemoglobin and other proteids thrown out of solution by the chloroform. It is also stated in this paper that oxy-hemoglobin is only slowly and incompletely precipitated by the action of chloroform at room temperatures. E. Salkowski, in his later paper,§ states that blood kept at a temperature of 40° C. for 24—48 hours in presence of chloroform changes to a thick mass, but found that the precipitation of the hemoglobin was not complete under such circumstances. Regarding the action of chloroform on serum, he states that serum * ‘ Zeitsch. £. physiol. Chem.,’ vol. 29, 1900, p. 416. + Quoted by Formanek, loc. cit. t ‘ Zeitsch. f. physiol. Chem.,’ vol. 29, 1900, p. 416. § ‘Zeitsch. f. physiol. Chem.,’ vol. 31, 1900, p. 329. 1904.] Properties of Solutions of Chloroform in Water, ete. 387 can be preserved in contact with chloroform for years without pre- cipitation at room temperatures, and finds this in agreement with Formanek’s results, who found, in presence of an alkaline reaction, no precipitation of serum by chloroform even at a temperature of 50—55° C. Formanek does not state whether his alkaline reaction is the natural alkaline reaction of the serum. Salkowski also found a precipitating action of chloroform upon solutions of albumose and casein. Our experiments were conducted with hemoglobin and serum obtained from pig’s blood. The serum used was obtained from clotted blood and was thoroughly centrifugalised before use. The hemoglobin was in all cases obtained by centrifugalising the blood corpuscles three times with normal saline, and then laking with distilled water and making up to the same volume as that of the blood taken. In the case of serum we found that the fluid acquired, with less than 1 per cent. of chloroform (and greater quantities up to satura- tion), a peculiar opalescent and fluorescent appearance, but remained quite transparent to transmitted light. On the addition of over 2 per cent. of chloroform, there is a tendency to precipitation even in the cold, and at the end of 24—48 hours there is a slight precipitate present, but the effect is much hastened on placing the mixture in an incubator at 40° C., so that it fecomes impossible to determine the maximum solubility of chloroform in serum at body temperature. Both in obtaining precipitation in the,cold and more rapidly at 40° C. in presence of the natural alkaline reaction of the fluid, our results are at variance with those of Formanek and Salkowski. The results were obtained several times in succession. The marked opalescence in the serum was obtained in preparation of solutions of known concentration in chloroform for purposes of measurement of their vapour pressures, and led us to doubt at first whether we were not dealing with a fine emulsion of chloroform in the serum. Since this point was of vital importance to our experiments on vapour-pressure, we investigated it as completely as possible. In the first place, examination with the microscope of the opalescent fluid showed no visible globules of chloroform, even with the highest powers. To make certain of the matter, a current of air from an aspirator was bubbled first through chloroform contained in a Woulff’s bottle, afterwards through a similar bottle containing water, and then, at the same temperature, was sent through a third Woulff’s bottle containing serum. By this procedure the serum never came in contact with fluid chloroform, nor with air more highly charged with chloroform vapour than corresponded to the saturation of the air in contact with it or passed through it. There could, hence, be no condensation of chloroform and no means EE SESS On 388 Prof. B. Moore and Mr. H. E. Roaf. [Apr. 12, by which an emulsion of chloroform, finer even than could be seen with a microscope, could be formed. Very soon, however, after the chloroform vapour began to pass through, a distinct difference in appearance was observable between the serum and a control placed alongside, and after a time the serum charged with chloroform in this manner was as opalescent as the specimens made in the usual way by shaking with weighed quantities of chloroform, and gave similar results with regard to vapour pressure. These results show that the marked opalescence is not due to an emulsion of chloroform, and further that it is not due to precipitation of proteid in the ordinary sense of the word, for no precipitate can be seen with the microscope in the opalescent fluid. Further, experiments on the vapour pressure show that the value of this is a long way from the maximum value, which it must obviously possess if an emulsion was present. A similar opalescence is obtained on adding to serum other organic liquids which possess anesthetic properties, thus we have obtained it on saturation of serum with ether,* benzol and xylol, but have not followed the matter up as we have done with chloroform. In the case of hemoglobin solutions, we have not been able to observe an opalescence, similar to that seen in the case of serum, up to the strength at which precipitation begins. In order to study where precipitation commences it is necessary to keep the hemoglobin and chloroform constantly stirred, or otherwise before the chloroform has dissolved the lower layer of hemoglobin solution in contact with the liquid chloroform becomes precipitated. When precautions are taken to prevent this occurring, precipitation is found to take place when about 2 per cent. of chloroform has been added, that, is long before saturation is reached (wde infra). This precipitation prevents both the determination of the solubility of chloroform in hemoglobin solu- tions, and the observation of the vapour pressure with high con- centrations. | With concentrations of 1 per cent. or under no precipitation whatever was found to occur. Contrary to Salkowski and Formanek, we found that no raising of the temperature was required to cause precipitation of the hemoglobin ; in fact, with strong solutions, the precipitation occurs within a few hours, even in the ice chamber. With time the precipitation is complete, and the precipitate is insoluble in water or saline, but in dilute sodium carbonate it is easily soluble, and we have found that the solution then shows the spectrum of alkaline hematin and not that of oxy-hemoglobin as was found by Formanek. * The opalescence with ether is not nearly so well marked, but increases on standing, and the solution in time becomes extremely viscid. 1904.] Properties of Solutions of Chloroform in Water, ete. 389 Il.—On the Relative Solubility of Chloroform in Water, Saline, Serum, and Hemoglobin Solution. In so far as we have been able to discover, no attention has been paid by previous experimenters to the maximum amount of chloroform capable of solution in the blood or serum as compared with that taken up by water or saline solution isotonic with blood. When any reference is made to the matter it has been usually assumed on general principles that the serum or plasma will behave like a saline solution of equal concentration and dissolve somewhat less chloroform than water.* In other words, that there is no specific action of the proteids or other substances in the plasma. This supposition we have found experimentally to be entirely erroneous, for both serum and solutions of hemoglobin dissolve much more chloroform than water or normal saline solution. This fact is of importance in regard to the mode of action, as it definitely points to an interaction between the chloroform and the proteid present. The presence of fats would of course increase the apparent solubility of chloroform in serum, and hence it is necessary in all cases to use perfectly clear serum, free from suspended fat ; this precaution we have always been careful to observe, and in addition the serum has always been centrifugalised. In this connection it may be added that the hemoglobin solutions which we have employed could not contain fatty matter, and hence the high solubilities which we have observed could only arise from chemical interaction between the hemoglobin and the chloroform. Methods for Determining Maximum Solubility. Three methods have been used in the determination of the maximum solubility of chloroform in the solvents mentioned above, which have given concordant results and shown that the solubility in proteid solution is much higher than in water or saline. | In the first method we have determined the amount of chloroform dissolved by obtaining the product of volume and vapour pressure at low pressure and with a small volume of fluid, so that practically all the chloroform was simply pumped off into the vacuum. In this method the volume of fluid experimented with is necessarily small, and this gives rise to experimental error of measurement, which is added to by the volume measured being large and pressure small, so that the results are only approximative, yet it is observable that they confirm those obtained by the more accurate methods described below. The details of the method are described in the succeeding section on the relationship between vapour pressure and. concentration of * Overton, ‘ Studien tiber die Narkose,’ p. 93. VOL. LXXIII. 4k 390 Prof. B. Moore and Mr. H. E. Roaf. [Apr. 12, chloroform, and it need only be mentioned here that we have obtained solubilities of 0°95 per cent., in normal saline (0°75 per cent.), 3°33 per cent. in serum and 4:42 per cent. in whipped blood, by this method. The second method employed consists in weighing out known amounts of chloroform into water and serum and hemoglobin solution respectively, and then determining by direct observation that con- centration in each case at which the chloroform ceased to be dissolved. This method of observation is made easy in the case of chloroform by the high specific gravity of that fluid, as a result of which on inverting the flask in which the determination is being made even minute globules of undissolved chloroform can be seen falling through the fluid. The determinations of solubility by this method are made on the following plan. Pure chloroform is dropped from a fine capillary pipette into a tared graduated flask of 25, 50 or 100 c.c., and carefully weighed to definite amount, corresponding, when the flask has been filled by the solvent under experiment, to a definite percentage of chloroform. A series of such flasks is prepared, and immediately after each flask is filled with the desired solvent and _ either shaken thoroughly by hand until solution is complete, or placed - on a rotary shaking machine. After the lapse of several days, during which time the flasks are never opened and are kept shaken ‘up, it is noticed at what level of concentration the chloroform ceases to be completely dissolved, and so the solubility is determined. The lower strengths of known value short of saturation, and also the saturated solutions so prepared, were kept and used also for the experiments on vapour pressure at varying concentration described in the next section of this paper. The following results were obtained by the application of this method at room temperatures, approximately (13° C.), for the percentage by weight dissolved :— Water, 0°8 per cent. dissolved, 0°9 per cent. dissolved, 1 per cent. not dissolved completely. Estimated solubility 0°95 per cent. Saline Solution (0°75 per cent. sodium chloride in water), 0-7 per cent. dissolved, 0°8 per cent. dissolved, 0°9 per cent. not dissolved. Estimated solubility about 0°83 per cent. Serum, 3 per cent. dissolved, 3°5 per cent. dissolved, 4 per cent. all dissolved save a few small globules. | Hemoglobin Solution or Blood.—Over 6 per cent. by weight is taken up when chloroform is shaken with blood or hemoglobin solution of equal strength to the blcod, prepared from blood by centrifugalising several times with saline, and subsequent laking with distilled water, and no globules of chloroform can be seen on careful examination with the microscope. But the solution rapidly changes in colour, and a 1904.] Properties of Solutions of Chloroform in Water, ete. 391 precipitate is thrown out on standing, as above described, which is quite insoluble in water or saline, but easily soluble in dilute sodium carbonate solution, and then gives the spectrum of alkaline hematin. The blood begins to give this precipitate when about 1°5 per cent. of chloroform has been added at room temperature, but with a lower concentration, and more rapidly when heated to body temperature in the incubator. Two per cent. of chloroform gives a precipitate in the: cold, and on heating to 40° C. a red flocculent precipitate leaving a clear colourless fluid above. The third method for determining the solubility of chloroform in the. fluids experimented with consists in shaking up thoroughly for several hours with an excess of chloroform, and then pipetting off and determining the amount of chloroform in the solution. The difficulty here is a rapid and accurate method of determining the amount of chloroform contained in a measured volume of the given. saturated solution. The procedure finally employed for this purpose, which is also being experimented with as a method for quantitative estimation of chloroform in blood and serum at lower values reaching to the anzsthetising value, was as follows, and led to very accurate results. A measured volume (usually 10 ¢.c.) of the fluid saturated with chloroform is placed in a flask fitted airtight with a double bored cork,. and a stream of hydrogen is aspirated through the solution, the oxygen. present in the flask and connections is absorbed by passing through alkaline pyrogallate, and the mixture of hydrogen and chloroform is. then burnt by passing over heated palladium asbestos placed in a very short combustion tube. All the chlorine in the chloroform is thus: burnt to hydrochloric acid, and the amount of this absorbed in standard alkali is then estimated, first by back titration against standard acid,, and then further checked, either by volumetric titration with standard. silver nitrate solution, or by gravimetric determination as silver chloride. The serum used in these determinations was examined for chloroform. emulsion by the microscope, but no undissolved chloroform in. suspension was observed. ‘The precipitate In serum at atmospheric temperature obtained by this method of shaking up with excess of chloroform was very dense, so that the serum became quite opaque. The results obtained by employing this method were as follows :— Distilled water, dissolved 0°95 per cent., and serum, dissolved _ 9°08 per cent. IlI.—On the Vapour Pressure of Chloroform Dissolved in Varying Concentra- tion in Water, Saline, Serum, and Hemoglobin Solutions respectively. A determination of the vapour pressure of an anesthetic in solution at varying concentrations in serum, in hemoglobin, or in blood, is of | 2H 2 392 Prof. B. Moore and Mr. H. E. Roaf. [Apr. 12, high practical importance, since it is upon the relationship of this vapour pressure to the concentration of the solution that the amount of anesthetic taken up by the blood circulating through the lungs depends. It has hitherto been taken for granted that the Dalton-Henry law can be applied, and that the amount of anesthetic taken up is strictly proportional to, and varies directly with, the ina of the vapour of the aneesthetic in the inspired air. This has never, however, to our knowledge, ‘been experimentally tested, and it seemed to us desirable to attempt such a determination. We have investigated from this point of view solutions of chloroform in serum, hemoglobin solution (of equal ‘strength in hemoglobin to the blood from which the hemoglobin was prepared) and whipped blood, and have contrasted the pressures obtained with those of solutions in chloroform, in water, and normal saline at equal concen- trations. The vapour pressures have been measured corresponding to concen- trations ranging from considerably below the anesthetising values for chloroform vapour pressure in air (viz., 8—10 mm.) observed by Paul Bert, up to the saturation points in most cases. Apparatus. The instrument employed for this purpose was a form of “ differential densimeter,” which, after passing through many modi- fications, took the form represented in the accompanying sketch (fig. 1), which is drawn approximately to a scale of 4 The two tubes shown are exactly similar, and are graduated in cubic centimetres and tenths in the upper portion, and in centimetres in the lower and wider portion. The tubes are connected as shown by means of thick-walled rubber tubing and a glass Y-piece to a stout glass mercury receiver capable of holding more than enough mercury to fill both tubes and their connections. The tubes are held in a vertical position by clamps attached to the strong vertical iron bar of a massive retort stand, and each tube is capable of being moved in its clamp vertically up and down for purposes of adjusting the mercury levels. In order to keep a constant temperature (in the case of the experiments carried out at body temperature) the upper portion of each tube, from about the middle of the wide part to the level of the stopper at the top, was encased in a hot-water jacket of the form shown in detail in fig. 2, and omitted for clearness from fig. 1. It was found convenient to use for the outer glass tubing of this jacket the largest size of a common variety of paraffin-lamp chimney, a0 < 393 1904.] Properties of Solutions of Chloroform in Water, ete. - which measured 8 cm. in diameter above, bulged out as shown to 9 cm., and then narrowed below to 6 cm. The wider top and bulge were found very useful in facilitating the introduction and vigorous movements of the bar electro-magnet used as . CEO / seqinn Psd ony Rn tli rr Wa tal! thi i Fas = hy Ai yp | | & SS ooo Cr wl e) ee Come W Fig. 1.—Diagram of differential Densimeter. a stirrer in connection with the iron stud (seen in figs. 1 and 2, at the top of the mercury), which was dropped into each tube, and, during an experiment, agitated up and down so as to thoroughly mix the fluid under experiment, and so bring it into equilibrium more rapidly with the vapour in the space above it. 394 Prof. B. Moore and Mr. H. E. Roaf. | [Apr. 12, The hot-water jacket was made watertight below by means of an india-rubber cork, as shown (fig. 2); the rubber cork was also bored in each case for two narrow glass tubes, for the purpose of carrying ra 7 | POO a yyy TT an \| Vl ', yy drinlieunlle ntl yl yi! |! AHI Tia Wty! Ve Te eee faclan of : Loo A __ Outflow to ot water WASREUBiGe! from bath. wate ies Fie. 2.—Section taken through the upper portion of one tube of the differential Densimeter and the hot-water jacket, showing the inflow and outflow tubes. Seale 4. water to and from the jacket. ‘The two tubes carrying the water in, stopped about 2 cm. above the upper surface of the rubber cork in each tube, so as to prevent blocking by mercury accidentally run over from the top when the inner tubes were being cleared out at the termination of a measurement. The outer ends of these two tubes x 1904.] Properties of Solutions of Chloroform in Water, ete. 399 were attached by means of narrow rubber tubings and a glass Y-tube to a bath of hot water, placed at a higher level, and a screw-down clip on each rubber tube regulated the flow until a thermometer placed in the corresponding hot-water jacket showed the desired temperature. A constant level of water was kept up automatically in the warm supply bath, and its temperature was regulated so as to lie 2—3° above that of the jackets. The two outflow tubes passed -up, as shown, inside the jacket to the level of the ground in glass stopper, and their outside ends were connected by means of rubber [tubing to the waste pipe. In fig. 1 the upper portion of the left-hand tube is shown in section, and that of the right-hand one in outline. The ground in stoppers shown were found, when sealed with mercury, to be much more effectual against minute leakages, which entirely vitiate the results, than any form of tap, and they are also much more convenient for introducing the solutions to be experimented upon, and for cleaning out the apparatus. Further, since they do not require to be operated between the commencement and termination of each determination of vapour pressure, they are better adapted to their particular purpose than a tap. In the course of our experiments, we also found it necessary to be able to dilute a solution with more of its solvent without allowing it to come in contact with any appreciable volume of air, and for this purpose found the stopper arrangement most convenient. The side tube shown at the lower end of each main tube was designed to trap. air which was found to slowly leak in through the rubber pressure tubing, when a vacuum was established by lowering the mercury receiver, and for a long time was a source of annoyance. We subsequently learned that the device had been first introduced by Lord Rayleigh. At the end of an experiment, any air which has collected is discharged by raising the mercury-holder, opening the screw-clip shown, and allowing enough mercury also to pass through to form a seal in the rubber tubing above the clip. The mercury-holder was suspended by means of a ring and loop ot wire attached around the bulb from a vertical rod, swivel, and hook, possessing a slow screw movement in a block attached to a horizontal rod fixed in a clamp, which could be moved up and down on a heavy retort stand. For large movements, the clamp was slid up or down the retort standard, and for fine movements the screw was raised or lowered in its block. In using the apparatus, the two vertical tubes are first placed at the same level, the mercury-holder is filled with mercury, and, with the two glass stoppers out, the whole apparatus is filled with mercury, the two stoppers are next inserted, enough mercury being left above them to form a seal, and the mercury-holder is then lowered until the two vertical tubes become evacuated. The receiver is then raised ee ee ee ee > a 7 SeTeeec te Se i - = ie iio “ ; 396 Prof. B. Moore and Mr. H. E. Roaf. [Apr, 12, again to the level of the stoppers, and any bubble of air found is discharged. The apparatus is now ready for an experiment, and, with stoppers out, the levels of mercury are adjusted until there is an equal volume left above the mercury on each side. A given volume of the solvent (say 5 ¢.c.) is now introduced on the one side (say left), and an equal volume of the solution of chloroform in the same solvent on the other side. In each case, immediately after the fluid has been introduced, the stopper is inserted, care being taken to prevent any air being included, either as a bubble at the mercury surface, or between the surface of the introduced fluid and the stopper. To achieve the latter end, we have almost always introduced above the mercury 2 or 3 c.c. more than the required quantity, so that 1t stood in the neck and slightly above, and then, by easing the stopper and gently adjusting the level of the mercury-holder, have brought the level of the mercury in the tube to the desired volume mark. After the solvent on the one side, and the solution of chloroform of the desired strength on the other side, have been successfully introduced in equal volume, and without any bubble of air, the mercury-holder is lowered until a space containing vapour has appeared on each side. The level of the mercury will be found to be lower on the chloroform side, and it is obvious, the instrument being independent of variations in atmospheric pressure, and the only different factor being the added chloroform on the one side,* that the difference in pressure will give the vapour pressure directly for that strength of chloroform soiution at that particular temperature. , There is hence no need to determine pressure due to dissolved gases on the two sides,t or pressure of aqueous vapour, since these balance, and the quickness with which readings can be directly obtained makes it possible to carry out a long series of determinations at varying strengths, without the proteid solutions having time to undergo bacterial change. Certain precautions have to be taken, however, and corrections made which may here be mentioned :— 1. Before taking a reading it is essential to move the tubes vertically and adjust the levels until the volumes of the vapour spaces above the upper aqueous solution meniscus in each case are exactly equal, otherwise inequality in pressure of gases pumped off on the two sides gives rise to an error, which is greater the smaller the vapour space. * There will be a smail difference in the pressure of water vapour on the two sides, due to there being a stronger solution onthe chloroform side, but this is in all cases too minute compared to the pressure of the chloroform vapour to make any appreciable error. + Slight differences in dissolved gases gave a disturbance with very dilute chloroform solutions, and this was later obviated by pumping the gases off (vide infra). 1904.] Properties of Solutions of Chloroform in Water, ete. 397 2. Before taking a reading, there must be certainty that each fluid is in equilibrium with its vapour space. This is shown by absence of variation when the apparatus is left at rest. For rapid and accurate working the mechanical stirring by means of the studs and magnet is indispensible, for even after the lapse of an hour when at rest the solution has not completely discharged its proper amount of chloroform into the vapour space. When once the control, containing, of course, no chloroform, has been thoroughly stirred it remains constant, and need not be changed at the end of each deter- mination, but can be used throughout an entire experiment. By vigorous stirring, peaineaeaet can be attained in 5—10 minutes, and the level does not afterwards change no matter how long stirring and observation be kept up. This important experimental observation we have taken occasion to verify several times during our experiments. 3. For very accurate working, especially with the dilute solutions and low pressures, it is necessary in the case of serum and hemoglobin to pump off the dissolved gases by means of a Tépler pump, otherwise these come off unequally from solvent and solution and disturb the results at the low pressures. The chloroform solutions are then made up from the pumped-out solvent, which also must be used for control and for making the dilutions. 4. The temperature must be the same in the jackets surrounding each tube at the time when each reading is taken, and in a series of determinations at varying strength and a constant temperature, that temperature must be closely maintained throughout. The temperature error is a maximum when the solutions are near saturation, for then the variation in vapour pressure per degree is very large; fortunately here the differences in level under observation are also very large, which diminishes the percentage error arising from small deviations in temperature. At concentrations away from saturation, the variations arising from small differences in temperature approximately obey the gas law, and under the conditions of our experiments become quite negligible. 5. A correction must be made in all cases, upon the concentration of the solution introduced into the tube, for the amount of chloroform pumped off from the solution into the vapour space. This correction is, of course, larger in the case of the more concentrated solutions with high vapour pressures. This has been done in the experiments of which records are given below, and accounts for the concentrations not being exact percentages or small fractions of exact percentages. The amount of chloroform in the vapour space is readily calculated from the product of the observed vapour pressure and the volume of the vapour space, and this amount deducted from the quantity contained in the chloroform solution when it was introduced, gives the necessary 398 Prof. B. Moore and Mr. H. E. Roaf. [Apr £2, datum for calculating the concentration in chloroform of the solution corresponding to the observed vapour pressure in the vapour space. The ratio of the vapour concentration in the fluid to the concentra- tion in the vapour space gives the coefficient of distribution (coefficient de partage, Theilungscoefficient): this should remain constant if the absorption of the chloroform vapour by the liquid were normal and strictly proportional to the vapour pressure, and if it varies it points to a physical or chemical aggregation or compound between the chloro- form and the fluid or its constituents (vde infra). Method of Reading.—The readings were taken with a cathetometer,* placed about 4 feet from the tubes, both for greater accuracy in reading than direct measurement would give and to avoid changes in temperature. Two Methods of Experimentation—We have employed two different methods of experimentation In investigating the variation in vapour pressure with varying concentration. It is obvious that the concen- tration of a measured volume of a strong solution introduced into the densimeter may be diminished by pumping off more and more chloro- form, by increasing the volume of the vapour space above the solution. A series of readings of differences in pressure may thus be obtained in which the vapour space on the two sides is kept equal and of known and increasing value throughout the series. This method we have called the method of “ variable vapour space.” On the other hand, a series of solutions of known and _ steadily diminishing or increasing concentration may be introduced into the densimeter and measured one after another as to their vapour pressure, in each case with a known fixed volume of vapour space. As stated above, the concentration at which each vapour pressure in the series is measured is then accurately oe This method we have called the method of “ constant vapour space.” Method of ‘* Variable Vapour Space.”—In ie method, unless the volume of the tubes of the densimeter is very large, the volume of solution and solvent respectively introduced must be very small. We have usually taken $ c.c. on each side, either of a saturated solution or of a very strong solution of known strength, and by altering the levels of the tubes and mercury receiver, have taken a long series of readings, in each case with equal volume of vapour space on each side, at every increasing volume of vapour space, until the difference in pressure became of small value, and the product of volume and vapour pressure became approximately constant, showing that practically all * The instrument used was made by Pye and Co., of Cambridge, and, by means of a vernier and divided screw-head, read to He, of a millimetre. We have frequently observed that we were able to take readings within five divisions, that is mm., which is far within the accuracy of other portions of our determi- Bi Le 2 ZT) nations. 