PROCEEDINGS .A/ EOYAL SOCIETY OF LONDON. From May 10 to. June 21, 18D4. VOL. LVI. LONDON: HARRISON AND SONS, ST. MARTIN'S LANE, in ©rbinarjr to $er MDCCCXC1V. Q m v-st IONDON : HARRISON AND SONS, PBINTEE? IN ORDINARY TO HER MAJESTY, ST. MARTIN'S LANE. CONTENTS. VOL. LVI. No. 336.— May 10, 1894. Pajre List of Candidates recommended for Election 1 The Composition of Atmospheres which extinguish Flame. By Frank Clowes, D.Sc., Lond., Professor of Chemistry, University College, Nottingham 2 Preliminary Eeport on the Results obtained with the Prismatic Camera during the Total Eclipse of the Sun, April 16, 1893. By J. Norman Lockyer, C.B., F.E.S 7 Researches on Modern Explosives. Preliminary Communication. By William Macnab, F.I.C., F.C.S., and E. Ristori, Assoc. M. Inst. C.E., F.R.A.S 8 On the Leicester Earthquake of August 4, 1893. By Charles Davison, M. A., Mathematical Master at King Edward's High School, Birming- ham 19 The Total Solar Eclipse of 16th April, 1893. Report on Results ob- tained with the Slit Spectroscopes. By E. H. Hills, Capt, R.E 20 The Stresses and Strains in Isotropic Elastic Solid Ellipsoids in Equili- brium under Bodily Forces derivable from a Potential of the Second Degree. By C. Chree, M.A., Fellow of King's College, Cambridge, Superintendent of Kew Observatory , 28 List of Presents .., 28 May 24, 1894. Researches on the Electrical Properties of Pure Substances. No. I. The Electrical Properties of Pure Sulphur. By Richard Threlfall, M«A., Professor of Physics in the University of Sydney, Joseph Henry Drapier Brearley,, Deas-Thomson Scholar in the University of Sydney, and J. B. Allen, Exhibition Commissioners' Scholar of the University of Adelaide, South Australia 32 On the Dynamical Theory of Incompressible Viscous Fluids and the Determination of the Criterion. By Osborne Reynolds, F.R.S., &c. .. 40 On certain Functions connected with Tesseral Harmonics with Appli- cations. By A. H. Leahy, M.A., late Fellow of Pembroke College, Cambridge, Professor of Mathematics at Firth College, Sheffield 45 On the Measurement of the Magnetic Properties of Iron. By Thomas Gray, B.Sc., F.R.S.E., Professor of Dynamic Engineering, The Rose Polytechnic Institute, Terre Haute, Indiana 49 a 2 IV Page On the Influence of certain Natural Agents on the Virulence of the Tubercle-Bacillus. By Arthur Ransome, M.D., F.R.S., and Sheridan Delepine 51 On some Voltaic Combinations with a Fused Electrolyte and a Gaseous Depolariser. By J. W. Swan, M.A 56 Measurements of the Absolute Specific Resistance of Pure Electrolytic Copper. By J. W. Swan and J. Ehodin 64 List of Presents •. 81 May 31, 1894. On the Electrification of Air. By Lord Kelvin, P.E.S., and Magnus Maclean, M.A., F.RS.E 84 On the Effect of Magnetisation upon the Dimensions of Iron Rings in Directions perpendicular to the Magnetisation, and upon the Volume of the Rings. By Shelford Bidwell, M.A., LL.B., F.R.S 94 Note on the Possibility of obtaining a Unidirectional Current to Earth from the Mains of an Alternating Current System. By Major P. Cardew 99 The Effect of Mechanical Stress and of Magnetisation on the Physical Properties of Alloys of Iron and Nickel and of Manganese Steel. By Herbert Tomlinson, B.A., F.R.S 103 Propagation of Magnetisation of Iron as affected by the Electric Currents in the Iron. By J. Hopkinson, F.R.S., and E. Wilson 108 On Rapid Variations of Atmospheric Temperature, especially during Fohn, and the Methods of observing them. By J. Y. Buchanan, F.R.S 108 The Root of Lyginodendron Oldhamium, Will. By W. C. Williamson, LL.D., F.R.S., and D. H. Scott, M.A., Ph.D., F.L.S., F.G.S 128 List of Presents ., 128 No. 337.— June 7, 1894. Election of Fellows 130 On the Newtonian Constant of Gravitation. By C. V. Boys, F.R.S., A.R.S.M., Assistant Professor of Physics, Royal College of Science, South Kensington 131 On the Recurrent Images following Visual Impressions. By Shelford Bidwell, M.A., LL.B., F.R.S , 132 Niagara Falls as a Chronometer of Geological Time. By J. W. Spencer, Ph.D 145 The Influence of Intra-Venous Injection of Sugar on the Gases of the Blood. By Vaughan Harley, M.D., Teacher of Chemical Pathology, University College, London, Grocer Research Scholar 148 Contributions to the Life-History of the Foraminifera. By J. J. Lister, M.A., St. John's College, Cambridge 155 List of Presents .. 160 June 14, 1894. Pajja The Molecular Surface-energy of the Esters, showing its Variation with Chemical Constitution. By Professor W. Ramsay, Ph.D., F.R.S., and Miss Emily Aston, B.Sc 162 The Complexity and the Dissociation of the Molecules of Liquids. By Professor W. Ramsay, Ph.D., F.R.S 171 The Molecular Surface-energy of Mixtures of non-associating Liquids. By Professor W. Ramsay, Ph.D., F.R.S., and Miss Emily Aston, B.Sc. 182 Flame Spectra at High Temperatures. Part II. The Spectrum of Metallic Manganese, of Alloys of Manganese, and of Compounds con- taining that Element. By W. N. Hartley, F.R.S 192 Flame Spectra at High Temperatures. Part III. The Spectroscopic Phenomena and Thermo-Chemistry of the Bessemer Process. By "W. N. Hartley, F.R.S., Royal College of Science, Dublin 193 On a Method for determining the Thermal Conductivity of Metals, with Applications to Copper, Silver, Gold, and Platinum. By James H. Gray, M.A., B.Sc., 1851 Exhibition Scholar, Glasgow University. 199 List of Presents .. 203 June 21, 1894. Researches on Explosives. Preliminary Note. By Captain Sir A. Noble, K.C.B., F.R.S., M.I.C.E., £c 205 Measurement of Colour produced by Contrast. By Captain W. de W. Abney, C.B., D.C.L., F.R.S 221 On some Phenomena in Vacuum-tubes. By Sir David Salomons, Bart., M.A., V.-P. lust. Elec. Engrs 229 On an Instrument for Indicating and Measuring Difference of Phase between E.M.F. and Current in any Alternating Current System. By Major P. Cardew, R.E 250 On the Difference of Potential that may be established at the Surface of the Ground immediately above and at various Distances from a buried Mass of Metal charged from a High Pressure Electric Light Supply. By Major Cardew, R.E., and Major Bagnold, R.E. .. 252 On the Viscosity of Water as determined by Mr. J. B. Hannay by means of his Microrheometer. By Robert E. Barnett, A.R.C.S 259 The Rotation of the Electric Arc. By Alexander Pelham Trotter, B.A. 262 The Electric Strength of Mixtures of Nitrogen and Hydrogen. By Miss P. G. Fawcett 263 The Asymmetrical Probability-Curve. By F. Y. Edgeworth, M.A., D.C.L 271 The Differential Covariants of Twisted Curves, with some Illustrations of the Application to Quartic Curves. By R. F. Gwyther, M.A., Fielden Lecturer in Mathematics, Owens College, Manchester 272 On the Singular Solutions of Simultaneous Ordinary Differential Equa- tions and the Theory of Congruencies. By A. C. Dixon, M.A., Fellow of Trinity College, Cambridge, Professor of Mathematics in Queen's College, Galway 277 The Spectrum Changes in 0 Lyrse. Preliminary Note. By J. Norman Lockyer, C.B., F.R.S 278 VI Page On the Photographic Spectrum of the Great Nebula in Orion. By J. Norman Lockyer, C.B., F.R.S 285 On the Absorption Spectra of Dilute Solutions. By Thos. Ewan, B.Sc., Ph.D., 1851 Exhibition Scholar in Chemistry in the Owens College.... 286 Researches on the Structure, Organisation, and Classification of the Fossil Reptilia. Part IX. Section 4. On the Gomphodontia. By H. G. Seeley, F.R.S 288 Researches on the Structure, Organisation, and Classification of the Fossil Reptilia. Part IX. Section 5. On new Cynodontia. By H. G. Seeley, F.R.S ." 291 Researches on the Structure, Organisation, and Classification of the Fossil Reptilia. Part IX. Section 6. Associated Remains of two small Specimens from Kilpfontein. By H. G. Seeley, F.R.S 295 On the Evolution of the Vertebral Column of Fishes. By H. Gadow, Ph.D., F.R.S., and Miss E. C. Abbott 296 On the Structure and Affinities of Heliopora cotrulea, Pall., with some Observations on the Structure of Xenia and Heteroxenia. By Gilbert C. Bourne, M.A., F.L.S., Fellow of New College, Oxford 299 Degenerations consequent on Experimental Lesions of the Cerebellum. By J. S. Risien Russell, M.D., M.R.C.P., Assistant Physician to the Metropolitan Hospital 303 A Contribution to the Study of (i) some of the Decussating Tracts of the Mid- and Inter-brain, and (ii) of the Pyramidal System in the Mesencephalon and Bulb. By Hubert Boyce, M.B., Assistant Pro- fessor of Pathology in University College, London 305 A Magnetic Survey of the British Isles for the Epoch January 1, 1891'. By A. W. Riicker, F.R.S., and T. E. Thorpe, F.R.S. [Title only] 307 On the different Forms of Breathing. By William Marcet, M.D., F.R.S. •[Title only] 307 List of Presents 307 No. 338. Third Report to the Royal Society Water Research Committee. By Percy F. Fraukland, Ph.D., B.Sc., F.R.S., Professor of Chemistry in Mason College, Birmingham, and H. Marshall Ward, D.Sc., F.R.S., F.L.S., F.R.H.S., Professor of Botany, Royal Engineering College, Cooper's Hill 315 No. 339. Report of Magnetical Observations at Falmouth Observatory for the Year 1893 ., 557 Obituary Notice : — Frederick Le Gros Clark i Index v Errata ... x PROCEEDINGS OF THE ROYAL SOCIETY May 10, 1894. The LORD KELVIN", D.C.L., LL.D., President, followed by Sir JOHN EVANS, K.C.B., D.C.L., LL.D., Vice-President and Treasurer, in the Chair. Professor Dmitri Ivanovitch Mendeleeff, who was elected a Foreign Member in 1882, signed the obligation in the Charter Book and was admitted into the Society. Mr. Benjamin Neeve Peach (elected 1892) was admitted into the Society. A List of the Presents received was laid on the table, and thanks ordered for them. In pursuance of the Statutes, the names of the Candidates recom- mended for election into the Society were read from the, Chair as follows : — Bateson, William, M.A. Boulenger, George Albert. Bradford, John Rose, M.D. Callendar, Professor Hugh Long- bourne. Cheyne, Professor William Wat- son, M.B., F.R.C.S. Froude, Robert Edmund. Hill, Professor M. J. M., M.A., D.Sc. Jones, Professor John Viriamu, M.A., B.Sc. The following Papers were read :- Love, Augustus Edward Hough, M.A. Lydekker, Richard, B.A. Penrose, Francis Cranmer, M.A., F.R.A.S. Scott, Dukinfield Henry, M.A., F.L.S. Smith, Rev. Frederick John, M.A. Swan, Joseph Wilson, M.A., F.I.C. Veley, Victor Herbert, M.A., F.C.S. VOL. T.VI, Prof. F. Clowes. The Composition of [May 10, I. " The Composition of Atmospheres which Extinguish Flame." By FRANK CLOWES, D.Sc., Lond., Professor of Chemistry, University College, Nottingham. Communicated by Pro- fessor ARMSTRONG, F.R.S. Received March 14, 1894. 1, Introductory Remarte, A. study of the experiments which have been made to determine the composition of atmospheres, which act extinctively upon flame, shows that in many cases the atmosphere under examination was in contact with water. The solvent action of water on the carbon di- oxide present seems in such cases likely to disturb the composition of the mixture. In other cases, only the proportion of oxygen in the extractive atmosphere was noted, and the nature of the diluent gas or gases was not taken into consideration. Experiments were also limited to the flames of a few combustible substances, or where a wider range of different flames were tried, the results reported were only of an approximate and relative nature. The experimental work, the results of which are summarized in this communication, was undertaken in order to supplement the deficiencies referred to above, with the view of drawing further generalisations, and of furnishing support to those already drawn from previous experiments. 2. Method of Experimenting. The mixtures of air with the extinctive gas were made in a glass cylinder, which was closed by a ground glass plate. A measured quantity of water, equal in volume to the percentage of extinctive gas to be mixed with the air, was first poured into the glass cylinder. The cylinder was then closed by the plate and in- verted in a vessel of water. A light xylonite ball . of known volume was then passed up, and the extinctive gas was introduced in sufficient quantity to fill the cylinder. The cylinder was then closed and its contents were mixed by the movement of the ball. In order to test the accuracy with which any desired mixture of gases could be prepared by this method, two mixtures of air with carbon dioxide were submitted to analysis. They furnished respec- tively 9'8 instead of 10 per cent., and 69'7 instead of 70 per cent, of carbon dioxide. The experimental flames used were 0'75 in. in height and were gradually lowered into the cylinder, the top of which was finally covered by the plate. The gases were burnt from a platinum jet I mm. in diameter. 1894.] Atmospheres which extinguish Flame. 3 The gaseous mixture was considered to be in extinctive proportions if the flame was extinguished during its downward passage, or imme- diately upon attaining its lowest position in the cylinder. The mixture was considered to contain the minimum necessary quantity of extinctive gas, when another mixture containing 1 per cent, less of the extinctive gas allowed the flame to continue burning in it for a few seconds only. The limiting differences between the results of repeated trials corresponded to 1 per cent, of the extinctive gas in the air. This minimiim necessary percentage of extinctive gas is recorded below in tabulated form. It was considered necessary to take the immediate extinction of the flame as the criterion of extinctive power, since the composition of the atmosphere was rapidly affected by the combustion of the flame. 3. Influence of the Size of the Flame. As a matter of convenience, the flames were, in all cases, set to a height of 0'75 inch. But a series of experiments was undertaken with the same flame of varying size, in order to ascertain if the pro- portion of extinctive gas necessary to extinguish the flame varied with the size of the flame. The results of these experiments with flames of hydrogen and alcohol, varying from 0'4 in. to 1*5 in. in height, show that the vary- ing dimensions .of the flame, within the wide limits included in the trials, are without influence on the proportion of carbon dioxide in the air necessary to produce extinction. 4. Method of Preparation of Gases Used. The carbon dioxide employed for the experiments was prepared in the usual way by the action of dilated hydrochloric acid upon marble. It was washed with water, and was proved to be practically free from air. The nitrogen was prepared by heating an aqueous solution contain- ing potassium nitrite, ammonium chloride, and potassium dichromate. An analysis of the resulting gas proved that it contained 99'7 per cent, of nitrogen. 5 Residts obtained by the Experiments. In the following table the number entered is the average of numerous closely concordant experimental results. The percentage volume of nitrogen in air is taken as 21. Cu a Prof. F. Clowes. The Composition of [May "8 o c ^ at oo cp o I;- 9 9 X 1- 1 .2 fa '« CO (-N CC CC 00 00 X X X GO CO -t IN C T. o x x x x 1 h o -£ .. '3 C '2 bb = o £ _2 r-1 CC 5^1 W ^ "^ M ^ -t N CO C °3 Ifi 1 o ts t- o o cc O »n i^ ec »— I8 2 | g B i be • .s "H^S ^ 00 CO 1?1 OJ o x »~ i^ cr "S 21 » !> 3 w --S II •^ -g 03 Li •s -s i >» -§§S OS CD -*1 CS >ra »0 i— i 00 i> *A c* 1O 00 •*$< CS 10 o CO 00 lol •31 10 oo *>• t- i— i 0 i— i o •—I * lO t> C^l O o to 10 •*? ^| i— 1 r-4 i— 1 p CO CO C5 i— 1 cs o O rH o CO IM s ^ 1° o b CO i— 1 r§ .5 «rT W c* fcd B 1 0 2 ^ "S O o" ^ aT ."t^ S, 8L p *:; ."2 _^ 8 « 11 W TO H PH § "g o O 1 Is PQ 1894.] Researches on Modern Explosives. 15 •^ . c^> £ a "5 ||1 Tp O5 £1 **f* . -H ip 1 -f ,'^ 0 0 bO JO rH M 71 rH IN "a 05 Tf. O ^ O ao a - • H 1 rfi 00 X US -N O o _o J .10 (NO rH rH lO o D 1 o o o c O o o •g g «8 S5! H|S-SJ O O! *J ' r* bO^ 0 O S 2 o ?% o "' s ft. a 16 Messrs. W. Macnab and E. Ristori. [May 10, the products of the number of calories by the volumes of gas, the last three figures being suppressed in order to simplify the results. In the case of EC and SS a certain amount of mineral residue was left, but this was not determined. Troisdorf leaves a slight, and Sifleite and BN a considerable, carbonaceous residue, part of it adhering so tenaciously to the bomb that an exact determination was not made. In the other experiments recorded in Tables I and II the degree of accuracy of the results may be gauged by the fact that the average weight of the products of explosion, calculated from the results found, amounts to 99' 7 per cent, of the weight of the explosive fired, the extreme limits being 10O5 and 98'9 per cent. In Table II the comparison of the pairs of results from explosives made with lower and more highly nitrated nitro-cellnlose shows that the use of the highly nitrated cellulose increases the quantity of heat developed, and diminishes the volume of gas. The composition Table III. — Showing the Heat developed by Explosives containing Nitro-glycerin and Nitro-cellulose in different proportions. Composition of explosires. Calories per gram. Nitro-cellulose (N = 13'3 per cent.) Mtro-glycerin. 100 per cent, (dry pulp) 100 (gelatinised) 90 80 0 0 10 per cent. 20 , 1061 922 1044 1159 70 30 , 1267 60 50 40 40 , , 50 , , 60 , 1347 1410 1467 0 „ „ 100 „ „ 1652 1062 1288 1349 1405 1134 1280 Xitro-cellulose (N = 12'24 per cent.) Nitro-glycerin. 80 per cent. 60 „ „ 50 „ „ 40 „ „ 20 per cent. 40 „ „ 50 „ „' 60 „ „ Nitre-cellulose (N = 13'3 per ^r ,. ,T., , . \ V aselme. lutro-glycenn. 55 per cent. 5 per cent. 40 per cent. 35 „ „ 5 „ „ 60 „ „ 1894.] Researches on Modern Explosives. 17 of the permanent gases is also altered, as might be expected, there being an increase in carbonic acid and decrease in carbonic oxide and hydrogen. The similarity in the volumes of gas produced and the composition of the permanent gases in the case of experiments F and G is worthy of note when the great difference in the original component ingre- dients of the explosives is borne in mind. Table III shows clearly the increase of heat due to increased per- centage of nitro-glycerin, as well as the difference of heat evolved from explosives containing nitro-cellulose of different degrees of nitration. The diminution in quantity of heat (about 200 calories) which the replacement of 5 per cent, of nitro-cellulose by vaseline makes is also very striking. . Table IV. — Showing the Heat developed and the Analysis of the Permanent Gas produced in a closed Vessel from which the Air has not been exhausted — the Explosive being in every case Ballistite of Italian Manufacture. Charge. Calories per gram. Analysis of the permanent gas. CO2. CO. H. N. 1587 1485 1446 1415 1380 37-0 36-4 36-2 36-2 36-3 17'6 22-0 24-6 26-0 27-0 3-2 4-6 6-1 7-2 7'9 42-2 37 '0 33-1 30-6 28'6 3 4 ;; ..:: 5 , 6 „ Traces of CH4 were found, but in this series of experiments the quantity of this gas was not determined. Table IV shows the part played by the oxygen of the air in the bomb ; when a smaller proportion of explosive in comparison with the air is present the combustion is more complete, and the heat evolved is greater, and the composition of the gases is correspond- ingly modified. In Table V the elementary percentage composition of some of the explosives, along with the percentage composition of the products of explosion by weight, is given. The composition of the samples has been calculated from the " bomb " analyses ; as an example, one of the explosives and its decomposition may be represented approximately by the following equation. We have assumed the nitro-cellulose to consist of a mixture of di- VOL. LVI. c 18 Researches on Modern Explosives. [May 10, ""3 SH . ^0 t- CO rH O 8 N <•* 0 00 OS O Cp 1 (5^ (N i-H 6 IN as rH 00 rH fX> rH CS rH oo os rH rH > ^^ • • <4H O 4* fc 1 g" • 00 CO 00 ip CO O ? 9 OS rH 00 re O hP tC oo co to CD ^P 1O us »o »o fl CS ^ >» 2 - co CO t^ 'C Q d _o c * |« "? CO O $• t- op -^ O *j (rH1 ** 0 0 O rH 0 o O 0 ,^> r» W bo a r- P o • M ~ o Ho ^* 1 1 1 1 1 1 1 I r^ CQ £ £ o 42 H O (N Or! o a 1a~bj ^ 9 8 IO O 9 IN (N IN CO rH "3 ^ 1 lo o o o o 0 0 O '~! W P-I ?J V o "c o ••"" »* (N 00 CD S d o • »O rH 9 f «0 cp CD -t^* | -^ S c3 O la CO Cl iH ao CO 0-1 CO co IM CO CO N £5 tn (B O .2 CO rH O -+3 '3 •• • CO IN O CO os •^ O .t- i-l * "H O ^0 "S ^^ t* os rH XO 00 CO CO ••H rH O OQ a to IN ^ "* IN CO CO CO-* OH o o '•£ 43 o - CO CO ? IO cp O 00 9 OS O rH 00 c3 «5 1 .t! § r?5 oo co »0 CD •* 0 U5 us ia JH '1 * |^jf M " O rO O /H ^5 d 6 co ep t^ CO OS 00 OS rH 00 op 10 oo OS CD 88 JD* •2 rH ^^ IN N OJ IN IN IN IN IN IN +» W (n *OQ ^1 1 rf 00 O «3 CO CO N 9 S 0 IN CO t» op GQ H 0 * a, g a & o o a O CO t^ co »o 8 CD QO »o oo »0 OS lO I>O US CD 3 r"~l H ^ 0 | o h 8 O 00 t* us «5 ^ CO rH IN rH ^^ os ^ 'rf a, PH d »o ^ rH IN rH (N 0 IN §5 IN IN IN rH (N IN •S § O o || '~^~~V^ -^ — , ID o c -^ , 11s CM a V §d * ^H /-s'S r2 x-s'§ J^.S .0 11 r-M o % hO S _.£ *U " r3 Jj QJ S .-> CJ S * (JJ e?5 ^"^ •4^ d O ^H £3 ^"* ^™ ^H "~* ^ ^^ c ^ r^ -H C3 -j GQ bo .g _-S g 05 f^^ a> ^ § ® ^ o g Jj» g S -^ fl • iH £ "& 1J ^ 1 III ti A ® 2 P--S ill 2 o g's .-§ ^.13 yi II O rfl 4) '« T3 O e ° .S « SH O i«^ o'S 53 ^'S CO |'3 ^3 S SO if S*9 CO -g 11 « " II ^ lyl'l sll ft! 0 g Jj'' ^a ^Pn cl^&p- — eL 'HS I ^"^ • h^ >FH * o^^o o ^^o o"^o 0*^0 10 O 83 " ctf % fi?4 »O W5 US IO -v 1 00 IN 00 IN CO — v ' OH ^ 4* O « H * d MM 1894.] On the Leicester Earthquake of Auyust 4, 1893. 19 and tri-nitro-cellulose in proportion corresponding to the nitrogen as found by analysis. The equation for Experiment C may be taken as follows : — 50 per cent. 50 per cent, nitro-cellulose nitro-glycerin. (N = 12'3 per cent.). 6[C3H5(N03)3] +2[C6H7(N03)iA] +3[C.H8(N08),O,] = 25COa+28CO+8H+30N+30H3O. The composition of this explosive, calculated from the foregoing formula and found by analysis, is as follows : — c Formula. 21-2 Analysis. 21-15 o 60-8 60-67 H 2-5 2-67 N. 15-5 15-58 100-0 100-07 These are some of the principal features noticeable in a preliminary survey of these experiments. We are continuing our investigations on the lines indicated in the paper, and are especially endeavouring to measure the actual temperature of explosion under varying con- ditions, and it is hoped that the results obtained will throw some light on the chemical and physical properties of many gases at b igh temperatures and under considerable pressures, and, at the same time, be useful in the practical application of explosives . IV. " On the Leicester Earthquake of August 4, 1893." By CHARLES DAVISON, M.A., Mathematical Master at King Edward's High School, Birmingham. Communicated by Professor J. H. PoYNTiNG, F.R.S. Received February 28, 1894. (Abstract.) « On August 4, 1893, at 6.41 P.M., an earthquake of intensity nearly equal to 6 (according to the Rossi-Forel scale) was felt over the whole of Leicestershire and Rutland and in parts of all the adjoining counties. The disturbed area was 58 miles long, 46 miles broad, and contained an area of about 2066 square miles. The direction of the longer axis (about W. 40° N. and E. 40° S.) and the relative position of the isoseismal lines show that the originating fault, if the earth- quake were due to fault- slipping, must run in about the direction indi- cated, passing between Woodhouse Eaves and Markfield, and heading C 2 20 Capt. E. H. Hills. [May 10, towards the north-east. The anticlinal fault of Cham wood Forest, so far as known, satisfies these conditions, and it is highly probable that the earthquake -was caused by a slip of this fault. The beginning of the sound preceded that of the shock in all parts of the disturbed area ; the end of the sound followed that of the shock in the central district and in the neighbourhood of the minor axis, but preceded it near the end of the major axis. Thus the sound ap- parently outpaced the shock in the direction of the major axis, but not in that of the minor axis. These time-relations of the sound and shock can be readily explained if the area over which the fault-slip took place were several miles in length, for the sound in all prob- ability is due to small and rapid vibrations proceeding chiefly from the margins of that area. The intensity was greatest at and near Woodhouse Eaves, and it is probable that the fault-slip began in the neighbourhood of this place, gradually diminishing in amount in either direction, rather rapidly towards the north-west, and much more slowly towards the south-east ; the rate at which the slipping advanced being greater than the velocity of the earth-wave. The total length of the fault-slip may have been as much as 12 miles or even more, and there can be little doubt that it was continued for some distance under the Triassic rocks on which Leicester is built. V. " The Total Solar Eclipse of 16th April, 1893. Report on Results obtained with the Slit Spectroscopes/' By E. H. HILLS, Capt. R.E. Communicated by the Joint Solar Eclipse Committee. Received March 7, 1894. The parties in Brazil and Africa were both supplied with these instruments, two being sent to each station. The instruments were arranged to take one photograph only during the eclipse with an exposure as long as possible. It was considered that the amount of light available would not allow of more than one successful exposure being made. Of the four resulting photographs, one of those taken in Brazil was unfortunately not finished before the sun reappeared, whilst the 'other shows a faint corona spectrum with a strong sky spectrum on both sides, and a considerable amount of general fog over the plate. I have been able to detect nothing of interest in this photograph, for the Fraunhofer. lines overlap the corona spectrum to such a degree that it is impossible to distinguish any bright lines with certainty. The instrument employed in Africa consisted of two spectroscopes, on one equatorial mounting. The first spectroscope had two prisms, 1894.] The Total Solar Eclipse o/16th April, 1893. 21 each 1*75 in. height and 2'5 in. in base, with refracting angles of 62°, and the second spectroscope had one prism 2' 6 in. both in height and base. Condensing lenses, 3'5 in. aperture and 17'5 in. focus, and of 3 in. aperture and 14'5 in. focus, were used with the two instruments re- spectively. Both spectroscopes were fixed on stout mahogany base-boards, and were completely adjusted before leaving England. To attach them to the mounting, a mahogany tube, about 6 in. square and 2 ft. long, was bolted to the top of the declination axis, and the base-boards of the spectroscopes were screwed on either side of it. A small telescope of 2^ in. aperture was attached on the other side of the tube to act as a finder and for purposes of adjustment. The mounting was one that was made for the eclipse of 1886. It con sisted of a tripod stand composed of pieces of angle iron with the polar and declination axes, and circles of the Corbett equatorial. It was found to be easy to set up and rigid. On arrival at Fundium, a site was selected, and a concrete base was formed. On this the instrument was set up, and no trouble was experienced in getting it into adjustment. The slits of the two spectroscopes were placed parallel to each other and tangential to a circle of declination, and were adjusted so that they cut across oppo- site limbs of the sun, that of the two-prism spectroscope being across the upper or western limb, and that of the one-prism spectro- scope across the eastern limb. For several days before the eclipse, trial plates were taken, in order to obtain reference spectra, and for getting the focus as perfect as possible, as well as for the sake of practising the development of the plates. The plates used were Cadett's most rapid make, and various de- velopers were tried, but no special peculiarities of behaviour were noticed ; pyrogallic acid was used for the eclipse plates. Before leaving England the plates were backed with a solution of asphalt in benzole, for the purpose of destroying the halation or reflection from the back surface of the glass. At the eclipse the shutters of the two cameras were opened about ten seconds after the commencement of totality, and closed about ten seconds before the end, giving a total exposure of three minutes fifty seconds. During the progress of the eclipse I observed the corona and the upper or western limb of the sun through the small telescope with a magnifying power of 40. The corona in this region showed very faint radial markings and several rosy-pink prominences were seen. The largest of these was one at the W. N. W. limb, which is the one of which a strong spectrum was obtained with the two- prism spectroscope. The plates were developed the same evening on 22 Capt. E. H. Hills. [May 10, board the " Alecto." The resulting photograph, in the case of the two-prism spectroscope shows a prominence spectrum on both sides of the dark body of the moon, and outside these a corona spectrum with a faint solar (dark line) spectrum on its extreme edge. The H, K and some other lines extend over the dark moon and on both sides beyond the limits of the corona spectrum. That of the one- prism spectroscope shows the same general character, but there is a prominence spectrum on one side only. Both these photographs were over-exposed, better results would have been obtained if two or even three exposures had been made in the same time. Measurement of the Photographs. The following is the method of measurement adopted. A very accurate micrometer by Hilger, reading to O001 mm., was employed throughout. The large number of bright lines in the prominence spectrum rendered the use of the reference spectra unnecessary. The hydrogen series, together with the lines at wave-lengths 4215'3, 4471'2, and the 6 group gave a sufficient number of fixed points through which to draw an interpolation curve. The micrometer readings of these lines having been taken with the greatest possible accuracy, an interpolation curve was constructed on a large scale, two curves being- drawn for each photograph as a check on each other. The micro- meter readings of the remaining prominence lines were then deter- mined and their wave-lengths taken from the curves. The micrometer readings of the corona lines were next taken. It was impossible to get both sides of the photograph in the field of the microscope at the same time, so each side was taken separately, thus getting four series of scale-readings representing possible coronal lines. The wave-lengths corresponding to these scale-readings were then determined from the interpolation curves, and lists were made — first, of lines common to both photographs ; second, of lines occurring on both sides of one photograph ; third, of lines which had been ob- served in previous eclipses. New measurements of the photographs were again made, with the same care as the first, and all lines in the lists were struck out which were not plainly visible in this second scrutiny. It is possible that this final list may contain some wave-lengths of lines due to accidental marks ; this must be rare, however, as any mark so treated must have been parallel to the lines. A comparison of the measurements of the two photographs will give a good idea of the limits of accuracy of these results. 1894.] The Total Solar Eclipse of 16th April, 1893. 23 The Prominence Spectrum. This list gives the wave-lengths of all the lines in one prominence from each photograph. The second prominence on the two-prism spectroscope plate is of a similar character to the one given, bui contains fewer lines. The intensities of the lines are given approximately by the numbers from 1 to 6. The most interesting feature of this spectrum is the extended hydrogen series. There seems no reason to doubt that the lines at wave-lengths 3692'5, 3687, 3682, 3678, 3675, 3672, 3669'5, and 3667 are members of it. M. Deslandres has obtained a photograph showing five hydrogen lines beyond the one at wave-length 3699 ; this photograph carries the series three lines further. The line at 3680 is the iron line, whose wave-length is given by Cornu as 3680'3, and by Hartley as 3679'5. The new notation for the hydrogen series has been used as con- venient. HJ3 is F, £[7 the line near G, and so on, consecutively. The Corona Spectrum. This is the final adopted list, as described above. It is almost impossible to estimate the intensity of these feeble lines by eye, so no attempt has been made to do so ; but in the column headed "intensity" is placed the number of occurrences of the line in the two photographs, the maximum number being four, viz., on each side of both photographs. Opposite each line in the table is placed the corresponding line that has been noted in previous eclipses. For a complete list of the ob- served corona spectrum, see Dr. Schuster's report on the eclipse of 1886 ('Phil. Trans.,' vol. 180 A, p. 335). It will be observed that the 1474 K, or so-called corona line, is placed in the prominence and not in the corona spectrum. This line is shown very faintly on the extreme limit of one photograph, in which it certainly appears to belong to the prominence. It is true that it extends into the corona, but at the same time it also extends in the opposite direction, over the dark body of the moon. Its appear- ance is somewhat similar to that of the strong hydrogen lines, whose apparent extension into the corona spectrum is probably due to atmo- spheric haze. This region of the spectrum has never been specially photographed with the slit spectroscope during an eclipse, and I think a serious! attack on it should most certainly be made at the first opportunity, by using plates which can now be prepared, which are specially sensitive1 to this region of the spectrum. 24 Capt. E. H. Hills. [May 10, Prominence Spectrum from Slit Spectroscope Photographs. Intensity. 2-prism spectroscope. 1 -prism spectroscope. Reference. 1 3667 -0 1 3669 -5 1 3672 -0 1 3675 -0 1 3678 -0 4 3680-0 3680 -0 2 3682-0 2 3687 -0 2 3692 -5 3692-2 3 3699 -0 3699 -0 H£ 2 3700-0 2 3701 -0 4 3707 -5 3707 -5 H7 3 3715 -5 3715 -1 4 8716 -9 4 3718 -0 3718 -0 B> 1 3718 -5 1 3724-0 1 3725 -3 4 3730 -0 3730 -0 H\ 2 3732 -8 1 3737-3 2 3741 -3 5 3745 -5 3745-5 He 3 3746-8 5 3755 -3 3755 -0 5 3757 -4 3757 -3 1 3759 -8 1 3764 -0 5 3767 -5 3767 -5 Hi 5 3795 -0 3795 -0 H0 1 3813 -5 3 3817-7 1 3822 -5 1 3823 -6 2 3827 -5 2 3828 -5 3 3830 -8 3830 -7 5 3834 -0 3834 -0 Hfj 5 3836 -9 3836 -5 1 3839 -6 2 3855 -8 2 3858 -8 1 3866 -5 1 3877 -1 1 3880-5 1 3882 8 6 3888 -0 3888-0 H£ 1 3894 -8 1 3900-0 1 3913 -6 6 3934-0 3934 -0 K 1 3944-5 1 3961 -5 6 3969 -0 3969 -0 'lit 1894.] The Total Solar Eclipse of 16th April, 1893. Prominence Spectrum from Slit Spectroscope Photographs (continued). Intensity. 2-prism spectroscope. 1 -prism spectroscope. Reference. 2 3986 -9 3 4026-6 4026-5 1 4047 -5 3 1 4078-2 4092 -5 4078-4 Ca (Lockyer, 4078'2) 5 4101 -2 4101 -2 H8 3 1 6 4215 -3 4227 -0 4340-0 4215 -3 4226 -5 4340-0 Ca (Thalen, 4215'3) Ca (Huggins,4227; Thalen, 4226'3) 4 4471-2 4471-2 V 6 4860 -7 4860-7 H/3 3 5015 -0 1 5169 -1 J* 1 5173 -6 w ^ b2 2 5184 -2 j" 1 5316 -0 •• 1474 K Corona Spectrum from Slit Spectroscope Photographs. Intensity. 2-prism spectroscope. 1 -prism spectroscope. Corresponding lines observed in previous eclipses. 1886. 1883. 1882. 2 3977 -6 3977 -0 3 3982 -6 3983 -0 2 3986 -4 . . 3986 -0 3986 2 3988 -8 1 3990 -0 . . 3990-0 2 3992 -5 3993 -2 9 w , . 3992 2 3994-2 3995 -0 2 3998 -8 3998 -2 3998-4 3998 3 4012-6 4011-9 2 4015 -6 4015-8 , , 4016 4015 3 4022-0 4022-0 2 4023 -0 4023 -8 1 4031-6 . . 4029-7 4031 2 4039-3 4040 ;0 . . 4037 2 4054-0 4054-7 4054-8 4056 4057 2 4067-5 4067-8 4067-7 4064 4067 3 4070 -5 4071-0 4071-0 4071 2 4144-5 9 m 4144-2 4144 2 4167 -2 ,. 