1904.] Properties of Solutions of Chloroform in Water, ete. 399 the chloroform had been pumped off from the solution. This constant product then gave the necessary datum for calculating the concentration of the original solution introduced into the densimeter, the product of vapour pressure and volume at each stage gave the datum for calcu- lating the quantity of chloroform pumped off from the solution, and therefore for deducing the corresponding concentration of solution. Or, also, by plotting vapour pressures as abscisse, and the product of vapour pressure and volume of vapour space as ordinates, the ratio of vapour pressure and amount of chloroform absorbed at each stage could be shown. The method of “variable vapour space” has, however, two working disadvantages, which caused us in the end to abandon it and replace it by the method of ‘‘ constant vapour space.” The first objection is that the amount of solution taken is small, hence it is difficult to measure it with accuracy, and to make it equal on the two sides. The second objection is that at the low concentrations the increase of volume for a small fall in pressure is very large, and hence the deter- minations become inaccurate, a small error in pressure reading making a large deviation. The values for high pressures are also inaccurate, but for a different reason ; these readings are taken with small volumes of vapour space, and unless the vapour spaces are accurately equal on the two sides, there is a large disturbance due to inequality in pressure of the previously dissolved gases pumped off on the two sides. The results, however in the intermediate pressures are accurate and are given below, as they confirm those given by the other method. Method of ** Constant Vapour Space.”—In using this method we have always introduced a volume of 5 c.c. of the solvent on one side, and 5 ¢.c. of a solution of known strength on the other, and have invariably adjusted the levels so that at the temperatures of observation there was a vapour space of exactly 5 c.c. on each side. In some cases we have started with a saturated solution of chloroform and have then made dilutions of different percentages of that solution, in the manner described below. In the later experiments we found it, however, more expedient, on account of knowing the exact concentra- tion directly, to prepare a solution of known strength, say 1 per cent., and for the more dilute solutions to use various percentage dilutions of this stock solution. The more concentrated solutions were obtained by making up solutions varying by 1 per cent. in strength, and + per cent. differences were got by mixing these with each other in equal proportions. For percentages less than 0-1 per cent., a 0-1-per- cent. solution was first prepared by making a ten-fold dilution of the l-per-cent. solution, and this 0:l-per-cent. solution was then diluted similarly to the 1-per-cent. solution. = ~~ ? i Ney 400 Prof. B. Moore and Mr. H. E. Roaf. 3 Agmeee Precautions in Preparing, Preserving, in Unaltered Strength, and Diluting by known Amounts, of the Solutions Used. In working with the solutions, it is indispensable that due precau- tions be taken against loss by escape of chloroform into the air during the various manipulations. 7 If a portion, for example, of a stock solution of known strength be pipetted off for use in the densimeter, and then the stock bottle be merely stoppered, the air space over the portion of solution left in the bottle rapidly becomes charged with chloroform vapour at the expense of the stock solution, and a second sample taken later from the same bottle will be found to be weak and give a wrong result. The way to guard against this is to fill the bottle with mercury up to the neck immediately after drawing off, and at once stopper up. Similar devices were employed in all dilutions to complete the process out of contact with air. The stock solutions were made by direct weighing, by dropping from a fine pipette, into graduated glass-stoppered flasks of 25, 50, or 100 c.c. capacity, according to the amount required, immediately filling up to the mark with the particular solvent, and setting at once upon a slowly rotating disc, driven at such a rate that the chloroform globules have just time to fall through the solution each time the flask is inverted. In this way, arapid solution is effected, so saving much time. Further, in certain cases such shaking is indispensable, for a solution of hemoglobin left in contact with even a small amount of chlorotorm without continuous shaking soon forms a precipitated layer at the bottom. The stock solutions were kept carefully stoppered, unless at the moment of introducing the pipette to remove a portion for experiment, and then the space within the bottle was always filled to the stopper with mercury in the manner above described. We found much difficulty owing to leakage of chloroform during dilution in the case of the strengths intermediate between the stock strengths, until it occurred to us to make these dilutions in the densimeter itself. For this purpose we always placed the mercury level at 10 c.c. below the stopper, on the side of the densimeter in which the mixing was to be carried out, then the proper amounts ot the two solutions necessary to give the desired strength was drawn up in two graduated pipettes and run into the densimeter tube, the stopper was then inserted, including only a minute bubble of air, and now by thorough agitation of the iron stud through the fluid by means of the electro-magnet a thorough mixing of the two fluids was attained, then by raising the mercury-holder and slightly easing the stopper 5 c.c. were allowed to escape, the mercury level being placed accurately at 5 ¢.c. A seal of mercury was finally dropped in above the stopper, and so the dilution was effected. 1904.] Properties of Solutions of Chloroform in Water, ete. A401 For example, to obtain a dilution of 1°5 per cent., 5 c.c. of 1-per-cent. solution and 5 c.c. of 2-per-cent. solution were drawn off into pipettes, placed in the densimeter and mixed as above described ; to obtain a 0-4-per-cent. solution, 6 c.c. of the solvent were taken and 4 c.c. of a 1-per-cent solution, and similarly treated; to obtain a 0-03-per-cent. solution, 7 c.c. of solvent and 3 ¢.c. of 0°l-per-cent. solution were taken, and so on. Certain of our experiments were carried out at room temperature and others approximately at body temperature (40° C.) ; the following protocols and accompanying curves show some of the typical results obtained, which have been confirmed in most cases by duplicates :— Variable Vapour Space. Experiment 1.—Distilled water containing approximately 0°78 per cent. of chloroform. Half a cubic centimetre was introduced into each tube of the densimeter, of the chloroform water on one side and of the same distilled water without chloroform on the other. The temperature at which the experiment was carried out was 17° C., and the volume at which readings of pressure were taken varied from 2—50c¢.c. The percentage of chloroform in the water was not known directly, but was calculated by extrapolation of the curve showing Weteecee Curve t, fic. 3: The following table gives the results of the experiment, which are also shown graphically in Curve 1, fig. 3, in which the abscisse show Fie. 3. A eS Sa + 32 a : Ss 0 ris lel eae Fon ] re = =<) | | | | =< 2e¢ | | | | | | | i = => | oie eee oo SS SS c5) =) i | | | = 24 | | | | Yeti | Sg OQ $ + : 1 ee eee Se | Bae ea) 3350 ell 1 — ee ee ee a ee l I : ae 8 | | | =a Se = | >o sO ==37 2ESo S| B © 09 | 2S oO S| aie On oC. 225 3S5° === 2tz20 i] = 22 ° | OL | | } } Sone = Ni T 1 z | } | S52. | [sol | | = Sm) | SQ “te | | =53 | Seeee | | TS] (ear Seca) | asenl | {ol aie D rve |, (es | Ss € | gust ata Pe (Reeueal ist wa.ter) Hie 3 3a (Curveg,Serumy | = | | | | | | } fel ae ean ey | | | eeelewa el pn) Gr 20". 0m a0 50 60 70 80 30 100 10 20 | ~<—Pressure in millimetres of mercury.— m.m. of mercury. +/ 402 Prof. B. Moore and Mr. H. E. Roaf. [Apr. 12, pressures of chloroform vapour and the ordinates the product of the volume and pressure of vapour. solution. mM In S by weight of chlorofot Percentage 1-0 O7 08 O09 (aap OK OS Olay BOel (7h 10) Q Fie. 4. iO 20 50 40 ~- -~ - j -- = 50 60 70 (Curve 2, Serum) i, Distilled water) In Curve 1, fig. 4, the same experi- WA { | | t \ { | f | i | 80.90 100 Il0 120 <—Pressures in m.m.of Mercury —> . (Proportion to percentage by weight of chloroform in vapour space.) ment is shown with pressures of chloroform vapour as abscisse, and percentage of chloroform dissolved at each pressure as ordinates :— 1904.) Properties of Solutions of Chloroform im Water, ete. Experiment 1.—Distilled Water. Temperature 17° C. of vapour space. (© CONT OD OU bo Volume | | Pressure Percentage | Gre Clailoyeo™ || Renn mirorets) form in by wera | | of chloro- ead form in space in aor mm. of space. mereury. | 74°61 | 0 °04957 60 °62 0 :04.027 51°29 | 0:038408 47°69 | 0°03168 | 43°08 | 0-02862 38°91 | 0:°02585 35°98 | 0:023890 34°36 | 0°02283 28°88 | Q-01915 | 26°16 | 0:01738 22 “62 0 -01503 17°65 0:01173 15°84 | 0-01052 | 14°37 | 0:00954 | 12°43 | 0-00835 11°87 =| 0:00787 10°54 | 0:00700 Weight in grammes ot chioro- form in vapour space. “00099 00121 00170 ‘00190 ‘00197 00207 00215 "00228 "00230 °00263 “00301 ‘00293 "00316 ‘00334 00334 "00354. 00350 SOQOQOOCOOOooeoeQeeeeS Weight in Percentage grammes | by weight of chloro- of chloro- form in form in sulvent. | solvent. 0 :00293 0°586 0 00271 0° 542 0 °00222 O. 44-4 0 °00202 0-404 0 °00195 0-390 0 ‘00185 0-370 0 :00177 - 0 354 0 00164 0-328 0 :00162 0 324 0°00129 | 0°238 0:00091 | 0O:182 0:00099 | 0-198 0 :00076 0 152 0 -00058 0-116 0 -00058 0°116 0 00038 0-076 0 :00042 0:084 403 Coefficient of dis- |, tribution | between vapour space and solvent. pela : 13 : 13 : 12 5 133° a jl4be : 14° : 14° : 16° : 14° : 12 : 16 + Ar: : 12 a 113) ° 2 19s +2 ee SASNNOHDWOSKRDHWRHISOCL OD A404 a Prof. B. Moore and Mr. H. E. Roaf. Apres a2, Experiment 2.—Serum containing approximately 1°65 per cent. of chloroform, the amount being determined by extrapolation of Curve 2, fig. 3. Experiment conducted in all respects similarly to Experiment 1. Graphic records of results shown by Curve 2, in figs. 3 and 4. Experiment 2.—Serum. Temperature 18° C. | Pressure | ot chloro- Volume | form in of vapour} vapour space. space in mm. of mercury. 1 115 °63 eS 111 °85 | 2 104 °27 3 98°15 4 89 °89 5 87°05 6 79°70 a 75-14 8 71°46 9 67°16 10 63°24 | 12 58°05 Oey cts 52°31 | 20 43 “64, 20 34°97 30 32 °38 35 28 °20 40 25 °44 45 22 °40 50 21 03 60 alyerey f 70 16 -03 80 13°93 90 V2 2 100 10 ‘70 125 Said 150 igo Percentage by weight of chloro- | form in vapour space. "07655 | "07405 "06498 05763 05277 04975 | ‘04730 04487 | 04447 04187 | "03843 | ‘03463 | “2889 02315 "02144: -01867 "01684: "01483 | "013892 701186 | ‘01061 "0092 ‘COSO ‘0071 “0058 “0052 QO seoeoooggoeeeoegeooee oe © fe Weight in grammes of chloro- form in vapour space. 00077 “00111 "00138 00195 00237 "00264 "00299 ‘00331 ‘00359 "00402 "00419 “00461 "00519 ‘0057 °00579 00643 ‘00653 00674 “00668 “00696 00712 00743 ‘00737 00722 ‘00708 ‘00726 ‘00786 QO oQocreoqoeTooeooqgooaqgoqooeSe Weight in grammes ot chloro- form in solvent. 00746 00712 “00685 00628 ‘00586 "00559 "00524 00492 ‘00464 00421 00404 00362 00304: ‘00246 "0024.4: 00180 °00170 "00149 "00155 "00127 “OO11i ‘00080 “00086 “00101 ‘00115 ‘00097 ‘00037 VIO] OWS OBDOOAS OS OOo Sooeaoeee& Percentage by weight | of chloro. | form in solvent. i We) LX) OCOCTCOCCFCCCOCOCCOCCOC OOOH HHP HHH be ep | Coefficient of dis- tribution between vapour space and solvent. : 19 : 19 :19 : 19 : 20 : 21 1 on : 20 : 20 : 18 : 19 :18 : 17 gil 7 : 21 : 12 : 18 iyi : 20 WER OWEHSHMSTWHH SASH SUH TMIL dow ck 1904.] Properties of Solutions of Chloroform in Water, ete. 405 Experiments with Constant Vapour Space. Experiment 3.—Saline solution containing 0°75 per cent. of sodium chloride, and approximately 0°95 per cent. of chloroform, determined from product of pressure and volume after pumping off. Five c.c. of the saline introduced on one side as a control and 5 c.c. of the chloroform solution on the other, and a vapour space of 5 c.c. being formed on each side. Temperature of experiment 15° C. The results are shown graphically in Curve 1, fig. 5. Fria. 5. __ _ JSS eens Ee aia J Tee Bible a 32 1S i ae li ae Ei babe me [allee| See a ie VOL. LXXIII. 2 F 406 Percentage of chloroform by weight in solu- tion originally introduced. "0024 “0048 "0095 "0190 0286 “0381 ‘0476 0571 0666 "0762 "0857 "0952 "1904 "2856 °3808 “4.760 5712 ‘6664 °7616 8568 ‘9520 QeearaeaaQgooeeoeooeoqoooeoe Prof. B. Moore and Mr. H. E. Roaf. ixperiment 3.—Saline. Temperature 15° C. Pressure of chloroform in vapour space in mm. of mereury. D6 “78 Percentage by | weight of chloroform pumped off into vapour space. “0004 "0012 “0018 “0030 "0038 "0045 "0049 °0056 “0068 0075 “0082 “0096 0172 "0256 "0321 0402 "0483 "0534 ‘0593 "0663 0738 QSOQQoeeseoqqeeeeoqeeoegoqqee © Percentage of chloroform by weight remaining in solution. SS) & =) iS) (op) © pa (ep) j=) Qooocecocoooooooeoose e . ° us ° we ° . e ° ° ° ° ~I Oo bo { Apr. 12, Coefficient of distribution | between. vapour space and solvent. | | a DOONDOMIAMA WwW MNODSNHFONANANINAMWSS ee pe 1904.| Properties of Solutions of Chioroform in Water, ete. 407 Experiment 4.—Serum containing approximately 3°33 per cent. of chloroform determined as in Experiment 3. Experiment performed similarly to Experiment 3 and temperature also 15° C. Results plotted in Curve 2, fig. 5. Percentage by weight of chloroform originally introduced. Se ee WONNNFEFrFODOCOOCOOCOoCOCO°C°CeO "0083 0167 0333 "0666 0999 1332 "1666 "1999 2332 "2665 *2998 3331 “6662 "9993 "3324 6655 “U986 3317 "6648 9979 °3310 Experiment 4.—Serum. Temperature 15° C. Pressure of chloroform in vapour space | in mm. of mercury. 0°66 1°82 2°74: 6 66 8°59 10°85 13°28 15°54 16 °96 18 °387 1S) SI) 22 -49 43°02 63°15 Percentage by weight of chloroform pumped off into vapour space. Percentage by weight of chloroform remaining in solution. 0079 "0155 ‘O315 "0621 0942 "1259 "1577 1895 *2219 "2943 "2870 "3181 6381 9571 "2774 6017 ‘9306 "2577 5877 *9123 *2500 ONNNW—HHrFODOOOOGOCOCOOCO0O0°O Coefficient of distribution between vapour space and solvent. fe pe ee et S) 2 BPE AAR RMT eT wWHE Ne nwa 408 Prof. B. Moore and Mr. H. E. Roaf. [Apr. 12, Experiment 5.—Saline containing 0°75 per cent. of sodium chloride, and the amounts of chloroform shown in the table, the maximum being 08 per cent. originally in solution. In this and succeeding experi- ments the amount of chloroform was known directly by weighing out as described above. The temperature in this and succeeding experi- ments was 40° C. The results of this experiment are shown in Curve 1, fig. 6. Fia@. 6. Hesiiacs Hite “Ee i a cE : fat a {Ginied Haehoiobin aay See ee Ee, or me on hee 4 aie pnewacelt| UAT | Tider o ist F8 | |S cal | ra Ee | ais SI y weight in solution. —> Ee c a sss a lied os ee ee Me 7] iS i zeal] sl Se N ES \ , Nie SS SI FAECES . fe) Bar RS heal q Ls \ \ y \ UES Ee 2 | PEE [PES | | |= LN SINES ig | : No ci CI se ee HE ree ea El ‘EEE oO iS HH of Ie <—Percentage of chloroform b (e} SE ma ial Rae Ss 50 100 50 <—Pressure in m.m. of Mercury.— 550 2 8 1904.] Properties of Solutions of Chloroform in Water, ete. Experiment 5.—Saline (0°75 per cent.). Temperature 40° C. 409 Percentage | by weight of chlorotorm originally introduced. SOororo aS Sle io > Pressure of chloroform in vapour space in mm. of mercury. Percentage by weight of chloroform pumped off into vapour space. QAOQOOQOQOeeoee2 e O) ° . OE ° ° ° ° lop) = ON Percentage by weight of chloroform remaining in solution. 0328 "0473 ‘0657 "0819 1596 "2385 *3205 3978 "4.793 5621 “6384: SQQoeegqgogeqgee Coefficient of distribution between vapour space and solvent. ° eel etl ee Oe el ol! BE RWERWEREBRWE SH OBSHONABDDH Experiment 6.—Serum containing amounts of chloroform shown in table by direct weighing. Results shown in Curve 2, fig. 6. Percentage by weight of chloroform originally introduced. =) o>) = ne oS DAN a ON EWWNMNEHOOOCOCOCOCOCOCCoO SHONONASCHZAUAK HD Experiment 6.—Serum. Temperature 40° C. Pressure of chloroform in vapour space in mm. of mercury. 6°80 8°51 ‘64: 09 “74 12 27 43 “46 “31 b4: °22 Percentage by weight of chloroform pumped off into Vapour space. SIeieoeiolo oe Clo SO DOO OOS S ° . . e e e bd Ore e . ° ° e ° ice) bo = | | Percentage by weight of chloroform remaining in solution. WONNFRrHOCOCOOCOCOCCCoCoO e e s e e « s Sc < ° . e e . rs oy) J iv) “i 2.2) j=) Coefficient of distribution between vapour space and solvent. DAANMNMNAIARDMIAIAWITIT OS MO BEEBE ee Ee ee ee ee ee WN ADSOUTIHWARME DEH NMAWAN 410 Prof. B. Moore and Mr. H. FE. Roaf. [Apr. 12, Experiment 7.—Solution of hemoglobin, of equal strength in hemoglobin to blood from which obtained, and containing the amounts of chloroform shown in the table. The results are shown graphically by Curve 3 of fig. 6. Experiment 7.—Hemoglobin Solution. Temperature 40° C. | Percentage | Pressure of poe Percentage Coefficient of ‘Gnlereform |wapour apace, | choweform | VS et |G originally in mm ie Berne or remaining in | vapour space Bae: : ed into vapour ee P I introduced. mercury. space solution. and solvent. | 0 02 2°85 0 -0018 0 -0182 1: d0e8 0-04. 7°31 0 -0045 0 -0355 Le aga 0-06 14°52 0 -0089 0 0511 1s fess | 0°08 1 aS 0 -0110 0 -0690 1: (68 | Oval 22°34 0 -0188 0 :0862 1638 0°2 43°85 0 -0270 0 °1730 1: 6°4 0°3 62°01 0 -0382 0 -2618 Le brs 0 *4 93°15 0 0573 0°3427 1: -5°9 0°5 114 °91 0°0707 0 4393 1s” 62 0°6 136 47 0 -0840 0 °5160 | 12 62 0°8 197 -10 0°1213 0 6787 1: bb 1°0 222 -64 0°1371 0 -8629 1: 653 1°5 262 °26 0 -1614 1 *3386 Ls 9Sss 2-0 | 268 -00 0 -1650 1 -8350 1: ee IV .—Solubilities of Gases in Serum and Hemoglobin in Presence of Chloroform, It was thought that such compounds as are shown by the above experiments to be formed between chloroform and serum or hemoglobin solution, might interfere with the carriage of oxygen and carbon-dioxide by the blood. Accordingly, experiments were carried out upon serum and hemoglobin solutions to test this point. A volume of about 500 ¢.c. of serum or of hemoglobin solution obtained as for the experiments in Section 3, was completely deprived of gases by exhaustion with a Topler pump at 40° C., afterwards saturation with air or air and carbon-dioxide, was carried out upon two equal volumes contained in similar bottles to one of which a sufficient quantity of chloroform was added to make a 1]-per-cent. solution, while the other served as a control. The gases dissolved in each case were then collected by means of the Topler pump as before, and analysed. The results throughout were negative, and thus proved a fortzorz that at the anesthetising values chloroform, does not depress the solubility of the respiratory gases in the blood. | 1904.] Properties of Solutions of Chloroform in Water, etc. 411 As an example, an experiment with serum shaken up with a mixture of air and carbon-dioxide may be quoted :— A yolume of 500 ¢.c. of serum was exhausted as above described. (a) A volume of 150 c.c. of this serum was poured. into a 500 c.c. stoppered bottle, and shaken up with a mixture of air and carbon- dioxide. Exhaustion and analysis of the gases in 70 c¢.c. gave the following results at 14° C. and 759 mm. :— SOo ce © — Cece, N= 54 ec. (b) A second volume of 150 c.c. of the exhausted serum treated exactly similar, but with 1°5 grammes of chloroform added, gave the following results from 70 c.c., at the same temperature and pressure :— COz 41:4 ¢¢,O=18 ec. N = 6:4 ce. There was obviously a slight leakage of air, but the figures are sufficient to show that there is no aaqmecnelils change 1 in the ) 9 O as 2ST > I a of length. 10-millionths Fie. 5.—Curve (G) shows changes of thermoelectric force for an unstressed iron wire; (I) same when the wire was loaded with 1620 kilogrammes per sq. cm. Curves (e) and (f) show changes of length for iron loaded with 351 and 1600 kilogrammes. Crosses on (F) and (G) indicate (e) and (f) corrected for mechanical stress. * © Roy. Soc. Proc.,’ vol. 47, p. 474, fig. 2, 1890. 428 Dr. 5. Bidwell. On the Changes of [Apr 17) changes of length undergone by an iron wire when loaded with weights of 351 and 1600 kilogrammes per sq. cm., the curve for the first-named load being practically the same as for an unloaded wire. Curves (F) and (G) indicate the changes of thermoelectric force with field for an iron wire, first when loaded with 1620 kilogrammes per sq. cm., and afterwards without load, the points of observation being »marked by dots. The crosses on (G) and (F) indicate the course of the curves (¢) and (f) after correction for mechanical stress; the correspondence is so close that it was impracticable to draw the corrected curves separately. The vertical scale of the corrected curves is not the same in the two cases; that for the upper one is given in the right-hand margin, and that for the lower in the middle of the diagram. Data for the construction of these curves were obtained in the same manner as before, and are given in the annexed table :— Table ITI. P = tractive force in grammes per sq.cm. E = elongation. | | pes oe E, P/200 + B, | P/200 + B, = | 15s aay | Curve (e). |Curve (f).| Ourve (e). | Curve (f). 40 6.550 33°0 | 26 = 3 59 0 30 °0 50 7,000 35 °0 28 — 4 63°0 31-0 70 7,700 38 °5 oT — 6 65 °d Byes 100 8,500 42-5 | 3 —{1 BB) UE: Silay 120 9,000 45-0 | 20 —14 65 °O 31°0 150 9,500 AT °5 | 16 = 118) 63-5 29-5 200 10,250 51-0 10 — 25 61°0 26:0 300 11,100 Beas dy me — 40 53-0 15°5 375 | 11,500 57-5 = 10m) | 2858 Ay 5 onan 5-5 | It is to be noted that the three sets of experiments—for change of thermoelectric force, change of length, and lifting power—were made with three different samples of iron wire. The effect of annealing upon the changes of length and of thermo- electric force is illustrated in fig. 6. Curve (g) shows the elongation of a wire in the state in which it was bought,* while (4) indicates its behaviour after it had been carefully annealed. This operation was performed by enclosing the wire in an iron tube which was placed in a hot fire and allowed to cool gradually as the fire died out. Curve (H) shows the changes of thermoelectric force in a piece of the same kind of wire when in its original state, and (K) the modification which resulted from heating the wire to redness in a Bunsen flame and cooling it in air. In both cases the effect of annealing is to depress the curve. * “Roy. Soc. Proe., vol. 55, p. 230) fig: A, 189. 1904.] Thermoelectric Power produced by Magnetisation. 429 b Microvolts. y llionths of length é Cr Magnetié Field. a 20 jOo-mi Fie. 6.—Curve (H) shows changes of thermoelectric force in an unannealed iron wire; (K) changes in the same wire when annealed. Curves (g) and (4) show changes of length in another wire when unannealed and after annealing. A glance at any of the curves of thermoelectric force fer iron will show how easily errors might arise if it were assumed that a wire which had been subjected to a magnetic field became perfectly demagnetised when the field was withdrawn. In all my experiments the curves would have been much lower but for the demagnetisation by reversals before every observation. The dimensional ratio of the short wires which I used was generally so small, and the self-demagnetising consequently so great, that an apparent reversal of thermoelectric force in strong fields did not often occur. But in the case of a wire 17°5 cm. in length and 0:026 cm. in diameter, this spurious reversal appeared in fields above 500. When the demagnetised wire was subjected to a field exceeding this strength, there occurred a galvanometer deflection to the right, indicating a genuine increase of thermoelectric force ; but when the magnetising current was interrupted, the spot of light, instead of going back again, went still further to the right, the thermoelectric force due to the residual magnetism being greater than that due to the strong field. When the magnetising circuit was again closed before the wire was demagnetised, the spot of light, of course, moved to the left, and if the residual magnetism were disregarded, it would naturally be supposed that a reversal of thermoelectric force was indicated. The spurious reversal was also very conspicuous in the case of the steel wires. VOL. LXXIII. : 2H 430 Dr. 8. Bidwell. On the Changes of [Apr. 11, Nickel.—Since the effect of magnetisation upon a nickel rod or wire is to shorten it, while, if the “ correction” be admissible, iron is always lengthened by magnetisation, it was to be expected from analogy that the effects upon thermoelectric power would be opposite in iron and in nickel. In iron the thermoelectric power of the magnetised with respect to the unmagnetised metal is positive; in nickel, therefore, it should be negative. ‘This view, though in accordance with what is generally accepted as the fact, is, however, at variance with the results of all my experiments. Five different samples of nickel were used, curves for four of these (L), (M), (N), (O), being given in fig. 7. Pa OC 5S. oO 23 I50E 6 fe) foe} 3 8 =0 = 100E § = S55 = 20 32 fe) = 35 i 3) as 0 a lo-millignths. 00 Fie. 7.—Curves (L) and (M) show changes of thermoelectric force for pure nickel ; (N) and (O) the same for impure nickel. Curves (£) and (Z) give changes of length, and correspond to (L) and (O); (m) shows mechanical compression ; the dotted curve (z) is (4) inverted and plotted to the scale of ordinates in the right-hand margin. Curve (L) relates to a piece of thick wire 2°95 mm. in diameter, bought of Messrs. Johnson and Matthey as pure. The retraction curve* for another piece of the same wire is that marked (4). Curve (M) was given by a wire 0°65 mm. in diameter; this was also supplied by Messrs. Johnson and Matthey, and described as being ‘as pure as conveniently possible.” Curve (N) shows the result of an experiment with a sample of wire supplied by Messrs. J. J. Griffin and Sons, which did not purport to be pure and probably contains a considerable proportion of iron. . # ©Phil, Trans.,’ vol. 179, p. 228, fig. 6, 1888. 1904.] Thermoelectric Power produced by Magnetisation. 431 Curve (O) was obtained from a strip about 1:4 mm. wide and 0-75 mm. thick, which was cut from a rolled sheet purchased at a metal warehouse. Curve (/) shows the retraction of a strip of the same metal 9 mm. in width.* The fifth specimen was a wire taken from a piece of nickel gauze. With this, too, the magnetised was found to be always thermoelectrically positive to the unmagnetised metal. [A sixth specimen, consisting of a wire 3°5 mm. in diameter, has been recently tested, with the same result.—May 23, 1904.] The form of curve (N) which rises to a maximum at about H = 150 suggested a possible source of error by which Thomson may have been misled, and I therefore repeated his experiment. The arrange- ment which is shown in fig. 8, is essentially the same as that employed by Thomson. A piece of Griffin’s impure wire was bent into the shape of a horse-shoe, as shown, one of the limbs passing through the small magnetising coil, PP; the ends of the horse-shoe were connected by brass binding-screws to wires leading to the galvanometer T. Heat was applied at or near Q by touching the wire with a hot copper rod, and when the magnetising circuit was closed, the galvanometer T indi- cated a thermo-current which usually had the same direction as in the other experiments—from unmagnetised to magnetised through hot. It Fie. 8. was, however, found possible to adjust the strength of the field and the position of Q so that a thermo-current flowed in the opposite direc- tion. This was the case, for example, when the field-strength was 400 and the distance of Q from the coil 8 cm. Ifa slight change were made in the position of Q, the application of the same field again pro- duced a current in the normal direction. This deceptive effect, which, of course, really occurred between more strongly and less strongly magnetised portions of the metal, could not be obtained with the pure nickel wire, nor with the wire taken from the gauze, which were the only other specimens tested. With an iron wire, however, it was quite easily produced. Curve (m), fig. 7, shows in ten-millionths of length the mechanical compression due to magnetisation ; in comparison with that for iron, * Loe, cit., p, 214, fig. 4. Payal 432 Dr. 8. Bidwell. On the Changes of [Apras. curve (d), fig. 4, 1t is seen to be quite insignificant. Since no experi- ments have been made to determine the lifting-power of nickel, this was calculated from the expression (271?+ HI)/g. Values of I were obtained from a table given by Ewing, based upon results of an experi- ment by Du Bois ;* the specimen used was an ovoid instead of a cylindrical wire, but the values might be increased or diminished by 10 per cent. without influencing the corrected curve to any appreciable extent. Young’s modulus M was taken as 2:2 x 10°, approximating 2175 x 10°, the value foundy for nickel by H. Tomlinson; this also need not be very accurately known for the present purpose. Table IV shows how the ordinates of the compression curve (P/220), and those of the corrected curve (R- P/220), were determined. Table IV. P = (2r12+HI)/g. C = Px 107/M = P/220. BR = retraction in ten-millionths. { H. i | P C R. B&G: | —|—-- — ————. _ a ee z ey 100 313 659 3:0 136 133 200 375) | 977 4°5 181 177 300 AOG) | 1180 5:3 210 205 4.00 A287) Salts 6-1 224, 218 500 4A1 1470 6°7 232 225 600 450 | 1572 7-1 239 232 700 456 | 1657 75 242 234. | 800 459 | 1724 7:8 244, 236 | 1200 471 | 1997 9-0 245 236 Curve (n) which nearly coincides with the thermoelectric curve (L) is the uncorrected retraction curve (/) inverted and plotted to the scale of ordinates in the right-hand margin. If the corrected curve were plotted to the slightly different scale just within the margin, it would be substantially identical with (nm), which may therefore be regarded as representing either the corrected or the uncorrected curve of change of length. The small difference noticeable in the forms of the two curves (OQ) and (1) for the two nickel strips is just such as would be caused by the difference between their dimensional ratios. As to the curves in fig. 9, which show a remarkable qualitative correspondence with regard to the complex effects of tensile stress, nothing need be added to what has already been said. The dotted curves{ are of course inverted. * ‘Magnetic Induction,’ 3rd edit., p. 164. + ‘Roy. Soc. Proc.,’ vol. 37, p. 390, 1884. t ‘Roy. Soc. Proc.,’ vol. 47, p. 475, fig. 3, 1890. 