4166 -0 2 • • 4169 -0 4169 -7 4169 4168 3 4175 -0 4174 -0 4173-6 9 0 4173 2 4181 -2 4182 -0 4183 -5 4185 4179 4 4191 -0 4190-3 4189 -2 4192 4195 2 4202 -1 4201 -8 2 4204 -4 4203 -5 26 Mr. C. Chree. Elastic Solid Ellipsoids under [May 10, Corona Spectrum from Slit Spectroscope Photographs (continued). Intensity. 2-prism spectroscope. 1 -prism spectroscope. Corresponding lines observed in previous eclipses. 1886. 1883. 1882. 2 4213 -5 4212 '2 4211 -8 4213 4212 2 9 9 4224-4 4222-6 4227 4224 2 4267 -5 4269 -2 4268 -5 . . 4267 2 4279 -7 4280-5 4280-6 4279 2 4295 -0 4295 -3 4293-9 4291 2 4299-5 4298 -7 4301 -0 2 4328-3 1 m 0 4331 -5 4332-1 4330 3 4353 -0 4552 -7 4354 -7 4353 3 4366-2 4364-9 4365-4 4363 (±3) 4 4372-1 4372-4 4372 -2 4370 4370 1 4378 -5 . . 4378 -1 4377 3 4386 -7 4386 -5 4387 -6 3 4389 -5 4390-2 4389 -0 3 4395 -2 4394-4 4395 -8 . . 4395 2 4447-5 f . 4445-8 4449 2 4454-7 4455-3 4452-9 4 4465-4 4465 -7 . . 4465 3 4468-8 4469-0 4468-5 3 4494-0 4494-3 4493-4 4490 3 4516 -0 4516-3 4515 -6 4518 2 4530 -2 4530 -0 4530 -0 2 4536 -0 .. 4536 -1 2 4550 -0 , . 4550 -0 4546 1 4554-3 , . 4557 '2 4555 (±3) 2 •• 5020-0 (±2) VI. " The Stresses and Strains in Isotropic Elastic Solid Ellip- soids in Equilibrium under Bodily Forces derivable from a Potential of the Second Degree." By C. CHREE, M.A., Fellow of King's College, Cambridge, Superintendent of Kew Observatory. Communicated by Professor W. G. ADAMS, F.R.S. Received March 2, 1894. (Abstract.) If a system of bodily forces whose values per unit mass are derived from the potential V — 1 cp ^~ 2 \ acts on an ellipsoid 1894.] Forces derivable from a Potential of the Second Degree. 27 whose density p is uniform, the statical resultant reduces to a couple whose components about the axes of x, y, z are respectively 4irabc (62— c2) /»S/15, 47rabc (c2— a2) //T/15, and 47rabc (a2— fc2) />TJ/15. These components vanish in the case of a sphere, but in an ordinary ellipsoid equilibrium will not exist unless S, T, U all vanish. The problem solved in the present memoir, viz., that of an isotropic elastic solid ellipsoid under the action of bodily forces derived from a potential is thus, for an ordinary ellipsoid, the most general case of equilibrium under forces derived from a potential of the second degree. The above potential covers forces arising from mutual gravitation or from rota- tion about a principal axis in an ellipsoid of any shape. The method of solution reverses the usual order of procedure, the stresses being first determined and then the strains and displace- ments. The solution obtained satisfies without limitation or assump- tion of any kind all the elastic solid equations. Unless the ratios a:b:c are assigned definite numerical values, the constant coeffi- pients in the expressions for the stresses and strains are of course somewhat cumbrous ; but for any specified case, whether of gravita- tion or rotation, or both combined, the solution becomes easily manageable. It enables the variation in the effects of gravitation and rotation with the change of shape of the ellipsoid to be completely traced. The comprehensiveness of the problem solved forbids more than a brief consideration of the general solution with illustrations of its application to a few of the more interesting special forms of ellipsoid. The results obtained for the very oblate and very oblong forms seem to show that in many cases of bodily forces the assumptions usually made in the treatment of thin plates and long rods would not be justified. By comparison with the author's previous researches, a close similarity is shown to exist between the phenomena in rotating flat ellipsoids and thin elliptic discs on the one hand, and rotating elongated ellipsoids and long elliptic cylinders on the other. Various results confirmatory of the accuracy of the present solution are obtained by the application of the general formulae for the mean strains in elastic solids. It is also shown that some of the results may be arrived at by the use of approximate but simple methods. The Society adjourned over the Whitsuntide Recess to Thursday, May 24. 28 Presents. [May 10, Presents, May 10, 1894. Transactions. Baltimore : — Johns Hopkins University. Studies in Historical and Political Science. Series 12. No. 3. 8vo. Baltimore 1894. The University. Berlin : — Gresellschaft fur Erdkunde. Verhandlungen. Bd. XXI. No. 4. 8vo. Berlin, 1894; Zeitschrift. Bd. XXIX. No. 1. 8vo. London 1894. The Society. Boston : — American Academy of Arts and Sciences. Proceedings. Vol. XXVIII. 8vo. Boston 1893. The Academy. Chapel Hill : — Elisha Mitchell Scientific Society. Journal. Vol. X. Part 1. 8vo. Raleigh, N.C. 1893. The Society. Cracow : — Academic des Sciences. Bulletin International. Mars, 1894. 8vo. Cracovie. The Academy. Hertfordshire : — Hertfordshire Natural History Society and Field Club. Transactions. Vol. VII. Parts 8 — 9. 8vo. London 1894. The Club. London: — British Astronomical Association. Journal. Vol. IV. No. 5. 8vo. London 1894; Memoirs. Vol. III. Part 1. 8vo. London 1894. The Association. Geological Society. Quarterly Journal. Vol. L. Part 2. 8vo. London 1894. . The Society. Institute of Brewing. Transactions. Vol. VII. No. 6. 8vo. London 1894. The Institute. Mathematical Society. Proceedings. Vol. XXV. Nos. 475 — 480. 8vo. London [1894]. The Society. Photographic Society. Journal and Transactions. Vol. XVIII. No. 8. 8vo. London 1894. The Society. Royal Medical and Chirurgical Society. Medico- Chirurgical Transactions. Vol. LXXVI. 8vo. London 1893 ; Catalogue of the Library. Supplement 7. 8vo. London 1893. The Society. Royal Microscopical Society. Journal. 1894. Part 2. 8vo. London 1894. The Society. Royal United Service Institution. Journal. Vol. XXXVIII. No. 194. 8vo. London 1894. The Institution. Society of Biblical Archeeology. Proceedings. Vol. XVI. Parts 5—6. 8vo. London 1894. The Society. Zoological Society. Transactions. Vol. XIII. Part 8. 4to. London 1894 ; Proceedings. 1893. Part 4. 8vo. London 1894. The Society. Lyons : — Universite. Annales. Tome III. Fasc. 1 — 2. Tome VI. Fasc. 4. 8vo. Paris 1892-94. The University. 1894.] Presents. 29 Transactions (continued) . Mexico : — Sociedad Cientifica " Antonio Alzate." Memorias y Revista. Tomo VII. Nos. 7—10. 8vo. Mexico 1894. The Society. Naples : — Reale Accademia di Archeologia, Lettere e Belle Arti. Atti. Vol. XVI. 4to. Napoli 1894 ; Rendiconto. Anno VII— VIII. Gennaio e Febbraio. 8vo. Napoli 1893-94. The Academy. Newcastle-upon-Tyne : — North of England Institute of Mining and Mechanical Engineers. Transactions. Vol. XLIII. Part 4. 8vo. Newcastle-upon-Tyne 1893; An Account of the Strata of Northumberland and Durham as proved by Borings and Sinkings. L-R, S-T. 8vo. Newcastle-upon-Tyne 1887-94. The Institute. New York : — American Museum of Natural History. Bulletin. Vol. VI. Pp. 81—96. 8vo. {New York] 1894. The Museum. Philadelphia: — Geographical Club. Bulletin. Vol. I. No. 2. 8vo. Philadelphia 1894 ; Charter, By-Laws, List of Members. 8vo. Philadelphia 1894. The Club. Santiago : — Sociedad Nacional de Mineria. Boletin. Ano XI. No. 64. 4to. Santiago de Chile 1894. The Society. Stockholm : — Kongl. Vetenskaps Akademie. Of versigt. Arg. LI. Nos. 2—3. 8vo. Stockholm 1894. The Academy. Toulouse : — Faculte des Sciences. Annales. Tome VIII. Fasc. 1. 4to. Paris 1894. The Faculty. Utrecht : — Physiologisch Laboratorium der Hoogeschool. Onder- zoekingen. Deel III. Afl. 1. 8vo. Utrecht 1894. The Laboratory. Vienna : — Kais. Akademie der Wissenscbaften. Sitzungsberichte (Math.-Naturw. Classe). Bd. Oil. Abth. 3. Heft 8—10. 8vo. Wien 1893; Sitzungsberichte (Phil.-Hist. Classe). Bd. CXXIX. 8vo. Wien 1893 ; Denkschriften. Bd. XLII. 4to. Wien 1893; Almanach. 1893. 8vo. Wien; Mittheilungen der Prahistorischen Commission. Bd. I. No. 3. 4to. Wien 1893. The Academy. K.K. Geologische Reichsanstalt. Abhandlungen. Bd. VI. Halfte 2. Bd. XV. Heft 6. 4to. Wien 1893 ; Jahrbuch. Bd. XLIII. Heft 3 — 4. 8vo. Wien 1894 ; Verhandlungen. 1894. Nos. 1—4. 8vo. Wien. The Institute. Zurich : — Naturforschende Gesellschaft. Vierteljahrsschrift. Jahrg. XXXIX. Heft 1. 8vo. Zurich 1894. The Society. 30 Presents. [May 10, Observations and Reports. Calcutta : — Meteorological Department, Government of India. Meteorological Observations recorded at Seven Stations in India. November, 1893. 4to. Monthly Weather Review. November, 1893. 4to. Calcutta 1894. The Department. Edinburgh : — Royal Observatory. Circular. No. 42. 4to. [Sheet.] 1894. The Observatory. India : — Archaeological Survey Circle, North- Western Provinces and Oudh. Annual Progress Report. 1893. Folio. RoorJcee. The Survey. Geological Survey. Palaeontologia Indica. Series IX. Vol. II. Part 1. Folio. Calcutta 1893; Records. Vol. XXVII. Part 1. 8vo. Calcutta 1894; a Manual of the Geology of India. Second edition. 8vo. Calcutta 1893. The Survey. London : — Meteorological Office. Daily Weather Reports. July — December, 1892. January — June, 1893. 4to. [London 1892-93] ; Meteorological Observations at Stations of the Second Order. 1889. 4to. London 1893. The Office. Sydney : — Department of Agriculture. Plant Diseases and their Remedies, by N. A. Cobb. Diseases of the Sugar-Cane. 8vo. Sydney 1893. Dr. Cobb. Washington: — U.S. Coast and Geodetic Survey. Report of the Superintendent. 1891. Part 2. 8vo. Washington 1892. The Survey. U.S. Department of Agriculture. Report of the Ohio Weather and Crop Service. February, 1894. 8vo. Norwalk, Ohio 1894. The Department. Journals. Agricultural Gazette of New South Wales. Vol. V. Part 3. 8vo. Sydney 1894. Department of Agriculture, Sydney. Boletin de Minas Industria y Construcciones. Ano X. No. 2. 4to. Lima 1894. Escuela Especial de Ingenieros, Lima. Horological Journal. Vol. XXXVI. No. 429. 8vo. London 1894. British Horological Institute. Morphologisches Jahrbuch. Bd. XXI. Heft 2. 8vo. Leipzig 1894. Prof. Gegenbaur, For Mem. R.S. Nature Notes. Vol. V. No. 53. 8vo. London 1894. Selborne Society. Revue Medico-pharmaceutique. Annee VII. No. 3. 4to. Con- stantinople 1894. The Editor. 1894.] Presents. 31 Chauveau (A.), For. Mem. R.S. La Vie et 1'Euergie chez 1' Animal. 8vo. Paris 1894. The Author. Clark (A. P.) [Nine Excerpts. 8vo. 1887-93.] The Author. Dane (L. W.) Customary Law of the Main Tribes in the Gurdas- pur District. 8vo. Lahore 1893. India Office. Ferree (B.) The Chronology of the Cathedral Churches of France. 8vo. New York 1894. The Author. [Hermite (C.), For. Mem. R.S.] Jubile de M. Hermite. 1822—1892 (24 Decembre). 8vo. Paris 1893. Comite Hermite. Hooker (Sir J. D.), F.R.S. The Flora of British India. Part 20. 8vo. London 1894. India Office. Macfarlane (A.) The Principles of Elliptic and Hyperbolic Analysis. 8vo. Boston, Mass. 1894. The Author. Maumene (E. J.) Extrait d'une Etude de 1'Acide tetrafhique (tar- trique) et de ses Composes Salins. 8vo. Chateauroux 1894. The Author. Ricco (A.) La Lava Incandescente nel Cratere Centrale dell' Etna e Fenomeni Geodinamici concomitant]. Folio. Roma 1894 ; [and Six other Excerpts. 8vo. and 4to.] The Author. Vallin (A. F.) Discursos leidos ante la Real Academia de Ciencias Exactas, Fisicas y Naturales en la Recepcidn Piiblica de A. F. Vallin. 8vo. Madrid 1893. Sr. D. Vallin. Wolf (R.) Astronomische Mitteilungen. No. 83. 8vo. [Zurich] 1894. The Author. Bronze Copy of the Medal struck in honour of the 70th Birthday of M. Charles Hermite, For. Mem. R.S. Comite Hermite. 32 Prof. Threlfall and Messrs. Brearley and Allen. [May 24, May 24, 1894. The LORD KELVIN", D.C.L., LL.D., President, in the Chair. Professor fileuthere filie Nicolas Mascarfc, who was elected a Foreign Member in 1882, signed the Obligation in the Charter Book, and was admitted into the Society. A List of the Presents received was laid on the table, and thanks ordered for them. The following Papers were read : — I. " Researches on the Electrical Properties of Pure Sub- stances. No. I. The Electrical Properties of Pure Sulphur." By *RIOHARD THRELFALL, M.A., Professor of Physics in the University of Sydney; JOSEPH HENRY DRAPIER BREARLEY, Deas-Thomson Scholar in the University of Sydney, and J. B. ALLEN, Exhibition Commissioners' Scholar of the University of Adelaide, South Australia. Communicated by Professor J. J. THOMSON, F.R.S. Re- ceived April 19, 1894. (Abstract.) Since there appears to be no definite information as to the elec- trical properties of pure elemental substances which are not metals, an attempt has been made to provide the necessary data in the case of sulphur. This element was chosen as being capable of easy puri- fication, and because it can exist in a variety of forms, from the comparison of the electrical behaviour of which some information was expected to be obtained. The experimental work was begun in 1886, and some preliminary results were published by one of the authors and Mr. J. A. Pollock in the ' Philosophical Magazine ' for 1890. These results referred to the construction of galvanometers for high resistance measurements, the reliability of the Clark cell as a source of small constant currents, and a method of using the gal- vanometer in resistance measurements in such a way that no galvanometer law of current measurement needs to be assumed. * Part I by Professor Threlfall and Mr. Brearley. Part II by Professor Threl- fall and Mr. Allen. 1894.J On the Electrical Properties of Pure Substances. 33 This method consists in observing the galvanometer indication of a carrent passing through the substance of high resistance under a known voltage, and subsequently causing the galvanometer to give the same deflection, by supplying it with a known fraction of the voltage of a Clark cell, and allowing this to act on the galvanometer when the latter is in series with a wire megohm standard. The dis- cussion of this method, depending as it does on the behaviour of Clark cells from which currents are being taken, shows that it is reliable ; but it is not intended here to go over the preliminary ground covered by the papers referred to. A considerable portion of the investigation of which the paper is an account, and which extends from 1889 to the present time (October, 1893), was made conjointly with Mr. Pollock. The first part of the paper deals with the purification of sulphur as obtained from several sources, with the result that, in the end, the following method was exclusively adopted. This method is based on the use of sulphur recovered by the Chance process, which comes into commerce as pure to at least one part in ten thousand, and results from burning hydrogen sulphide from alkali waste with insufficient air for complete combustion. The commercial product is melted, and filtered through glass wool and platinum gauze. It is then twice distilled, in such a manner as to be free from exposure to dust : and sometimes it was subsequently freed from gas, by heating in a vacuum to near the boiling point. The purity of the resulting sulphur is tested by the following criteria. It must be free from smell. It must leave no residue on evaporation from a platinum dish. When cooled suddenly from a high temperature, it must remain of a clear yellow colour : when perfectly crystalline it must be absolutely soluble in carbon bisulphide. The absence of arsenic and selenium from the sulphur employed, was proved to about one part in a million by burning the sulphur to trioxide and applying the appropriate tests. The reaction by sulphur dioxide test for selen- ium, when properly carried out, is more delicate than the Marsh's test for arsenic, even when the smell of the hydrogen is adopted as a criterion. If a perceptible mirror of arsenic is to be accepted as a criterion, the arsenic test is still less delicate. A number of experiments are described, tending to show that neither arsenic nor selenium can possibly exist to any appreciable extent in alkali waste produced in the Leblanc process, so that the Chance sulphur is probably more free from these impurities than the limit we can reach by analysis. All other known impurities are got rid of by distillation and exhaustion in vacuo. Section 3 of the paper deals with a discussion of various methods of measuring high resistances, and gives the detail of the method adopted by us for rapidly effecting alternate measurements of VOL. LYI. D 34 Prof. Threlfall and Messrs. Brearley and Allen. [May 24, resistance and capacity. The method of producing films of pure sulphur between aluminium electrodes is also explained. It is neces- sary to perform the melting, &c., in a gold vessel. Section 4 of the paper deals with the method of constructing gal- vanometers of high sensitiveness and resistance. In order to observe as small currents as possible, advantage was taken of every circum- stance, both of observation and construction, likely to lead to enhanced sensitiveness. For instance, instead of observing the steady deflection, we habitually observed the throw of the needle on reversal of the current through the galvanometer. The steady de- flection was only observed as a check. The highest degree of sensitiveness we ever found it necessary to use, was such that the throw on reversal was 1 micrometer division for a current of 1'5 X 13~13 amperes, with a period of vibration of about 25 seconds. 1 micrometer division is divisible into five parts, so that the sensitiveness for least observable throw on reversal is 3 X 10~u amperes. This sensitiveness, however, cannot be taken ad- vantage of, except with very elaborate contact keys, and under rare conditions of magnetic steadiness. We adopted the Kelvin type of instrument. We consider that the problem of sensitive galvano- meter building has not hitherto been approached in the proper manner. Almost any first-rate instrument will give enormous sensi- tiveness on occasion : but this sensitiveness is, in general, accom- panied by instability, and is useless in practice, on account of zero changes. The really important matter is to ensure the presence of high sensitiveness with ease and certainty, not after hours of adjust- ing, but immediately on the necessity arising ; in this we have been perfectly successful. Success in this matter depends entirely on a large number of details, for a discussion of which the paper must be consulted. Exact drawings are also provided, both of the instrument as a whole, and of the more important subsidiary parts. The follow- ing notes must suffice here. 1. It is essential that the coils shall be adjustable to the magnetic system after the latter is mounted. 2. Astaticism of sufficient perfection can only be secured by the simultaneous magnetisation of all the members of the magnetic system when they occupy their final relative positions. This necessi- tates special appliances. 3. If copper wire is used for the coils, no other metal must be included in the circuit or connections of the instrument, otherwise thermo-electric effects cannot be avoided. 4. The instrument has four tiers of coils and magnets, whereby improved electromagnetic conditions are obtained. 5. The most important part of the instrument is that which belongs to the adjusting of the magnetic control. This must be 1894.] On the Electrical Properties of Pure Substances. 35 •exceedingly stiff and well made, supported quite independently of the astatic system, and capable of the finest adjustment. 6. Stability of zero depends chiefly on. the uniformity of the con- trolling field all over the suspended system ; this is, perhaps, best obtained by using very large and symmetrically disposed magnets Above and below the suspended system. 7. When this is attended to, there is no advantage in using a •" tail " magnet. 8. The chief remaining difficulty is found to be due to continual small changes taking place in the direction of the earth's horizontal field. This is best overcome by attending to the astaticism of the magnets, and using a fairly strong uniform controlling field opposed to that of the earth. 9. It is essential that the instrument be entirely surrounded by massive iron screens. 10. A novel method of illumination has been worked out, ensuring uniformity of brightness of the scale images, without any appre- ciable heating of the galvanometer. The transparent divisions of an opaque scale are caused to give rise to interference fringes, which are then observed in a telescope with a micrometer scale in the eye- piece. Readings of the position of the magnetic system to one second of arc can be easily and certainly made. 11. The most important improvements we have made relate to the insulation of the instrument, the minute adjustment of the con- trolling field, the recognition of the necessary conditions for high sensitiveness combined with stability, and the method of optical magnification. The instrument can now be used at the sensitiveness mentioned with all the ease and certainty which is generally attained with a millionfold less sensitiveness. 12. Further improvements can be made by using some material of greater strength than glass for the mirror, and by improved magnetic screening. Our screens were of cast iron, and weighed about 300 pounds ; the screening was not nearly sufficiently perfect. Section 5 contains an account of a large number of experiments •extending over three years on the phenomena of conduction in sulphur — of which the following are the chief results. Crystalline sulphur, whether monoclinic or " aged " monoclinic — {which we have ventured to distinguish as a distinct variety, since it preserves the melting point, but is divested of the crystallographic properties of fresh 'monoclinic sulphur) has a specific resistance of 1038 C.Gr.S. units as a minimum. By exposure to the air of a room the sulphur condenses moisture, which reduces its apparent specific resistance, but not nearly so much as in the parallel case of glass. The total residual charge is either absent, or less than four parts in ten thousand of the original charge, when a film of sulphur about a D 2 36 Prof. Threlfall and Messrs. Brearley and Allen. [May 24,. quarter of a millimeter thick is charged for ten minutes with about 300 volts. By very careful drying we have succeeded in reducing the residual charge with a film of mica 0*2 mm. thick to about 1 per cent, of the original charge under similar circumstances. In view of the want of homogeneity in the crystalline sulphur film this freedom from residual effect is noteworthy, and is perhaps to be explained by the entire absence of conductivity. Crystalline sulphur has an electric strength which is more than enough to support 33,000 volts per centimeter — how much more we do not know. At 75° C the specific resistance with 285 volts per quarter millimeter falls to about 6'8xl025 C.G.S. The specific inductive capacity increases slightly as the temperature rises. As the temperature of the sulphur rises the conductivity increases slightly up to the melting point, when there is an enormous increase. When a film containing about 5 per cent, of insoluble sulphur pro- duced by cooling rapidly from a temperature abov? 170° C is examined, it is found to have a sensible conductivity which is not due to surface action, for it is not altered by fusing quartz rods into the exposed part of the surface, nor by blowing air saturated with' water vapour against the surface. The conductivity depends on the exact composition of the mixture of soluble and insoluble sulphur, but may be taken at from 1025 to 1026 C.G.S. units for a film contain- ing from three to six per cent, of amorphous unstable sulphur at ordinary temperatures. This conductivity is always greater when the voltage of about 300 volts on a film a quarter of a millimeter thick is first applied, or reversed. It is established that the in- creased conductivity occurs after the sulphur has rested — whether the voltage is applied for the first time, or whether it has been applied before in either direction. When the voltage is reversed this effect is more strongly marked, and the conductivity only settles to a steady value after a considerable time. The conduction, eifchei- when the current is steady, or when it commences or is reversed,, does not obey Ohm's law either for small voltages (say eight volts) or large ones (say 300) when the film is 0'25 mm. thick. The deviation is, however, greater at high voltages, and greater when the "commencing" or "reversing" effects are taking place than when the conduction is steady. The deviation is always in the direction of making the conduction at high electromotive intensities- greater than at low. The specific inductive capacity of a mixture of soluble and insoluble sulphur is markedly higher than that of purely crystalline sulphur. We have some evidence that the changes occur- ring during the first few days after the film is made lead to art increase of specific inductive capacity. The temperature coefficient of the specific inductive capacity is positive, and of the order 2 x 10"5" pev degree between 20° and 70° C. 1894.] On the Electrical Properties of Pure Substances. 37 The residual charge is larger than when the sulphur is purely soluble, and with about 400 volts per millimeter is about 187 per cent, of the initial charge after 10 minutes' charge ; the condenser being discharged for a fraction of a second and left for 10 minutes. At lower electromotive intensities it is rather greater in comparison with the initial charge, sufficiently so to be distinctly noticeable. On heating the sulphur the conduction increases from about 50° C, in fact whenever the process of annealing takes place. When the annealing change (destruction of amorphous sulphur, and formation of soluble sulphur) is taking place rapidly the conduction is consider- able. Many attempts made for the purpose of deciding whether the increase of conductivity depends on the mere proportion of insoluble sulphur present, or whether it depends on the rate at which the conversion process is taking place, yielded no absolutely certain results, but the evidence, such as it is, points to the latter as being probably the most important — but we do not consider that it can explain the conductivity at low temperatures. This conductivity is essentially discontinuous, and in this resembles the conduction through moisture films condensed on glass, ebonite, and sulphur itself. Several of the above-mentioned peculiarities of sulphur conduction were observed by Quincke in the case of insulating liquids, and were ascribed, in part at all events, to the action of dust motes in the liquids. There is no doubt, however, that in the case of sulphur these effects are inherent to the process of conduction, for they were as strongly marked in what we consider to have been our purest film (as tested by the colour) as in the least pure one. There is some evi- dence that mixtures of insoluble and soluble sulphur show a maxi- mum conductivity when the sulphur contact is between 5 and 3 per cent. All the phenomena of conduction are also noticed — the specific resistance being about the same — when we examine films containing 88 per cent, of insoluble sulphur, produced by applying enormous pressures in a testing machine to the insoluble sulphur formed on suddenly cooling sulphur from a high temperature. The sulphur, which was originally plastic, was exhausted with carbon bisulphide, and the residue treated with sulphur chloride to obtain stability. A pressure equal to the weight of 100,000 pounds on an area of say 25 square inches, causes about 12 per cent, of the insoluble sulphur to become soluble, whether it has been treated with sulphur chloride or not. Check experiments on soluble sulphur showed that the very pure benzene used to moisten the sulphur for the purpose of com- pression produces no subsequent change in the conductivity. The pressures were applied for from five to 10 minutes. No change in the conductivity of mixed films was produced by stressing in alternate directions with a frequency of, say, five per second, and a voltage of from 100 to 200 volts per quarter millimeter. 38 Prof. Threlfall and Messrs. Brearley and Allen. [May 24,. A very large quantity of exact numerical data referring to these points is contained in the paper. Part II of the paper is interposed as the facts disclosed bear on the general argument. This part of the paper bears on several corre- lated questions. Section 1 deals with the contact force in air between purely soluble sulphur and mixtures of insoluble and soluble sulphur containing about 10 per cent, of the former. The result of some rather interesting work on this point by the electrometer needle method, shows that when soluble and mixed sulphur is in contact (produced by melting the parts together), there is a contact force of the order of from one to two volts between them. The positive charge is on the insoluble sulphur. These experiments were made by using a double sulphur needle over metallic semicircles, and also by using the ordinary metallic electrometer needle over sulphur quadrants. The latter gives the best results. The phe- nomena are very complicated, and require to be carefully sifted ; for an account of the very considerable difficulties the paper must be consulted. Section 2 deals with the question as to whether light has any effect on the conductivity of sulphur. Monckman (' Boy. Soc. Proc.,' vol. 46, 1890) considers that he has discovered such an. effect. A very large number of experiments, however, on mixed and crystal- line sulphur cells, failed to indicate to us any such peculiarity, and we consider that Monckman must have been mistaken in this matter. Section 3 deals with the qualitative phenomena of conduction in sulphur cells containing from 5 per cent, to 20 per cent, of insoluble sulphur. The general results agree with those already described, although the methods of preparing and quenching the viscous sulphur were different. The electrodes were also of platinum, wire instead of aluminium plates. The temperature resistance changes are treated rather fully in this section, and bring into prominence the enormous influence of the annealing process. Section 4 deals with a determination of the specific inductive capacity of sulphur by the method of weighing, and contains an account of the different sources of error to which we discovered the method was subject. Several ways of obtaining the required potential difference were investigated, with the result that the most satisfactory is by the use of an alternator giving a frequency of about sixty, and an induction coil used as a transformer. This avoids the difficulty which occurs when the sulphur plates get charged in virtue of their conductivity, and is noticed whenever (1) a continuously directed P.D., or (2) an unequal alternating one (as by a coil with hammer or mercury break) is used. 1804.] On the Electrical Properties of Pure Substances. 39 Part I is then continued. § 6 deals with an investigation of the specific inductive capacity of various kinds of sulphur by the method of weighing, advantage being taken of the laborious investigation of the method dealt with in Part II. Various other matters came to light, and we furnish a drawing of suitable apparatus and describe the necessary course of procedure to make the method accurate and satisfactory ; in particular the proper way of preparing plates of crystalline and friable substances. The results for the specific induc- tive capacities are as follows, at a temperature of 14° C. " Aged " monoclinic sulphur K = 3'162 Ditto with 1'43 per cent, insoluble unstable sulphur K = 3-510 Ditto with 3 per cent, insoluble unstable sulphur K = 3-75 An experiment on purely amorphous sulphur is not yet ready for publication ; but the above results will go a long way to clear up the great differences in the hitherto published values of this constant for sulphur. They also serve as a check on our observations on thin films, and show that our measurements of film thickness — a grave difficulty — were moderately successful. A number of experiments bearing on a theory of conduction which we venture to suggest are also included in this part of the paper. This is followed by an account of the theory to which our experi- ments led us, and which is briefly as follows. Sulphur in either of the extreme conditions does not conduct ; we can only examine the purely soluble state, for the other is not sufficiently stable for a satisfactory investigation ; however, we may say that changing the content of amorphous unstable sulphur from 3 per cent, to 88 per cent., produces little or no change in the conductivity. Taking this and other facts into consideration, we believe that what we have called mixtures of the two kinds of sulphur are really compounds, and that the conduction is electrolytic. We have framed what we believe to be a novel theory of electro- lysis, which explains all the facts which we have observed, and which has the peculiarity of introducing the idea of an electrolytic con- vection current, in connection with which the resulting changes of specific inductive capacity allow of all the phenomena of conduction observed taking place, though the charges may never really reach the electrodes. It will be seen that the effects of fatigue-reversal and the phenomena of discontinuous conduction are well accounted for by this theory. The only objections we have to it are that it is based on a molecular theory of matter, which we are persuaded requires to be remodelled, if it is to afford any real explanation of things as they are. A theory of residual effect based on the theory of conduction is 40 Prof. 0. Reynolds. On the Dynamical [May 24, also proposed. This differs from Maxwell's theory in that tlie latter merely postulates changes of specific resistance and specific induc- tive capacity from point to point of the dielectric, while our theory is distinctly chemical. We consider that our results on mixed films are best explained by the theory we propose, though the difficulty of dis- proving Maxwell's theory is almost equal to the difficulty of establish- ing it, and we do not wish to imply that some sort of explanation on this theory may not be constructed to fit in with our observations. This is a point, however, on which we are still engaged. The matter may, perhaps, be best summed up in the statement that the evidence we have against Maxwell's theory is nearly worthless; but that we do not consider this theory necessary if our theory of conduction be accepted. II. " On the Dynamical Theory of Incompressible Viscous Fluids and the Determination of the Criterion." By OSBORNE REYNOLDS, F.R.S., &c. Received April 25, 1894. (Abstract.) The equations of motion of viscous fluid (obtained by grafting on certain terms to the abstract equations of the Eulerian form so as to adapt these equations to the case of fluids subject to stresses depend- ing in some hypothetical manner on the rates of distortion, which equations Navier* seems to have first introduced in 1822, and which were much studied by Cauchyf and PoissonJ) were finally shown by St. Yeuant§ and Sir Gabriel Stokes, || in 1845, to involve no other assumption than that the stresses, other than that of pressure uniform in all directions, are linear functions of the rates of distortion with a co-efficient depending on the physical state of the fluid. By obtaining a singular solution of these equations as applied to the case of pendulums in steady periodic motion Sir G. Stokes^f was able to compare the theoretical results with the numerous experi- ments that had been recorded, with the result that the theoretical calculations agreed so closely with the experimental determinations as seemingly to prove the truth of the assumption involved. This was also the result of comparing the flow of water through uniform tubes with the flow calculated from a singular solution of the equations so long as the tubes were small and the velocities slow. On the other * ' Mem. de 1' Academic,' t. ri, p. 389. t ' Mem. des Savants Etrangers,' t. 1, p. 40. J ' Mem. de 1' Academic,' t. \, p. 345. § ' B.A. Keport,' 1846. || ' Cambridge Trans.,' 1845. H ' Cambridge Trans.,' vol. ix, 1857. 