1904.] Thermoelectric Power produced by Magnetisation. 433 40 400 (P) . ate Bee) a @ 5 Ze pe 2 30; ee 5 a GB 3 5 = = O = 320 200 = : = = Q = 10 100.8 iD s ) oe 5 300 600 O oe Magnetia Field. Fic. 9.—Curves (P) and (Q) show changes of thermoelectric force in an annealed nickel wire when loaded respectively with 970 and 447 kilogrammes per sq. cm. Curves (0) and (p) show changes of length for an unannealed wire when loaded with 980 and 420 kilogrammes per sq. cm. Cobalt.—Experiments are made with two different specimens of cobalt, the results being given in fig. 10. 10-— 00 (S) | (R) = 2 = ae | | eae So fe) all eee 505 S 5 3 | eam a eee ea Fee 22 S) (Got ie eee a omen sa = as Uses (7) ee 5 ‘o) yo cae if | oz se 1 200 400 BHO _-- 860 000 00 1400 a) | BSeu ~~) Magnetic |Field. = | TS we & : = : 50 : | J Fie. 10.—(R) and (S) show thermoelectric changes for cast and rolled cobalt respectively ; (¢) shows changes of length for the cast cobalt; (7) is the curve of mechanical compression. Curve (R), showing thermoelectric changes, relates to a cast rod supplied by Messrs. Johnson and Matthey; the change of length curve (7) was obtained with another piece of the same rod.* This latter curious curve is quite unlike the thermoelectric curve (R), * ‘Phil. Trans.,’ vol. 179, p. 228, fig. 6, 1888. ae 434 Thermoelectric Power and Magnetisation. [Aor aot, and evidently cannot be made to resemble it by applying the correction for mechanical compression. The compression curve (7) was constructed, like that for nickel, from values of I furnished by Ewing; Young’s modulus was taken as 2x10°, the value found by Tomlinson* for unannealed cobalt being 2005 x 10°. The thermoelectric curve (S) was given by a strip of rolled cobalt, for which I am indebted to the kindness of Messrs. Henry Wiggin and Co., of Birmingham. It is of interest as indicating the similar behaviour of a very different sample of the metal. If there is any relation between the thermoelectric and the strain phenomena in cobalt, it is obviously disguised by some cause which has yet to be discovered. [Note added May 23, 1904.—Some further experiments have been made with specimens of cobalt which had been very thoroughly annealed. For cast cobalt the change-of-length curve is entirely altered by annealing, becoming, at least in fields up to 1360 units, a straight line represented by h = 0-056H, where R is the retraction in ten-millionths. The thermoelectric curve is considerably lowered, but in other respects is not much affected. No relation between the thermoelectric and the strain phenomena could be traced. From an examination of the curves in fig. 10 it appears that the thermoelectric power of the unmagnetised with respect to the magnetised cast cobalt is proportional to the compressive stress, and consequently to the square of the magnetic induction. | * Roy. Soc. Proc.,’ vol. 39, p. 530, 1885. ita - 1904.] Hxpervmental Determinations for Saturated Solutions. 435 “ Experimental Determinations for Saturated Solutions.” By the EARL OF BERKELEY. Communicated by F. H. NEVILLE, F.R.S. Received March 28,—Read May 19, 1904. (Abstract. ) The object of this research is the experimental determination of those physical constants of concentrated solutions, which are necessary for the tentative application of the gas-law equations. Saturated solutions were chosen because, presumably, dissociation is relatively at a minimum. This part ot the work deals with the densities and solubilities of the chlorides, sulphates, and nitrates of sodium, potassium, rubidium, cesium, and thallium, and also with their respective alums, except that of sodium. The densities were determined by means of a small pipette-shaped pyknometer, of about 5 c.c. capacity, the lower end of which was turned upwards and the upper, 120 mm. long, was graduated. ‘This was filled with the saturated solution and weighed, and from the known capacity of the pyknometer, together with the weight of solu- tion it contained, the density was calculated. The solubility was obtained by washing out the contents of the pyknometer and evapo- rating to dryness, the weight of salt left giving the solubility. The densities and solubilities were determined in two ways. In one the saturated solution, which was in contact with an excess of salt and continuously stirred, was cooled to the temperature of observation and the density and solubility determined. In the other an unsaturated solution was raised to the temperature of observation, being continuously stirred in contact with an excess of salt (in both cases the solution is kept at the temperature of observation by means of a thermostat), and the density determined at intervals of about 12 hours until constant. This constant density and resulting solubility was averaged with the.density and solubility obtained in the first method, and the mean was assumed to be the true density and solubility of the saturated solution. The observations were made at intervals of 15°, in this manner, between 0° C. and 90° C. The constants were also determined at the boiling point of the saturated solutions in an apparatus in which steam was caused to bubble vigorously through the solution with excess of salt, until the temperature became constant, this constant temperature being assumed to be the boiling point. The boiling point itself was not accurately determined, as it was 436 Earl of Berkeley and Mr. E. G. J. Hartley. [Apr. 21, found that no emergent column correction could be satisfactorily applied ; the pressures, however, under which the saturated solutions boiled were recorded. The results are given in a tabular form at the end of the paper. “A Method of Measuring directly High Osmotic Pressures.” By the Earl of BERKELEY and E. G. J. HarTLEY. Communicated by W. C. D. WuetTHAM, F.R.S. Received April 21,—Read June 2, 1904. This paper gives an account of some preliminary experiments made in furtherance of a scheme of work outlined by one of us in a communication to the Royal Society.* The ordinary method of determining osmotic pressures, 7.¢., that adopted by Pfeffer, Adie,t and others, is evidently not suitable for high pressures; the difficulty of attaching the manometer to the porous pot in a manner such that it will not move at the junction is practically unsurmountable. It seemed likely that if a porous plate were tightly squeezed between two hollow hemispheres, the necessary conditions of stability might be attained, and the first apparatus carried out this idea. It was made by Messrs. Miiller in 1901, and, being of glass, was designed only to stand moderate pressures, though it was hoped that, ina way described below, it might be used to measure indirectly the osmotic pressures of concentrated solutions. It consisted of two glass globes A and B (see fig. 1), holding the porous plate C between them. The plate was glazed round the edge, and carried the semi-permeable membrane of copper ferrocyanide on the face adjacent to the solution. A rubber ring on either side of C and between it and the glass flanges of A and B, served to form a watertight joint when A, B and C were strongly pressed together by means of a suitable brass fitting, which, however, is not shown in the figure. It had been intended to put solutions of different concentrations in the two vessels, and measure the difference between their osmotic pressures ; 1t was found that, although it would be possible to obtain the desired result, yet the time taken for the pressure to develop was too long. An experiment with a solution of 114°7 grammes of sugar in the * Karl of Berkeley, ‘‘ Experimental Determinations for Saturated Solutions,” read May 19, 1904, see p. 435. + W. Pfeffer, ‘Osmotische Untersuchungen,’ Leipsic, 1877. R. H. Adie, ‘ Jl. Chem. Soce.,’ 1891, p. 344. 1904. | Method of Measuring High Osmotic Pressures. 437 litre, gave an osmotic pressure of 8°13 atmospheres, the theoretical value, derived from the gas laws, is 7°89. This apparatus was discarded, and attempts were made to determine osmotic pressures in a quicker way by a method somewhat similar to that which Tammann” tried, 2.¢., by directly applying to the solution a gradually increasing pressure until the osmotic pressure has been reached, and meanwhile noting the change in the volume of the solvent. We replaced the two glass globes by iron cylinders, and connected one of them, B, which was filled with the solution, to a iw 1 --Manometer Rubber powerful pressure apparatus. Water filled the other, A, and also a graduated glass capillary, connected to A by a perforated rubber stopper. Pressure was applied to the solution and the rate at which the water rose in the capillary was noted. The rate was found to be, apparently, slightly different before the osmotic pressure had been reached to that which it was after that pressure had been passed. The phenomenon was so complicated, however, by the volume effect due to the compression of the washers placed on either side of the plate, that no very clear results could be obtained. * G. Tammann, ‘ Zeit. f. Phys. Chem.,’ vol. 9, p. 97. 438 Karl of Berkeley and Mr. E. G. J. Hartley. [Apr. 21, Final Form of Apparatus.—We had foreseen this defect, and a means of overcoming it, should it mask the desired result, was suggested by. Dr. Burton, of the Cambridge Scientific Instrument Company. It con- sisted in replacing the plate by a porcelain tube; surrounding the e SSS aes cS s y Ny bad B NN 2G AY LLL are middle portion only by the solution and leaving the ends free. Thus, when the pressure was applied to the solution, the compression of the ‘“‘dermatine” rings, which confine it, would have no effect on the volume of water inside the tube. This is embodied in the final form of apparatus which is shown, in vertical section, in fig. 2. 1904. | Method of Measuring High Osmotic Pressures. 439 The porcelain tube, AA, was made of Dr. Puckal’s paste by the Royal Berlin Factory, and of a porosity slightly less than that of the porous pots they make for Dr. Thorpe’s arsenic apparatus. It is about 15 c.m. long, and has an outside diameter of 2°5 ¢.m., and an inside diameter of 1 cm. The ends are glazed and the membrane was formed on the whole length of the outer surface. The outside brass easing BB, contained the solution, to which the pressure was brought by the steel pressure tube C. The porcelain tube is secured to B by the “‘dermatine” rings DD, and the latter are strongly compressed, between B and the metal sleeves HE, by screwing down the nuts FF, the lateral expansion causes them to grip the tube tightly. GG are holes perforated through B, and through the sleeves HH, the use of which will be explained later on. H is another perforation through B which allows the apparatus to be emptied without taking it to pieces ; in several experiments a Schaffer and Budenberg pressure-gauge was attached here, ordinarily it is closed by a screw-down metal plug. The ends of the porcelain tube are closed by pieces of thick-walled rubber tubing II, through which the brass tubes JJ pass; the length of rubber is such that they extend further into the tube than the distance between the ‘‘dermatine” rings and the ends of the tube. A watertight jomt between A and J is obtained by compressing the rubber tubes between the metal washers KK, and the screw sleeves LL. The brass tubes JJ are joined, in one case to a glass tap, and in the other to a glass capillary by rubber tubing. The glass capillary is graduated in millimetres and was calibrated; 1 cm. of the bore contains ‘0042 c.c. : The operation of measuring the osmotic pressure of a solution consisted in filling the inside of the porcelain tube with water, and the surrounding vessel B with the solution ; then noting the rate at which the level of the water in the graduated tube moves while the pressure is being gradually increased on the solution. Theoretically, as long as the osmotic pressure has not been reached, the level of the water in the capillary should fall; when the osmotic pressure is exceeded it should rise and the “turning point” should give the osmotic pressure. In either case the rate at which it moves is a function of the difference between the osmotic pressure and the pressure on the solution. In the actual experiments, although the level rose and fell, the “turning point” was at some other pressure higher than the osmotic pressure; the chief cause of this difference was that however tightly the ‘“dermatine” rings were compressed against the porcelain, the sugar solution leaked past them. ‘This was doubtless because the membrane could not be formed quite on the outer surface of the tube, and consequently a very narrow ring of the tube was lett open. The leak was so small that no difficulty was experienced in keeping the pressure up, even at 50 atmospheres, but Bid. 440 Earl of Berkeley and Mr. E. G. J. Hartley. [Apr. 21, the effect of the leak was gradually to saturate with solution the surface of the exposed portion of the tube, and thus abstract water from the inside through the membrane. We call this leak the guard-ring leak, and it was hoped that the following method gave us a means of estimating it. Guard-Ring Leak Correction.—The holes GG (see fig. 2), were bored, and a stream of the solution whose osmotic pressure was to be determined, was directed through them; and to insure that the surface was thoroughly saturated, the whole apparatus was placed in a bath of the same solution. The inside of the tube together with the capillary having been filled with water, and the temperature of the bath and apparatus having become constant, the rate of fall of the level in the capillary was noted. It should be mentioned that the space enclosed by B was also filled with water, and C was replaced by an open glass tube. The level of the water in this open tube was kept at the same height as that in the capillary, so that any small change in the temperature of B did not alter the rate of fall of the level in the capillary. The guard-ring leak correction was found to be the larger the more concentrated the solution, and it also varied shghtly with the height of the water in the capillary. This correction was of a magnitude such that it rased the “ turning point,” with reference to the pressure, by an amount equal to from 5 to 10 per cent. of the osmotic pressure. The ‘Solution-Leak Correction.—We have hitherto been unable to make semi-permeable membranes completely impervious to sugar; on testing the water from the inside of the tube a trace was always indicated. It was, therefore, considered advisable to determine the amount of sugar which had come through in each experiment. At the end of the experiment the inside of the tube was washed out without taking down the apparatus, and its content of sugar analysed by means of Fehling’s solution. On the assumption that the resulting quantity of sugar indicated that a corresponding amount of solution had come through the membrane, and on the further assumption that the rate at which this solution came through was proportional both to the time during which it had been subjected to pressure and to the amount of that pressure, the displacement of the “turning point” was calculated. It was found that, with a good membrane, the solution leak correc- tion was of a magnitude such that it lowered the “ turning point,” with reference to the pressure, by an amount equal to from 2 to 3 per cent. of the osmotic pressure. In this connection it may be pointed out that a check on the algebraical sum of these two corrections may be obtained from the experiments themselves. For by beginning an experiment at a 1904. | Method of Measuring High Osmotic Pressures. 44] pressure on the solution less than its osmotic pressure, and gradually increasing the former until the latter is exceeded, then gradually reducing the pressure until again below the osmotic pressure, the application to the corresponding capillary readings of a formula involving the guard-ring and the solution-leak rates will enable one to calculate the sum of the two. This sum, deduced in this manner from the capillary readings in an experiment with a solution containing 180 grammes in the litre, was 5:36 cm. The separately determined “ guard-ring leak” correction was 10°17 cm., and the difference, 4°81, represents the calculated “solution leak.” The observed ‘solution leak,” determined by analysis, and calculated on the assumptions enumerated above, gave 5:03 cm. Some importance is attached to this method of working as it appears to afford a means of determining the osmotic pressures of solutions of substances for which no truly semi-permeable membranes have yet been found. The Pressure Apparatus—The pressure is obtained by means of a vertical steel plunger working in asteel cylinder. ‘The plunger is forced into the cylinder by aniron lever, at one end of which weights are hung, and the bottom of the cylinder is connected with C (see fig. 2). The cylinder and pressure tube are filled with the solution whose osmotic pressure is being determined, and the plunger is made to work almost pressure tight by a “dermatine” ring, sleeve and nut, similar to that shown at D, E and F, in fig. 2. A horizontal lever is attached to the plunger and is worked to and ine, at intervals, so as to keep the pressure on. The pressure ppplicd when different weights were placed at the end of the lever was determined by connecting the press-tube to a Schaffer and Budenberg standard gauge. The Formation of the Senu-Permeable Membranes.—Numerous different ways of forming the copper ferrocyanide membrane were tried; the most successful was by first depositing the film by diffusion in the manner Pfeffer* recommends and then finishing it by the electrolytic methods due to Morse and Horn.t The porcelain tubes were immersed in the copper sulphate solution (50 grammes of CwSQ,,5a¢ in 1000 c.c.) in a desiccator, and exhausted free of air; they were then taken out and their inner and outer surfaces dried with filter paper. The endg were then closed by rubber plugs and they were allowed to dry in the air for from 4 to | hour, after which they were quickly plunged into a solution of potassium ferrocyanide (42 grammes in 1000 c.c.). When the membrane was seen to be fairly uniform in colour, the tubes were transferred to the electrolytic cell. This consisted of a beaker contain- ing the ferrocyanide solution in which a platinum foil electrode. pe loc. cit. + H. N. Morse and D. W. Horn, ‘ American Chemical Journal,’ vol. 26, p. 80. 449 Method of Measuring High Osmotic Pressures. [Apy. 21, connected with the negative pole of the battery, was suspended. This electrode was surrounded by a porous pot so as to prevent the solution, when it had been rendered alkaline by the passage of the current, from reaching the membrane on the tube. As mentioned by Morse, we found that the alkaline solution acted injuriously on the membrane. The porcelain tubes were filled with the copper solution, and were furnished with copper electrodes. ‘The voltage used was 100; and when the resistance of the tubes had risen to a constant value, they were taken down and soaked in distilled water for several days. They were then again set up, and the current passed until the resistance was again constant, generally at a higher value than before, upon which they were taken down and washed. ‘This process was repeated till no further change in resistance took place. The highest resistance obtained was 170,000 ohms ; but out of some fifty tubes only eight or nine reached this figure. These latter were those selected for the experiments, and it. was found advisable to remake the membranes electrolytically after they had been subjected to considerable pressures. [May 21.—We have lately found that the membranes are greatly improved if they be re-made electrolytically under high pressure. The pressure should be applied to a very strong solution of sugar in which the potassium ferrocyanide has been dissolved and which surrounds the outside of the tube ; the inside of the tube being filled with the usual copper sulphate solution and remaining under atmospheric pressure. | ?esults.—The following results with cane sugar were obtained ; the experiments were made more for the purpose of testing the method than to get accurate observations of the osmotic pressures; but we think they are within 10 per cent. of the true values :— Pressure when Concentration in| Pressure at which | “ turning point” Py Peace . aavney) ressure deduced | grammes per the “turning point ’’| corrected for guard- from Borie litre. occurred. ring and solution y . leaks. 120°7 10 °5 atmos. | 9-5 atmos 8-4 atmos 180-0 Lora erie | 14 A. 12 See 249 -O I) 2m ak MONS ae i6-7 360 -0 % | BO.’ 25: 420 0 45:9, Kawe cn Chaie: 29°2 Me Advantages of the Method.—It may be of use briefly to point out what seem to be the advantages of this method :— * In this experiment the actual “turning point” was not reached. 1904. ] Colours in Metal Glasses and Films. 443 1. The membrane, owing to the fact that there is no glazed cap to the tube, goes right up to the ends.* 2. There is no necessity to have pressure-tight joints. 3. Greater speed in working; the actual experiment takes from 2 to 3 hours only. 4. The form of tube used is such that it will withstand very high pressures—it may be of interest to mention that one of the tubes and membrane stood a pressure of 120 atmospheres without apparent harm. We publish this preliminary notice as it will be some time before the experiments can be continued—a new apparatus has to be cast, and new porcelain tubes are required. We hope, by means of the new apparatus, to reduce greatly the guard-ring leak. We are glad to avail ourselves of this opportunity to thank Mr. W. C. D. Whetham for the kindly interest he has taken in the research and Mr. H. Darwin for designing the pressure apparatus. “ Colours in Metal Glasses and in Metallic Films.” By J. C. MAXWELL GARNETT, B.A., Trinity College, Cambridge. Com- municated by Professor Larmor, Sec. R.S. Received April 19,—Read June 2, 1904. ( Abstract.) The first part of the paper is devoted to coloured glasses. The phenomena which it seeks to explain were observed by Siedentopf and Zsigmondy.t Expressions are first obtained for the electric vector of the light scattered from a small metal sphere when a train of plane polarised light falls upon it, the investigation following Lord Rayleigh.t By means of these expressions it is proved, from the diagrams and statements given by Siedentopf and Zsigmondy, that the metal particles which they observed in gold glass are spherical in shape when the diameters are less than 10°° cm. The fact that such particles are spherical throws light on the manner in which metals erystallise out of solution, the particles taking first a spherical form under the action of surface tension, and later, when they become too large for the forces of surface tension to overcome the crystallic forces, becoming amenable to the latter. Mr. G. T. Beilby has ‘previously arrived at similar conclusions.§ An investigation into the optical properties of a transparent medium * Cf. Adie, loc. ett. + ‘Ann. der Phys.,’ January, 1903. t ‘Phil. Mag.,’ vol. 44, 1897, and ‘ Collected Papers,’ vol. 4, p. 305. § ‘Brit. Assoc. Report,’ Southport, 1903. 444. Colours in Metal Glasses and Films. [Apr os containing metal spherules, so that the average distance between two neighbouring spheres is considerably less than a wave-length of light, is next undertaken. It is shown that every such medium has a pertectly definite colour by transmitted light, depending only on the optical constants of the metal of which the spheres are made, on the refractive index of the substance in which they are embedded, and on the quantity of metal, but not on the size or distance apart of the spheres. The intensity of the absorption of light of each colour is pro- portioned to p», the volume of metal per unit volume of medium. It is calculated, by means of the metal constants given by Drude,* that, with glass of refractive index 1:56, gold glass is more red than yellow, silver glass a little more yellow than red, copper glass con- siderably more red than yellow, and ‘“ potassium-sodium ” glass much more blue than yellow, provided always that the average distance between two neighbouring particles of metal in the glass be consider- ably less than one wave-length; in which case, as stated above, the particles must be spherical. Metal glasses for which this provision is satisfied will be called “regular.” It is next proved that the presence of metal spheres accounts for the optical properties of regular gold ruby glass, and that the irregularities in the effects of colour and polarisation, sometimes exhibited by gold glasses, are due either to excessive distance between adjacent gold particles or to excessive size of such particles—the latter, however, involving the former. Experiments are described, proving that this regular colour can be produced in a colourless metal glass, containing the metal in solution (which is the state in the manufacture of gold or copper ruby glass before the second heating) by the f-radiation from radium. Thus, a piece of clear gold glass and a piece of clear soda giass were exposed to the emanation for two days, when the gold glass had acquired an unmistakable pink tint, while the soda Bes had turned an intense blue-violet. In the second part of the paper, the optical properties of media built up out of metal spheres as before, but now so that the volume of metal may have any value between zero and unity, instead of remaining very small, as in metal glasses, are investigated. The changes of colour, and the final change to almost complete trans- parency, observed by Mr. G. T. Beilbyt in gold and silver films, are accounted for. Explanations are also given of the changes of colour on heating, observed by Professor R. W. Wood}, in potassium and sodium films deposited on the insides of exhausted glass bulbs. The * «Phys. Zeitschrift,’ January, 1900. + ‘Roy. Soc. Proc.,’ vol. 72, 1908, p. 226. * ‘Phil. Mag.,’ 1902, p. 396. ge 1904. | The General Theory of Integration. 445 increase in strength of colour, which was generally observed in the light transmitted through these films when the plane of polarisation of obliquely incident light was changed from that of incidence to a perpendicular position is accounted for. In Part III some evidence is brought to show that the allotropic silvers obtained by Carey Lea* are particular cases of the media which have been considered in the second part. “The General Theory of Integration.” by W. H. Young, Se.D., St. Peter’s College, Cambridge. Communicated by Dr. E. W. Hopson, F.R.S. Received April 25,—Read May 19, 1904. (Abstract. ) The paper begins with a recapitulation of the well-known definitions of integration and of upper and lower integration (intégral par excés, par défant; oberes, unteres Integral). The theorem on which the Darboux definition of upper (lower) integration is founded is stated and proved in the following form :— Given any small positive quantity ¢, we can determine a positive quantity ¢, such that, if the fundamental segment 8 be divided up in any manner into a finite number of intervals, then, provided only the length of each interval is less than ¢, the upper summa- tion of any function over these intervals differs by less than ¢ from a definite limiting value (the upper integral). Next follows a discussion as to whether it is admissible to adopt a more general mode of division of the fundamental segment than that used by Riemann, Darboux and other writers, when forming summa- tions (upper, lower summations), defining as limit the integral (upper, lower integral), of a function over the fundamental segment. It is shown by examples first that the restriction as to the finiteness of the number of intervals into which the fundamental segment is divided cannot be removed without limitations; but that it can be removed, provided the content of the intervals is always equal to that of the fundamental segment. Secondly it is shown that the error introduced by taking the summation over an infinite number of intervals whose content is less than that of the fundamental segment, is not in general corrected by adding to the summation the content of the points external to the intervals multiplied by corresponding value (upper, lower limit) of the function. Similarly it is shown that the more * © Amer. Journ. of Science,’ 1886. VOL. LXXIII. 21 446 Dr. W, H. Young. [Apr. 23, general division of the fundamental segment into component sets of points, whose content plays the part of the length of the intervals in the original definitions, leads to summations which do not, in general, | have a definite limit even for integrable functions. The lower limit of such generalised upper summations is shown to be not less than the upper limit of such generalised lower summations ; but it is shown that in general, only in case of upper continuous functions does the former give us the upper integral, and in the case of lower semi-continuous functions does the latter give us the lower integral. In general, introducing the terms outer and inner measure of the integral for these limits, the lower integral is less than the inner measure, which is less than the outer measure, which is less than the upper integral. The property of semi-continuous functions just mentioned leads to a new form of the definition of the upper (lower), integral in this case, namely, as follows :— Divide the fundamental segment S into a finite or countably infinite number of measurable components, multiply the content of each component by the upper (lower) limit of the values of an upper (lower) semi-continuous function at points of that component and sum all such products; then the lower (upper) limit of all such summations for every conceivable mode of division is the upper (lower) integral of the semi-continuous function. Introducing upper and lower linuting functions,* we then have the following theorem :— The upper (lower) integral of any function is the upper (lower) integral of its associated upper (lower) semi-continuous function. This leads to a new definition of upper and lower integration, which is as follows :— Divide the fundamental segment into any finite or countably infinite number of measurable components, multiply the content of each component by the upper (lower) limit of the maxima (minima) of the function at poimts of that component and sum all such products ; then the lower (upper), limit of all such summations for every conceivable mode of division is the upper (lower) integral of the function over the fundamental segment. This gives us also a definition of the integral in the case when it exists, that is, when the upper and lower integrals are equal. This form of the definitions is at once extendable to the case when the fundamental set S is any measurable set whatever, we merely have to replace the word segment by set, or more precisely by measurable * “On Upper and Lower Integration,” ‘ Lond. Math. Soc. Proc.’ 1904. ] The General Theory of Integration. 