1894.] Theory of Incompressible Viscous Fluids, fyc. 41 hand, these results, both theoretical and practical, were directly at variance with common experience as to the resistance encountered by larger bodies moving with higher velocities through water, or by water moving with greater velocities through larger tubes. This dis- crepancy Sir G. Stokes considered as probably resulting from eddies which rendered the actual motion other than that to which the sin- gular solution referred and not as disproving the assumption. In 1850, after Joule's discovery of the Mechanical Equivalent of Heat, Stokes showed, by transforming the equations of motion — with arbitrary stresses — so as to obtain the equation of (" Vis-viva ") energy, that this equation contained a definite function, which repre- sented the difference between the work done on the fluid by the stresses and the rate of increase of the energy per unit of volume, which function, he concluded, must, according to Joule, represent the Vis- viva converted into heat. This conclusion was obtained from the equations irrespective of any particular relation between the stresses and the rates of distortion. Sir G. Stokes, however, translated the function into an expression in terms of the rates of distortion, which expression has since been named by Lord Bayleigh the Dissipation Function. In 1883 the author succeeded in proving, by means of experiments with colour bands — the results of which were communicated to the Society* — that when water is caused by pressure to flow through a uniform smooth pipe, the motion of the water is direct, i.e., parallel to the sides of the pipe, or sinuous, i.e., crossing and recrossing the pipe, according as Um, the mean velocity of the water, as measured by dividing Q, the discharge by A, the area of the section of the pipe, is below or above a certain value given by KfijDp, where D is the diameter of the pipe, p the density of the water, and K a numerical constant, the value of which according to the author's experiments and, as he was able to show, to all the experiments by Poiseuille and Darcy, is for pipes of circular section between 1,900 and 2,000, or, in other words, steady direct motion in round tubes is stable or unstable according as or >2000 the number K being thus a criterion of the possible maintenance of sinuous or eddying motion. The experiments also showed that K was equally a criterion of the law of the resistance to be overcome — which changes from a * ' Phil. Trans.,' 1883, Part III, p. 935. 42 Prof. 0. Reynolds. On the Dynamical [May 24, resistance proportional to the velocity and in exact accordance with the theoretical results obtained from the singular solution of the- equation, when direct motion changes to sinuous, i.e., when In the same paper it was pointed out that the existence of this; sudden change in the law of motion of fluids between solid surfaces when proved the dependence of the manner of motion of the fluid on a relation between the product of the dimensions of the pipe multiplied by the velocity of the fluid and the product of the molecular dimen- sions multiplied by the molecular velocities which determine the value of fju for the fluid, also that the equations of motion for viscous fluid contained evidence of this relation. These experimental results completely removed the discrepancy previously noticed, showing that, whatever may be the cause, in those- cases in which the experimental results do not accord with those obtained by the singular solution of the equations, the actual motions. of the water are different. But in this there is only a partial explanation, for there remains the mechanical or physical significance of the existence of the criterion to be explained. In the present paper the author applies the dynamical theory to- the motion of incompressible viscous fluids to show — (a.) That the adoption of the conclusion arrived at by Sir Gabriel Stokes, that the dissipation function represents the rate at which heat is produced, adds a definition to the meaning of u, v, 10 — the components of mean or fluid velocity — which was previously wanting ; (6.) That as the result of this definition the equations are true, and are only true, as applied to fluid in which the mean-motions of the matter, excluding the heat motions, are steady ; (c.) That the evidence of the possible existence of such steady mean-motions, while at the same time the conversion of the energy of these mean-motions into heat is going on, proves the existence of some discriminative cause by which the periods in space and time of the mean-motion are prevented from approximating in magnitude to the corresponding periods of the heat motions ; and also proves the existence of some general action by which the energy of mean-motion is continually transformed into the energy of heat-motion without passing through any intermediate stage ; That as applied to fluid in unsteady mean-motion (excluding; 1894.] Theory of Incompressible Viscous Fluids, $-c. 43 the heat-motions), however steady the mean integral flow may be,, the equations are approximately true in a degree which increases with the ratios of the magnitudes of the periods, in time and space, of the mean-motion to the magnitude of the corresponding periods of the heat-motions ; (e.) That if the discriminative cavise and the action of transformation are the result of general properties of matter, and not of properties which affect only the ultimate motions, there must exist evidence of similar actions as between mean-mean-motion, in directions of mean flow, and the periodic mean-motions taken relative to the mean-mean- motion but excluding heat-motions. And that such evidence must be of a general and important kind, such as the unexplained laws of the resistance of fluid motions, the law of the universal dissipation of energy and the second law of thermodynamics ; (/.) That the generality of the effects of the properties on which the action of transformation depends is proved by the evidence that resistance, other than proportional to the velocity, is caused by the relative (eddying) mean-motion. (' and 1 ; £' denoting the colatitude and q the longitude referred to a plane irough the two poles, it is shown that ,W,H/) cos 7717' = cos m<*[\ uM o • P» (") + 22; — — j «„„• . P (",?'>") cos rq > (. n + r\ ) -\- sin m-i . 2 ^ - - — vmr • P 0 V> '') sin rq, n + rl 4(3 Prof. A. H. Leahy. On certain Functions [May 24, -j wHir.P(n,r,i>) cos rq j- — cos ?w-7 . 2^ - — —' vmr • P (n,r,v) sin rq, n + r I n — ' where P(n,m,u) is the "associated function" — , , • (l — ct/t '" •v are put for cos ft', cos £', 7 is the longitude of the second pole referred to the original pole and plane, and umr vmr are the functions of |3, the angular distance between the poles whose properties are discussed in the paper. When m is zero, i.e., when P(n,m,fi) is a zonal harmonic, the function umr reduces to P(n,r,fi), if ft is put for cos ft, and vmr is zero. The general equations connecting the functions, and the values of the functions for general and for particular values of TO and r are investigated. If f>' and e' are the colatitude and longitude of a point referred to the second pole and any plane through the pole, the integral of the product of any two tesseral harmonics both of the nih order over the surface of a sphere can be expressed concisely in terms of the functions Umr, vmr. The result is (J P (n,m,/u!) cos (iwy'+a) . P (n,m,v) cos (re' + p) dS = — {cos (7717 + «) cos (rc + p) umr(ft) — sin (m/y + «) sin ( if 7 is the longitude of the second pole referred to the plane through the first, e the longitude of the first pole referred to the plane through the second, ft is the angular distance between the poles, and «, p are constants. The functions umr, vmr are connected by several equations, bearing a great resemblance to equations connecting tesseral harmonics of the same order. They are of course functions of. n, and should be written, when n may have different values, in the form u^m ,r,v,i>m ,r, but the n is omitted for brevity in most of the results. Some of the most important results are the following, the dashes denoting differential coefficients — •!*",„,. sin ft + u'mrcosft + {n(n = 2mrvmrcotft ...... (21); «»t,H-i + 2 u'mr— (n + r) (n—r + 1) «mr_j = 0 ...... (24) ; (n—r + 1) um>r-\ = 2mcosec(3vmr ...... (27). 1894.] connected ivith Jesseral Harmonics, with Applications. 47 All the relations connecting u,,,,-, vlltr, &c., are duplicate ones, similar relations being obtained by interchanging u and v. The differential equation satisfied by either function is of the fourth •order, the two functions being different solutions of this equation. 'The two remaining solutions of the equation have also been obtained, and called " functions of the second kind." The equation of finite differences satisfied by the functions is also of the fourth order. The general value of uMr is — 2U,nr — i (^n. ( \ Y .rm (m + r—2k) P (n,m + r—2k, ,m — k + s—l ! m + r—k — 1 ! n + k—s ! r— k+s— l! k— 1 ! m — k ! m + r — k — s ! n — k + s ! r — k \ s — 1 ! k — s ! s ! ^?\ — 2k) n + m ! n — -,-,., o7 ^ P(n,m + r — 2k,ft) *~^~ , ^ —k — lln+r—k—slr— m—k\ m+r—k—HYn—r+k+s ! r— k >-s ! s ! k 1 s-^1 ! , , > — -r-,, -. -f(-l)r. —- P («,m-r,/t), n—m I n + m — r ! •where I(r/2) is the greatest integer in r/2. The value of vmr is given by vmr .sinp = £ ( — : fc=0 -— 2k — 1) P(n,m + r— 2k— 1,«) I ! m+ r — k— 1 ! n+k — s ! r — k + s — 1 ! k ! m—k — 1 ! m + r — k — s — 1 ! n — k + s ! r — k — 1 ! s \ k—s ! s ! w— m! n+m— l ! P( +r_2fe_1 } ^Q m— A; — i ! m+r— /c— s -1 ! »— r + k + s + 1 ! r— A;— T— i UT^l Simpler values for umr vmr are given for general values of m from r = 0 to r = 6 inclusive. The values of the functions umr, vmr are of a simpler form when ft is 48 On Functions connected with Tesseral Harmonics. [May 24r a right angle, and can be expressed by a single series. When n— r is even, the series n— r,n + r w-_r - j } ml ml _ 2 l ? - - -» . ( n + 2m—r—2t\n + r—2t\ is the value of umr( T/2) when m is even, and of vMr(w/2) when m is odd ; the series being continued until one of the factorials in the denominator becomes negative ; and n being supposed greater than 2m. When n is less than 2m, the lower limit of t is m — ^(» + r). A similar series gives the values of «,rtr(7r/2) when m is odd, and °f 2w(""/2) when m is even for the case when n — r is odd. The values of umr (IT 1 2) when m is even and of «Hir(7r/2) when m is odd, are in this case equal to zero. When n—r is even, the values of wwr(7r/2) when m is odd, and of t-WJr(5r/2) when m is even, are also equal to zero. The value of umr is in all cases equal to urm, and the value of vmr is equal to vrm. This result gives several algebraic identities, using general values of umr. Since 7t0,r = P(w,r,yw,), we have by this result umi0 = P(«,m,/t), whence we get the result that f P («,m,;/) cos w/dj = 2 7rP,j (v) . P (n,m,fji) cos 7717. Thus the line integral of a Laplace's function referred to the first pole along a small circle described about the second pole at angular distance £ from it is the value of the function at the second pole multiplied by 27rPn(i>) . sin 8, where v is cos 8. Equations can also be obtained connecting «„_„,.,. and un+im,r, where the ns are different. The most useful result is n (n—m + 1) (n—r+l) un+i>m>r— (2« . .... (45), and a similar equation obtained by interchanging u and v. From this equation a table of the functions for different values of n can be cal- culated, and is given from n =• 0 to n = 4. Since vm,0 = v0im = 0, the- number of the functions for any given value of n is (w + l)2 + w3. Two physical applications of the results are given. The first is an application of the result of a line integral of a Laplace's function re- ferred to one pole along a small circle described about another. The result is employed to establish that the law assumed by Boltzmann and Maxwell for the number of particles which have a given velocity in 1894.] Measurement of the Magnetic Properties of Iron. 49 an irregular system of moving molecules (or a "disturbed gas") is unaltered in form by collisions between the molecules. In the second application the functions are used to find the mutual potential energy of two layers of gravitating matter on two spheres, the density at any point on each sphere being expressed in terms of spherical harmonics referred to fixed coordinates upon it, and the spheres having any position with reference to the line joining their centres. The case of two ellipsoids not differing much from spheres is also worked out numerically, and the stable positions discussed. A stable orbit is possible with the major axes of the ellipsoids constantly in a straight line. If one ellipsoid is fixed and the other projected so as to de- scribe a nearly circular orbit about it, with its major axis initially pointing to the centre of the other, the orbit will be possible if in a plane perpendicular to the least axis of the greater, but the devia- tion of the major axis of the second from the line of centres will con- tain a term which to the first approximation is secular, and may ultimately cause this axis to deviate from its initial position. There are three stable positions for the second ellipsoid if the first ellipsoid is fixed and the centre of the other fixed. These positions will iii general be with the major axis of the second pointing towards the centre of the first, and in a line with the major, mean, and least axes of the first, but if c, the distance between the centres, is so small that / 9 1 1\ /1 9 n ^ ^ / 1 9 *? ^' - / -5 -L J-\0 / 1.6 t O \ o t ~t / 1.4 I O Pf — jr, jj- ) c < I TT,— •— o)ai> or than I— r n\* a.] 3 di 2/ \ai2 ai J 0,1 / \a\ ~ aj 2 i rhere a^a/a/' are the least, mean, and greatest axes of the first sphere, the stable positions will be different. Thus the stable positions fill always be with major axis of the second in the line of centres if 2/ai2 is greater than 7/5. The " functions of the second kind," which are the two remaining- alutions of the differential equation of the fourth order satisfied by vmr, are also briefly investigated. [V. "On the Measurement of the Magnetic Properties of Iron.'' By THOMAS GRAY, B.Sc., F.R.S.E., Professor of Dynamic- Engineering, The Rose Polytechnic Institute, Terre Haute. Indiana. Communicated by Lord KELVIN, P.R.S. Received April 6, 1894. (Abstract.) This paper gives the results of a continuation of the investigation rhich formed the subject of a paper communicated to the Royal Society in 1892, and published in the 'Philosophical Transactions,' VOL. LVI. E 50 Measurement of the Magnetic Properties of Iron. [May 24, vol. 184, A, pp. 531 — 542. The results now given have been to a large extent obtained by the same method, namely, from the curves giving the relation of the current flowing in the circuit to the time measured from the application or the reversal of the impressed E.M.F. on the circuit. In this case, however, the personal element has been eliminated from the curves by the application of the autographic recorder referred to as under construction in the previous paper. This apparatus, which is a modification of the " Thomson siphon-recorder," has been found to work satisfactorily, and has considerably increased the ease and the accuracy with which the curves can be produced. A description of the apparatus and specimens of the curves drawn by it are included in the paper. There is also included in this paper a description of the apparatus and method of experiment in the application of a wattmeter to the determination of the energy dissipated by transformers under E.M.F.'s of different frequency of alternation. The accuracy of the measurements so made were checked by comparison with the results of measurements made by Joubert's instantaneous contact method. The apparatus and method of experiment adopted for the application of this method were to some extent different from those commonly employed, and they are therefore described. The results of some further experiments on the large electromagnet used in the previous experiments, and described in the paper above referred to, are given, but a large part of the results quoted in this paper refer to closed circuit transformers of the types manufactured by the Westinghouse and the General Electric Companies. The experi- ments have been chiefly directed to the following points : — 1. A Comparison of the Total Energy required to produce Different Magnetic Inductions, and the Corresponding Dissipation of Energy. — In connection with this, the effect of air gap in the magnetic circuit has been investigated somewhat more fully. It is shown that, by introducing a moderate air gap, the energy dissipated for a given induction through the coils may be reduced one-third. 2. The Law of Variation of Hysteresis with Variation of Induction. — The experiments indicate that, although for any special case the energy dissipated can be approximately expressed by an equation of the form E = ABa, that both A and a. are different for different kinds of iron. It seems probable, also, from the results obtained, that a. is not absolutely constant for any one iron, but that it increases with increase of B. 3. The Effect of Increased Frequency of Cyclic Variation of Mag- netism on the Dissipation of Energy. — In this investigation a trans- former, the iron case of which was made up of very thin sheets, was used. The thickness of the sheets was about 16-100ths of a millimetre, and the sheets were insulated from each other by means of thin 1894.] Influence of certain Agents on the Tubercle-Bacillus. 