447 set. A particular form of division of S, analogous to that by means of intervals of the same content as the fundamental segment, is shown to lead infallibly to the upper and lower integrals of any function with respect toS; this mode of division is called division of S by means of segments (é, é), it is such that each component lies inside a correspond- ing interval of length less than ¢, the content of these intervals being less than S + ¢’, and the poimts of S which are not internal to the intervals forming a set of zero content. Based on this division of the fundamental set, we have an alternative definition of upper and lower integration with respect to a fundamental set, which is more nearly allied to the Darboux definitions for the case when the fundamental set is a finite segment. This is as follows :— Let the fundamental set, excluding at most a set of points of zero content, be enclosed in or on the borders of a set of non-over- lapping segments each less than ¢, and of content less than S+e. Then let the content of that component of 5S in any segment be multiplied by the upper (lower) limit of the values of the function at points of that component, and let the summation be formed of all such products. Then it may be shown that this summation has a definite limit when ¢ is indefinitely decreased, independent of the mode of construction of the segments and the mode in which e approaches the value zero. This limit is called the upper (lower) integral of the function with respect to the fundamental set S. In the case when the upper and lower integrals coincide, the func- tion is said to be integrable with respect to 8, and the condition of integrability is found in a form agreeing completely with Riemann’s condition in the case when S is a segment. ‘To prove this the theorem is required that the swm of any finite number of upper integrals of upper semi-continuous functions with respect to a fundamental set S is the upper integral of their swm, and the proof of this theorem is given. It is then shown that, except in the case of upper (lower) semi- continuous functions, the upper (lower) integral over the fundamental set S is not necessarily equal to the sum of the upper (lower) integrals over any set of components of 8, but that this is the case when S§ is divided by means of segments (¢, ¢’). A function which is integrable with respect to S is shown to have the following properties :— (1) It is integrable over every component set of S. (2) The integral of the integrable function is equal to the sum of the integrals over every finite or countably infinite number of components into which S may be divided. (3) The sum of the integrals of any finite number of integrable 212 Pada | 448 Dr. W. H. Young, [Apr. 23, functions over S is equal to the integral of the sum of those functions over S. In § 21 the calculation of upper and lower integrals with respect to any fundamental set S is reduced to a problem of ordinary integra- tion. The formule, which are similar in form to those already given by the author for the case when § is a finite segment, in a paper presented to the London Mathematical Society, are as follows :— The upper integral of any function with respect to a measurable set S is Kr KS+ Idk, JK where K is any quantity not greater than the lower limit, and K’ not less than the upper limit of the function for points of S, I being the content of that component of S at every point of which the maximum of the function is greater than or equal to £. The lower integral is K KS - Jdk, JK J being the content of that component of S at every point of which the minimum of the function is less than or equal to &. These formule lead to certain theorems with respect to the distribu- tion of the values of an ordinary continuous function and of an integrable function. The remainder of the paper is devoted to the discussion of the inner and outer measures of the integral of any function, and in the case when they are equal of the generalised integral of a function, which is, in this generalised sense, integrable. In particular it is shown that such functions are none other than the functions which Lebesgue has named summable, and the generalised integral is shown to be identical with the Lebesgue integral in the case when S is a finite segment; a geometrical interpretation of the integral, similar to that used by Lebesgue, is given in the general case. Contrasting the first definition given of the generalised integral with the geometrical definition, it is seen that they stand to one another in the same relation as the ordinary definition of integration, say of a continuous function, to its definition as a certain area. Just as, however, the mathematical concept of area is more complex than, and, indeed, depends on that of length, so does the theory of the content of a plane set of points depend naturally on that of a linear set. Just as the determination of area requires the application of the processes explained in the first definition of integration of continuous functions, so with the content of a plane set. ‘Thus the comparative simplicity of the geometrical definition is only apparent. 1904. ] The General Theory of Integration. 449 Lebesgue’s theorem that the sum of two summable functions is a summable function and its integral is the sum of their integrals is then proved by geometrical considerations, and a more general theorem is given, Viz. :— If X° and X! be the outer and inner measures of the content of the ordinate section of a measurable set by the ordinate through the point z, X° and X* are both summable functions, and the generalised integral of either is the content of the measurable set. It is here assumed that the content of the set got by closing the measurable set is finite. The content of any measurable set, with this restriction only, is thus obtained in the form of a generalised integral and, therefore, of an ordinary integral ; in fact— The content of any measurable set (provided the set got by a it has finite content) is | Idz. Here I is the content of the component of the fundamental set at which the inner (or the outer) measure of the content of the ordinate section of the given set is greater or equal to /. It is to be remarked that though in this abstract reference has only been made to linear and plane sets and to the corresponding integrals, the arguments are perfectly general and apply to space of any number of dimensions. For instance the concluding result is as follows :— To find the content of a measurable n-dimensional set, take any hyperplane section and project the whole set on to this hyper- plane. Any measurable set containing this projection we take as the fundamental set 8. Divide S up in any way into a finite or countably infinite set of measurable components, and multiply the content of each component by the upper (lower) limit of the values of the (linear) inner or outer content of the corre- sponding ordinate sections of the given set, summing all such products, the lower (upper) limit of all such summations is the content of the given set. 450 Messrs. D. McIntosh and B. D. Steele, [Apr. 26, “On the Liquefied Hydrides of Phosphorus, Sulphur, and the Halogens, as Conducting Solvents.—Part I.” By D. McIytosu and B. D. STEELE. Communicated by Sir WILLIAM Ramsay, K.C.B., F.R.S. Received April 26,—Read May 19, 1904. Ammonia, water, and hydrofluoric acid, are the only hydrides of the elements which have been systematically investigated with respect to their solvent properties, and in particular with respect to their power - of forming electrically conducting solutions. With the object of extending our knowledge of the properties of these hydrogen compounds, when liquefied, and in the hope that more light might thereby be thrown on the question of ionic dissociation, the investiga. tion, of which this is a brief abstract, has been undertaken. The following hydrides of the fifth, sixth, and seventh groups have been examined, hydrogen phosphide, sulphide, chloride, bromide, and iodide. Since it had been found by preliminary experiments that, with the exception of hydrogen phosphide, all these possessed the power of conducting the current when certain substances were dissolved in them, a number of physical constants were measured before proceeding to the systematic study of the conductivity of solutions. A brief summary of the results found hitherto is contained in the following tables, in which the temperatures are given to the nearest- tenth of a degree. 1.—The Vapour Pressure Curves. These were determined by the method described by Travers, Senter, and Jaquerod,* and used by them for the measurement of the vapour pressures of liquid oxygen and hydrogen. Hydrochloric Acid. ie Leg a. 138 Ee he = ANSE 141°0 oi | 3160 —85'9 648 — 104°5 198:°0 —92°9 430-0 —83°2 748 —101°3 245°0 —898 522°0 —80°5 868 Hydrobromic Acid. -— 104°2 96:0 —89°3 245-0 —76°7 501 —100°7 142-0 —87'1 284:°0 —74:0 575 — 96°3 185-0 —§3:0 357-0 —~707 682 a SEs 214:0 —79°3 431-5 —684 775 = “Phil, Trans) a7 1902) vol! 200) p. hes: 190+. | On Liquefied Hydrides as Conducting Solvents. , 451 Hydriodic Acid. dis iP 1 P Ay P — 779 74:0 —54°8 303°5 —43°5 530 — 735 92°0 —514 369°0 41°78 — 68°4 126°0 —50°0 376°0 —39'4 644 — 63°5 185°5 —47°T7 438:0 —369 713 — 995 224:0 —46°1 474-0 —35°9 769 Sulphuretted Hydrogen. —~ 84:0 193-0 —75:6 314-0 —69:1 456 org 2200 73-30) 3640 —66:1 538 mT 8s4. 270:0 eG 1. 400-0 —62:2 676 Phosphoretted Hydrogen. — 1059 237°0 —97°7 393-0 —886 644 —101:2 319-0 —93'1 498-0 —866 719 The melting and boiling points as read from the curves are as follows :— HCl. HBr. HI. H.S. HP. Meg: .....: pie 860 508 ss aun 135) eee — 82°9 — 68°7 — 3d°1 —6071 . —86°2 2.—The Densities. The densities of the pure liquids were determined over a wide range of temperature, and the values at the boiling point are given in the following table :— HCl. HBr. HI. HS. H3P. Density at b. p. ..... 1:195 2°157 2°799 0964 0-744 3.—TLhe Molecular Surface Energy. The surface energies were measured over a considerable range of temperature, using a modification of the method of Ramsay and Shields. In the table the value of the molecular surface energy (MV) at various temperatures is given. The values of d\(MV)3/dT and of the association factor « are tabulated separately. From the results it will be seen that of the substances examined, hydrogen, bromide, iodide, and sulphide occur as simple molecules ; whilst hydrogen chloride and phosphide are more or less associated. en =) "Tr wr Px 452 Messrs. D. McIntosh and B. D. Steele. [Apr. 26, Hydrochloric Acid. ayn d (MV). ay. \ (MY)3. ae A (MV)3. 163°1 263-7 175°8 244°8 187-2 229°3 168°5 255°9 180°1 239°0 189°9 223°6 1WAcs 250°8 183-2 233°6 192°6 221:0 Hydrobromic Acid. 181°9 330°1 “"188°9 314°6 198-2 294°8 184°8 3256 193-4 307°3 200°5 292°2 186°1 320°1 195°3 299°6 203°9 283°8 Hydriodic Acid. 225°3 367-0 230°9 309°3 235°0 348°0 227°1 362°8 232°9 301° 236°5 344°6 229°3 3D8°5 = at = aes Sulphuretted Hydrogen. 189-1 349-5 ogee, 334 2039 824-7 191°3 345°3 oom 328°3 206-9 316-7 1946 338-0 20is, 3266 2108 3086 Phosphoretted Hydrogen. GZal 287°2 175-4 273-4 <= = 171°8 279°6 LO 265-4 = —= Variation of molecular surface energy with temperature and the association factor— HCl. HBr. inde HS. H3P. dd (MV)3/dT... 1°47 2-03 1-99 1-91 1-70 ree ue Sites 15 1-0 1-0 1-1 1-4 4,.—The Viscosities, and Viscosity Temperature Coefficient. The viscosities of the pure liquids were determined by comparison with that of distilled water at 22°. The object of the measurements was to procure data for the comparison of the temperature coefficients of viscosity and of electrical conductivity. Viscosity Temperature Coefficient, dy/dT. Hydrochloric Acid. Hane 160°8 166°7 aile'7 147-0 183°2 188°2 Ts ovate ves 0:590 0:569 0:530 0-514 0°493 0-477 dn[dT = 0°88. 1904. | On Liquefied Hydrides as Conducting Solvents. 4.53 Hydrobromic Acid. ae 186°8 188°8 190°8 is)7) 1903 199-4 He: 0-911 0-902 0-890 0-877 0857 0-851 ania We— 03571. Hydriodie Acid. ere da 225°0) 22772), 7 2306. 231°5 23379, 23864 a: hiss AT Ole ody ASH a cA 26 14020 3 labs dy/dT = 0°70. Sulphuretted Hydrogen. tase sie: LO LSO 193°3 9832 201°2 206-1 209°8 PS ae 3.303, 0547 0528 0-510 0°488 0-470 0-454 dnfaT = 1-10. 5.—wNSolubilities and Conductivities. Of the substances examined at this stage, the organic ammonium salts were found to be readily soluble, and to give conducting solutions. Some doubt existed as to whether any metallic salts were dissolved ; if so, none were found to conduct the current. The only readily soluble inorganic substances were hydrogen chloride and bromide dissolved in sulphuretted hydrogen. It is somewhat remarkable that these solutions are perfect non-conductors. The conductivity of a few substances was accurately determined ; as, for instance, that of solutions of triethyl ammonium chloride in hydrogen bromide and sulphide. It was found that the molecular conductivity of these two solutions, and of all the others which were examined, increased enormously with increasing concentration, instead of showing a slight decrease, as in the case of aqueous solutions. The further study of the solubilities and conductivities forms the subject of another paper.* * Infra, p. 454. 454 On Liquefied Hydrides as Conducting Solvents. [Apr. 26, “On the Liquefied Hydrides of Phosphorus, Sulphur, and the Halogens, as Conducting Solvents.—Part II.” By E. H. ARCHIBALD and D. McIntosH. Communicated by Sir WILLIAM Ramsay, K.C.B., F.R.S. Received April 26,—Read May 19, 1904. In continuation of the preceding investigation,* a very large number of substances have been examined with regard to their solubility in the four substances—hydrogen chloride, bromide, iodide, and sulphide. The following is a brief summary of the results which have been so far obtained :— No salt of the metals has been found to dissolve in more than traces in either solvent, and in no case was it certain that such substances dissolved at all. We cannot, therefore, confirm the observation of Helbig and Fausti,t who state that stannic chloride dissolves in liquid hydrogen chloride, but does not form a conducting solution. On the other hand, many organic substances were found to be readily soluble; as for example, the amine salts, acid amides, and certain alkaloids among compounds containing nitrogen; and alcohols, ethers, ketones, phenols, and some organic acids and esters among compounds containing oxygen. In every case where a conducting solution was formed the dissolved substance was one containing an element the valence of which might be increased, thus dyad oxygen or sulphur becoming tetrad, or triad nitrogen becoming pentad. The conductivity of a Jarge number of solutions was measured, and in all cases the molecular conductivity increased enormously with concentration. No case has been met with in which the molecular conductivity varies in the same way as in aqueous solutions. The substances rarely conduct better than n/25 KCl. A few of the measurements are given in the following table, in which the concentration C is given in gramme molecules per litre, and the conductivity in reciprocal ohms x 10°. Substance. Solvent. C. Conductivity. Acebamide (24... 0 Gee eer HBr 0-011 65:2 ROU Ss Ge ek ur oe O-713 3155-0 Acetonitrile... 2 eee eee HCl 00463 1512:0 anti so lets ater Fi 1232 9636°0 Hithyl oxide... 02). ae HI 0-10 19°5 RIL 2 Geille sages es Ae 2208:0 Triethyl ammonium chloride... H2S 0:014 1170 Py as nk i 0-401 1580°0 * Ante, p. 450. + ‘Zeit. fir Angewandte Chemie,’ vol. 17, 1904. 1904.] On the Lymphatic Glands in Sleeping Sickness. 455 A large number of temperature coefficients have been measured. These were found to be in the majority of cases positive, 4 i.e., the conductivities increase with rise of temperature. All the experiments wnich have been hitherto carried out lead to the conclusion that it is the dissolved substances (7.¢., the acetamide, etc.) which carries the current and not the halogen hydride. In other words, we are dealing with solutions in which the organic and not the inorganic substance undergoes electrolytic dissociation. Further experiments are at present in progress, having for their object the measurement of the molecular weight of the dissolved substances (McIntosh and Archibald) and the determination of the transport numbers (Steele). Discussion of the results so far obtained is deferred until these experiments are completed. “Note on the Lymphatic Glands in Sleeping Sickness.” By Captain E. D. W. Greic, I.M.S., and Lieutenant A. C. H. Gray, R.A.M.C. Communicated by Colonel Bruce, F.ES., at the desire of the Sleeping Sickness Commission. Received and Read May 5, 1904. Captain Greig, in a letter dated March 17, 1904, writes that following a suggestion of Dr. Mott, they have examined the contents ot lymphatic glands during life from fifteen sleeping-sickness patients. In all of them actively motile trypanosomes were very readily found in cover-glass preparations taken from the cervical glands. They were also present in other glands such as the femoral, but were not nearly so numerous. They found the trypanosomes to be far more numerous in the glands than in the blood or cerebro-spinal fluid, and believe that the examination of fluid removed from lymphatic glands will prove to be a much more rapid and satisfactory method of diagnosing early cases of sleeping sickness than the examination of the blood. At first the glands were excised, but this was soon found to be unnecessary, as it is easy to puncture a superficial gland with a hypo- dermic syringe and suck up some of the juice into the needle and blow this out on a slide. The actively moving trypanosomes were readily found aiter a short search in these slides, when a prolonged search in similar preparations of the blood from the finger failed to discover them. In stained specimens, in addition to well-formed trypanosomes, there exist many broken-down remains, which suggests that a destruc- tion of the trypanosomes takes place in the glands. ; , Sa 456 On the Lymphatic Glands in Sleeping Sichivess. [May 5 The authors also examined the cervical lymphatic glands of the five natives suffering from trypanosomiasis who have been under observa- tion for the past year, and found actively motile trypanosomes in the liquid withdrawn from the glands in all of them. Tabula, one of these patients, is employed in the hospital, and the dispenser reports he is getting very stupid. The lymphatic glands were also examined for streptococci by staining and culture, but in every case were found to be sterile. Some of the cases, the glands from which were examined for streptococci, were very far advanced. The streptococcus invasion must, in the opinion of the authors, be a very late one and only occur shortly before death. Observations made upon the blood show a constant increase in the percentage of lymphocytes, but the total leucocytes are not increased. The authors consider that these observations throw a new light upon the glandular enlargements which have been so constantly noticed in sleeping sickness, and that the disease is essentially a polyadenitis brought about by the arrest of the trypanosomes in the glands where many of them are destroyed, but whence some escape from time to time into the blood stream and thus occasion the increase which has been observed in the peripheral circulation. They regard their observations upon the presence of trypanosomes in number ,in the lymphatic glands of both early cases of trypano- somiasis and advanced cases of sleeping sickness, as affording important evidence of the unity of these diseases, and further proof that the trypanosomes are the essential cause of sleeping sickness. “he ~-1904.] Short-Period Atmospheric Pressure Variation. 45 “The Rehaviour of the Short-Period Atmospheric Pressure Varia- tion over the Harth’s Surface.” By Sir Norman LOCKYER, KCBS LD. ERS; and’ WititaAM J: S. Lockyer, Chief Assistant, Solar Physics Observatory, M.A. (Camb.), Ph.D. (Gott.), FR.A.S. Received April 13,—Read April 28, 1904. [Prates 12 anp 13.] In a paper which we communicated to the Society in June, 1902,* we drew attention to the fact that in our investigation dealing with the percentage frequency of prominences and changes in atmospheric pressures, we found that the pressures in India and Cordoba behaved in an opposite manner, the short period variations of one being the inverse of those of the other; both, however, were closely associated with the prominence frequency. In a subsequent papert we showed that these two regions, in which these inverse pressure-variation conditions were clearly distinguishable, were, as far as had then been investigated, of considerable extent, the Indian region extending to Ceylon, Java, Mauritius and Australia, and that of Cordoba to the southern part of the United States. The facts there collected were stated to be so suggestive that the inquiry was being continued by collecting and discussing observations made in other areas on the earth’s surface, so as to note the extent of these similar pressure areas. The present communication contains the results that have so far beer: obtained. ; The greater portion of the facts here collected has been discussed some time, but as it was considered desirable, before communicating the present paper, to include as many regions on the earth’s surface as could be obtained, a longer delay than was anticipated has taken place ; even now there are many regions which we have been unable to include. The regions for which further observations are desired include the west coast of Africa, the northern part of South America, and the north-western portion of North America, and Polynesia in the South Pacific Ocean. In our previous papers we have pointed out the advisability of dividing the year into groups of months according as the pressure is above or below the mean value for the year. In this way the high or the low pressure months can be dealt with separately, if necessary, and any excess or deficiency from a mean value exhibited in either or both of these from year to year can be closely followed. Such a division of the year can be accurately determined for places * ©Roy. Soc. Proc.,’ vol. 70, p. 500. t ‘ Roy. Soe. Proc.,’ vol. 71, p. 134. 458 sir N. Lockyerand Dr. We Jos: Lockyer. [Apreits: which have a regular and pronounced annual pressure variation, such as India, and where the yearly barometric range is of far greater magnitude than any other aperiodic fluctuation. In those regions where the mean yearly curve is more misleading than otherwise, the division, according to the two seasons included in the two groups of months, April to September and October to March, is best adapted.* The system adopted in the present investigation was to take the pressure variations over India and Cordoba as the chief types of each region, denoting those of the former by the symbol (+), and those of the latter by (—). The pressure curve of any other place was then taken and compared with each. If, for example, it was found that the curve extending over several years exhibited an excess pressure at those epochs when the Indian pressure curve was in excess, then it was classified as being similar to the Indian type and represented by a (+ ). If it was seen that although it was more like the Indian curve than that of Cordoba, but yet not quite the exact counterpart of India, then it was denoted by (+%). Ina similar way pressure curves like Cordoba were classified as (—), and those more like Cordoba than India as (— %). In some regions the pressure variation curves were distinctly a mixture of both the Indian and Cordoba types, and it was difficult to classify them satisfactorily by the above method. The symbol adopted for these cases was (+1). Again, there were further some curves in which even this mixed type of symbol was not sufficient to exhibit the relationship of their variations to the other curves, so a special symbol (1) denoting ambiguity was used. In the present investigation of this similarity or dissimilarity of atmospheric pressure changes over large areas, it was found that the special types were apparent sometimes in the yearly curves, sometimes in those for one or other of the high or low pressure groups of months, or sometimes in both of these. It did not, however, appear to follow that, because the type was distinguishable in the yearly curves, it was necessarily apparent in both the curves of the high and low pressure months. : The accompanying table, although yet somewhat incomplete, gives a tabulated statement of the data employed in the present survey. The table explains itself, but it may be remarked that in Columns 6 * To show the misleading nature of the mean annual pressure-variation curve over, for example, the British Isles, it is only necessary to plot the actual monthly values of pressure for any one year on this mean curve and draw a curved line through them, when it will be seen that there is practically no relationship what- ever between the two curves. If, on the other hand, the actual monthly pressure values during any one year be plotted on the mean annual pressure-variation curve for India, the former follow very closely the swing and amplitude of the latter. ‘Uns 0} Jo sataydsturaty T40q Fo svore Aprep uvoTR 94} MOE pomUIIOJOp sv BUTUIM pur vuUTxem godsuns jo sqooda aif} oJOUWapP soUT, TeoII0A Wayotq pue snonuryu0D oyg—‘azoyr ‘*T “Sy ur dew oT} 0F UOTAL[AI UT parpnys oq plNoys oyv[g siyy, *(—) eqoprop pur (+) erpuyz : AjoweU ‘suoere, ornssorg jo sodéy, urepy20my ay} pue ‘edorng uto9se A UT seovyd OFZ SUOTYBIIBA SANSSOG POLog yLOY oY} Usemyoq Arysuoyvpoy oy SuIMeEYS—'ZT GLVTG 006! O68) O88 | OZ8I |? j (das -udv) wwdWW'S , VHOAHOO SSYOZV nm» (a d= 499) Lez. WEWOd 006! Ossi 088| 0281 ¢ 4 a ae oa “904 ‘005 “hoay ‘wahiyoo"T pun salyoory — Lockyer und Lockyer. Koy. Sve. Proc, vol. 73, Plate 12. 1870 1880 1830 1900 | | BOMBAY | ‘75 INDIA. l (APR-SEP) rh I 7 | * Se GODTHAAB ,, GREENLAND (octT- mar) STYKKISHOLM ICELAND ° (OCT-MAR) 745 VALENCIA IRELAND (octT- ee : | | | VIENNA | AUSTRIA | | (Ss—EP-FEB) | ) | | ) eo a ee | MADRID | SPAIN (sep-FeEB) 7 | DELGADA : AZORES . (oct-apr) 76! | | CORDOBA ” ! S.AMER* (APR- SEP) ; l | 1 1870 1880 1890 —«*t 900 Prats 12.—Shewing the Relationship between the Short Period Pressure Variations for places in Western Europe, and the two)Main Types of Pressure Variations, namely : India (+) and Cordoba (—). This Plate should be studied in relation to the map in fig. 1. Nofe.—The continuous and broken vertical lines denote the epochs of sunspot maxima and minima as determined from the mean daily areas of both hemispheres of the sun. "ZL 94e[q Ul sv oures SOUL] [VOT4LOA—'970AT “POSHOAOT TO9Y SVT PANO PUOIES ot[4 o8ed YORE UT *(SatozV) Bpesjaq pur (pupooyT) paolynaz0eg sv ‘tayqo TOVE TvOM SUOT}RIS OMY ye puR “(eT}O0G BAON) AoupAG puR (eITVAIsNY ‘g) oprepepy ‘(wolmouy *g) eqopatoM pur vIpuy se Tons ‘Ajpeotydeisoes pojeredas ATOpIM SUOT]RAS ye SUOMIPUOH OANSSeIgG OSI9AOY oFVASNI[I OF soAdND Fo sated oatyT—el VIG 0061 068i O88! OZ8I Z9Z Q3aLHy3ANi | : BAYND \ f 9 (ud -190) > SANOZV _, WavOTAG BL (aww — 190) ~~ al. a ee i ri oC ssl * ‘OT Mg CL, “Jor “Tong “o0g' “how ‘wahyoorT pun tahyoo'T Lockyer and Lockyer. Roy. Soc. Prog., vol. 73, Plute 13. 1870 1880 1890 1900 . u BOMBAY « ! INDIA 4 (APR- SEP) 3 29-70 S.AMER* ° (APR-SEP) 6 CURVE INVERTED. OW CORDOBA S.AUSTR* '° (YEAR) 3005 ADELAIDE . NOVA SCOT.” (YEAR) 95 CURVE INVERTED. 29:98 “89 SYDNEY | ICELAND °° (oct - MAR) 74.8 58 BERUFJORD DELGADA AZORES ° (OCT-APR) 64 CURVE INVERTED 767 ; i t 1870 1880 1890 1900 Prave 13.—Three pairs of Curyes to illustrate Reyerse Pressure Conditions ut stations widely separated geographically, such as India and Cordoba (S. Americw), Adelaide (S. Australia) and Sydney (Nova Scotia), and at two stations near each other, as Berufjord (Iceland) and Delgada (Azores). In each case the second curye has heen reversed. Nofe.—Vertical lines same as in Plate 12. 1904. ] Short-Period Atmospherie Pressure Variation. 459 and 7 are given the groups of months (whether high or low pressure months) in each case for which the pressure curves examined afforded the greatest resemblance to the types which form the basis of the classification. In cases in which the curves of the mean yearly values alone have been utilised, these columns have been leit blank and reference made in the column headed ‘‘ General Remarks.” Although the above classification gives a very fair idea on the whole of the types of pressure variations from one region to another, minor: peculiarities have been met with which have tended to add a certain amount of difficulty. These remarks apply principally to places in the more northern latitudes. Thus, for instance, Greenland and Iceland have been classified as of the (+ 7%) type, the British Isles, Germany, and Spain of the (+ 2) type, and the Azores of the (—%) type. From a glance at the accompanying plate (Plate 12), which shows the relation to each other of some of the pressures over these areas, the changes from a(— 1%) to a (+1) type can be observed. It will be seen that although practically the same groups of months have been taken in each case, pressure in excess of the mean value in Greenland or Iceland corresponds to a deficiency of pressure over the area covered by Great Britain, Austria and Spain, the curves being in the main the reverse of each other. Again, the pressure curve for the Azores follows more nearly the (—) type, as will be seen by comparing it with the Cordoba curve, but it has a certain similarity to those of Madrid, Vienna, etc., to which it must therefore be closely connected. While the western portion of Europe is of this (+1?) type, the eastern portion gradually assumes the (-?%) type, and this region extends not only probably to Norway and Sweden, but right across European and Asiatic Russia. The European Russian type of curve has an undoubted similarity to those of more Western Europe, but there are variations which indicate that the type is more like that of Cordoba than India. Again, another region in which rather mixed types of pressures are met with is that of Eastern and North-eastern Canada. Curiously enough Prince Edward Island and Sydney (Nova Scotia) correspond very closely to the (—) type, if allowance be made for the differences about the year 1877. The inverted curve for the latter with the Adelaide (Australia) pressure curve for comparison is shown in an accompanying plate (Plate 13). In addition to illustrating this reversal between Adelaide (+) and Sydney (Nova Scotia) (— 1%), this plate shows also, to serve as examples, curves for two other sets of reverse pressure conditions. Thus Le) sie 460 Sir N. Lockyer and Dr. W. J. S. Lockyer. [Apr 13, Analysis of a ar eons © i Country. Station. Type. Remarks on type EOS e observations. ee | MiraGieties Ay: IBOMIMDAY: cc. aiepee “ee + | Chief (+) type,see Plate 13 1860—1903 (Madras io. 025,200 eek ee 46 1861—1902 Caleutta: 9. ocieeisic 1 vas : 1860—1902 Nagpur............./ + . 1869— 19038 Dar(ecling \c.. - iee || : 1875—1901 Ceylon)... Colombo... <<... stair | : 1881—1900 | i Persia ....... Pushires- neces ee + | : 1879—1887 Aaa ee is LACE aaa een Sia miee 2 | = 1883—1899 Indian Ocean | Seychelles...........| + - : 188E—1898 Rodirivteg 2. ts t4 Jered ane : 1885—1898 Mies ae 10a 5 34) ates caper, al ba 1875—1901 Hast anGtes.c< | GRAUAWAG, “x << 01s « crete ere] ett: 1866—1898 Philippimesss.) Manila i20..2.o5 eee sie 1883—1901 | Malay Penin. | Singapore.......--<.|, & 1869 —1902 UAvgistre ayer oct Merl: Pats tobe tees ater Mee Slight differences .......| 1876—1900 A delaide’.' slo itsrec se crt ae Remarkably pronounced, | 1876—1899 see Plate 18 Seb R Roos soars oo fa ae ac 1860—1899 New Zealand | Dunedin. t Difficult to classify ......| 1866—1898 | Auckland . ? ‘ oak eee 1867—1903 | Africa.......| Zanzibar (Island) .. SPY 1880-3—1891-7' Durbaw |) js s-saeeepeen, ot 1884— 1903 Capetown... cae naceeye to : ae 1860—1899 Cairo. eee ain 2 e ee 1869—1898 Alexandria .........| +? : 1870—1896 Biskra wees. eee ae se 2 : 1874—1884 Sierra Leone. . =F ee 1876—1900 St. Thomas (Island). Me Ce 2) bd 1874—1883 Kimberley . : 1876—1899 St. Paul de Toman +? 5 1880—1895 South America| Cordoba .. — | Chief (—) type,see Plate 13} 1873--1902 MueUuMen . 6. aie = Distinctly like Cordoba 1885—1897 COWES iole, fm +e, adie as — | Exactly like Cordoba....| 1877—1897 Santiago eee aH ee ee 1874—1884 Verde PAaneiLO: \s.siemes see - 1860—1899 AW estulindacs sid MUGRIMMAICD aoc «core wine eie's —? 1866—1900 Barbadoes @eeoseceeeeere ? e e } 1865—19C0O 1904.] Short-Period Atmospheric Pressure Variation. Pressure Types. 461 | Months in which types are | most conspicuous. Apr.—Sept. 33 33 Apr.—Sept. Oct. —Mar. Apr.—Sept. May— Oct. June—Oct. Apr.— Sept N ov.— Apr. Oct.—Mar. Apr.—Sept. Oct.— Feb. Nov.—May May—Oct. May—Sept. VOL, LXXIII. ) Low press. 3? 29 Low press. High press. Low press. High press. 2) High press. 33 3) Low press. 33 High press. Low press. High press. Low press. High press. Source of data. Indian Monthly Weather Reviews | Mauritius Met. Obsns........ | Meteorological Office ... | | Report of Philippine Com. | | Met. | Met. Obsns. at Adelaide..... | Met. ”? mission Met. Obsns. at Straits Settle- ments Obsns. district Obsns. Wales Reviews Report of Govt. Astronomer, Natal Meteorological Office | Met. Report of Abassia Obsy., | Met. Zeitschrift, 1897 | Kong. Sv. Vet. | Supplement to Cairo eeesveo se lingar, vol. 29, No. 3 | Army Medical Dept. me | Lisbon Met. Obsns. . of and Annals Lisbon Observatory Loanda Met. Obsns. | Anales de l Oficina Met. Argen- tina, vol. xiii Anales del’ Oficina Met. Argen- tina Kong Sv. Vet. Ak. Handlingar, vol. 29, No. 3 Meteorological Office . Army Medical Dept. Reports at Perth and | in New South | | New Zealand Statistics (Met. q | Met. of New Zealand .... | Indian Monthly Weather Ak. Hand. | General remarks. Type equally prominent in curve for year. 3) ” bP) 9) Yearly curve only examined. oP) 9 Type equally prominent in eurve for year. Record very short. Hqually prominent in yearly and Oct.—Mar. curves. | Yearly curve alone examined. | Short and uncertain record. Yearly curve examined. . | Prominent in yearly curve. Pe] 39 | Very prominent in yearly curve. 39 9 1877 pulse absent. | Record very broken and short. Broken record, 1871—1877. Record too short. Yearly curve alone examined. Record too short. Yearly curve alone examined. 99 2) Broken record. Continuous good record. Yearly curvealone examined. Two short breaks in record, 2) 462 Sir N. Lockyer and Dr, W. J. 8. Lockyer. [ Apr. 13 Analysis of Pressure | : | Period of | Country. | Station. Type. | Remarks: on type. absiee ee North America| Jacksonville.........) — Some slight differences .. | 1873—1899 | Mobile 4. is-.60<0e a B. 1873—1899 ‘Pensacold.;.. dessus ade “s ie -.|: 1880—1899 Nashville .........--| — ?| Marked differences......| 1873—1899 Nan Diego .,........| — | Slight differences .. -| 1873—1899 St. Louis............| —?| Marked differences......| 1878—1899 Bismarck ee) ey ‘ Hy 1875—1899 Kansas) City). ..<)-- soe! 1890—1899 LEONG Pa Dome oa oc | | =k 2 — 1878—1890 Salt Lake Cit by cece ss} oe? | Slisht difieremees) care 1874—1899 Santa Fé. ; ite Close resemblance ...... 1873—1899 | Denes bees oe —? me 4: 1868—1899 | Portland . | x P| Equally like both types..; 1873—1899 Galveston! ye serene eS Some slight differences...| 1873—1899 | Alpena ...........-.| 4/?| Difficult to determine....| 1873—1899 Buffalo . |) oP? _ » ....| 1873—1899 -| Pikes Peak . at 6 D 1874—-1887 Ft. St. Michaels) +? : 1874—1884 (Alaska) Atlantic Ocean Bermuda......+-...-| — | Undoubtedly this type...| 1866—1899 @anadarts< sonuqlt LOvOMEO rs Ys sis 4 2 peste eee | sa ae 1874—1890 Fort atl (Maitobay i? 1874—1891 | Nova Scotia ..| Sydney . =e ee Plate 138. dis 1874—1898 Sandwich Isles Honolulu . = 7 Indian pulse present 1873—1901 Tahiti .......| Papeete .. 25 5 Like India, one year in| 1876—1891 advance Greenland . Jacobshayn ..... Ls fae wis 5- 1873—1897 Godthaab.. | + ?| Very like Jacobshavn, see| 1873—1895 | Plate 12 Ieeland,.....| Stykkisholm ........; + ?]| Very slight differences,| 1874—1901 | | 1877present, see Plate 12 Becurjord 32... - + P| Very slight differences, | 1874—1901 877 present, see Plate 13) LOPES 5.0.4 ai COMBO chs. eprnterMamees (fo it <- 1874—1898 | Vardores oct) sacs Oe en eee 1s70—1898 Aalesand,..-.. sixisakelee P || ae =i 1861—1898 | Stornoway...... + ?| 1877 pulse absent.......| 1863-—1903_ | Aberdeen .. + ?/| Very slight differences,| 1869—1899 1877 absent ; Armach. 42°. cas «x 3s SEN - = 1869—1903 Stonyhurst .-....-.-ally =e? 1885— 1902 | Valeneta :.)..0.) 2.08 sel eee 2 |) Seerelatest? oe... 1868—1899 Palermo... 229 ws 1874—1884 Madrid .... | +? | Remarkable likeness to| 1860—1899 | (—) at times,see Plate 12) . Gibraltar ..0. Ea kesOrks —Apr. a | 1864—1900 US Oy air «| see 1860—1896 Vienna..... P| ee See Plate ]2 .>...ua+-:| 1860-1emm Constantinople....... P 7 ‘ | 1874—1884 1904] Short- Period Atmospheric Pressure Variation. Types—continued. Months in which types are most conspicuous. Source of data. 463 General remarks. (Mar.— Aug.) (Low press.) Nov.—Feb. Oct.—Mar. Nov.—Mar. Sept.—Feb. Nov.—Apr. Apr.—Sept. Oct.— Mar. May— Oct. Dec.—-May June— Novy. Oct.—Mar. Dec.— Mar. Aug.—N ov. Aug.—Feb. Oc vide. Apr.—Sept. 3) Sept.—Mar. Apr.—d uly Apr.— Sept. 3) Apr.—Sept. Mar.—Aug. Apr.—Sept. High press. Low press. High press. Low press. High press. High press. High press. Hich press. Low press. Low press. High press. 9 Low press. 39 High press. High press. 99 High press. Low press. Low press. Report of Chief of Weather Bureau, U.S.A. Kong. Sv. Vet. Ak. Hand- lingar, vol. 29, No. 3 Army Medical Dept. Reports Report of Met. Service of Dom. of Canada 3) oP) Met. Olan. Horotata SI AS Met., Zeitschrift, 1892 ... Danish Met. Aarbog......... Met. Aarbog Norsk 3) 39 MSS. Met. Office Records sare 99 be) 99 >) Results of met. and mag. observations, Stonyhurst MSS. Met. Office Records ....| Kong. Sv. Vet. Ak. Hand- | lincar, vol. 29, No. 3 | Madrid Met. Sinsne Army Medical Dept. Reports. Lisbon Met. Obsns..... Denkschritten iKaiserlichen | Ak. de Wissenschaften Kong. Sy. Vet. lingar, vol. 29, No. 3. Ak. Hand. | Short record. by) Record broken, 1882—1885. Yearly mean curve alone examined. 9) 3) 99 3) Short record. Yeurly curve alone examined. Broken record. Record somewhat broken. Yearly curve alone examined, +) 39 Record broken after 1891. Record twice broken. Yearly mean curve examined. Record broken, 1883—1887. 4 1874—1875. Yearly curve alone examined. Short record of definite type. 2 Keg 464 Sir N. Lockyer and Dr. W. J. 8. Lockyer. [Apr. 13, Analysis of Pressure : | Period of Country. Station. Type. | Remarks on type. | observations. | | | North Atlantic) Azores..............| — ?} Remarkable, undoubted| 1874—1903 © | (—), see Plates 12 and 13 | Las Palmas..........| —? i or | 1882—1892 . Russia.......| St. Petersburg.......| —P oe oe 1870—1899 : Moscow. .......-. oe P 30 He 1870—1899 | Mar claat uti. Ja o- sreuetat ey Difficult to classify......| 1874—1884 | Wugansk 2. «6.6 secsnelee © 7 ae 1872—1899 . Qrenboure.icse cele dis re 1870—1899 Catherinbourg...... or ? 1870—1899 ; Arkangel .....cese0.) =? 1871-—1899 |} NIDEHIG Mais. ois || Comp . 6) << diem) serrate ene 1874—1899 Barnaoul ianastcrace —? 1870—1899 | INertehimsleis 2 oss scrum oe bie te 1870—1899 Chimay. secs Pekin c.. tease P i wh 1874—1881 || Zi=Kka-Wel .. . <0) + + ire ae | i - 1874—1884 Hong-Kong (Island)..; — | Exactly like Cordoba....) 1884—1903 JAVA elerels ole =| MLOKTOL . oo). i ieee =k P a Su 1873—1895 | Kioto eecvseeee eoveve +?) ee ee 1883—1899 Bombay (+) is compared with the Cordoba (—) pressure curve (inverted), and is an example of the adopted types of pressure varia- tion. Iceland is compared with that of the Azores (inverted), and shows the reverse conditions that prevail between a (+1) type and (— 1) type. | A fact to which attention was very often drawn in attempting to classify the pressure curves was that some curves after following very closely for many years the Cordoba ( — ) or,Indian (+) type of pressure, as the case may be, would revert back to the opposite type for a period of years. Thus to take the case of one station alone, namely, Sydney (Nova Scotia) as an instance, the pressure curve follows very closely that of India from 1875—1882, after which up to 1890 it has a very close resemblance to the Cordoba type. The behaviour of this Sydney (Nova Scotia) pressure curve can be compared with the Adelaide (Australia) curve in Plate 13, but it must be noticed that the former has here been inverted. There is another important fact which this study has brought to 1904. Short-Period Atmospheric Pressure Variation. Types—continued. 465 Months in which types are most conspicuous. Source of data. General remarks. Observations made at Hong- Kong Observatory Met. Zeitschrift, 1899........ Report of Cent. Met. Obsy. of Japan. | Oct.—Apr. Low press. | Annales de l’Obsy. do Infante | Prominent in yearly curve. D. Luiz Resumé de las Obs. Met. de | 1886 no record. Provinces (Spanish) ee Annales de ’Obsy. Cent. Phys. | Yearly curve alone examined. de Russe Sept.—Feb. | High press. 5 ” a ce Kong. Sv. Vet. Ak. Handlingar, vol. 29, No. 3 Oct.—Mar. | High press. Meteorological OUTED cocc nen 36 ae Annales de l’Obsy. Cent. Phys. Break in record 1875—1887. de Russe Yearly curve alone ex- amined. ve a6 5 45 Yearly curve ant examined. Jan.—June | High press. | Meteorological Office ........ Oct.—Mar. FA % x Prominent in yearly curve. ” +B) ” 9 29 ae Apr.—Sept. | Low press. 9 . a3 Annales de l’Obs. Cent. Phys. de Russe Oct.—Mar. | High press, | Kong. Sv.Vet. Ak. Handlingar,| Short record. Prominent in vol. 29, No. 3 yearly curve. Apr.—Sept. | Low press. | Met. Zeitschrift, 1886, and| Visible in yearly curve. All the curves for the half- yearly and yearly values show similar variations. light and which plays most probably an important réle with regard to the pressure variations at places which exhibit a mixed type of pressure. ‘The earth’s surface as has been shown may be divided mainly into two regions, one portion showing excess pressures at certain epochs, while the other shows deficient pressure at the same epochs. If the former region exhibits a greater excess than usual (as an example, the Indian region in 1877), then the region over which this type of pressure occurs may probably be more extensive, and the boundary dividing the two chief types of pressure will necessarily be pushed away from this region. Stations, therefore, that were just on the fringe of this boundary may at these epochs become enveloped in this more extensive high-pressure area, and will exhibit the Indian type of pressure variation. Should the Cordoba region become more extensive than usual owing to a similar cause, then the border stations will assume the Cordoba type of pressure variation. It is not proposed to enter here into detail on this point, as the subject requires very close examination, but 466 Sir N. Lockyer and Dr. W. J. 8. Lockyer. [Apr. 13, mention may be made of the very great area which was covered by the continuous excessive high pressure that. prevailed over the Indian region from the end of 1876 to about the middle of 1878. On fig. 1 is given a map of the world on which are marked the types of pressure variations in each region which is included in this barometric survey. An attempt has been made by means of a neutral line to show approximately the mean lines of separation of these two chief pressure types, although it must be remembered that this line is liable to a probable small oscillation about its mean position. As far as can at present be determined, one line commencing to the west of Alaska, separating this region from Siberia, passes easterly along about the 60° parallel of latitude and runs in a south-easterly direction between South-west Greenland and North-east Canada. It then crosses the North Atlantic, passing to the north of the Azores, and skirts the south-western portion of Portugal. It then strikes down towards the Equator, cutting North-west Africa, as far as can be judged from the scant pressure values available, through the middle of the Sahara. It leaves Africa near the Gold Coast,. passes into the South Atlantic, where it cannot be traced further owing to lack of observations in this southern ocean. The other boundary or neutral line passes to the north-east of Greenland and north of Iceland, crosses the southern portion of Norway and Sweden, and traverses Southern European Russia. It then takes a course somewhat more easterly, skirting the northern part: of the Caspian Sea and Turkestan, passes between Tibet and Mongolia, and through China. It then leaves the Continent a little to the south of the Yellow Sea, and passes into the North Pacific Ocean. Here its path cannot be traced, but it evidently passes well to the east of the Philippine Islands, and Solomon Islands, takes a new south-westerly course, skirting the eastern side of Australia and passing between Tasmania and New Zealand. Its track is then again lost in the Southern Pacific Ocean. Although too much weight must not at present be given to the positions of these neutral lines throughout their whole length, it is interesting to note that they are fairly symmetrical to one another although no attempt has been made to make them so. Both lines apparently cross the equator at about antipodal points, and both appear to have a similar trend in northern and southern latitudes. We seem then to be in presence of a general law relating to the pressures which occur simultaneously in two different regions of the globe, separated and defined more or less by a neutral line, this neutral line forming a fulerum about which see-saws of pressure from one region to another take place. Special cases of such reverse pressure: variations have been previously detected. 467 Short-Period Atmospheme Pressure Varvation. 1904.] “AFISSBTO 09 pue BIpUy jo omngxta (¢ F) “eIpuy uvyy eqoptog oy, etom (¢—) ‘eqoptog oy, (—) “eqopsop ueyqy eIpuy oy ext, (+)—! Ajrusis spoquds oyy, ‘suoyviaeA ommssorg omeydsomyy jo sedAy, wey, OW} OY} Jo MoOIyNGIajysIq] oYy 8 _o8 q{norgip “(¢) “eqopaop I] odour (¢ +) “erpuy UlyBa}SN[ [J] — [ “OT 468 sir N. Lockyer and Dr. W. J. 8S. Lockyer. {| Aprids; Thus as long ago as 1879 Blanford,* from a discussion of the secular variations of barometric pressure over the wide area covering Siberia, Indo-Malaysia and Australia, pointed out that there existed a kind of long-period see-saw of a character that, while the pressure at the tropical stations was low, that in Siberia was high, and vice versé. This fact, it will be seen, is quite in harmony with the pressure-type distribution, as shown in the accompanying map (fig. 1). Hildebrandssont has discussed the relation between the pressure variations of numerous places mainly situated in the chief centres of action of the atmosphere widely distributed on the earth’s surface for the period 1874—1884. In this valuable communication, some of the chief results which he was led to deduce were that there were several regions which exhibited opposite types of pressure variations. The following places are those to which he calls attention, and for comparison we give the types in brackets which have been allotted according to the method adopted in the present paper; where no type is added the region has not been examined :— The Azores (—%) and Iceland (+ %); Siberia (— ?) and Alaska (+ 4), especially in winter; Tahiti (+?) and Tierra del Fuego; India (+) and Siberia (—?); Greenland (+?) and Key West (Florida) (—); Buenos-Ayres (— ?) and Sydney (Australia) (+). It is interesting to note that these results agree well in the main with the present distribution of the regions which have been examined. Again Hann{ has recently drawn attention to the fact that there exists a see-saw between the Azores and Iceland, and he showed that in 80 per cent. of cases the largest positive pressure variations at Stykkisholm (Iceland), corresponded to negative pressure variations at Ponta Delgada (Azores), and that the largest negative pressure variations at Stykkisholm were in 87 per cent. of cases positive varia- tions at Ponta Delgada. This result obtained from the observations extending from 1846— 1900 endorses Hildebrandsson’s previous conclusion deduced from observations over the period 18741884, and confirms the position of the neutral line shown on fig. 1, dividing the two large types of pressure areas. Quite recently Professor Bigelow§ has published a map of the world on which he has indicated the distribution of the pressure types according as they follow the Indian (or direct type, as he calls it) or the Cordoba (indirect) pressure variations. Professor Bigelow has also found that there are many regions in * ‘Report of the Meteorology of India in 1878,’ pp. 2—385. t+ “Quelques Recherches sur les Centres d’Action de ]’Atmospkére,”’ ‘ Kongl. Svenska Velenskaps-Akad. Handlingar,’ vol. 29, No. 3. { ‘ Kaiserliche Akademie der Wiss. in Wien,’ January 7, 1904. § ‘Monthly Weather Review,’ p. 509, November, 1903. 1904. ] Short-Perwod Atmospheric Pressure Variation. 469 which it is very difficult to say exactly which type is followed, and as he says there may be “ differences of opinion as to the assignment of some of these curves, but the reader can make any different arrangement that he prefers.” In most of the main features, however, his map suggests a somewhat similar distribution of these pressure types to that given here. Thus, he finds that “the region around the Indian Qcean gives direct synchronism, South America and North America give inverse synchronism, while Europe and Siberia give an indifferent type. Greenland and Iceland seem to have direct type like the Indian Ovean,...'. “The eastern hemisphere tends to direct synchronism, except in Europe and Russia where the indifferent type prevails, and the western hemisphere to the inverse type.” It may be further pointed out that regions which are the reverse of one another as regards these secular pressure variations should very probably experience opposite kinds of abnormal weather, while those over which the same type of pressure variation exists should have weather of an abnormal but similar nature. That this is inclined to be so as regards the latter statement has been recently* very forcibly pointed out by Sir John Ehot with respect to the. Indian area. He writes :— “The drought of 1895—1902 was a more or less general meteoro- logical feature of the whole area, including Abyssinia, Hast and South Africa, Afghanistan, India, probably Tibet, and the greater part or whole of Australia.” The whole of this region, as will be seen from the accompanying map (fig. 1), is embraced by the (+) type of pressure. In the light, therefore, of the existence of these large regions of opposite pressure types, it is vital in the interest of long-period fore- casting that observations from all portions of the globe should be included in any discussion. Several years ago Eliott drew attention to these oscillations of pressure of long period, other than the diurnal and annual oscillations in India. In this important memoir he pointed out that “they are directly related to the largest and most important features of the weather in India, viz., the character and distribution of the precipita- tion of rain and snow in the Indian monsoon area.” There is reason, therefore, to believe that this short period pressure variation will in the future be of considerable assistance in helping * ‘Broad Views,’ p. 193; ‘The Meteorology of the Empire during the Unique Period 1892—1902,’ by Sir John Eliot, K.C.1I.E., F.R.S. + “A Preliminary Discussion of certain Oscillatory Changes of Pressure of Long Period and of Short Period in India,” ‘ Indian Met. Memoirs,’ vol. 6, part 2, 1895. 470 Sir W. Ramsay and Prof, J. N. Collie. [May 18, meteorologists to form a more definite idea of the prospects of approaching seasons. We wish to express our thanks to Dr. W. N. Shaw, F.RB.S., who. | has kindly assisted the work by permitting us to utilise the valuable collection of pressure data deposited in the archives of the Meteoro- logical Office. We also owe a debt of gratitude to Messrs. W. Moss and T. F. Connolly, who have shown great zeal in completing the necessary computations and drawing the numerous curves which were required for the different stations that have been investigated. “The Spectrum of the Radium Emanation.” By Sir WILLIAM Ramsay, K.C.B., F.R.S., and Professor J. NORMAN COLLIE, F.R.S. Received May 18,—Read May 19, 1904. Attempts have been made since July, 1903, to see and map the spectrum of the emanation from radium, for at that date the con- version of the emanation into helium was observed by Ramsay and Soddy, and during the first discharge of the induction current through the emanation, it was believed that a peculiar spectrum was noticed ; indeed, three lines were persistent, and were mentioned in the communication on the subject in these ‘ Proceedings.’ But such attempts have uniformly failed; at the first moment of the discharge, indeed, a brilliant spectrum has twice been observed, which soon became confused and indistinct. It faded before it was possible to map it, and owing to the presence of impurities, generally carbon monoxide, nitrogen, or hydrogen, the special spectrum was obscured. All that could be said was that it appeared to present some brilliantly green lines. These experiments, however, have not been fruitless ; they have led to better knowledge of the precautions which it is necessary to take to eliminate impurities. The arrangement of the apparatus, too, has been simplified, and the manipulation made easier. As it is possible that others may wish to repeat the experiments, and may perhaps have even better success in mapping the spectrum, we think it well to enter into the details of the manipulation somewhat minutely, and to give a woodcut of the apparatus employed. The stock of radium bromide (about 109 milligrammes) dissolved in about 10 ¢.c. of water in two small bulbs was attached by sealing to a small Tépler’s pump. Between the pump and the bulb there was a stop-cock, greased, of course, to insure freedom from leakage ; but in order to prevent the long contact of the emanation with the stop-cock, and its possible contamination with carbon dioxide, the mercury from 1904. | The Spectrum of the Radiwm Emanation. 471 the pump was caused to flow past the stop-cock by raising the reservoir of the pump and closing the exit tube at its lower end; the mercury slowly leaked past the valve of the pump, passed the tap (which was then shut), and so confined the space above the radium bromide by means of mercury. As radium bromide yields electrolytic gas, containing an excess of hydrogen, the pressure gradually rose; the mercury in contact with this gas remained perfectly bright, and showed no tendency to adhere to the glass; the presence of ozone thus appears to be excluded, but this excess of hydrogen will form the subject of a future communication. The emanation was allowed to accumulate for 14 days. The pump was exhausted until no trace of a bubble passed down the capillary exit tube. But as even then a trace of air must have remained Bre in the barrel, the tap leading to g the bulbs containing the radium | bromide was turned rapidly, so as to admit a trace of the electrolytic To pump. } c gas into the pump and ‘wash it na out.” This gas was rejected. The I remaining electrolytic gas with the JA 4 emanation was collected in a tube < which had previously been heated to redness, and then twice washed J out with pure oxygen. The mer- R.0s ‘ B cury in the collecting tube was e 5 then boiled, and the bubble of gas removed. It was hoped thereby to have eliminated every trace of nitrogen. The gas was then intro- duced into the gas-burette, shown in the figure, through the inverted syphon. All the mercury was freshly filtered and pure. The apparatus, too, was freshly con- structed and heated to redness to burn out traces of dust. The gas- burette had been washed out with alkali and with nitric acid, and then with a stream of distilled water; it was dried by drawing through it a stream of dust-free air. Some slightly moist caustic potash was melted on to the glass, near the sparking wires ; this was intended to absorb any trace of carbon dioxide which might have chanced to be formed during the explosion 0 the electrolytic gas by the burning of dust. The rubber tube was cemented AT2 Sir W. Ramsay and Prof. J. N. Collie. [May 18, on to the burette; the burette was washed out twice with oxygen, and by lowering the reservoir several times the upper end was made a torricellian vacuum ; it was left thus for some time, so as to insure the removal of adhering nitrogen from the walls of the tube. The electrolytic gas was then introduced, and exploded. As the explosion-burette was graduated, the total volume of the gas, as well as that of the residual hydrogen, was read. There were 16°43 c.c. of gas ; the residual hydrogen measured 1:01 c¢.c. at normal temperature and pressure, and thus amounted to 6°18 per cent. of the total. The volume of this gas was increased by lowering the pressure so that it was in contact with the fused potash. It was left for more than an hour ; the potash, of course, was wet with the water formed by the explosion. The capillary tubes above the stop-cock of the gas-burette, which had been twice washed out with oxygen, were pumped as empty as possible, until the vacuum-tube showed only the yellow and green lines of the mercury spectrum, and the faintest trace of a hydrogen spectrum. A strong current was passed between the electrodes so as to heat them and expel occluded oxygen. After this process had been repeated as long as was thought safe, until, as remarked, the hydrogen spectrum was extremely faint, the tap to the pump was closed. The hydrogen containing the emanation was then admitted from the explosion- burette ; it was dried by passage through the narrow tube B filled with phosphoric anhydride, and it entered the bulb C, and the vacuum- tube D. This vacuum-tube was made of lead glass, with electrodes of aluminium. It was 2°5 cm. long, with a capillary of about 1 cm. in length. The aluminium electrodes were closely surrounded with glass, fused round them, so as to limit the capacity of the tube as much as possible ; it was probably under one-twentieth of a cubic centimetre. Liquid air was next poured into the jacket surrounding the bulb ©, and the reservoir was raised and lowered half a dozen times, so as to convey all the gas into contact with the cooled bulb. The mercury was then raised to the level a, and the tap to the pump opened; and while the jacket was kept replenished with liquid air the hydrogen was pumped off, until its spectrum had almost entirely disappeared, the red line being hardly visible. The tap to the pump was then closed, the level of the mercury was raised to 5, and the liquid air allowed to evaporate. The bulb was so bright that it was easy to read the time on a watch. The mercury was then raised to the level ¢, and the current passed. The spectrum was very brilliant, consisting of very bright lines, the spaces between them being perfectly dark; it hada striking resemblance in general character to the spectra of the gases of the argon group. A direct-vision spectroscope, made to special design by Heele, with an illuminated scale for reading, had immediately before been standardised by noting the position of the leading lines of helium and 1904. | The Spectrum of the Radium Emanation. 473: hydrogen. They were found to lie exactly on a scale which had previously been constructed. The new lines were read as rapidly as possible, an operation which required about half a minute. During a second reading many of the lines had faded, and the secondary spectrum of hydrogen began to appear, and rapidly grew stronger. It was identified by throwing into a field a hydrogen spectrum through the small prism ; and it soon became so powerful as to mask the spectrum of the emanation completely. In order to attempt to recover it, the mercury was again drawn down to a, and liquid air again poured into the jacket; the emanation again condensed, and the tap to the pump was opened, and the hydrogen removed by the pump until its spectrum was again hardly visible. On repeating the series of operations already described, the spectrum of the emanation was seen a second time, but it was so transient that only the position of some of the lines could be confirmed. Next day, only the spectrum of hydrogen was visible; its secondary spectrum was strong. The day after, the same was the case: but interposing a jar and spark-gap brought out two lines which had previously been mapped; they were very feeble. In the table which follows, all the strong lines which were read are given ; the degree of coincidence of those which are of known wave- length shows the approach to accuracy obtained ; the error is probably less than five, Angstrém units. Wave-length. Remarks. 