51 paper. The full load capacity of the transformer was about 6,000 watts. The range of frequency (including the autographic recorder, the wattmeter and the Joubert's instantaneous contact method experi- ments) was about from 3 per minute to 8,000 per minute. The results indicated that, throughout this range, there is no variation in the dissipation of energy per cycle when, the inductions are equal. Data deduced from these experiments as to the magnetic qualities of the iron used in the different transformers are given in the paper. V. " On the Influence of certain Natural Agents on the Viru- lence of the Tubercle-Bacillus." By AUTHUR RANSOME, M.D., F.R.S., and SHERIDAN DEL^PINE. Received May 1, 1894. Three years ago Dr. Ransome communicated to the Society the results of some experiments, carried out in concert with Professor Dreschfeld, of Owens College, " On certain conditions that modify the virulence of the bacillus of tubercle." The tendency of these researches was to prove " that fresh air and light, and a dry and sandy sub-soil, have a distinct influence in arresting the virulence of the tubercle-bacillus ; that darkness some- what interferes with this disinfectant action ; but that mere exposure to light, in otherwise bad sanitary conditions, does not destroy the virus." The following table gives the results of similar experiments by ourselves. Experi- ment No. 3. 1. 7. 2. 8. 3. 11. 4. 12. 5. 58. 6. 59. 7. Table I. Rabbit inoculated in peritoneum with fresh sputum. Killed 55 days after. Showed well-marked tuberculosis. Rabbit. Sputum exposed to light and air 45 days in June and July. Showed no tuberculosis after 86 days. Rabbit. Sputum exposed in air-shaft in dusk at the same time. Showed slight tuberculosis after 86 days. Guinea-pig. The same sputum exposed at the same time, in air and light, inoculated under the skin. Showed no distinct tubercle in 80 days. Guinea-pig. Same methods, only in dusk. Showed advanced tuber- culosis in 80 days. Guinea-pig. Another sputum exposed in April for 16 days to little or no air, in darkness. Gave well-marked tubercle after 42 days. Guinea-pig. Ditto, ditto. E 2 52 Dr. A. Ransome and Mr. S. Delepine. On tie [May 24, We have now carried the enquiry a little further, and, amongst other objects, have endeavoured to determine how short a period of exposure to air and light would suffice to destroy the poisonous action of the microbe. We selected guinea-pigs as the most susceptible animals to test this question. In the first instance pure cultivations of the bacillus were pre- pared, and were found to be active by frequent inoculations. Small portions of this material were 'spread in a thin layer, upon pieces of sterilized paper. They were arranged in circles of about 2 mm. in diameter, so as to give every opportunity for the action of the elements. They were then exposed in a glass-room, with free access to air and light, z.e.,?close to open windows, for diminishing periods of time, viz., 14, 10, 6, 4, and 2 days respec- tively. Contemporaneous daily records were kept of temperature, maximum and minimum, and of the amount of sunshine taken through the glass roof, by means of one of Negretti and Zambra's sunshine recorders. The following table (Table II) gives the results of the meteorol- ogical observations, but as will be seen presently, only those for the first few days are of importance. Xo result from the other papers ; the control experiments showing that the bacilli used after this date had lost their virulence. Even the results of Experiments 97 and 98 are doubtful on that account, but Experiment 85 was made with a very virulent specimen, as was proved by the inoculation of two other guinea-pigs, with paper in- fected with the same quantity of the same cultivation, and kept the same length of time, but not exposed to sunlight. In both these cases advanced tuberculosis was produced in 44 days. It may be noted that only one of these experiments can be en- tirely relied upon, and that in this case, after 4 days' exposure to air and 12^ hours of sunshine, there was no result from the inocu- lation. These observations, though not in any way conclusive,, are in accord with those of Professor Koch,* and they encouraged us to believe that even short exposures of the tubercle-bacillus, even in sputum, to air and light, might render it innocuous. In the next series of observations, it was determined to allow * Koch, ' Yerhandlungen des International en Medicinisclien Congresses ' (Berlin, 4th to 9th August, 1890), vol. 1, p. 35. Koch says that for some years it has become known that light could kill bacteria. He alludes, no doubt, to the ex- periments of Downes and Blunt, Arloing, Boux, and others. Marshall Ward's experiments with the Bacillus Anthracis are still more recent. Koch had been able to confirm this with regard to the tubercle-bacillus, of which cultivations exposed to sunlight might be killed in a space of time varying from a few minutes to some hours. When exposed to cliff use daylight in a room they were killed in from five to seren days. 1894.] Influence of certain Agents on the Tubercle-Bacillus. 53 94 0J> us °O us rH .-:!•» i— 1 si PlH •4 0US US °O US 1— 1 •)« i-H i-l Oo o ~4et CO rH IN o CO (N ' CO °O5 *Q ^ OS co ^? ~W IN OS US rH 00 0QO CO IN °OS US i— ( IN rH 1-H CO O IN -H US CO O-l a^ bD d O 00 US 0IN CO •In I ^ O "H ^ •> C IN X>» US CO . SS-rj •H O5 OH ^ ^ „; -S US ^ S * ® §» g CM CD O •as CD ]Ml l&IOi. si r§r§^S£^ 00 S °£ g rH * ^ -3 >> ^ COr§ <0 ^ ^ O H o*° O O) US ^ -4* £» 0 S"5 & -~ o '"' £ &1 13 H rH '• o . | ,_| CO C IN IN O us 03 C.' S1^ eo ^ ft ,CD O °05 US cs ^i co ^^ r-^j 1 a 3 3 a g K pj jj i § . S J O 1 ft £ a • 03 -3 3 E | £ o H « a 3 i 54 Dr. A. Ransome and Mr. S. Delepine. On the [May 24, tuberculous sputum to dry (a) in air and light, (6) in air and dark- ness, (c) in a close cupboard. Fresh sputum, rich in bacilli, was obtained and exposed in watch glasses. Specimen (a) was dry in four days ; specimen (6) in eight days ; and specimen (c) in 19 days. Specimens (a) and (6) were closed up as soon as they were dry and kept until specimen (c) was ready, and then portions of the sputum were scraped off the glasses and inoculated into guinea-pigs directly after scraping. Table III gives the results. Table III. No. 117. 1.1 (Sputum (a) inoculated subcutaneously into two guinea-pigs, 118. 2.) (.killed 64 days afterwards, gave well-marked tuberculosis. 119. 3.") f Sputum (b), similarly used, showed no results 120. 4.J 153 days after. 126. 5. Sputum (c) gave well-marked tuberculosis 50 days afterwards. The results of these experiments are somewhat anomalous. The sputum was in rather thick masses and thus dried slowly, and would with difficulty be affected by the natural agents to which they were exposed. This fact would probably account for the continued viru- lence of sputa of 1 and 2, but the immunity from sputa 3 and 4, after eight days exposure to a current of air in darkness, is hardly likely TO be due to this exposure ; we can, therefore, draw no decided con- clusion from this series of experiments. In the fourth series of observations, the sputum was spread upon paper, and was thus more rapidly dried at the ordinary tempera- tures, about 24 hours sufficing. It was then in most cases scraped and thus partly converted into " tuberculous dust " before being exposed to the same conditions as before. In this way it might be expected to be more readily affected by the elements. An attempt was made to measure the amount of air as well as light, an anemometer and a sunshine-recorder being placed near the sputum exposed at the open window. Only a rough guess could thus be made as to the quantity of air passing over the sputum, however, for the papers had to be loosely covered with thin gauze to prevent the " dust " from being carried away by the wind, and the anemometer recorded currents in both directions. Three sets of experiments were made. 1. Papers were placed, to be used for control experiments, in the dark, close cupboard. 2. Papers were placed in the air-shaft of a draught- closet in dim light, pure air only passing through it. 1894.] Influence of certain Agents on the Tubercle-Bacillus. 55 3. Papers were exposed to air and light for three days, February 20, 21, and 22. The rate of air current was about 1,000 ft. per hour, and the sunshine recorded was one hour. Others were exposed for a longer period. The amount of tuberculous dust was so small that portions of the paper were inserted together with it, under the skin. Table IV. No. 151. 191. 192. 160. 170. First Set of Experiments. 1. Sputum kept only one day in a closed, dark cupboard, after drying on paper, produced well-marked tuberculosis in 31 days. 2. Sputum kept under the same conditions, but exposed to a little air for 35 days, produced distinct local tuberculosis in 23 days. 3. Idem. Second Set of Experiments. 1. Sputum kept in the draught closet for three days in a current of air (about 1,000 cubic feet per hour) in darkness, at the ordinary tem- perature, gave well-marked tuberculosis in 32 days. 2. Sputum under exactly the same conditions gave well-marked tubercu- losis in 24 days. Third Set of Experiments. 156. 1. Sputum exposed to light for three days, during which there was one hour of sunshine. Ventilation good. Temperature, maximum 50° and minimum 38° F. No tuberculosis after 46 days. 157. 2. Sputum under the same conditions as the last, except that it had not been- reduced to dust, gave the same negative results after 50 days. 189. 3. Sputum exposed to light for seven days ; 15 hours of sunshine ; brisk ventilation. Temperature, maximum 88°, minimum 29° F. No tuberculosis after 22 days. 190. 4. Sputum exposed to light for two days (after being kept dry for four weeks) ; short exposure to sunshine (not many hours) ; ventilation slight. Temperature, maximum 60°, minimum 22°. No tuberculosis after 22 days. It will be noted that in all the specimens exposed in the dark, tuberculosis was the result, but it must be observed, that in the case of those exposed in the draught-closet, only three days were allowed to pass before they were removed from the influence of the air- current. On the other hand, all the specimens exposed to both air and light, whether for two, three, or seven days, were found to have en- tirely lost their power for evil. The specimen exposed for two days only, had, however, been kept for four weeks before being exposed to these influences, and it had thus lost a portion of its virulence. These researches have an important bearing upon the question of the limits of the infectiveness of tubercle. o<) Mr. J. W. Swan. On some Voltaic Combinations [May 24, It has long been known that the disease is most common in the dirty, ill- drained, ill-ventilated dwellings of the poor, and, even in records intended to prove the contagiousness of phthisis, there are few, if any, of transmission of the disease in clean, well-lighted, well-ventilated houses or hospitals, even those for consumption. Long before Koch's discoveries, and before the disinfection of sputum was practised as it is now, the conveyance of the disease, under these conditions, was recognised by many to be one of the rarest events. If the results that we have obtained with sputum are confirmed by others, as we trust they will be, they will afford some explanation, of these facts. So far as they extend at present, they show (1) that finely divided tuberculous matter, such as pure cultures of the bacillus, or " tuber- culous dust," in daylight, and in free currents of air, is rapidly de- prived of virulence, (2) that even in the dark, although the action is retarded, fresh air has still some disinfecting influence, and (3) that in the absence of air, or in confined air, the bacillus retains its power for long periods of time.* VI. " On some Voltaic Combinations with a Fused Electrolyte and a Gaseous Depolariser." By J. W. SWAN, M.A. Com- municated by LORD RAYLEIGH, Sec. R.S. Received Feb- ruary 28, 1894. It is well known that fused salts behave in many respects like electrolytes in solution, and that voltaic combinations analogous to well-known voltaic cells may be formed with fused electrolytes. The experiments of Brownf have recently illustrated this subject in relation to the Daniell type of cell. For various reasons it appeared to the writer desirable to ascertain the behaviour of a cell with fused electrolyte and a gaseous depolariser, and corresponding in this last particular to the Upward cell. The following is chiefly a record of some of the experiments made in connection with this research. A cell of this kind may be looked at from a theoretical point of view as follows : — A rod of metal, M (fig. 1), is immersed in a fused chloride of the same metal, MCI, and a chemically inactive conductor, C, is also immersed in the fused salt ; when M and C are connected with an electrostatic volt-meter, the metallic chloride is immediately * A portion of the expenses of this research has been defrayed by a grant from the British Medical Association. f ' Roy. Soc. Tr jc.,' vol. 52, pp. 75—91 . 1894.] with a Fused Electrolyte and a Gaseous Depolariser. 57 Fia. 1. polarised, and an E.M.F. calculable from the heats of combination, M/C1 = MCI, is developed. If M and C are metallically connected, a, momentary current passes, but the combination is immediately polarised by the opposing couple, formed by the cathion of the electrolyte M and the pole C. To prevent this polarisation, chlorine has to be supplied at this pole. Complete depolarisation should occur if the pole C consisted of a solid rod of chlorine. This is im- possible, but gaseous chlorine, used as a depolariser, can be made to effect more or less complete depolarisation, and should, theoretically, yield as the result of its heat of combination with lead an E.M.F. of 1'7942 volts. In experiments made with a view to realise as nearly as possible the ideal condition for preventing polarisation the cathode was always molten lead. It was found that hard gas-retort carbon had very little action upon molten alkaline chlorides and on chloride of lead, at the temperature required for their fusion. Carbon was, therefore, employed as the anode or conducting pole in most of the combinations. The electrolyte used was either the molten chlorides of sodium and potassium mixed, or chloride of lead. As there is a continuous form- ation of PbCl2 during the action of the cell, and as it is a good con- ductor, it alone was finally adopted as the electrolyte. As a depolariser, chlorine gas was used. Many experiments were made to find a suitable way of applying the chlorine. The following are details of some of the most suggestive of them. Exp. 1. — A cell was constructed as shown in fig. 2. The arrange- ment consists of an outer iron vessel, with a stratum of molten lead 58 Mr. J. W. Swan. On some Voltaic Combinations [May i FIG. 2. — ci: covering the bottom to some depth, over the lead is a layer of NaCl and KC1 previously fused, into which is immersed the lower and closed end of a carbon tube, which forms the + pole. The mouth of the carbon tube is closed by a fire-clay lid luted on, and through which pass two small clay tubes for the inlet and outlet of chlorine. The whole was heated in a small gas furnace. A binding screw on the iron vessel, which served as a connection with the lead, was used as the negative terminal, and another screw fixed on a copper ring surrounding the carbon tube, served as the positive pole connection. The first trial was made without chlorine. Short circuited through 1,000 ohms the cell developed an E.M.F. of 0'3 volt. A momentary current of more than one ampere was observed when the cell was short circuited through a low resistance ammeter. Chlorine was then passed through the tubes inside the carbon pole, but no depolar- ising effect was observed, even when the chlorine had a slightly higher pressure than the atmosphere, yet the gas passed through the exposed sides of the carbon tube and through the cement at the top. This experiment was repeated several times with carbon tubes of the smallest possible thickness, and always with the same result. It is evident, therefore, that an absorption of chlorine similar to that which takes place in the Upward cell does not occur when a molten electro- lyte of the kind employed in this experiment is used. Exp. 2. — As this method of applying chlorine was unsuccessful, another form of apparatus was adopted. The poles were of the same material as in the previous experiment, but the carbon pole 1894.] ic ith a Fused Electrolyte and a Gaseous Depolarizer. 