6567 Hydrogen C; true wave-length, 6563; observed each time. 6307 Observed only at first ; evanescent. 5975 ” oD ” 5955 ” ” 5805 Observed each time ; persistent. 5790 Mercury ; true wave-length, 5790. 5768 ‘s > AE 5769. 5725 Observed only at first ; evanescent. 5595 Observed each time; persistent and strong. 5465 Mercury ; true wave-length, 5461. 5105 Not observed at first ; appeared after some seconds ; persisted, and was visible during the second examination. 4985 Observed each time ; persistent and strong. 4865 Hydrogen F'; true wave-length, 4861. 4690 Observed only at first. 4650 Not observed when the emanation was examined again. 4630 Ditto. 4360 Mercury ; true wave-length, 4359. ATE © Sir W. Ramsay and. Prof, J. N. Collie. - [May is; When the spectrum was examined two days later, besides the hydrogen and mercury lines, there were seen :—5595, feeble; 5105, feeble ; 4985, very feeble; this was with a jar and spark-gap inter- posed ; the ordinary discharge showed only the primary and secondary spectra of hydrogen, and that of mercury. Eleven days later, the emanation from the same stock of radium bromide was collected, and treated in exactly the same manner. ‘This time, however, an excess of conscientiousness made us continue to extract gas with the pump from the bulb C containing the frozen emanation, surrounded by liquid air, for too long a time. Every two or three strokes of the pump collected a minute bubble, occupying about the tenth of a millimetre in length of the very narrow fall-tube of the pump, which was really a fine-bore capillary. The yield of gas appeared to be continuous; and when these bubbles were examined in the dark they were brilliantly luminous. This gas was really the emanation, which possesses a feeble vapour pressure even at the temperature of liquid air. Needless to say, on attempting to examine the spectrum, little was seen, for the pressure of gas in the vacuum- tube was too low. | The tube was therefore washed out with the gas which had been pumped off, and the process was repeated. The minute bubbles which passed down the capillary fall-tube of the pump were examined, and pumping was stopped when they showed a very faint luminosity in the dark. On compressing the emanation into the spectrum-tube, the spectrum was again brilliant, and measurements were made. It was found possible to read the lines several times, for although the spectrum faded in less than a minute, it appeared to recover on ceasing to pass the current. But this recovery soon failed and, as before, nothing could be detected after 5 minutes but the primary and secondary spectra of hydrogen. Now the tube was practically vacuous before warming the bulb containing the emanation; no current would pass; but it is, of course, possible that the gas carrying the emanation had net been perfectly dried in passing through the tube B, containing phosphoric anhydride; any water-vapour would have condensed in the cooled bulb C, and would only slowly have vaporised into the vacuum-tube. On arriving there, it would give the hydrogen spectrum. Another possibility is that it may have come out of the electrodes; for it has been frequently noticed in glowing out a vacuum-tube with aluminium electrodes that even after all trace of hydrogen has been removed by passing the discharge so as to heat the electrodes, and by pumping, the hydrogen spectrum has reappeared on admitting a trace of one of the gases of the argon group, and passing the discharge for a longer time ; but the intensity of the spectrum which replaced that of the emanation may perhaps warrant the supposition that hydrogen as well as helium is one of the products 1904. ] The Spectrum of the hadium Emanation. . 475 of the disintegration of the emanation. This, however, is very doubtful, and judgment must be suspended until more satisfactory evidence is forthcoming. The lines read were :— Wave-length. Remarks. 6350 Not observed before ; faint. 5975 Observed before ; faint. 5959 %) ” 5890 Not observed before ; faint. 5854 99 ” 9 5725 Observed before ; fairly strong. 5686 Not observed before ; faint. D595 Observed before ; strong and persistent. 5580 Not observed before ; faint. 5430 2) %) 5393 %» %3 %) 5105 Bright ; persistent ; observed before. 4985 . 4966 Not observed before ; bright, but transitory. 4640 Transitory ; possibly 4650 and 4630, which were seen before as distinct lines. The line 4966 was particularly brilliant at first; but it soon assumed secondary importance. Some lines which had previously been observed were not seen; they are 6307, 5805, 5137, and 4690. An attempt was made to obtain the spectrum with a jar and spark-gap; but only hydrogen and mercury were to be seen. The resistance soon became very high, and there was danger of piercing the vacuum-tube. Previous attempts in conjunction with Mr. Soddy gave lines with wave-length 5725 (jar), 5595 (no jar), 5105 (no jar), 4985 (no jar) ; the line 5585 was observed three times, and 5105 twice previously. The lines 6145 and 5675 mentioned in our last paper (April, 1904) were not seen, unless the latter is identical with 5580. It may perhaps be mentioned that the line 5595 was seen by Pickering in the spectrum of lightning, and was not identified with a line in the spectrum of any known gas; it is said to have been a very strong line, of intensity 30.* There can be no doubt that the lines given are the chief lines in the visible spectrum of the emanation; as for the pressure, the volume of emanation was about 1/30,000th of a cubic centimetre, and the capacity of the vacuum-tube, say, 1/20th; this would make the pressure about 1/10th of a millimetre. It may have been twice as much, for the numbers given are merely estimates. It may be remembered that, at the Chemical Congress held in Paris * © Astrophysical Journal,’ 1901, vol. 14, p. 368. A476 The Spectrum of the Radiwm Emanataon., [May 18, in 1900, it was suggested that no element should receive a name until its spectrum had been mapped. Of course, the converse does not follow, that, after the spectrum of an element has been mapped, it should receive a name. The “emanation from radium,” however, is a cumbrous expression, and sufficient evidence has now been accumu- lated that it is an element, accepting that word in the usual sense. It is true that it is only a transient element, and ought in justice to be called a compound; but of what? It stands on a wholly different plane to any known compound in the amount of heat with which it parts during its spontaneous change, and in the peculiar electrical phenomena which accompany its transformation. It is a gas; it follows Boyle’s law; as Rutherford and Soddy have shown, it resembles the gases of the argon series in its indifference to chemical reagents, for it not merely withstands the prolonged action of mag- nesium-lime at a red heat, but also, as Ramsay and Soddy . have proved, prolonged sparking with oxygen in presence of caustic potash. Its molecular weight has been found to be nearly 200, and, if it is monatomic, that number would also express its approximate atomic weight. Now, it appears advisable to devise a name which should recall its source, and, at the same time, by its termination, express the radical difference which undoubtedly exists between it and other elements. As it is derived from radium, why not name it simply “exradio”? Should it be found that the emanation, which is sup- posed to be evolved from thorium, is really due to that element, and not to some other element mixed with thorium in exceedingly small amount, a similar name could be given, namely, ‘“exthorio.” If the existence of actinium as a definite element is established, its emana- tion would appropriately be named “exactinio.” It is unlikely that others will be discovered, but, if they are, the same principle of nomen- clature might be applied. It should be stated, in conclusion, that Mr. Soddy collaborated in the experiments preliminary to this successful mapping of the spectrum ; had he not been obliged to leave England, he would, no doubt, have shared whatever credit may attach to this work. 1904. ] Notes on the Statolith Theory of Geotropism. AT7 “Notes on the Statolith Theory of Geotropism. J. Experiments on the Effects of Centrifugal Force. II. The Behaviour of Tertiary Roots.” By Francis DARWIN, F.R.S., and D. F. M. Pertz. Received May 30,—Read June 9, 1904. if. According to the statolith theory,* there are in plants certain cells specialised to act as organs of orientation in space; organs, in fact, functioning like the otocysts of certain animals. In both cases the sense of verticality is believed to be the result of the pressure of certain heavy bodies (usually starch grains in the case of plants), on a sensitive surface, namely, the lining membrane of the otocyst in the case of animals, or in the case of plants the protoplasm lining the statolith-containing cells (statocytes). In a papery dealing with the arguments for and against the theory, Jost brings forward as the most serious objection the behaviour of plants to centrifugal force. He found that plants, subjected to centrifugal force equal to from 0°02—0-05 g.,t exhibited curvature, but that the starch-grains were uniformly distributed throughout the statocytes, not, as should be the case according to our theory, resting on the cell walls furthest removed from the axis of rotation. Jost sees in these results an absolute proof that, in the cases investigated by him, the starch grains do not function as statoliths. It seemed to us that this conclusion was a somewhat hasty one and we determined no longer to delay the investigation of Knight’s experiment in relation to starch grains, which we had previously recognised as a necessary part of the statolith question. Our experi- ments were carried out on seedlings of Sefaria and Sorghum, in which the statoliths are in the cells of the cotyledon (coleoptile), and in which the position of the movable starch can easily and rapidly be determined by splitting the cotyledon longitudinally and examining the two halves mounted in iodine solution. The experiments were directed to two points, viz., the centrifugal force needed to produce (a) geotropic curvature, (6) movement of starch-grains. It will be seen that our results are directly opposed to those of Jost, inasmuch as, according to us, the lowest effective centrifugal force is about the same in the two sets of experiments. The centrifugal apparatus was driven by a hot-air engine regulated by one of Griffiths’s gas-regulators * Noll, ‘ Heterogene Induction,’ Leipzig, 1892; Haberlandt and N&mee, ‘ Ber. Deutschen Bot. Gesellsch.,’ 1900. + “Die Perception des Schwerereizes in der Pflanze,” ‘ Biolog. Centralblatt,’ vol. 22, March 1902, p. 161. t g. being the acceleration of gravity. VOE. LX XII. Dae 478 Mr. F. Darwin and Miss D. F. M. Pertz. [May 30, made by the Scientific Instrument Co. In the experiments on curvature the seedlings were cut and fixed by melted cocoa-butter to cork supports in small metal boxes, in which the air was kept damp. They were either placed tangentially at right angles to the axis of rotation, or else parallel to the axis. In the experiments on the distribution of the starch the seedlings were cut and placed in grooves in a sheet of cork, being kept in place by damp filter-paper and a second sheet of cork firmly fixed. In the majority of experiments the tip of the seedling was also fixed to a little bar of wax. The seedlings were fixed radially, the apices of some being inward and of others outward. This arrangement gives a striking result in successful experiments, for the starch travels to the apical end of the cells in the specimens whose apices point outwards, while it remains basal in the others. (a) Experiments on Curvature. We give our results in the form of a summary instead of publishing the details of each experiment. It will be seen that there was a good deal of irregular nutation; this is a drawback to the use of Sorghum and Setaria, but these plants being otherwise convenient we continued to employ them. Adding together the results obtained with centrifugal forces of 0:02 g., 0:03 g. and 0:04 g. we find that 85 seedlings (68 per cent.) curved to the centre (apogeotropically). 18 ,, (14:4 per cent.) did not curve at all. 22 4, (176 per cent.) curved away from the centre (pros- geotropically). We conclude from these results that seedlings of Sorghum and Setaria are to some extent stimulated geotropically by a centrifugal force of from 0:02—0-04 g. The average amount of apogeotropic curvature to the centre is only 20°, and as this is the result of about 22 hours stimulation, we are justified in believing that under our conditions a definite geotropic curvature cannot with any certainty be produced with centrifugal forces of much less value than 0-02 g. The fact that observers working with different plants and by other methods have found curvature with considerably weaker centrifugal force, does not concern us, since our investigation is a comparative one. (b) The Behaviour of the Statoliths to Centrifugal Force. Horizontal Ams. In the first series the seedlings were all placed radially with the apex outwards for 22—24 hours. The behaviour of the starch is not uniform in the cells of the cotyledons, so that it. is only possible to give a general impression of the results. 1904. | Notes on the Statolith Theory of Geotropism. 479 Thus “scattered and apical” means that the starch is to a great extent diffused through the cells, but that ina good many cases the starch has accumulated at the apical ends of the cells. On the other hand “apical and scattered” means that the starch is apical rather than scattered. Table I. Plant. i ae Temperature. Position of starch. net cts) |) ie ae ee Ee g. Ce -Setaria. .... | 0 -02 14 Scattered and apical. Reure. te..'| 0-02 16 Scattered, apical, and some basal. Pee ois he | 0-02 17 Scattered. Peters: . SOR 0:0082 BO oa Or 0:°0457 a8. ,, 9. 3870 0°0447 The same result is brought out in Tables III and IV, which apply to somewhat more concentrated solutions of milk sugar. 5 grammes Milk Sugar per 100 c.c. Table III. Table IV. 100 c.c. Enzyme Extract. 20 c.c. Enzyme Extract. t. el ts K. t. ice K. Lhr.. 13°% 0:0640 1 hr. 1°6 0-00700 2hrs, 22°) 0°0543 2hrs. 3:2 0:00705 Sita) aaa a 0-0460 Bey, 5-4 0-00602 D ay) abr O 0:°0310 QBe Ugg, Dae 0°00455 24, ys) YOLEO 0-0129 2925. 2 oed 0:00430 Bi oe 000370 i Sere 000365 144 ,, 49°6 0-00207 That similar results are obtained on working with different materials at different times is shown in Tables V and VI, which refer to experi- ments made at times 4 months apart. 10 grammes Milk Sugar per 100 c.c. Table V. Table VI. Strong Enzyme Extracts. te L. K. t. 5 K. lair = ae 00605 MEET a5 TEES Th 0:0560 2eins.) 22-1 0:0543 2 hrs.) L820 00431 Ae, eee A 0:0446 3-4, / 22°0 -) OgCsas (oe gorse Y 0°0346 Aion lie Oa 0°0314 74: SN nn eS ES) 0°0146 23 55 SACLE 0- 0119 48 ,, 64°95 0-0094 28) oe 20 0:°0117 AT. ys) 61-0, 00087 | 1904. ] Studves on Enzyme Action. 507 Emulsin.—Tables VII—X refer to experiments made with emulsin ; the conclusions to be drawn from them are precisely similar to those deduced from the experiments with lactase. It is to be noted that emulsin acts much less rapidly than lactase. 2 grammes Milk Sugar per 100 c.c. AJAY og a © © K. "0218 °0169 "0095 "0081 “0069 "0064 "0059 ‘0057 "0055 K. OLaY *0102 “00896 -00790 -00667 °00478 °00404 Table VII. Table VIII. 0:2 gramme Emulsin. 0-4 gramme Emulsin. t. £. K. t. X. hr. 3°2 0:0282 ihre 4°9 ee, 4°8 0:0214 2 hrs. 7°5 2hrs. 6:4 0°0143 4h, 9°4 aa 7°6 0-0114 6 ,, § 1O°6 4s 5, 9:0 0:0091 23. 80:5 oo. | TOO 0:0091 BO 3a Zar, LOT 0:0041 AS, AES Be 22 O 0:0037 53>, 50r0 ae 5, 290 0:0031 144 ,, 84:0 B3,, 30° 7 0-0030 44, = 622 0-0029 206 3, THD 00024 ! 5 grammes Milk Sugar per 100 c.c. Table IX. Table X. 0-4 gramme Emulsin. 0-1 gramme Kmulsin. t. GI) K. t. He Pr: I@ 0-00440 tlie, 3°1 Zens. 1*8 0:00395 2hrs. 4:6 eh 3-9 Gh 00352 Sut G0 On 4°5 0-00333 Ao 120 Za 5; 15:0 0°00320 4; 8°8 46 ,, 25°5 0-00277 22 4, 21-5 LO, 30 °6 0:00271 46 ,, 34°7 Lote 3 «O47 0-00206 TO toe 0 Soo eo 2 2 © © "00382 Maltase.—Tables XI and XII refer to experiments with maltase and maltose and show that here again the same conclusions hold. K decreases steadily: when the proportion of enzyme is small the change is at first linear. It is of interest as confirming these results that when K is calculated from the value given in Croft Hill’s tables* a similar series of decreasing values is obtained. * Chem. Soc. Trans.,’ 1898, p. 634. 508 Dr. E. F. Armstrong. [Apr. 5; Table XI. Table XII. Maltose, 5 per cent. Maltose, 10 per cent. i 2. KE iz a vig 1 hr. ie3 0:0329 ihr: ho 0:0209 2uhrs), 1379). 00325 3 hrs. hale 0:0180 4 ,, 24:4. 0:0304 5 LSS 0:0170 74 ,, Sud: 0:0229 eae 235 °9 0:0052 23.,, 35°2 0:0082 28 ,, 25:0 0:0045 47 31°4 0°0035 2) Concentration of Hydrolyte.—Although the experiments recorded in Tables I—XII furnish evidence that the rate of change decreases as the concentration of the sugar solution is increased, while the actual weight of sugar hydrolysed increases; as the experiments were made at different times and with different materials, it was necessary to carry out a strictly comparable series in which the amount of sugar present was the only factor varied. Table XIII shows the results obtained in experiments with. relatively concentrated solutions of milk sugar and lactase in which the amount of enzyme present was quite small. As the concentration of sugar was increased, the rate of change diminished so that the fraction of sugar hydrolysed in a given time was inversely proportional to the amount of sugar present—so that a constant weight of sugar was changed independently of the con- centration—a result which is in agreement with Adrian Brown’s observations with invertase. Table XIII.—Amount of Sugar Hydrolysed. 24 hours. 46 hours. 144 hours. Solutions containing— =| : - Propor- | Weight.) TTP? | weight.) *Z0P™ | Weight. | 10 per cent. ......| 14-2 1-42 22 -2 2 -22 33 °4 3°34 ZAR aie | 7:0 1-40 10°9 2°18 16°9 3°38 BU 4.°8 1°44 | Fe: 2°21 11:0 3°30 Table XIV relates to a series of experiments made with emulsin and milk sugar in which, however, the proportion of emulsin taken was not particularly small. Although the rate of change decreased as the concentration increased, the experiments afforded no evidence that a constant weight of sugar was hydrolysed in a given time inde- pendently of the concentration. In a second series, however, in which a considerably smaller proportion of enzyme (see Table XV) was used, 1904. | Studies on Enzyme Action. 509 it was found that an approximately constant weight of sugar was hydrolysed whatever the concentration of the sugar. Table XIV.—Amount of Sugar Hydrolysed. 22 hours. 46 hours. 70 hours. * Solution containing— oe P P por- ‘ ropor- ‘ ropor- : eo eee at. | © Weight.) “j-5,, | Weight. )10 percent... ...... 22 Ve 29 2-9 39°6 3°96 gee SS. 13 2-6 20 4-0 25:1 | 5:0 30 3 Le a eee 10 3°6 15 4°5 19:0 4°7 Table XV.—Amount of Sugar Hydrolysed. 23 hours. 48 hours. 92 hours. Solution SSS containing— i; Propor- | Weaight.| 'T°R°™- | wWeight,| FTOP°™ | Weight tion. oan tion. = brome Oe ek ( | 10 per cent....-. LO sly) 29 °8 2-98 35 °7 3°57 20) 2a eee 10°6 | 2-12 15°3 | 3:06 23°3 | 4-66 aU. e 7-0 3-06 16°3 | 4°89 2-10 NOS 4 if The conclusions drawn from the experiments recorded in the three preceding tables apply only to the concentrated solutions in which the proportion of enzyme present was small; in very dilute solutions, on the other hand, quite another effect is produced by changing the concentration. Table XVI relates to experiments in which the pro- portion of enzyme present was large relatively to that of sugar. It will be seen that on increasing the amount of sugar present there was nearly a proportionate increase in the amount hydrolysed, though the proportion hydrolysed, as well as the value of K, remained constant, a result in agreement with the law of mass action. It should be mentioned, however, that the error affecting titration in these experiments is somewhat large, owing to the influence of the proteids on the determination of the end point. Table XVI. | Milk sugar per | Amount changed | K 100 c.e. in 3 hours. i | ate | na 1°O gms, 0 *185 0 -0296 0° 0-098 0 :0298 0°2 0 °0416 0 0337 510 _ Dr. E. F. Armstrong, [Apr. 5, Concentration of Enzyme.—Lastly, Table XVII shows the effect of varying the amount of enzyme present. Experiments in this direction: are limited by the uncertain nature of the material, as well as by the fact that it is the concentration of enzyme relatively to that of sugar which must be studied. ‘The solution contained 5 per cent. of sugar. Solutions containing varying amounts of enzyme were obtained by diluting portions of one and the same stronger solution. The weight hydrolysed in a given time by varying amounts of enzyme was approximately proportional to the amount of enzyme, provided that the amount was not too large and also that the comparison was made during the earlier stages of hydrolysis before the secondary products began to exert a marked influence. Table XVII.—Proportions Hydrolysed in 100 c.c. of a 5 per cent. Solution. Solution containing |1°5 hours. | 20 hours. | 25 hours. _ 45 hours. | 68 hours. | 1 c.c. lactase...... 0°15 Zak 2G; 23) 3°9 4°8 2-5 c.c. lactase.... 0 °4, oro. || GS - i More 12 °6 TOY ehh Relates 23-3 =) ih e8@RG 48 °5 3°2 45 °8 54°5 — == | 20. ;, »” ! Very small quantities of the enzymes lactase and emulsin were found to be capable of hydrolysing only a small amount of sugar: their action then ceased. This result affords very definite evidence in favour of the view that the products of hydrolysis—in this case glucose and galactose—are capable of combining with enzymes and of removing them from the sphere of action. Table XVIII.—Proportions Hydrolysed in 100 ¢.c. of a 5 per cent. Solution. a | | Solution containing | 24 hours. 144 hours. | | 0°66 c.c. lactase.... 2°3 2°3 TiO): epee 32 35 2:0) Pes At 6°3 7 °4 520)! aoe 15-4 34-0 Summary and Discussion of Results—Nature of Enzyme Action. From the results recorded in the foregoing tables, it is clear that two periods may be distinguished in the course of change—an earlier period, during which the change is a linear function of the time; and a later period, during which change proceeds at another rate (Tables ) 1904.] Studies on Enzyme Action. 511 Il, IV, VII, XI). The duration of these periods is conditioned by the relation which the amount of enzyme present bears to the sugar, the linear period disappearing when the amount of enzyme present is relatively considerable (Tables I, III, VIII). It will be noted that in the eases studied in which the action proceeds comparatively slowly, e.g., hydrolysis by lactase, emulsin or maltase—K falls rapidly in value ; whereas when the action proceeds rapidly, ¢.g., hydrolysis by invertase or diastase—K increases. To understand the origin of these differences, it is desirable to consider the subject of mass action somewhat in detail, even at the risk of repetition. Changes such as that which cane sugar undergoes in an aqueous solution under the influence of a catalyst, in which, on account of its large relative mass, the part played by the solvent may be disregarded —known as changes of the first order—are assumed to follow the logarithmic law of mass action: it is supposed that the one substance only undergoes change and that the same fraction of the residue is changed in successive equal intervals of time, 7.¢., the factor K deduced rom the equation K =. log a is a constant. But this is an ideal conception, as it involves the assumption that the amount of catalyst functioning is so small as to be negligible. In actual practice, not only is the amount of enzyme used not incon- siderable but it is known that the rate of change is affected both by alteration in the concentration of the hydrolyte and in the proportion which the hydrolyte bears to the catalyst ; moreover, the products of change are known to exercise an influence and in concentrated solutions reversion may take place. In view of the proof afforded by E. Fischer’s researches that a close relationship exists in configuration between the hydrolyte and the enzyme which alone affects it, there can scarcely be any doubt that hydrolysis by enzymes, in the first instance, involves the combination of the enzyme with the hydrolyte, as Horace Brown and Glendinning have assumed to be the case. On this assumption, the rate at which the change proceeds will depend on the extent to which the combination of enzyme with sugar takes place as well as on the degree of readiness with which the compound breaks down. The proportion of the total sugar present in solution combined with the enzyme and undergoing change at any one moment may be regarded as the active mass of the sugar. The conception of an active mass has already been introduced by Arrhenius* but in another connection—to account for the very rapid rise in the value of K, when cane sugar is hydrolysed by acids, occasioned by a rise in temperature, the increase * ‘ Zeit. Phys. Chem.,’ 1889, vol. 4, p. 226. 512 Dr. E. F. Armstrong. [Apr. 5, being far greater than can be accounted for by the increased “ionisation” or by the greater mobility of the molecules at the higher temperature. Arrhenius supposed that only a part of the total hydrolyte present. was really concerned in the hydrolysis: this he termed the “active part.” It was arbitrarily assumed that this was a fixed proportion of the total mass at any temperature and that the proportion increased rapidly as temperature rose. On the hypothesis that the enzyme combines with the sugar, the active mass of the sugar will be that portion s of the whole S which is in combination with an amount of enzyme e: it will be convenient. to speak of the combination s+e as the active system. It is to be supposed that several influences are at work in a solution containing enzyme and sugar: on the one hand, enzyme and sugar molecules seek to combine; but on the other, the water molecules also tend to unite with the sugar molecules—so that there is, so to speak, competition between the enzyme and the water for the sugar molecules, which results in the establishment of an equilibrium depending to: some, though probably a very limited extent, on the proportion rela- tively to enzyme in which the water and sugar molecules are present.* ‘The possibility that the products of change also compete for the enzyme must, however, not be left out of consideration. It is necessary to consider separately four sets of conditions, viz. :— Case I, in which, whatever the amount of sugar present, the quantity of enzyme is relatively small. Case II, in which there is a difference from Case I inasmuch as the quantity of enzyme is relatively considerable. Case III, in which the amount of enzyme diminishes as the action proceeds. Case IV, in which the amount of sugar present is varied. Case I.—As hydrolysis proceeds, assuming that the enzyme itself is not affected by the work it does, since the magnitude of the active system depends on the amount of enzyme present, it is obvious that in the initial stages, if the total amount of sugar present S be large com- pared with s, the enzyme will be in presence of enough sugar molecules. to establish the maximum possible number of effective combinations : or in other words the magnitude of the active system will remain constant and the change will be expressible, as Brown and Glendinning have pointed out, as a linear function of the time. As hydrolysis. * The enzymes, it is to be supposed, are colloids, 7.e., substances which have but little affinity for water: the stability of the combination s + e will therefore be but to a slight extent dependent on the proportion in which water is present ; whereas, probably, in the case of a combination of a crystalloid, such as an acid, with sugar, the proportion in which the components of the system are in equilibrium would vary to aconsiderable extent with the concentration. This is known to be the case.. The effect of acids is dealt with in a separate note. 1904. | Studies on Enzyme Action. 513 proceeds, the amount S of sugar present decreases until it is no longer negligible compared with that of the active part s and hence the enzyme will no longer effect the maximum possible number of com- binations: the proportion of sugar s undergoing change will then be a function of the total mass and the formation of active systems will be governed by the law of mass action. The rate of change will be a logarithmic function of the time. This explanation is fairly in accordance with the observed facts in the case of invertase and diastase, the only enzymes hitherto experimented with, which have always been used in very small quantities. Case IJ.—Ii, on the other hand, the quantity of enzyme used be relatively large, the active mass will be a function of the total mass from the very beginning of the experiment, so that the linear part of the change will escape notice. O’Sullivan and Tompson seem to have used a relatively large proportion of enzyme, and therefore it is easy to: understand why they found the action of invertase to follow the logarithmic law, whilst subsequent observers using relatively small quantities of enzyme have noted departures from this law. Case IJ[.