50 FIG. 3. was an open tube. The electrolyte was a fused mixture of equivalent proportions of NaCl,KCl. A porous pot was introduced in order to separate the products of electrolysis set free at the electrodes. The chlorine gas was supplied through a clay tube, which passed down the centre of the carbon tube. As before, an iron crucible was used as the containing vessel for the fused lead and the electrolyte, it also served as a means of electrical connection with the lead. In relation to the depolarisation effect of the chlorine, which it was the principal object of the experiment to observe, the interfering action of the iron was found by comparison with porcelain to be practically nil ; this no doubt is a consequence of its becoming coated by local action with a film of lead. The whole arrangement was heated in a reverberatory furnace. "When the electrolyte was perfectly fused, the element was short circuited through a volt-meter of 1000 ohms resistance. An E.M.F. of 0'3 volt was observed, the outside current being from the carbon to the lead. This was the E.M.F. after polarisation. A current of chlorine was then passed through the earthenware tube ; rhile the current of gas was slow there was no effect, but when the speed of the issuing gas was increased until the gas passed in bubbles ilong the side of the carbon, alternately surrounding it with chlorine and electrolyte, the E.M.F. rose to 1'25 volt. The action of the cell then similar to a completely depolarised cell. When short circuited through a low resistance ammeter, it produced a steady current of 1*0 ampere for three-quarters of an hour. The potential lifference between the poles was of course very small, while this current was passing, the exterior resistance being very small com- 60 Mr. J. W. Swan. On some Voltaic Combinations [May 24, pared with the interior. When, however, the circuit was opened, it almost instantly rose to 1 "25 volt. So far the experiments showed that, as chlorine is nearly or per- fectly insoluble in fused chloride of lead, or in fused chlorides of sodium and potassium, it is necessary in this case that the surface of the carbon pole on which the cathion is deposited be alternately exposed to the action of the gas and electrolyte. Many experiments confirmed this conclusion. The exposed surface of the carbon tube in this experiment amounted to only 10 or 12 sq. cm. Exp. 3. — As in the arrangement last described, the use of a porous pot and a clay tube was found to be objectionable, through the action of the electrolyte upon them, an arrangement was devised by which the use of the porous pot and tube were avoided. The details are seen in fig. 4. The carbon tube serves as an electrode, and also for conveying the chlorine to the electrolyte. To render it impervious to the gas, it was surrounded by a close-fitting porcelain tube. This tube was closed at the top by a paraffined cork, through which a glass tube in connection with the chlorine supply was passed. The remainder of the apparatus was the same as in Experiment 2, but FIG. 4. • \ B * + ^KGl ^M 'lOir. ~ \-~ — = — — -J *=, L- 1894.J iciili a Fused Electrolyte and a Gaseous Depolariser. (51 without the porous pot. This arrangement gave the following results. The E.M.F. when short-circuited through a voltmeter (300 ohms) gave 1'4 volt, the chlorine entering rapidly. When short-circuited through 1 ohm, it gave a constant current of 0'6 ampere with a P.D. of 0'9 volt. The rather large interior resistance of 1'3 ohm is due not to the electrolyte, but to the greater length of the carbon tube, and bad contacts produced by the corrosive action of the chlorine. No good results were obtained until the chlorine gas bubbled out of the carbon tube, thus realising the conditions before mentioned, as necessary for the production of any large electrical effects. Exp. 4. — With a view to obtain larger effects, another form of cell was tried, as shown in fig. 5. The carbon pole C consisted of a thin rod of electric light carbon, 5 mm. diameter and 15 cm. long. It was passed FIG. 5. through a cork fitted in a porcelain tube, 2'5 cm. diameter. A glass tube bent at right angles was passed through the same cork, to serve for the delivery oi the chlorine. The end of the carbon rod was a little short (about 3 mm.) of the end of the porcelain tube. The containing vessel was a Berlin porcelain crucible, 7 cm. diameter. The conductor from the lead was an iron wire, protected from the action of the electrolyte by a surrounding porcelain tube. The electro- lyte was chloride of lead. The whole was arranged as shown in fig. 5, When the chlorine issued from the porcelain tube, the necessary con- itions for depolarisation were in a large degree realised, the contact tween the electrolyte and the carbon being at times almost broken, <52 Mr. J. "W. SAvan. On some Voltaic Combinations [May 24, and at other times the chlorine forming a nearly complete envelope round the carbon, these conditions following in rapid alternation. The highest E.M.F. observed was 1'25 volt. The largest current given was 0'9 ampere with a P.D. of 0'25 volt. The current was fluctuat- ing, owing to the varying conditions at the carbon pole. Exp. 5. — The following figure shows a construction almost identical with the last, differing only in a few practical details, occasioned by this cell being made larger than the last. The porcelain tube pre- viously used was replaced by an inverted porcelain crucible, having two holes drilled in the bottom ; the larger hole afforded a passage for the conductor from the carbon pole : through the smaller one there passed a porcelain pipe for the chlorine supply. The carbon pole was composed of a disc of gas retort carbon, pierced with holes, as shown in the figure. In the middle of the disc was screwed a tube of carbon, which passed up through the larger hole in the bottom of the crucible and was secured in this position by nuts of retort carbon. The bottom of the carbon tube was filled with fused lead, into which dipped a thick copper wire that formed the positive pole connexion. FIGK 6. The connexions between the carbon and porcelain were luted with a mixture of borax and fire-clay fused at a bright red heat. The other vessel consisted of a short cylindrical plumbago crucible. The fused lead was, as usual, on the bottom, connexion being made with it by means of a heavy open iron ring, with its free end turned up and bent over the rim of the crucible. The internal resistance was small, owing to the improved contact of the leading wire with the carbon 1894.] with a Fused Electrolyte and a Gaseous Depolarizer. 63 pole. The electrolyte was fused PbCl2. The whole was heated in a gas furnace. When the electrolyte was fused, the chlorine was passed rapidly through, so that it issued from under the porcelain crucible. The E.M.F. was then between 0'94 and 0'96 volt., and never rose higher than O98 volt. The lower E.M.F. was evidently due to the fact that part of the surface of the carbon pole was not subject to the action of the chlorine, but remained polarised by deposition of lead. The behaviour was much like that of a constant cell with an E.M.F. of between 0'94 and 0'96 volt. The method of observation was to alter the exterior resistance and then read the current and P.D., then break the circuit and read the E.M.F. Calculated internal resistance. Amperes. 12-0 \ c&lculntccO y> y Ohm. 0-06 100 0-07 4-0 0-08 1-26 0-16 1-75 0-14 2-15 O12 2-50 0-12 oi 1 I From these observations it will be seen that the internal resistance was calculated, col. 3, in order to find whether polarisation is greatest when a small or large current is taken from the cell. From the results it is apparent that the internal resistance, and at the same time the polarisation, decrease when the current increases. This kind of cell, therefore, differs from those in which aqueous electrolytes are sed, inasmuch as the polarisation decreases with increased electrical 'utput. The observations of P.D. and E.M.F. were taken almost imultaneously, and the variation of resistance as the gas bubbles passed out was thus avoided. As the internal resistance was very small the whole time, and remained almost constant during a varia- tion of the current from 1'26 to 2'5 amperes, it may be said to be a constant battery, with an E.M.F. lower than the theoretical value. The reason of this lower E.M.F. is probably due to some part of the large carbon plate being covered with reduced lead, thus forming an opposing couple of smaller capacity and lower resistance than the primary elements, its effect being to reduce the main current. The E.M.F. of this opposing couple is of necessity the same as that of the main current, but, owing to its lower internal resistance, its P.D. is less ; if it were not so, the cell would yield no appreciable current. This reasoning explains why the results obtained with small cells were better than those obtained with large ones. Besides the experiments mentioned, trials have been made, with (>4 Messrs. J. W. Swan and J. Rliodin. Absolute [May 24, more or less success, of many other forms of this combination, includ- ing some in which very porous hollow carbon poles were used, and through which the chlorine was forced, but the effects obtained were less than those recorded. The research has proved that it is possible to form pyro-batteries of the Upward type, although it is extremely difficult to realise the conditions required for effective action. In a future communication I hope to record the results of experiments made, with a view to utilise oxygen as a depolariser in connexion with cells with fused electrolytes. VII. " Measurements of the Absolute Specific Resistance of Pure Electrolytic Copper." By J. W. SWAN and J. RHODIX. Communicated by Lord RAYLEIGH, Sec. R.S. Received February 28, 1894. At the beginning of 1893 it was resolved to make some very careful measurements of the specific resistance of pure electrolytic copper, drawn into wire without previous fusion. Researches made during the latter end of 1892 had shown that the specific resistance of electrolytic copper varies considerably. The resistance of about thirty wires of the same length and diameter, made from specimens of electrolytic copper, prepared in different ways in the laboratory, showed differences of resistance amounting to a maximum of 1'4 per cent., both when in a hard and when in a soft or annealed state, and measured at the same temperature. These preliminary measurements were made by means of a Wheat- stone's bridge, constructed for comparing the unknown resistances of short well-conducting wires with the resistance of a standardised platinoid wire, according to Thomson's method. The accuracy ob- tainable by this method was 0'25 per cent. The best specimens of wire were subjected to a further and still more accurate examination. The measurements of the specific resistance and temperature co- efficient of one of these wires, and of some wire made from the same copper, after undergoing a second electrolytic refining, form the subject of this paper. It was resolved to make measurements giving an ultimate accuracy of O'l per cent. As they were intended to be absolute, the first problem was the determination of the exact dimensions of the wires to be measured. The measurement of the length was made by means of direct comparison with a standard metre rule; that of the diameter was determined by the specific gravity method, which consists in finding the absolute weight of a known length of wire and its density or unit volume weight as determined from its specific gravity, and then calculating its average diameter. 1894.] Specific Resistance of Pure Electrolytic Copper. 65 In the determination of the specific gravity, both the hydrostatic balance method and the picnometer method were used. The latter method was found to give more accurate results. A point of great importance was the estimation of the temperature of the specimens during measurement. To obtain as great accuracy as possible in the temperature readings, an apparatus similar to a calorimeter was employed for enclosing the coil of wire whilst it was measured. A standard thermometer divided in tenths of degrees centigrade was used, placed in the vessel containing the sample of wire. It was read at a distance by means of a telescope. The arrangement is represented by fig. 1. It consists of three FIG. 1. cylindrical tinned iron vessels arranged concentrically one inside the other. The section of the two outer vessels (fig. 1) is distinguished by the shading from that of vessel 3. The vessels formed three water- VOL. LVI. F 6ti Messrs. J. "W. Swan and J. Rhodin. Absolute [May 24, tight compartments. Jackets 1 and 2 were filled with water, and the vessel 3 with paraffin oil having a flashing-point of 150° C. To stir the liquids, three circular rings of iron wire were enclosed, one in each of the compartments. These rings were suspended on the iron wires, sl, s2, &c. The iron wires themselves were fixed on a bar of brass, &. By means of the crank of an electromotor (?jt) and strings and pulleys, the bar b could be moved up and down, stirring each of the liquids simultaneously. To make the motion easy a balance weight, W, was used. The coil to be measured was enclosed in com- partment 3. This compartment was closed at the top by a hollow lid of tinned iron, pierced with holes to allow for the passage of the stirring rods, &c. This lid effectually protected the paraffin oil from surface cooling. The two holes, h and h', made to allow the wires connecting the ends of the coil to pass out, were lined with ebonite to prevent contact with the metal of the lid. The thermometer, T, was let down through a tube in the middle of the lid. As a proof of the effectiveness of the arrangement, it may be mentioned that a small Bunsen burner when burning at its full power, and placed under compartment " 1 " raised the temperature of the inner one " 3 " only 0'1° C. in thirty seconds, a length of time more than sufficient for making a resistance determination. When the temperature in "3" tad been raised to 92° C. and then allowed to cool (being constantly stirred), the temperature (in "3 ") only fell to 40° C. in twenty-four hours, notwithstanding the temperature of the laboratory was only 15° C. Another important detail was an arrangement for securing the wires whilst they were measured. Fig. 2 represents this. It consists FIG. 2. of a piece of ebonite tube 5 cm. diameter, with a deep double screw thread cut on the outside. It was pierced all over with large holes 1 cm. diameter, to allow the paraffin oil to freely circulate in the inside, where the thermometer was inserted. A short rod of ebonite (p) was put through and across the cylinder at the bottom. The wire to be measured was bent round this cross rod, so that equal 1894.] Specific Resistance of Pure Electrolytic Copper. 67 lengths were hanging dbvvn. The double bent wire was then wound up in the double screw thread, and secured at the top by means of string ; the influence of self-induction was thus avoided. Each wire had four terminals, the main current terminals, and the shunt ter- minals, which were soldered to pieces of stout copper, the distance between which determined the length of the wire under examination. These reels saved the wires from being hardened or otherwise in- jured. The length of the wire was taken both before and after the electrical measurement. In some earlier experiments when mica strips were used for coiling the wire upon, great differences were observed. When the ebonite reels were employed hardly any differ- ence could be observed. The electrical measurements were made by the fall of potential method, refined as much as is possible. A D'Arsonval galvanometer was employed as an indicator of potential difference. The " dead beat " property of this instrument is an advantage which in this class of measurement cannot be over-rated. Finding in the first experiments with the D'Arsonval galvanometer, that its sensitive- ness was not sufficient for our purpose, the upper suspension was lengthened very considerably. The coil was made to hang on the upper wire, the wire below the coil being left slack. By this means the required degree of sensitiveness was obtained. The galvanometer readings were made by means of a telescope and scale, a plane mirror being attached to the galvanometer coil. To gain the advantage of using very small deflections, the scale and galvanometer were widely separated, the distance between them was 8 m. The largest deflec- tions used were 600 divisions of the scale (each division = 1/30 in.), corresponding to (at that distance from the mirror) an angle of 1° 49', which falls within the limits for : — (1.) tan 2 angle = 2 tan angle + O'OOl x 2 tan angle. The readings of the deflections were found to be proportional to the P.D. by direct and carefully made measurements, they were always read in two directions, and the sum taken as the deflection ; the accuracy was thus doubled, and possible error by displacement of the zero point avoided. The limit of sensitiveness of the galvanometer was : — 1 scale division deflection = g-oo'oVoo'o amPere (approximately), or = 0-00000003101 ampere (accurately). The resistance of the suspended coil was 10 ohms. The time re- quired for the coil to come to rest was about two seconds. The accompanying diagram shows the arrangement for making the measurements. They were made by direct comparison between a standard ohm and the various coils to be measured, and in such a manner that error due to the resistance of the leads was eliminated. r 2 68 Messrs. J. W. Swan and J. Rhodin. Absolute [May 24-, FIG. 3. -8 metres The following is a description of the details : — B was a single lead and lead peroxide element, of a comparatively large capacity (30 ampere hours). It was found necessary to have the capacity of the cell tolerably large to avoid the necessity of making too large corrections for the fall of potential in the cell itself between different observa- tions. By means of a mercury contact key, K, the battery could be short-circuited through a nickel wire resistance r (of about 6 ohms), the standard ohm B and the resistance to be measured X. All these resistances were in series. (The current in the main circuit was only O2 or 0'25 ampere, and did not produce any sensible heating of the ,various resistances.) By means of the current reverser, C (a mercury switch), the current direction could be easily changed. The galva- nometer, Gr, was connected with another mercury switch, Ci, by means •of which it could be put in shunt to the standard ohm and X alter- nately in rapid succession. In the galvanometer circuit the resist- ance box, BB, was inserted in order to render it possible to make the deflections nearly equal when X differed considerably from R. 1894.] Specific Resistance of Pure Electrolytic Copper. 