—When the amount of the enzyme does not remain constant but for some reason decreases, the magnitude of the active system will not only be a function of the amount of sugar but also of that of the enzyme ; it will therefore be represented by an equation of the second order, in which both of two interacting substances decrease—as, for example, is the case in the interaction of an alkali and methylic acetate. Such an expression corresponds to a curve falling off from a logarithmic curve and therefore giving a series of decreasing values. for K when this is calculated for the simple logarithmic law. In such a case, the change in its early stages will still be a linear function of the time, as the diminution in the amount of enzyme will not at first. materially influence the magnitude of the active system. Stated shortly, the ordinary equation of mass action a = KS-2) where S is the total sugar and x the amount changed in time 7, is applicable: only to the period during which a constant relatively large proportion of enzyme is present together with a continually decreasing amount of sugar but uninfluenced by the products of change. During the final period, when the products of change exercise an. influence by withdrawing enzyme from the sphere of action, dx a= 56-9 E-») 514 - Dr. E. F. Armstrong. [Apr. 5, where E is the total enzyme, y the amount withdrawn in combination with the products in time ¢. During the period when the proportion of sugar present is very large, « becomes negligible compared with 8, so that dx Ga KS=h where & is a constant. The apparent duration of the linear period must be affected not only by « becoming no longer negligible compared with S but also by the extent to which the products of change make their influence felt. It may here be pointed out that Henri’s formula combines in a single expression the linear and logarithmic periods, but does not take into account the last period during which the products of the change exercise a retarding influence. All these points are brought out not only in the results given in this paper but also in those tabulated by Brown and Glendinning. Although these observers have called attention to the existence of a linear period followed by a “ logarithmic” period, they have failed to point out the meaning which may be attached to the decrease in the values of K observed during the latter stages of hydrolysis, when the products exercise a marked influence. Their experiments afford an example of such an influence. Thus in Table IV,* referring to the hydrolysis of a 3 per cent. solution of starch by diastase at 21°, the value of K will be seen to increase until about half the starch was hydrolysed and then to decrease. If the value after 60 minutes be made the new starting point and the values he recalculated in the manner adopted by Brown and Glendinning,t the following figures are obtained :— aR Old time units. ™ New time units. Ko. 60 minutes — aie SO Ue, 20 minutes 458 1OOs aa BO ea 461 120 aos, GbOse s. 448 140 5 80 . 410 It is evident that whereas K, should be a constant if the period were “logarithmic,” it begins to decrease after about 66 per cent. of the starch has been hydrolysed: showing that the influence of the products in removing enzyme has begun to make itself felt. The action of invertase appears to be much less affected by invert * Loc. ctt., p. 400. + Loe. cit., p. 393. 1904. ] Studies on Encyme Action. 515 sugar and that of diastase by maltose than is that of lactase, emulsin or maltase by the products to which they respectively give rise. Consequently, for these latter enzymes, the linear period is of short duration (Tables II, IV, VI, XI) and the logarithmic period is barely perceptible, owing to the rapid reduction in the rate. Case [V.—When the amount of enzyme and water is kept constant whilst that of sugar is increased, it may be supposed that the magnitude of the active system will increase until s+e reaches a maximum, a definite equilibrium being established between enzyme, sugar, and water, the whole of the enzyme, perhaps, becoming com- bined with the sugar. It may be assumed that if the amount of sugar be further increased, the equilibrium will remain unaffected, notwithstanding that an addition of sugar is practically equivalent to a withdrawal of water. But it s+e remain unaltered, whatever the proportion of sugar present beyond a certain minimum, a constant amount of hydrolyte will undergo change in a given time, although the proportion changed as also the value of K will decrease as the concentration is increased. This conclusion is entirely in agreement with the facts elucidated, especially by Adrian Brown, and with my own observations. Lastly, it should be pointed out that in discussing the action of enzymes, besides taking into account the conditions affecting the formation of an active system and the part which such a system plays in hydrolysis, it is necessary to consider the relative stability or tendency to break down under the influencezof water of the combina- tion of enzyme and sugar in connection with the very different rates at which different enzymes condition hydrolysis. Generally speaking, the observed differences in the rate at which hydrolysis is effected may be conditioned by— (a) Specific differences in the enzymes, ¢.g., lactase as compared with emulsin, which both hydrolyse milk sugar. . (>) Differences in the configuration of the hydrolyte, e.g., 6-methyl glucoside and 6-methyl galactoside, which are both hydrolysed by emulsin. (c) Differences in the stability of the hydrolytes, ¢.g., cane sugar as compared with milk sugar. But it is very difficult to institute just comparisons, for whereas, in the case of acids, the effect of only a single substance on the variety of sugars may be contrasted, in studying the hydrolysis of sugars under the influence of enzymes, it is necessary in most cases to use a different enzyme for each sugar, so that positive data are not easily obtained. Experiments made under comparable conditions with acids to test their action on different sugars show that these are hydrolysed at very different rates: thus, for instance, whilst cane sugar is hydrolysed nearly 1000 times as rapidly as maltose, this latter under- VOL. LXXIII. 20 516 Dr. E. F. Armstrong. [Apr. 5, goes change about 1°3 times as rapidly as milk sugar. It is scarcely possible to doubt that even greater differences exist both in the affinity of the enzymes for the sugar and in the degrees of readiness with which the enzyme sugar systems break down than are known to hold in the case of acid sugar systems. It is important to keep these considerations in view in discussing the rate at which different enzymes effect hydrolysis. It will be apparent from what has been said that there is no reason to suppose that the action of enzymes follows any other than a normal course ; the difficulties which have been met with in inter- preting such changes may be ascribed to the incomplete consideration of the numerous factors involved. [D 1810, 8010; Q 1230.] “ Studies on Enzyme Action. III.—The Influence of the Products of Change on the Rate of Change conditioned by Sucroclastic Enzymes.” By EpWArRD FRANKLAND ARMSTRONG, Ph.D., Salters’ Company’s Research Fellow, Chemical Department, City and Guilds of London Institute, Central Technical College. Communicated by Professor H. E ARMSTRONG, F.R.S. Received April 5,—Read April 28, 1904. D 1810 Hexoses—power of sucroclastic enzymes to combine with. D 8010 Enzymes—activity of correlated with configuration of hydrolyte. D 8010) Emulsin, lactase, maltase—varying influence of glucoses and Q, 1240 glucosides on their activity. In the previous paper, it has been shown that, in order to explain the action of sucroclastic enzymes, it is necessary to assume not only that the enzyme combines with the hydrolyte but that it is also more or less affected by—-and presumably combines with—the product of change. At present there is but little information available bearing on this latter contention. The experiments to be described have been made with the object of ascertaining by direct observation whether and to what extent the action of a given enzyme is affected by one or more of the products formed under its influence. They establish very clearly the existence of a close relationship between the configuration of the hexose and the enzyme in those cases in which a retarding influence is apparent: it is difficult to explain such a result except on the assumption that the enzyme and hexose combine together in some intimate manner. 1904] Studies on Eneyme Action, puley Historical.—The observation made by Wiirtz in 1879,* when studying the action of Papain on fibrin, that the enzyme was in some way retained, probably first gave rise to the conception that the primary action of enzymes was additive in character. The first definite evidence of combination, however, appears to be that advanced by O’Sullivan and Thomson in 1890.; It was shown by these observers that in the presence of cane sugar invertase will withstand without injury a temperature fully 25° higher than it will in its absence. They pointed to this as a very striking fact which was difficult to explain except on the assumption that the invertase enters into combination with the sugar ; they further supposed that it was capable of combining with: the invert sugar. Systematic experiments were made by Tamman in 1892,t who showed that the hydrolysis of amygdalin and salicin by emulsin is materially retarded in either case on adding any one of the products of their change in advance. The effect of glucose on the hydrolysis of maltose by maltase was studied by Croft Hill in 1898.§ The retardation observed was attri- buted by him, perhaps not quite logically, merely to a reverse action by which the glucose underwent conversion into maltose: it may be added that the investigation was undertaken by him from this point of view. A more definite step forward was taken by Henri in 1901, who showed that the retarding effect of invert sugar was mainly due to the fructose, glucose having little or no effect on the action of invertase on cane sugar. Evidence that the effect was due to a specific action of the invert sugar was adduced in the same year by Adrian Brown,|| who showed that whilst the rate of hydrolysis of cane sugar was materially reduced by invert sugar, the corresponding amount of milk sugar had little or no effect, thus precluding the conclusion that the retardation was due to increased viscosity of the liquid. In the course of his classic researches on the action of enzymes on the stereochemically related glucosides commenced by Emil Fischer in 1894, it was clearly established that the closest relationship exists: between the configuration of the hydrolyte and that of the particular’ enzyme which can affect it. In a course of lectures delivered at the Pharmaceutical Society in' 1892,** my father discussed fermentation phenomena generally from the point of view that hydrolysis was conditioned by the association of. * b A. 360. Sold by Harrison and Sons, St. Martin’s Lane, PHILOSOPHICAL TRANSACTIONS FOR 1901-3. Szrizs A. Containing Papers of a Mathematical or Physical Character. Vol. 201 (1903). With 3 Plates. Price £1-16s. 6d. Vol. 202 (1903). With 13 Plates. Price £1 13s. Od. Srrizs B. Containing Papers of a Biological Character. Vol. 194. With 49 Plates. Price £3 Os. Od. Vol. 195. With 16 Plates. Price £1 7s. Od. RECENT PAPERS IN THE ‘PHILOSOPHICAL TRANSACTIONS.’ . 354, 355. . 356. . 307. . 308. . 309. “20. . 218. HZo. . 220. . 221. 222, Published by Duzav & Co., 37, Soho Square. An Analysis of the Results from the Kew Magnetographs on “ Quiet”’ Days during the Eleven Years 1890 to 1900, with a Discussion of Certain Phenomena in the Absolute Observations. By CHARLES CHREE, Se.D.; LL.D., F.R.8., Superintendent of the Observatory Department of the National Physical Laboratory. Price 6s. On the Photo-Electric Discharge from Metallic Surfaces in Different Gases. By W. Mansercu Varuey, M.Sc. (Vict.), Ph.D. (Strasburg), B.A. (Cantab.). Late 1851 Exhibition Research Scholar, Emmanuel College, Cambridge; Assistant Professor of Physics and Electrical Engineering, Heriot-Watt College, Edinburgh. Price 1s. 6d. ! An Enquiry into the Variation of Angles observed in Crystals; especially of Potassium-alum and Ammonium-alum. By Henry A. Mizrs, D.Se., M.A., F.R.S., V.P.G.S., V.P.C.S., Waynflete Professor of Mineralogy in the University of Oxford. Price 33. 6d. Radiation in the Solar System: its effect on Temperature and _ its Pressure on Small Bodies. By J. H. Poyntine, 8c.D., F.R.S., Pro- fessor of Physics in the University of Birmingham, Price ls. Title, Contents, Index, &c., Vol. 202. Price 6d. On the Propagation of Tone over the Surface of an Elastic Solid. By Horace Lams, F.R.S. Price 2s. On the Structure of Gold-Leaf, and the Absorption Spectrum of Gold. By J. W. Mattet, F.R.S., Professor of Chemistry in the University of Virginia. Price 1s. On the Evolution of the Proboscidea. By C. W. ANDREWS, D.Sc., F.G.S., British Museum (Natural History). Price 1s. Experiments in Hybridization, with Special Reference to the Effect of Conditions on Dominance. By Leonarp Doncaster, B.A., King’s College, Cambridge. Price 3s. On Reptilian Remains from the Trias of Elgin. By G. A BouLzeneer, F.R.S. Price 3s. 6d. Studies in the Morphology of Spore-producing Members. No. V.— General Comparisons and Conclusion. By F. O. Bower, Sc.D., F.R.S., Regius Professor of Botany at the University of Glasgow. Price 2s. 6d. A Method for the Investigation of Fossils by Serial Sections. By W. J. Sonzas, D.Sc., LL.D., ¥.R.S., Professor of Geology and Paleontology in the University of Oxford. Price 1s. An Account of the Devonian Fish, Paleospondylus Gunni, TRAQUAIR. By W. J. Sounas, D.Se., LL.D., F.R.S., Professor of Geology and Paleontology in the University of Oxford, and IcrrNA B. J. Soztas, B.8c., Lecturer on Zoology in Newnham College, Cambridge. Price 2s. Just Published. Royal 4to. 428 pues with numerous Plates ue d Diagrams, sit and Separate Portfolio of Geological Maps and Sections. THE ATOLL OF FUNARUTE | BORINGS INTO A CORAL REEF AND THE RESULTS. Price 30s. To Fellows of the Royal Society, 20s. Published by the Roya Society, and sold by HARRISON AND Sons, St. Martin’s Lane. Just Published. Demy 8vo. REPORTS OF THE SLEEPING SICKNESS COMMISSION. No. I—Containing Communications 1 and 2. By Dr. CASTELLANI, Cou. Bruce, and Dr. NaBarro. Price 7s. 6d. No. Il.—Containing Communications 3—5. By Drs. Caristy, CASTELLANI, and Low. Price 7s. 6d, : No. II.—Containing Communications 6 and 7. By Dr. Caristy and Mr. THEOBALD. Price 6s. . No. IV.—Containing Communication 8. By Cot. Bruce, Dr. | NABARRO, cee Capr. Greic. Price 6s. 6d. REPORTS TO THE MALARIA COMMITTEE OF THE ROYAL SOCIETY. First Series, 1900. Demy 8vo., pp. os Price 2s. 6d. Second ,, 1900. ‘ 5 2. PAL Boa 7-2 hind. co woo. vA i A ts 2s: Fourth ,, 1901. eg e 20. 5) SS Bib POOR: - a 79. 1 eS. ae Sixth ,, 1902. a Bee 23. 2 Is. Seventh ,, 1902. Ms 52. ie 3s Highth ,, 1903. a oe rere bE 2s. Demy 80. pp- 160... Price 10s. REPORTS TO THE EVOLUTION COMMITTEE OF THE ROYAL SOCIETY. - Revorr I.—Experiments undertaken by W. Batson, F.R.S., and Miss E. R. SAUNDERS. Published and sold by HARRISON AND SONS, 45, 46, & 47, ST. MARTIN’S LANE, W.C., R. FRIEDLANDER & SOHN, BERLIN, AND ALL BOOKSELLERS. PROCEEDINGS OF THE ROYAL SOCIETY. VOL. LXXIIL ! No. 489. CONTENTS. ok, PAGE On the Acoustic Shadow of a Sphere. By Lorp Rayuriceu, O.M., F.R.S. With an Appendix, giving the Values of Legendre’s Functions from Po to Po) at Intervals of Five Degrees. By Professor A. LODGE. (Abstract) ° ; : . : “ : : rivawe : ~ pop The Significance of the Zoological Distribution, the Nature of the Mitoses, and the Transmissibility of Cancer. By EH. F. Basurorp, M.D., and J. A. Murray, M.B., B.Sc. Communicated by Professor J. Rosz BRADFORD, F.R.S. (Plate 2) . : : : é d : : st GB: Conjugation of Resting Nuclei in an Epithelioma of the Mouse. By E. F. BasH¥orD, M.D., and J. A. Murray, M.B.,B.Sc. Communicated by Professor J. Rose BRADFORD, F.R.S. . ‘ ‘ E : i RAC On the Part played by Benzene in Poisoning by Coal Gas. By R. STAEHELIN, \ M.D., Senior Assistant in the Medical Clinic at Basle. Communicated by Professor E. H. Stantine, F.R.S. . j é i k ‘: a Ce The “Islets of Langerhans” of the Pancreas. By H. H. Dats, B.Ch., George Henry Lewes Student. Communicated by Professor StaRLine, M.D., E.R.S. (Abstract) ° wy ae: ° ° ° e ° ° 84 The Reduction Division in Ferns. By R. P. Gregory, St. John’s College, Cambridge, University Demonstrator in Botany. Communicated by Professor H. MARrsHALL WARD, F.RS. . . HLS Rae oO Ee Sy The Secreto-motor Effects in the Cat’s Foot Studied by the Electrometer. By Aveustus D. Watter, M.D., F.B.8. win eis oie yan Us acl IN, ae On the Effects of Joining the Cervical Sympathetic Nerve with the Chorda Tympani, By J. N. Lanerey, F.R.S., and H.K. AnpERson, M.D. . 99. The Longitudinal Stability of Aerial Gliders. By G. H. Bryan, Sc.D., F.RS., and W. H. WituiamMs, B.Sc., University College of North Wales , . 100 Cultural Experiments with “Biologic Forms” of the Erysiphacee. By Hanest §. Sarmon, F.L.8. Communicated by Professor H. Marsuarn Warp, F.R.S. (Abstract) . s : 3 . Se ie 2 16 On the Origin of Parasitism in Fungi. By Grorer Masser, Principal Assistant, Herbarium, Royal Gardens, Kew. Communicated by Sir Wittiam T. Tuise.ron-Dyrr, K.C.M.G., C.LE., F.R.8. (Abstract) . 118 Minutes of Meetings of February 4, 1904, and February 11, 1904 . e ° = Price Two Shallongs, sonia. Fepruary 24, 1904. fart NOTICE TO AUTHORS. Authors of Papers are urgently requested to send in all Drawings, Diagrams, or other Illustrations in a state suitable for direct photographic reproduction. They should be drawn on a large scale, in Indian ink, on a smooth white surface. Where illustrations are numerous, much time would be saved if authors would indicate in advance those considered essen- tial to the understanding of their papers. NOTICES TO FELLOWS OF THE ROYAL SOCIETY. The Council have directed that the Minutes of the Meetings of the Society shall be sent out as an inset in the ‘ Proceedings,’ separately paged, and shall afterwards be republished in the ‘ Year-Book.’ The ‘ Proceedings’ are sent by post to every Fellow of the Society who resides within the limits of the Postal Union. The ‘Philosophical Transactions’ are delivered only to those Fellows who call for them, or who send a written applica- tion to the Assistant Secretary. Such an application may, if so desired, be filed as a, standing order. The separate Papers forming the ‘ Philosophical Transactions’ will be sent post free, as issued, to those Fellows who prefer to have them in that form. And, on application to Messrs. Haryvison and Sons, 46, St. Martin’s Lane, these will be bound in a cloth case for 2s. 6d., or the cloth cases for binding may be purchased, price Is. 6d. = 6 6 Demy 8vo, Cloth, 12s. 6d., with Plates and Woodeuts, 427 pp. THE RECORD OF THE ROYAL SOCIETY. SECOND EDITION, 1901. Containing ‘an account of the Foundation and Early History of the Society, its Benefactors, Trust Funds, &., with Chronological and Alphabetical LISTS: OF THE FELLOWS FROM THE FOUNDATION. Just’Published. Royal 4to. 307 pages and 27 Plates. Price 20s. To Fellows of the Royal Society, 15s. : REPORT TO THE GOVERNMENT OF CEYLON ON THE PEARL OYSTER FISHERIES OF THE GULF OF MANAAR. By W. A. HerpMAN, D.Sc, F.RB.S. With Supplementary Reports upon the Marine Biology of Ceylon by other Naturalists. Parr T, 1903, Published by the Royal Society and sold by HARRISON AND SONS, ST. MARTIN’S LANE. Sold by Harrison and Sous, St. Martin’s Lane. PHILOSOPHICAL TRANSACTIONS FOR 1901-3. Series A. Containing Papers of a Mathematical or Physical Character. Vol. 201 (1903). With 3 Plates. Price £1 16s. 6d. Vol. 202 (1903). With 13 Plates. Price £1 13s. Od. Serizs B. Containing Papers of a Biological Character. Vol. 194. With 49 Plates. Price £3 Os. Od. Vol. 195. With 16 Plates. Price £1 7s. Od. RECENT PAPERS IN THE ‘PHILOSOPHICAL TRANSACTIONS.’ A, 354, A, 355. Published by Duzat & Co., 37, Soho Square. An Analysis of the Results from the Kew Magnetographs on “ Quiet’’ Days during the Eleven Years 1890 to 1900, with a Discussion of Certain Phenomena in the Absolute Observations. By CHARLES CHREE, Se.D., LL.D., F.R.S., Superintendent of the Observatory Department of the National Physical Laboratory. Price 6s. On the Photo-Hlectric Discharge from Metallic Surfaces in Different Gases. By W. Mansercu Varury, MSc. (Vict.), Ph.D. (Strasburg), B.A. (Cantab.), late 1851 Exhibition Research Scholar, Emmanuel College, Cambridge; Assistant Professor of Physics and Electrical Engineering, Heriot-Watt College, Edinburgh. Price 1s. 6d. An Enquiry into the Variation of Angles observed in Crystals especially of Potassium-alum and Ammonium-alum. By Henry See sixth)... 1902. is is 22. 3 Is. Seventh ,, 1902 Hf AY He. 3 as. Highth ,, 1903. % i vd. he 2s. Demy 8vo., pp. 160. Price 10s. REPORTS TO THE EVOLUTION COMMITTEE OF THE ROYAL SOCIETY. » Report I.—Experiments undertaken by W. Batzson, F.RB.S., and Miss E. Be SAUNDERS, Published and sold by HARRISON AND SONS, 45, 46, & 47, ST. MARTIN’S LANE, W.C., R. FRIEDLANDER & SOHN, BERLIN, AND ALL BOOKSELLERS. : ik \e Sea Sse asta ae (Pde tone alata esate RS : : PROCEEDINGS OF EW ROYAL SOCIETY. VOL. LXXIIL. No. 492. CONTENTS. PAGE On the High Temperature Standards of the National Physical Laboratory : an account of a Comparison of Platinum Thermometers and Thermo- junctions with the Gas Thermometer. By J. A. Harxer, D.Sc., Fellow of Owens College, Manchester, Assistant at the National Physical Laboratory. Communicated by R. T. GuazeBRoox, F.R.S. (Abstract) 217 The Spectra of Antarian Stars-in Relation to the Fluted Spectrum of Titanium. By A. Fowuer, A.R.C.S., F.R.A.S., Assistant Professor of Physics at the Royal College of Science, South Kensington. Com- municated by Professor H. L. CannENDAR, F.R.S. (Plate 6) : . 219 The Specific Heats of Metals and the Relation of Specific Heat to Atomic Weight.—Part III. By W. A. TitpENn, D.Sc., F.R.S., Professor of Chemistry in the Royal College of Science, London. (Abstract) . . 226 Further Researches on the Temperature Classification of Stars. By Sir Norman Lockyer, K.C.B., LL.D., E.R. (Plates 7—9) s é L228 On the Construction of Some Mercury Standards of Resistance, with a Determination of the Temperature Coefficient of Resistance of Mercury. By F. E. Situ, A.R.C.Sc., Assistant at the National Physical Laboratory. Communicated by R. T. GuazEBROOK, F.R.S: (Abstract) 239 On Electric Resistance Thermometry at the Temperature of Boiling Hydrogen. By Professor James Dewar, M.A., LL.D., F.RS. (Plate 10). . 244 Physical Gunsant at Low Temperatures. (1)—The Densities of Solid Oxygen, Nitrogen, ec etc. By Professor JAMES Dewar, M.A., LL.D., D.Sc., F.R.S. : ’ : ; : é , : . 261 Minutes of Meetings of March 17, 1904, and March 24, 1904. k : Raab 8 ¢ oe APR 99 1904 OFfce Lib Libraty joni neti , Price Three Shallings. APRIL 9, 1904. NOTICE TO AUTHORS. Authors of Papers are urgently requested to send in all Drawings, Diagrams, or other Illustrations in a state suitable for direct photographic reproduction. They should be drawn on a large scale, in Indian ink, on a smooth white surface. Where illustrations are numerous, much time would be saved if authors would indicate in advance those considered essen- tial to the understanding of their papers. NOTICES TO FELLOWS OF THE ROYAL SOCIETY. The Offices and Library will be closed for the Easter Recess from (and including) Thursday, March 381, to the end of Easter week. The Council have directed that the Minutes of the Meetings of the Society shall be sent-out as an inset in the ‘ Proceedings,’ separately paged, and shall afterwards be republished i in the ‘ Year-Book’ — The ‘ Proceedings’ are sent by post to every Fellow of the Society who resides within the limits of the Postal Union. The ‘Philosophical Transactions’ are delivered only to those Fellows who call for them, or who send a written applica. tion to the Assistant Secretary. Such an application may, if so desired, be filed as — a standing order. ‘The separate Papers forming the ‘ Philosophical Transactions’ will be sent post free, as issued, to those Fellows who prefer to have them in that form. And, on application to Messrs. Harrison and Sons, 46, St. Martin’s Lane, these will be bound in a cloth case for 2s. 6d., or the cloth cases for binding may be purchased, price 1s. 6d. ! - Demy 8vo, Cloth, 12s. 6d., with Plates and Woodcuts, 427 pp. THE RECORD OF THE ROYAL. SOCIETY. SECOND EDITION, 1901. Containing an account of the Foundation and Early History of the Society, its Benefactors, Trust Funds, &c., with Chronological and Alphabetical LISTS OF THE FELLOWS FROM THE FOUNDATION. Just Published. Royal 4to. 307 pages and 27 Plates. Price 20s. | . To Fellows of the Royal Society, 15s. REPORT TO THE GOVERNMENT OF CEYLON ON THE PEARL OYSTER FISHERIES OF THE GULF OF MANAAR. By W. A. HERDMAN, D.Sc, F.RS. | With Supplementary Reports upon the Marine Pole of Ceylon by other Naturalists. PART i 1903. Published by the Royal Society and sold by HARRISON AND SONS, ST. MARTIN'S LANE. Sold by Harrison and Sons, St. Martin’s Lane, PHILOSOPHICAL TRANSACTIONS FoR 1903—4. Series A. Containing Papers of a Mathematical or Physical Character. Vol. 201 (1903). With 3 Plates. Price £1 16s. 6d. Vol. 202 (1903). With 18 Plates. Price £1 13s. Od. Szrizs B. Containing Papers of a Biological Character. Vol. 195. With 16 Plates. Price £1 7s. Od. Vol. 196. With 23 Plates. Price £1 16s. Od. RECENT PAPERS IN THE ‘PHILOSOPHICAL TRANSACTIONS.’ Published by Duzav & Co., 37, Soho Square. A. 356. An Enquiry into the Variation of Angles observed in Crystals; especially of Potassium-alum and Ammonium-alum. By Henry A. Miers, D.Se., M.A., F.RS., V.P.G.S8., V.P.C.S., Waynflete Professor of Mineralogy in the University of Oxford. Price 3s. 6d. A. 357. Radiation in the Solar System:. its effect on Temperature and _ its Pressure on Small Bodies. By J. H. Poynting, Sc.D., F.R.S., Pro- fessor of Physics in the University of Birmingham, Price 1s. A. 358. Title, Contents, Index, &c., Vol. 202. Price 6d. A . 359. On the Propagation of Tremors over the Surface of an Elastic Solid. By Horacr Lamp, F.R.S. Price 2s. A. 360. On the Structure of Gold-Leaf, and the Absorption Spectrum of Gold. / : By J. W. Matter, F.R.S., Professor of Chemistry in the University of Virginia. ~ Price 1s. A. 361. Mathematical. Contributions to the Theory of Evolution.—XIE. On a Generalised Theory of Alternative Inheritance, with Special Reference to Mendel’s Laws. By Kart PrEarson, F.R.S. Price 1s. 6d, A. 362. On the Acoustic Shadow of aSphere. By Lorp Ray1zieu, O.M., F.R.S. With an Appendix, giving the Values of Legendre’s Functions from > Po to Py) at Intervals of Five Degrees. By Professor A. Lopez. Price 1s. A. 363. On the Integrals of the Squares of LEllipsoidal Surface Harmonic Functions. By G. H. Darwin, F.R.S., Plumian Professor and Fellow of Trinity College, in the University of Cambridge. Price 1s. B. 219. On Reptilian Remains from the Trias of Elgin. By G. A Bovutenezr, E.R.S. Price 3s. 6d. B. 220. Studies in the Morphology of Spore-producing Members. No. V.— General Comparisons and Conclusion. By F. O. Bowzz, Sc.D., F.B.S., Regius Professor of Botany at the University of Glasgow. Price 2s. 6d. B. 221. A Method for the Investigation of Fossils by Serial Sections. By W. J. Sotzas, D.Sc., LL.D., F.R.S., Professor of Geology and Paleontology - in the University of Oxford. Price 1s. B. 222. An Account of the Devonian Fish, Paleospondylus Gunni, TRAQUAIR. By W. J. Sorzas, D.Sc., LL.D., E.RS., Professor of Geology and Paleontology in the University of Oxford, and Iezrna B. J. Sorzas, B.Sc., Lecturer on Zoology in Newnham College, Cambridge. Price 2s, B. 223. The Colour-physiology of Higher Crustacea. By FrEpERIcK Kzrstp, M.A., University College, Reading, and F. W. GaMBLz, D.Sc., Owens College, Manchester.. Price 12s. 6d. B. 224, Title, Contents, Index, &e. Vol. 196, Price 6d. B. 225. On the Fructification of Neuropteris heterophylla, Brongniart. By ROBERT Kipston, F.R.S8., L. and E., F.G.S._ Price 1s. Just Published. Royal 4to. 428 pages, with numerous Plates and Diagrams, and Separate Portfolio of Geological Maps and Sections. | THE ATOLL OF FUNAFOUTI, BORINGS INTO A CORAL REEF AND THE RESULTS. Price 30s. To Fellows of the Royal Society, 20s. Published by the RoyaL Society, and sold by HARRISON AND SONS, St. Martin’s Lane. Just Published. Demy 8vo. REPORTS OF THE SLEEPING SICKNESS COMMISSION. No. -I.—Containing Communications 1 and 2. By Dr. CASTELLANI, Cou. Bruce, and Dr. NABARRO. Price 7s. 6d. No. I[.—Containing Communications 3—5. By Drs. Curisry, CASTELLANI, and Low. Price 7s. 6d, if No. II1.—Containing Communications 6 and 7. By Dr. Cristy and Mr. THEOBALD. Price 6s. No. IV.—Containing Communication 8. By Cou. Brucs, Dr. NABARRO, and Capt. GREIG. Price 6s. 6d. REPORTS TO THE MALARIA COMMITTEH OF THE > ROYAL SOCIETY. First Series, 1900. Demy 8vo., pp. 76. Price 2s. 6d. Second ,, 1900. i. eS 22) 3 See Third. 1900: ‘ * Arpt he 2s. Fourth ,, 1901. yp A 20. ys OS ae arth: e OOr si ts 79. . .2siane Sixth ,, 1902. : a ge a Seventh ,, 1902 Bris Alt as 52. a 3s. Kighth ,, 1903. a ss The os 2s. Demy 8vo., pp. 160. Price 10s. REPORTS TO THE EVOLUTION COMMITTEE OF THE ROYAL SOCIETY. Rerort I.—Experiments undertaken by W. Barzson, E.R. S., and Miss E. R. SAUNDERS. Published and sold by HARRISON AND SONS, 45, 46, & 47, ST. MARTIN’S LANE, W.C., R. FRIEDLANDER & SOHN, BERLIN, AND ALL BOOKSELLERS. PROCEEDINGS OF DEE ROYAL SOCLETY. VOL. LXXIIL. No. 493. CONTENTS. PAGE On a Criterion which may serve to Test various Theories of Inheritance. By Kart Prarson, F.R.8., University College, London : ‘ ; - 262 Some Uses of Cylindrical Lens-Systems, including Rotation of Images. By Grores J. Burcu, M.A., D.Sc. Oxon, F.R.S., Lecturer in ae University College, iedln : : ‘ 2 : : : 281 On the Effect of a Magnetic Field on the Rate of Subsidence of Torsional Oscillations in Wires of Nickel and Iron, and the Changes Produced by Drawing and Annealing. By Professor ANDREW GRay, F.R.S., and ALEXANDER Woop, B.Sc. : d : : F : . 286 A Radial Area-Scale. By R. W. K. Epwarps, M.A. Communicated 2 Professor A. G. GREENHILL, E.RS. : : 5 : : A 292 On the Compressibility of Solids. By J. Y. Bucnanay, F.RS. . : . 296 Price One Shilling and Sixpence. May 7, 1904. f { NOTICE TO AUTHORS. La att Authors of Papers are urgently requested to send in all Drawings, Diagrams, or other Illustrations in a state suitable for direct photographic reproduction. They should be drawn on a large scale, in Indian ink, on a smooth white surface. Where illustrations are numerous, much time would be saved if authors would indicate in advance those considered essen- tial to the understanding of their papers. NOTICES TO FELLOWS OF THE ROYAL SOCIETY. The Council have directed that the Minutes of the Meetings of the Society shall be sent out as an inset in the ‘ Proceedings,’ separately paged, and shall afterwards — be republished in the ‘ Year-Book.’ The ‘ Proceedings ’ are sent by post to every Fellow of the Society who resides within the limits of the Postal Union. The ‘Philosophical Transactions’ are delivered only to those Fellows who call for them, or who send a written applica- tion to the Assistant Secretary. Such an application may, if so desired, be filed as a standing order. The separate Papers forming the ‘ Philosophical Transactions’ will be sent post free, as issued, to those Fellows who prefer to have them in that form. And, on ‘application to Wisesrsl Harrison and Sons, 46, St. Martin’s Lane, these will be hound , in a cloth case for 2s. 6d., or the cloth cases for binding may be purchased, price ls. 6d. Demy 8vo, Cloth, 12s. 6d., with Plates and Woodcuts, 427 pp. THE RECORD OF THE ROYAL SOCIETY. SECOND EDITION, 1901. Containing an account of the Foundation and Early History of the Society, its Benefactors, Trust Funds, &c., with Chronological and Alphabetical — LISTS OF THE FELLOWS FROM THE FOUNDATION. ae ust: Published. Royal 4to. 307 pages and 27 Plates, Price 20s. To Fellows of the Royal Society, 15s. hs REPORT TO THE GOVERNMENT OF CEYLON ON THE a PEARL OYSTER FISHERIES OF THE GULF OF MANAAR. At) By W. A. HERDMAN, D.Sc, F.R.S. With Supplementary Reports upon the Marine Biology of Ceylon: by other Naturalists. Part I. 1903. Published by the Royal Society and sold by HARRISON AND SONS, ST. MARTIN'S LANE. Sold by Harrison and Song, St. Martin’s Lane. PHILOSOPHICAL TRANSACTIONS FoR 1903—4. Series A. Containing Papers of a Mathematical or Physical Character. Vol. 201 (1903). With 3 Plates. Price £1 16s. 6d. Vol. 202 (1903). With 13 iriates. Price £1 13s. Od. Surres 8. Containing Papers of a Biological Character. Vol. 195. With 16 Piates. Price £1 7s. Od. Vol. 196. ‘With 23 Plates. Price £1 16s. Od. RECENT PAPERS IN THE ‘PHILOSOPHICAL TRANSACTIONS,’ Published by Dunav & Co., 37, Soho Square. A. 357. Radiation in the Solar System: its effect on Temperature and its Pressure on Small Bodies. By J. H. Poynrine, Se.D., F.R.8., Pro- fessor of Physics in the University of Birmingham. Price 1s. . 858. ‘Title, Contents, Index, &c., Vol. 202. Price 6d. . 859. On the Propagation of Tremors over the Surface of an Elastic Solid. By Horacr Lamp, F.R.S. Price 2s. A. 360. On the Structure of Gold-Leaf, and the Absorption Spectrum of Gold. 2 By J. W. Mater, F.R.S., Professor of Chemistry in the University of Virginia. Price Is. A. 361. Mathematical Contributions to the Theory of Evolution.—XII. On a Generalised Theory of Alternative Inheritance, with Special Reference to Mendel’s Laws. By Karu Pearson, F.R.S. Price 1s. 6d. A. 362. Onthe Acoustic Shadow of aSphere. By Lorp Rayurieu, O.M., B.R.S. With an Appendix, giving the Values of Legendre’s Functions from Po to Py) at Intervals of Five Degrees. By Professor A. Lope. Price ls. A. 363. On the Integrals of the Squares of Ellipsoidal Surface Harmonic Functions. By G. H. Darwin, F.R.S., Plumian Professor and Fellow of Trinity College, in the University of Cambridge. Price 1s. B. 219. On Reptilian Remains from the Trias of Elgin. By G. A BouLEencsr, F.R.S. Price 3s. 6d. - B. 220. Studies in the Morphology of Spore-producing Members. N 0. vo General Comparisons and Conclusion. By Ff. O. BowER, Sc.D., F.R.S., Regius Professor of Botany at the University of Glasgow. Price 2s. 6d. B. 221. A Method for the Investigation of Fossils by Serial Sections. By W. J. Sontzas, D.Se., LL.D., ¥.R.S., Professor of Geology and Paizontology in the University of Oxford. Price ls. B. 222. An Account of the Devonian Fish, Paleospondylus Gunni, TRAQUAIR. By W. J. Sortas, D.Sc., LL.D., F.R.S., Professor of Geology and Paleontology in the University of Oxford, and Ia¢zrna B. J. Sonuas, B.8ec., Lecturer on Zoology in Newnham College, Cambridge. Price 2s. B. 223. The Colour-physiology of Higher Crustacea. By FrupERIcK KEEBLE, M.A., University College, Reading, and F. W. GamBiez, D.Sc., Owens College, Manchester. Price 12s. 6d. . 224, Title, Contents, Index, &c. Vol. 196. Price 6d. B B. 225. On the Fructification of Neuropteris heterophylla, Brongniart. By RoserRT Kipston, F.R.S., L. and H., F.G.8. Price 1s. B. 226. On the Origin of Parasitism in Fungi. By Grorar Masszz, F.LS., Principal Assistant, Herbarium, Royal Botanic Gardens, Kew. Price 1s. B. 227. On the “Islets of Langerhans” in the Pancreas. By H. H. Dats, M.A, B.C., George Henry Lewes Student, University College, London. Price 1s. 6d. Just Published. Royal 4to. 428 pages, with numerous Plates and Diagrams, and Separate Portfolio of Geological Maps and Sections. THE ATOLL OF FUNAFOTL BORINGS INTO A CORAL REEF AND THE RESULTS. Price 30s. To Fellows of the Royal Society, 20s. Published by the RoyAL SocrErTy, and sold by HARRISON AND Sons, St. Martin’s Lane. Just Published. Demy Syo. REPORTS OF THE, SLEEPING SICKNESS COMMISSION. No. I.—Containing Communications 1 and 2. By Dr. CASTELLANI, Cou. BrucE, and Dr. NABARRO. Price 7s. 6d. No. II.—Containing Communications 3—5. By Drs. Curisty, CASTELLANI, and Low. Price 7s. 6d, ’ No. III.—Containing Communications 6 and 7. By Dr. Curisty and Mr. THEOBALD. Price 6s. No. [V.—Containing Communication 8. By Cort. Bruce, Dr. NABARRO, and Capt. GREIG. Price 6s. 6d. REPORTS TO THE MALARIA COMMITTEE OF THE ROYAL SOCIETY. First Series, 1900. Demy 8vo., pp. i 4 Price 2s. 6d. Second ,, 1900. be i Z: 53 . ee Third: <4; 1900: as ia ie i 2s. Fourth }. ,.4 190): Mt e 20. 7. oS oe Piih 4) POOL as id 79. Pe ee Sixth ,, 1902. B an 23. i Is. Seventh ,, 1902 i 3 52. - 35. Highth ,, 1903. Ms a TE BS 2s. Demy 8vo., pp. 160. Price 10s. REPORTS TO THE EVOLUTION COMMITTEE OF THE ROYAL SOCIETY. RELoORT dia MEE renents undertaken by W. Barzson, F.R.S., and Miss E. R. SAUNDEBS. Published and sold by HARRISON AND SONS, 45, 46, & 47, ST. MARTIN’S LANE. W.C., R. FRIEDLANDER & SOHN, BERLIN, AND ALL BOOKSELLERS. PROCEEDINGS OF THE ROYAL ‘SOCIETY. VOL. LXXIII. No. / CONTENTS. CRroonian LectuRE.—The Chemical Regulation of the Secretory Process. By W. M. Bavtuiss, F.R.S., and EH. H. Startine, F.R.S. Experiments on a Method of Preventing Death from Snake Bite, capable of Common and Easy Practical Application. By Sir Lavprr Bruyron, PAGE Me att > 10 M.D., F.R.S., Sir Joseps Fayrer, Bart., K.C.S.1., F.R.S., and Lzonarp Rogers, M.D., B.S., etc., Indian Medicai Service The Effects of Changes of Temperature on the Modulus of Torsional Rigidity of Metal Wires. By Franx Horton, D.Sc., B.A., St. John’s College, Cambridge; 1851 Exhibition Research Scholar of the University of Birmingham. Communicated by Professor J. J. THomson, F.RS. _ (Abstract) The Sparking Distance between Electrically Charged Surfaces.—Preliminary Note. By P. HE. SHaw, B.A., D.Sc. Communicated by Professor J. H. Poyntine, F.R.S. . ! d : 2 : i i ; Further Note on some Additional Points in Connection with Chloroformed . Calf Vaccine. By Anan B. Green, M.A., M.D. Nigala Communi- cated by Dr. W. H. Powzr, C.B., F.R.S. : 5 Further Experiments on the Production of Helium ioe Rudin: By Sir Witiiam Ramsay, K.C.B., F.R.S., and FRepeRick Soppy, M.A.. Corrigenda in Mr. W. Shanks’s Tables “On the Number of Figures in the Reciprocal of a Prime.” By Lieut.-Colonel Attan Cunninenay, R.E., Fellow of King’s College, London. Communicated by the Secretaries A Research into the Heat Regulation of the Body by an Investigation of Death Temperatures. By Eprep M. Cormer, M.A., M.B., B.C. (Cantab.), ¥.R.C.S., B.Sc. (Lond.), Surgeon to Out-patients, St. Thomas’s Hospital, and Assistant-Surgeon to the Hospital for Sick Children, Great Ormond Street, Hrasmus Wilson Lecturer, Royal College of Surgeons, and Jams E. H. Sawyer, M.A., M.D. (Oxon.), M.R.C.P., Honorary Anesthetist to the Ear and Throat Hospital, Birmingham, lately House Physician to St. Thomas’s Hospital. Communicated by Professor J. N. Lanauey, F.B.S. A Note on the Action of Radium on Micro-organisms. By Anan B. GrRusn, M.A., M.D. (Cantab.) Communicated a Sir Micuarn Fosrsr, K.C.B., E.R. S. (Plate 11) . i : : : : For continuation of Contents see second page Price Three Shillings k os May 28, 1904. Ny eal 323 334, 337 342 346 359 361 375 CONTENTS (continued). ? . PAGE On certain Physical and Chemical Properties of Solutions of Chloroform in Water, Saline, Serum, and Hemoglobin. A Contribution to the Chemistry of Anesthesia.—(Preliminary Communication). By BENJAMIN Moore, M.A., D.Sc., Johnston Professor of Bio-Chemistry, University of Liverpool, and Hrrzert E. Roar, M.B., Toronto, Johnston Colonial Fellow, University of Liverpool. Communicated by Professor C. S. SHERRINGTON, E.RS. . Z ‘ N i , : : . 882 Minutes of Meetings of April 28, 1904, and May 5, 1904 ; faites SPRY, anne NOTICES TO FELLOWS OF THE ROYAL SOCIETY. The Council have directed that the Minutes of the Meetings of the Society shall be sent out as an inset in the ‘ Proceedings,’ separately paged, and shall afterwards be republished in the ‘ Year-Book.’ The ‘ Proceedings’ are sent by post to every Fellow of the Society who resides within the limits of the Postal Union. The ‘ Philosophical Transactions’ are delivered only to those Fellows who call for them, or who send a written applica- tion to the Assistant Secretary. Such an application may, if so desired, be filed as a standing order. The separate Papers forming the ‘ Philosophical Transactions’ will be sent post. free, as issued, to those Fellows who prefer to have them in that form. And, on application to Messrs. Harrison and Sons, 46, St. Martin’s Lane, these will be bound in a cloth case for 2s. 6d., or the cloth cases for binding may be purchased, price is. 6d. Demy 8vo, Cloth, 12s. 6d., with Plates and Woodcuts, 427 pp. THE RECORD OF THE ROYAL SOCIETY. SECOND EDITION, 1901. Containing an account of the; Foundation and Early History of the Society, its Benefactors, Trust Funds, &¢., with Chronological and Alphabetical LISTS OF THE FELLOWS FROM THE FOUNDATION. ber ust: Published. Royal 4to. 307 pages and 27 Plates. Price 20s. To Fellows of the Royal Society, 15s. REPORT TO THE GOVERNMENT OF CEYLON ON THE PEARL OYSTER FISHERIES OF THE GULF OF MANAAR. By W. A. Herpman, D.Sc., FBS. With Supplementary Reports upon the Marine Biology of Ceylon by other Naturalists. ParTil, 1903. / Published by the Royal Society and sold by HARRISON AND SONS, ST. MARTIN’S LANE. . Sold by Harrison aud Sous, St. Martin’s Lane. PHILOSOPHICAL TRANSACTIONS | : FOR 1903-4. SERizes A. Containing Papers of a Mathematical or Physical Character. Vol. 201 (1903). With 3 Plates. Price £1 16s. 6d. Vol. 202 (1903). With 13 Plates. Price £1 13s. Od. Srrizs B. Containing Papers of a Biological Character. Vol. 195. With 16 Plates. Price £1 7s. Od. Vol. 196. With 23 Plates. Price £1 16s. Od. RECENT PAPHRS IN THE ‘PHILOSOPHICAL TRANSACTIONS.’ Published by Duzav & Co., 37, Soho Square. A. 357. Radiation in the Solar System: its effect on Temperature and _ its Pressure on Small Bodies. By J. H. Poynrine, Sc.D., F.R.S8., Pro- fessor of Physics in the University of Birmingham, Price 1s. . 358. Title, Contents, Index, &c., Vol. 202, Price 6d. . 359. On the Propagation of Hees over the Surface of an Elastic Solid. By Horacre Lams, 2.R.8. Price 2s. A. 360. On the Structure of Gold-Leaf, and the Absorption Spectrum of Gold. By J. W. Matzzt, F.R.S., Professor of Chemistry in the University of Virginia, Price Is. A. 361. Mathematical Contributions to the Theory of Evolution.—XII. On a Generalised Theory of Alternative Inheritance, with Special Reference to Mendel’s Laws. By Karu PEarson, F.R.S. Price Is. 6d. A. 362. Onthe Acoustic Shadow of aSphere. By Lorp RayirieH, O.M., F.R.S. With an Appendix, giving the Values of Legendre’s Functions from Po to Pop at Intervals of Five Degrees. By Professor A. Lopexr. Prices. A.i,363. On the Integrals of the Squares of Ellipsoidal Surface Harmonic Functions. By G. H. Darwin, F.R.S., Plumian Professor and Fellow of Trinity College, in the University of Cambridge. Price 1s. B. 222. An Account of the Devonian Fish, Paleospondylus Gunni, TRAQUAIR. By W. J. Sotzas, D.8c., LL.D., F.R.S., Professor of Geology and Paleontology in the University of Oxford, and IczERNA B. J. Sonzas, B.Sc., Lecturer on Zoology in Newnham College, Cambridge. Price Qs. B. 223. The Colour-physiology of Higher Crustacea. By FREDERICK KEEBLE, M.A., University College, Reading, and F. W. GamBLz, D.Sc., Owens College, Manchester. Price 12s. 6d. . 224, Title, Contents, Index, &. Vol. 196. Price 6d. B B. 225. On the Fructification of Neuropteris heterophylla, Brongniart. By Ropert Kipston, F.R.S., L. and W., F.G.S. Price 1s. B. 226. On the Origin of Palate in Hates By GrorGe Masszz, F.L.S., Principal Assistant, Herbarium, Royal Botanic Gardens, Kew. Price 1s. B. 227. On the “Islets of Langerhans” in the Pancreas. By H. H. Datz, M.A., B.C., George Henry Lewes Student, University College, London. Price 1s. 6d. B. 228. Experimental Researches on Vegetable Assimilation and Respiration. IiI.—On the Effect of Temperature on Carbon-Dioxide Assimilation. By GapRiztte L. C. Marruani, Fellow of Newnham College, Cam- bridge. Price 3s. 6d. B. 229. Cultural Experiments with “Biologic Forms”’ of the Hrysiphacee. By Hryest 8. Satmon, F.L.8. Price Ls. Just Published. Royal 4to. 428 pages, with numerous Plates and Diagrams, and Separate Portfolio of Geological Maps and Sections. pe ae THE ATOLL OF FUNAFUTI, BORINGS INTO A CORAL REEF AND THE RESULTS. Price 30s. ‘To Fellows of the Royal Society, 20s. Published by the Roya Society, and sold by HARRISON AND SONS, St. Martin’s Lane. Just Published. Demy 8vo. REPORTS OF THE SLEEPING SICKNESS COMMISSION. No. [—Containing Communications 1 and 2. By Dr. CASTELLANI, : Cou. BRUCE, and Dr. NABARRO. Price 7s. 6d. No. I.—Containing Communications 3—5. By Drs. CHRISTY, CASTELLANI, and Low. Price 7s. 6d, ; No. IIJ.—Containing Communications 6 and 7. By Dr. Curisty and Mr. THEOBALD. Price 6s. No. [V.—Containing Communication 8. By Cox. Bruce, Dr. NABARRO, and Capt. GREIG. Price 6s. 6d. REPORTS TO THE MALARIA COMMITTEE OF THE ROYAL SOCIETY. First Series, 1900. Demy 8vo., pp. ; i Price 2s. 6d. Second ,, 1900. 4 3 2: feels. Gd Third 4°53. ye: . i - al Ds. Fourth ,, 1901. ie a 20. | eSonoae Purth..2,.). 190% s iM 79. iO eee Sixth ,, 1902. : 83) Seventh ,, 1902 ip ds SPA Met 3s. Eighth ,, 1903. i Sk Tes ea er! Demy 8vo., pp. 160.. Price 10s. REPORTS TO THE EVOLUTION COMMITTEE OF THE ROYAL SOCIETY. Reronr?1.—Experiments undertaken by W. Bateson, F.R.S., and Miss ER SAUNDERS. Published and sold by 1IARRISON AND SONS, 45, 46, & 47, ST. MARTIN’S LANE, W.C.. R. FRIEDLANDER & SOHN, BERLIN, AND ALL BOOKSELLERS. ¢ PROCEEDINGS OF Wh ROVAL SOCIETY. VOL. LXXIII. No. 495. CONTENTS. PAGE On the Changes of Thermoelectric Power produced by Magnetisation and their Relation to Magnetic Strains. By SuHrtForD Brpwett1, M.A., Ne.D., F.R.S. . : : : : : : é : : s . 413 Experimental Determinations for Saturated Solutions. By the Earn or ~ BERKELEY. Communicated by F! H. Neviniz, F.R.S. (Abstract) . 435 A Method of Measuring directly High Osmotic Pressures. By the Hart or BERKELEY and EK. G. J. Harrinry. Communicated by W. C. D. Wuetuaw, E.R.S. 436 Colours in Metal Glasses and in Metallic Films. By J.C. MaxwEtt Garner, B.A., Trinity College, Cambridge. Communicated by Professor Larmor, Sec. B.S. (Abstract) : : : ( : é : é k . 443 The General Theory of Integration. By W. H. Youne, Se.D., St. Peter’s College, Cambridge. Communicated by Dr. BH. W. Hopson, F.RS. _ (Abstract) : : : d : 4 ; : : : t . 445 On the Liquefied Hydrides of Phosphorus, Sulphur, and the Halogens as Conducting Solvents.—Part I. By D. MocIntosH and B. D. STrets. Communicated by Sir Winn1am Ramsay, K.C.B., F.R.S8. . re . 450 On the Liquefied Hydrides of Phosphorus, Sulphur, and the Halogens as Conducting Solvents—Part IJ. By EH. H. Arcuipatp and D. McIntosH. Communicated by Sir Wint1amM Ramsay, K.C.B., F.R.S.. 454 Note on the Lymphatic Glands in Sleeping Sickness. By Captain E. D. W. Greie, I.M.S., and Lieutenant A. C. H. Gray, R.A.M.C. Communi- cated by Colonel Bruce, F.R.S., at the desire of the Sleeping Sickness Commission. . : J d é : : i : f : . 455 ' The Behaviour of the Short-Period Atmospheric Pressure Variation over the Earth’s Surface. By Sir Norman Lockyer, K.C.B., LL.D., F.R.S., and Wiitiiam J. S. Lockyer, Chief Assistant, Solar Physics Observatory, M.A. (Camb.), Ph.D. (Gétt.), F.R.A.S. (Plates 12and18) . 3 - 457 ‘The Spectrum of the Radium Emanation. By Sir Wintiam Ramsay, K.C.B., F.R.S., and Professor J. Norman Coxiis, F.R.S. ea tao) | qengenian lngtjp oe. yall “hy, : It 6 1904. ice Library For continuation of Contents see second page ofj Price Three Shillings. JUNE 22, 1904. CONTENTS (continued). Notes on the Statolith Theory of Geotropism. I. Expeiiments on the Effects | of Centrifugal Force. II. The Behaviour of Tertiary Roots. By Francis Darwin, F. RS., and D. F. M. Prertz. On the Electric Effect of Rotating a Dielectric in a Magnetic Field. By Harorp A. Witson, M.A., D.Sc., Fellow of Trinity College, Cambridge. Communicated by Professor J. J. Taomsoy, F.R.S. (Abstract) BaxkERiIan Lecturzt.—The Succession of Changes in Radio-active Bodies. By E. RurHerrorp, F.R.S., Macdonald Professor of ae McGill University, Montreal. (Abs. On the Electric Equilibrium of the Sun. By SvanTE eee Cont municated by Sir Witt1am Hueetns, Pres. B.S. Minutes of Meetings of May 19, 1904, and June 2, 1904 auc NOTICES TO FELLOWS OF THE ROYAL SOCIETY. The Council have directed that the Minutes of the Meetings of the Society shall be sent out as an inset in the ‘ Proceedings,’ separately paged, and shall afterwards be republished in the ‘ Year-Book,’ - The ‘ Proceedings’ are sent by post to every Fellow of the Society who restdes ea the limits of the Postal Union. The ‘ Philosophical Transactions’ are delivered only to those Fellows who call for them, or who send a written applica- tion to the Assistant Secretary. Such an application may, if so desired, be filed as @ standing order. The separate Papers forming the ‘ Philosophical Transactions’ will be sent post free, as issued, to those Fellows who prefer to have them in that form. And, on application to Messrs. Harrison and Sons, 46, St. Martin’s Lane, these will be bonne in a cloth case for 2s. 6d., or the cloth cases for binding may be purchased, price 1s. 6d. Demy 8vo, Cloth, 12s. 6d., with Plates and Woodcuts, 427 pp. THE RECORD OF THE ROYAL oe SECOND EDITION, 1901. Containing an account of the Foundation and Harly History of the Society, its Benefactors, Trust Funds, &c., with Chronological and Alphabetical LISTS OF THE FELLOWS FROM THE FOUNDATION. ‘Just Published. Royal 4to. 307 pages and 27 Plates. Price 20s. To Fellows of the Royal Society, 15s. : REPORT TO THE GOVERNMENT OF CEYLON ON THE PEARL OYSTER FISHERIES OF THE GULF OF MANAAR. By W. A. HerpMan, D.Se., F.RS. - With Supplementary Reports upon the Marine Biology of Ceylon by other Naturalists. } PART, 151903. Published by the Royal Society and sold by HARRISON AND SONS, ST. MARTIN’S LANE. Sold by Harrison and Sons, St. Martin’s Lane. . PHILOSOPHICAL TRANSACTIONS FOR 1903-4. Srrizes A. Containing Papers of a Mathematical or Physical Character, Vol. 201 (1903). With 3 Plates. Price £1 16s. 6d. Vol. 202 (1903). With 13 Plates. Price £1 13s. Od. Serres B. Containing Papers of a Biological Character. Vol. 195. With.16 Plates. Price £1 7s. Od. Vol. 196. With 23 Plates. Price £1 16s. Od. RECENT PAPERS IN THE ‘PHILOSOPHICAL TRANSACTIONS,’ Published by Duzat & Co., 37, Soho Square. . 358. Title, Contents, Index, &., Vol. 202. Price 6d. 359. On the Propagation of Tremors over the Surface of an Elastic Solid. By Horacr Lams, F.R.S. Price 2s. A. 360. On the Structure of Gold-Leaf, and the Absorption Spectrum of Gold. By J. W. Matzet, F.R.S., Professor of Chemistry in the University of Virginia. Price 1s. A. 361. Mathematical Contributions to the Theory of Evolution.—XII. On a Generalised Theory of Alternative Inheritance, with Special Reference to Mendel’s Laws. By Kart PEarson, F.R.S. Price 1s. 6d. A. 362. Onthe Acoustic Shadow of aSphere. By Logp Raytzteu, O.M., F.R.S. With an Appendix, giving the Values of Legendre’s Functions from i Po to Pop at Intervals of Five Degrees. By Professor A.Lope@x. Pricels. A, 368. On the Integrals of the Squares of Ellipsoidal Surface Harmonie Functions. By G. H. Darwin, F.R.S., Plumian Professor and Fellow of Trinity College, in the University ‘of Cambridge. Price 1s. A. 364. The Specific Heats of Metals and the Relation of Specific Heat to Atomic - Weight.—Part III. By W. A. Tinpen, D.Se., F.R.S., Professor of Chemistry in the Royal College of Science, London. Price 1s. > > B. 222. An Account of the Devonian Fish, Paleospondylus Gunni, TRAQUAIR By W. J. Soxzas, D.Sec., LL.D., F.R.S., Professor of Geology and Paleontology in the University of Oxford, and IczrNna B. J. Sorzas, B.Se., Lecturer on Zoology in Newnham College, Cambridge. Price 2s. » B. 223. The Colour-physiology of Higher Crustacea. By FrEpERick Kruste, M.A., University College, Reading, and F. W. Gamsuz, D.Sc., Owens College, Manchester. Price 12s. 6d. B. 224. Title, Contents, Index, &. Vol. 196. Price 6d. oe B. 225. On the Fructification of Neuropteris heterophylla, Brongniart. By Ropert Kinston, F.R.S., L. and E., F.G.S. Price 1s. B.*.226. On the Origin of Pameitiem in eee By GEorcr Masses, F.L.S., Principal Assistant, Herbarium, Royal Botanic Gardens, Kew. Price 1s. B.,227. On the “Islets of Langerhans” in the Pancreas. By H. H. Datz, M.A., B.C., George Henry Lewes Student, University College, London. Price 1s. 6d. B. 228. Experimental Researches on Vegetable Assimilation and Respiration. IiT.—On the Effect of ‘Temperature on Carbon-Dioxide Assimilation. By Gapgietiz L. C. Marruarr, Fellow of Newnham College, Cam- bridge. Price 3s. 6d. B. 229. Cultural Experiments with “Biologic Forms” of the Erysiphacee. By. Ernest 8. Satmon, F.L.S. Price 1s. Just Published. Royal 4to. 428 pages, with numerous Plates and Diagrams, and Separate Portfolio of Geological Maps and Sections. — THE ATOLL OF FUNAFUTI, BORINGS INTO A CORAL REEF AND THE RESULTS. Price 30s. To Fellows of the Royal Society, 20s. Published by the Roya Socrety, and sold by | HARRISON AND SONS, St. Martin’s Lane. Just Published. Demy 8vo. REPORTS OF THE SLEEPING SICKNESS COMMISSION. No. I—Containing Communications 1 and 2. By Dr. CASTELLANI, Co. BRucE, and Dr. NapaRro. Price 7s. 6d. No. II.—Containing Communications 3—5. By Drs. Curisty, CASTELLANI, and Low. Price 7s. 6d. No. III.—Containing Communications 6 and 7. By Dr. Curisty and Mr. THEOBALD. Price 6s. No. [V.—Containing Communication 8. By Cou. Bruce, Dr. NABARRO, and CAPT. GREIG. Price 6s. 6d. REPORTS TO THE MALARIA COMMITTEE OF THE ROYAL SOCIETY. First Series, 1900. Demy 8vo., pp. Me Price 2s. 6d. Second ,, 1900. ip a 2. 53 id Sade Third: .35 1906: ne - 2 4s 2s. Fourth ,, 1901. 5 cs 20. i. Cos eanee Hitch) 3. ele Oe. us % 79. 1c.) Seen Sixth 707 1909: # i ys: i Is. Seventh ,, 1902 i ‘ 52. i 35. Highth ,, 1903. a 4 hath 4 2s. Demy 8vo., pp. 160. Price 10s. REPORTS TO THE EVOLUTION COMMITTEE Le THE ROYAL SOCIETY. Revort I.—Experiments undertaken by W. Barzson, F.R.S., and Miss EH. RB. SAUNDERS. Published and sold by HARRISON AND SONS, 45, 46, & 47, ST. MARTIN’S LANE, W.C., R. FRIEDLANDER & SOHN, BERLIN, AND ALL BOOKSELLERS, PROCEEDINGS OF Pee OVAL SOGELE TY, VOL. LXXIII. | No. 496. CONTENTS. PAGE Studies on Enzyme Action. II.—The Rate of the Change, conditioned by Sucroclastic Enzymes, and its Bearing on the Law of Mass Action. By EDWARD FRANKLAND ARMSTRONG, Ph.D., Salters’ Company’s Research Fellow, Chemical Department, City and Guilds of London Institute, Central Technical College. Communicated by Professor H. EH. Arm- STRONG, F'.R.S. ee ot : é : A d : 4 : . 500 Studies on Enzyme Action. III.—The Influence of the Products of Change on the Rate of Change conditioned by Sucroclastic Enzymes. By EDWARD FRANKLAND ARMSTRONG, Ph.D., Salters’ Company’s Research Fellow, Chemical Department, City and Guilds of London Institute, Central Technical College. Communicated by Professor H. EH. Arm- sTRONG, F.R.S. BD ea ue 3 : : : Stan é oat) Paley Studies on Hnzyme Action. [V.—The Sucroclastic Action of Acids as con- trasted with that of Enzymes. By EDwarpD FRANKLAND ARMSTRONG, Ph.D., Salters’ Company’s Research Fellow, and RoBERT JOHN CALD- WELL, B.Sc., Clothworkers Scholar, Chemical Department, City and Guilds of London Institute, Central Technical College. Communicated by Professor H. H. Armstrone, F.R.S. . : 3 : ‘ : - 526 Euzyme Action as Bearing on the Validity of the Ionic-Dissociation Hypothesis and on the Phenomena of Vital Con By Henry. HE. PREMSTHONG ERD ERS. sy. SUM ng ry enna eat Oe eiindex: F e ‘ A s ‘ ’ : : : : : . 943 Title, Contents, &c. | Minutes of Meetings of June 9, 1904, and June 16, 1904 : 8 : ets oc eT ae eit ae MpIaRe ECHR. Wann re ansonian inStity, Uo, Office Libraty: Price One Shilling and Sixpence. JULY 7, 1904. ; NOTICE TO AUTHORS. Authors of Papers are urgently requested to send in all Drawings, Diagrams, or other Illustrations in a state suitable for direct photographic reproduction. They should be drawn on a large scale, in Indian ink, on a smooth white surface. Where illustrations are numerous, much time would be saved — if authors would indicate in advance those considered essen- tial to the understanding of their papers. NOTICES TO FELLOWS OF THE ROYAL SOCIETY. The Council have directed that the Minutes of the Meetings of the Society shall be sent out as an inset in the ‘ Proceedings,’ separately paged, and shall afterwards be republished in the ‘ Year-Book.’ The ‘ Proceedings’ are sent by post to every Fellow of the Society who resides within the limits of the Postal Union. -The ‘ Philosophical Transactions’ are delivered only to those Fellows who call for them, or who send a written applica- tion to the Assistant Secretary. Such an application may, if so desired, be filed ‘as a standing order. The separate Papers forming the ‘ Philosophical Transactions’ will be sent post free, as issued, to those Fellows who prefer to have them in that form. And, on application’ to Messrs. Harrison and Sons, 46, St. Martin’s Lane, these will be bound in a cloth case for 2s. 6d., or the cloth cases for ee may be purchased, price ls. 6d. - Demy 8vo, Cloth, 12s. 6d., with Plates and Woodcuts, 427 pp. THE RECORD OF THE ROYAL SOCIETY. SECOND EDITION, 1901. Containing an account of the Foundation and Early History of the Society, its Benefactors, Trust Funds, &c., with Chronological and Alphabetical LisTS OF THE FELLOWS FROM THE FOUNDATION. Just Published. Royal 4to. 307 pages and 27 Plates. Price 20s. To Fellows of the Royal Society, 15s. REPORT TO THE GOVERNMENT OF CEYLON ON THE PEARL OYSTER FISHERIES OF THE GULF OF MANAAR. By W. A. Herpman, D.Sc., FBS. With Supplementary Reports upon the Marine Biology of Ceylon by other Naturalists. Part I. 1903. Published by the Royal Society and sold by HARRISON AND SONS, ST. MARTIN’S LANE. Sold by Harrison and Sons, St. Martin’s Lane. PHILOSOPHICAL TRANSACTIONS FOR 1903-4. Srrizes A. Containing Papers of a Mathematical or Physical Character. Vol. 201 (1903). With 3 Plates. Price £1 16s. 6d. Vol. 202 (1903). With 13 Plates. Price £1 13s. Od. Series B. Containing Papers of a Biological Character. ne Vol. 195. With 16 Plates. Price £1 7s. Od. Vol. 196. With 23 Plates. Price £1 16s. Od. RECENT PAPERS IN THE ‘PHILOSOPHICAL TRANSACTIONS.’ _ Published by Duzav & Co., 37, Soho Square. . 358. Title, Contents, Index, &c., Vol. 202. Price 6d/ . 359. On the Propagation of cane over the Surface of an Elastic Solid. By Horacr Lams, F.R.S. Price 2s. A. 360. On the Structure of Gold-Leaf, and the Absorption Spectrum of Gold. By J. W. Mattet, F.B.S., Professor of Chemistry in the University of Virginia. Price 1s. A. 361. Mathematical Contributions to the Theory of Evolution—XII. On a oe Generalised Theory of Alternative Inheritance, with Special Reference to Mendel’s Laws. By Karu Prarson, F.R.S. Price 1s. 6d. A. 362. On the Acoustic Shadow of a Sphere. By Lorp RayteicH, O.M., F.R.S. With an Appendix, giving the Values of Legendre’s Functions from Po to Poy at Intervals of Five Degrees. By Professor A. LopaE. Price ise: A. 363. On the Integrals of the Squares of Ellipsoidal Surface Harmonie Functions. By G. H. Darwin, F.R.S., Plumian Professor and Fellow of Trinity College, in the University ‘of Cambridge. Price ls. A. 364. The Specific Heats of Metals and the Relation of Specific Heat to Atomic Weight.—Part ITI. By W. A. TrtpEn, D.Sc., F.R.S., Professor of Chemistry in the Royal College of Science, London. Price 1s. B. 223. The Colour-physiology of Higher Crustacea. By FREDERICK KEEBLE, M.A., University College, Reading, and F. W. GAMBLE, D.Sc., Owens College, Manchester. Price 12s. 6d. . 224, ‘Title, Contents, Index, &c. Vol. 196. Price 6d. . 225. On the Fructification of Newropteris heterophylla, Brongniart. By RoBeRT Kinston, F.R.S., L. and H., F.G.S. Price Is. 226. On the Origin of Parasitism in Fungi. By Gzorax Massez, F.LS., Principal Assistant, Herbarium, Royal Botanic Gardens, Kew. Price 1s. 227. On the “Islets of Langerhans” in the Pancreas. By H. H. Dats, M.A., B.C., George Henry Lewes Student, University College, London. Price 1s. 6d. B. 228. Experimental Researches on Vegetable Assimilation and Respiration. III.—On the Effect of Temperature on Carbon-Dioxide Assimilation. By GABRIELLE L. C. Marruant, Fellow of Newnham College, Cam- bridge. Price 3s. 6d. - B. 229. Cultural Experiments with “Biologic Forms” of the Erysiphacee. By Ervest 8. Sanmon, F.L.8. Price 1s. Ww WwW we B. 230. The Physiological Action and Antidotes of Colubrine and Viperine Snake Venoms. By Lzonarp Rogers, M.D., B.S.,M.R.C.P., F.R.C.S., Indian Medical Service. Price 3s. 6d. Just Published. Royal 4to. 428 pages, with numerous Plates and Diagrams, and Separate Pere of Geological Maps and Seca THE ATOLL OF FUNAFUTI, BORINGS INTO A CORAL REEF AND THE RESULTS. Price 30s, To Fellows of the Royal Society, 20s. Published by the Roya Soctrety, and sold by HARRISON AND Sons, St. Martin’s Lane. REPORTS OF THE SLEEPING SICKNESS COMMISSION. No. [.—Containing Communications 1 and 2. By Dr. CASTELLANI, CoL. Bruce, and Dr. NABARRO. Just Published. Demy 8vo. No. IJ.—Containing Communications 3—5. CASTELLANI, and Low. Price 7s. 6d, No. Ili.—Containing Communications 6 and 7 and Mr. THEOBALD. Price 6s. No. [V.—Containing Communication 8. NABARRO, and CAPT. GREIG. REPORTS TO THE MALARIA COMMITTEE OF THE First Series, 1900. Demy 8vo., pp. 76. 1900. - Second ,, i i 22. Third: |; .3.4. 1900: i i 45. Fourth ,,° 1901: * a 20. RitGh i al 90a: u § 19. Sixth ,, 1902. y Re DS: Seventh ,, 1902 x . Sy Highth ,, 1903. Fe ss hk: REPORTS TO THE EVOLUTION COMMITTEE OF THE Rerort IJ.—Experiments undertaken by W. Bateson, F.R.S., and Miss E. BR. 3 | * HARRISON AND SONS, 45, 46, & 47, ST. MARTIN’S LANE, W.C., R. FRIEDLANDER & SOHN, BERLIN, ROYAL SOCIETY. By Dr. CHristy Price 7s. 6d. By Drs. Curisty, By Cot. Bruce, Dr. Price 6s. 6d. Price 2s. 6d. Demy 8vo., pp. 160. Price 10s. ROYAL SOCIETY. SAUNDERS. Published and sold by AND ALL BOOKSELLERS. ls. 6d. Ws 3 9088 01306 0314