69 The method of making a determination was as follows : — 1st, C^ was placed in the position which made the galvanometer circuit a shunt of the standard ohm. RB was adjusted so as to make the de- flection of a convenient amount by placing in a resistance Rr (Rr having been determined by means of preliminary experiments). Then K was pressed down and a deflection of the galvanometer " a " ob- served. The current was then reversed, and a deflection " b " observed, (a + fc) wa.s then placed in the observation table under the heading GR. The key opened. The temperature of B observed and noted (TB). 2nd, Ci was then placed in the position which made the galvanometer circuit a shunt to X. RB adjusted to make the deflections of a suitable size by putting in a total resistance of Rx (determined by previous experiment). Then K was pressed down, and the deflection al observed, the current was then reversed, and a deflection bi observed, (a^ + bi) was then put down in the table under the heading Grx. The temperature of X observed and noted (Tx). 3rd observation. No. 1 repeated. 4th „ „ 2 A set of measurements like these were made in 20 seconds, thanks to the dead-beat quality of the galvanometer, and the easy manipula- tion of the mercury switches. It is evident that, if the galvanometer deflections were exactly proportional to the current, and if the resistance of the galvanometer circuit ivas sufficiently high to be neglected, and also if the potential difference of the accumulator did not alter between observations 1 and 2, the resistance X can be put CM - G5 S = resistance of the galvanometer + resistance of galvanometer leads. The proportionality of the galvanometer deflections with the limited range employed was, as previously stated, experimentally proved. The conductivity of the galvanometer as a shunt was negligible, as is evident from the fact that it was more than 2000 ohms resistance (when a minimum), and then diminished the total resistance of the main circuit about 1/8000 ohm (it was shunted round ^ ohm), the resistance of the main circuit amounting to 8 )hms. The fall of potential of the accumulator was very seldom appreciable during one set of measurements. This will be seen from the identity between observations 1 and 3 and 2 and 4 respectively in the tables. When appreciable a correction was applied. When all the arrangements were completed, a determination of the specific resistance of the best specimen of copper (marked " A ") was made. Origin of the Copper. — The copper was deposited in a large rocking 70 Messrs. J. W. Swan and J. Rhodin. Absolute [May 24, ^ -+j cS £* a> ^H a cS E-t be fl e o o 55 '5 o> OH CO .0 eB H il- 11 rH (N O5 o S 9 i— 1 9 C5 CO CO >0 O5 OJ SD 00 X X X S *J o ,3 . » S »c CO X X (N "_£3 bJD ^* ac MM F. O V3 'S '5 CO O5 CO £u ^ C ,_( (N •- 1 g ? til O rH O i— ( O rH O ^O 03 « C^ r_| CD * ^hi* c3 a w rH 1C 'S " •*•" S T1 J>- O5 CD ^ .2 + So o M i— 1 M 05 X (N . CO 05 -*j • QQ ^P X CM f| 3 'co a O co 2 rH CN) Tjt ' M sc o rH 05 (N (N C » ss g rH kC f-H E S >c KO kO lO c! (N !N IN 51 , So ^ " (N N ~ «_. o) to CO O CO o X *- .J3 S a «d C5 CO .2P " * d ip 1C K fi j> "" -i- So ^ rH 1C 05 CD CM S5 d J .-as o *2 IH o C j£ g rC C f~ '. 2Sj •T3 CD C3 'o o & -t> o 0 ^3 3 r~^ M |pd' "o 4> ft rH 1-1 -S c _o *o a 0 a O ^^^ •S ™ "S ' "o 2 ••Jj •^3 ^J V J s J •^ S W rH 1L" O CS fi o i— i hH M 1— ( 1— 1 rH 1894.] Specific Resistance of Pure Electrolytic Copper. 71 tank from ordinary sulphate of copper solution prepared from pure crystallised sulphate of copper, pure sulphuric acid, and distilled water. The anode was a large plate of ordinary commercial electro- lytic copper, and the cathode was a large polished plate of rolled copper. Before placing the cathode in the bath it was silvered by rubbing it over with a solution of cyanide of silver in potassium cyanide. This coating of silver was converted into iodide of silver by means of a solution of iodine in potassium iodide. As is well known, this treatment renders the stripping of the deposit from the cathode an easy matter. On this cathode copper was deposited to a thickness of 2'5 mm., and then the deposit was stripped off. Several deposits were made and tested roughly, as stated in the beginning of this paper. The best of them was one marked " A." Preparation of the Wire. — A strip was cut from the deposited sheet of copper, filed round, and then drawn through sapphire dies to a diameter of approximately 0'02 in. These determinations were made by means of the hydrostatic balance principle. A re-determination of the specific gravity of the copper I in the above table was made by means of a picnometer. Picnometer + distilled H20 at 15° C. = 91-2878 grams. „ + specimen + „ „ = 123'3562 „ Specimen = 36'1012 „ . ' . weight of displaced H20 = 4'0328 ,. Specific gravity — 8'9519. As is seen from the above numbers, the specific gravity of copper when it is pure varies very little with hardness and other conditions, the variations when at a maximum only amounting to 0'0004 of the Avhole. The mean of the above results may, therefore, be taken as the specific gravity of this copper at 15° C. The mean is 8'9oll = the specific gravity. This value is not the one required for the calculation of the dimen- sion of wires ; what is required is the density or absolute weight of an ideal cubic centimetre of the metal at 15° C. The weight of 1 c.c. of water at 15° C. is, according to Kohlrausch, ' Praktische Physik,' 0-99915 gram. The specific gravity of the copper divided by this figure gives the density : — Weight of 1 c.c. of copper at 15° C. — 8'9587 grams (8'959). Specific Resistance of Deposit "JL " (hard drawn) at different Tem- peratures "between 12'9° C. and 90'2° C. — A hard-drawn wire of the A deposit was measured by the apparatus described. For determining 72 Messrs. J. W. Swan and J. Rhodin. Absolute [May 24, the diameter a piece 300 cm. long was taken. The weight was found to be — 5'6009 grams. The ascertained density of the copper being 8*959 indicates an average diameter of — (3.) X = 2 5-6009 300 x 8-9597T = 0-05151 cm. at 15° C. This same wii'e was fitted with shunting terminals 250 cm. apart, and wound on one of the previously described reels, and then placed in the circuit. A large number of observations were made at different temperatures : they are arranged in Table II. The following are the headings of the columns : — Tx = temperature of the unknown resistance ; Gx = deflection (proportional to the current) with, the unknown resistance ; GB = deflection when standard ohm in circuit ; TR = temperature of standard ohm ; Gx = deflection of unknown resistance (supposing GB to be always 1240) ; Gx (6th column) is deflection with unknown resistance (X) in circuit corrected, so as to compare with GB at 1240 and TR at 15° C. ; the figures in this column multiplied by a constant give the resistance of the wire. Table II. Tx. Gx. G1{. (GB ="\240). T*. (GB = 1240 TII = 15° C.) C°. Ce. 12-9 1006 1250 998 13-5 997 16-0 1017 1249 1010 13-7 1010 17-2 1018 -5 1245 1014 13-9 1014 18-15 1023 -5 1246 1019 1019 19-6 1027 . 1245 1022 1022 20-3 1030 -5 1244 -5 1027 1027 20-9 1033 1244 1030 13'-8 1030 22-6 1039 1244 1036 1036 23-3 1043 1244 5 1039 ll'-o 1038 24-2 1045 -5 1243 1043 1042 25-2 1048-5 1242 1047 1046 26-2 1052 1242 1050 1049 27-3 1056 1211 1055 1054 28-4 1058 -5 1239 -5 1059 12" 5 1058 29-4 1060 1238-5 1061 1060 30-2 1065 1240 1065 1064 31-1 1069 -5 1239-5 1070 12>-1 1069 32-2 1073-5 12;<9 1074 1073 33-2 1077 1238 -5 1078 1077 34 2 1080 1237 -5 1082 12'-5 1081 35-3 10S3 1237 1086 1085 36-2 1086 -5 1236 1090 1089 37-2 10SO-5 1237 1093 ., 1092 1894] Specific Resistance of Pure Electrolytic Copper. 73 Table II — contimied. Tx. Gx. GK. (GB = 1240). TR. (GK = 1240 J?a =15 Cj. 38-2 1093 1236 1097 12-5 ' 1096 39-2 1097 1236 1101 1100 40-1 1101 1237 1104 ij 1103 41-2 1106 1237 1109 1108 42-2 1108 1235 1112-5 1111 43-5 1111 1234 1116-5 • 1115 44-2 1114-5 1234 1120 1119 45-2 1119-5 1235 •1124 • 1124 46-2 1123 j 1128 sj 1127 47-2 1128 i 1133 . 1132 48-2 1133 1236 1137 M 1136 50-2 1140-5 1144 12-8 1143 51-2 936-5 1013 1146 14-5 1146 52 939-2 „ 1150 )} 1150 53-1 943 . 1154 • 1154 54-1 945-2 1012 -7 1157-5 1157-5 55-2 949 1013 1162 1162 56-2 951-5 1012-5 1165 15'2 1165 57-2 955 1012 -5 1169-6 5> 1169 -6 58-2 958 1013 1173 1173 59-2 960-5 1011 -5 1177 • 1177 60-2 963 1011 1181 1181 61-2 966-5 1011 -5 1185 15-5 1185 62-2 969-2 1011 1189 1189 63-2 973 1193 ri 1193 64-2 976 >; 1197 1197 65-2 979 V 1200 „ 1200 66-2 982-5 1012 1204 ); 1204 67-2 985-2 1011 1207 -4 1 1207 -4 68-2 988-5 1011 1212-4 1212-4 69-2 991 1215-5 1215-5 70-2 994 M 1219 1219 71-2 997-5 1010 1225 1225 72-2 1000 1228 | 1228 73-2 1003 ^ 1231 -5 , 1231-5 74-2 1006 1235 1235 75-2 1010-5 M 1240-5 ( 1240 '5 76-2 1013 1244 124 1 77-2 1017 M 1249 1 1249 78-2 1020 M 1252 } 1252 79-2 1023 -2 1256 ( 1256 80-1 1027 1011 1260 1260 81-2 1031 1010-5 1265 j, 1265- 82-2 1034 1010-5 1269 1269 83-2 1037 1011 1272 ' 1272 84-2 1039 -5 1010 1276 1276 85-2 1043 1010 1280-5 1280 -5 86-2 1047 1010 1285 1285 87-2 1050 1009 -5 1290 1290 88-2 1052 1009 1293 1293 89-2 1056 1009 1298 )? 1298 90-2 1060 1009 1303 » 1303 74 Messrs. J. W. Swan and J. Bhodin. Absolute [May 24, The absolute specific resistance in C.G.S. units was calculated from the above numbers as follows. The resistance in ohms of the measured wire -at any temperature is found by using Equation No. 2 :— From the definition of the B, and that of specific resistance in C.G.S. units, formula 5 is deduced: — (5.) . = X.^109, where r = radius of the wire in centimetres. Z = length „ „ „ If the value of X as given in Equation No. 2 is substituted in Equation No. 5, the following is obtained : — (6.) a = < In the previous table if Gx is read as in the last column, the follow- ing will be the values for the various elements of Equation No. 6 : — RB = 8000 B.A. units for all observations. Bx = 2000 „ S = 17-3 „ B = 1 ohm. GK = 1240. Scale divisions for all observations. r At 15° C. But, according to Mat- 2r = 0'05151 cm. J thiessen, no correction is made I = 250* „ for the variation of these con- *- stants. In Equation No. 6 everything is constant except a and Gx, and the equation results in the following, when the numerical values as above are substituted for the general expressions : — (7.) a = 1-6906 log 1-6906 = 0-2280258. The specific resistance in C.G.S. units of the hard-drawn rocking tank A deposit at any temperature can therefore be calculated by making use of the values given in the above table. In order to find the specific resistance at 0° C. it is necessary to calculate the tempera- ture coefficient (A^) from the above observations. The method of 1894.] Specific Resistance of Pure Electrolytic Copper. 75 least squares is the one which naturally suggests itself in calculating this coefficient, but the calculation would be too laborious in com- parison with the value of the figure arrived at. Therefore, in order to ascertain which values could be used for calculating the temperature coefficient (A<), the above observations were plotted in the usual way, the co-ordinates being x = temperature arid y = specific resistance in C.G.S. units. As previously stated, the values in column 6 of Table II have to be multiplied by 1'6906 to give the specific resistance in C.G.S. units ; to save labour the multiplication was done graphically (Graphic Table No. 1). The resistances in column 6 of Table I reduced to C.G.S. units by means of the graphical Table 1 are tabulated in another table (Table III). The graphic Table No. 2 is a reproduction of Table III, and gives the results of the above measurements. As the values formed so very nearly a straight line, two only are selected : — 1707 C.G.S. units at 16-0° C. 1937 „ „ 5V2° C. These gave the following equations : — i707ss«(l+16y) 1937 = ff(l + 51-2y), from which .7- = 1603- C.G.S. units y = 0-004077. x is the specific resistance of the A deposit (hard drawn) at 0° C., and y is the temperature coefficient. By means of these values an extrapolation was made from 0° to 16° C. in the graphic Table No. 2. 76 Messrs. J. \V. Swan and J. Rliodin. Absolute [May 24, * "Q 1 <3 1 t>2 § 5 o— ojio 2 i r ^ I f 4 I Graphic Table 2. Spec, resistance of Rockingtank A. copper in C.G.S. units. oo i6K> 16201630 16« J50 1660 too i6ao 1690 wo mo reo 1730 1740 iTM 1760 i77o i78o 1790 BOO mo mww\mwmmmmwmw\m\wmmwmzmmivmmmwmimti»i)mmii / t t f f t I f (f f / jf lows the result of the measurements of the specific resistance of Rocking lank A copper ( observations reduced to the C.G.S. system. f f , : 2 / t K 1 / r y / r / / J / t i / [ / / 2 | 1 1 / j / / j2 - s 1 1 I h ; c , : ii 3 <: , t n c > •: i L PI T "f 5 ^ c? if i n /*c d ^ *g / 02 78 Messrs. J. W. Swan and J. llhodin. Absolute [May 24, Table III. Rocking Tank A Copper. Hard drawn. Specific Resistance. Tempera- ture. C.G.S. units. Tempera- ture. G'.G-.S. units. Tempera- ture. C.G.S. units. 16-0 1707 47-2 1913 79-2 2123 17-2 1713 48-2 1920 80-1 2130 18-1 1722 50-2 1932 81-2 2138 19-6 1727 51-2 1937 82-2 2145 20-3 1735 52-0 1944 83-2 2150 20-9 1741 53-1 1951 84-2 2157 22-6 1751 54-1 1957 85-2 2164 23-3 1754 55-2 1964 86-2 2172 24-2 176L 56-2 1970 87-2 2180 25-2 1768 57-2 1978 88-2 2186 26-2 1773 58-2 1983 89-2 2194 27-3 1783 59-2 1990 90-2 2203 28-7 1787 60-4 1996 29-7 1792 61-2 2003 30-2 1799 62-2 2010 31-1 1807 63-2 2016 32-2 1614 64-2 2023 33-2 1821 65-2 2029 34-2 1827 66-2 2035 35-3 1834 67-2 2041 36-2 1841 68-2 2049 37-2 1847 69-2 2055 38-2 1853 70-2 2061 . 39-2 1860 71-2 2070 40-1 1865 72-2 2076 41-2 1873 73-2 2082 42-2 1878 74-2 2088 43-5 1885 75-2 2097 , 44-2 1892 76-2 2103 45-2 1900 77-2 2111 , 46 '2 1906 78-2 2117 Measurements with Rocking Tank A Copper, Soft. A similar wire to that used in the previous measurements was annealed in a tube of hard glass, through which was passing a current of dry CO2 gas, in order to prevent oxidation. The following measurements were made with it : — Density at 15° C., 8'959 (see previous table). Absolute weight of 300 cm. of the wire, 5'5746 grams. Diameter. /5-5746 X = 2r = 2 A/ ^ 8-959 x 300 = 2 X °'025694 = 0'051388 cm. 1894.] Specific Resistance of Pure Electrolytic Copper. 79 250 cm. of this wire were compared with the standard ohm, according to the method described. The following are the observa- tions : — Table IV. Gx TX- GX- o*. (GB = 1240). TR. (GB = 124° TR = 15°C.). 16 -8° C. 820- 1023 994 20° C. _ 819-5 1023 993 , 994 19-95 829- 1020 1008 — . M 828-5 1020 1007 : 1008 48-0 916- 1019 1114 , — , " \ ,, 916- 1019 1114 , 1115 • From the numbers in column 6 the values of the specific resistance of this sample can be obtained by multiplication with the constant 1*6826 obtained in the same manner as the value for the hard variety, the difference being due to the difference in the diameters. All the other constants of the measurements were the same. The results by calculation give the following values : — Specific resistance of rocking tank A copper, annealed in C02 gas:— At 16-8° C. = 1672-4 C.G-.S. units. „ 19-95 = 1696*0 „ „ 48-0 = 1876-0 „ The temperature coefficient, calculated from the following for- mula : — gives the value 0*00418 at ordinary temperatures. Applying this value of x to the amount of specific resistance at 16*8° C., the following is obtained as the specific resistance of this special sample of annealed rocking tank A deposit : — gelser. Meddelelser om Gr<}>nland. Hefte 3. Fortsaettelse 1 — 4. Hefte 7—13. 8vo. Kjfbenhavn 1887-94. The Commission. Pulkowa: — Observatoire Central Nicolas. Observations de Poul- kovo. Vol. X. Folio. St. Petersbourg 1893; Publications. Serie 2. Vol. I. Folio. 8t. Petersbourg 1893; Russische Expeditionen zur Beobachtung des Venusdurchgangs, 1874. Abth. 1. 4to. St. Petersburg 1891 ; Tables Auxiliaires pour la Determination de 1'Heure par des Hauteurs Correspondantes de Differentes Etoiles. Construites par Th. Wittram. 8vo. St. Petersbourg 1892. The Observatory. Sydney : — Observatory. Meteorological Observations. November — December, 1893. 8vo. \_Sydney.~\ The Observatory. Washington : — U.S. Department of Agriculture. Monthly Weather Review. February, 1894. 4to. Washington; Report of the Ohio Weather and Crop Service. March, 1894. 8vo. Nonoalk, Ohio. The Department. Journals. Acta Mathematica. Bande XI — XVIII, Heft. 1. 4to. Stockholm. 1888-94. The Editor. Annaes de Sciencias Naturaes. Anno 1. No. 2. 8vo. Porto 1894. The Editor. Astronomy and Astro-Physics. May, 1894. 8vo. Northfield, Minn. The Editor. Epigraphia Indica of the Archaeological Survey of India. Vol. II. Part 14. Folio. Calcutta 1893. Revenue and Agricultural Department, Calcutta. Journal of Comparative Neurology. April, 1894. 8vo. Granville, Ohio. The Editors. Stazioni Sperimentali Agrarie Italiane. Vol. XXVI. Fasc. 3. 8vo. Modena 1894. R. Stazione Agraria, Modena. G 2 84 Lord Kelvin and Mr. M. Maclean. [May 31 T Journals (continued). Zeitschrift fur Naturwissenschaften. Bd. LXVX Heft 5 — 6. 8vo. Leipzig 1894. Naturwissenschaftlicher Verein, Halle, Dollen (W.) Stern-Epbemeriden auf das Jahr 1894. 8vo. Dorpat 1894. The Author. Girouard (D.) Lake St. Louis Old and New Illustrated and Cavelier de La Salle. 8vo. Montreal 1893. The Author. Hunt (A. R.) On certain Affinities between the Devonian Rocks of South Devon and the Metamorphic Schists. 8vo. London 1892 ; [and three other pamphlets. 8vo.] The Author. Hutchinson (Rev. H. N.) Creatures of other Days. 8vo. London 1894. The Publishers. Marsh (O. C.) Restoration of Elotherium. 8vo. [New Haven~\ 1894. The Author. May 31, 1894. • The LORD KELVIN, D.C.L., LL.D., President, followed by Sir JOHN EVANS, K.C.B., D.C.L., LL.D., Vice- President and Treasurer, in the Chair. A List of the Presents received was laid on the table, and thanks ordered for them. The following Papers were read : — I. " On the Electrification of Air." By LORD KELVIN, P.R.S.T and MAGNUS MACLEAN, M.A., F.R.S.E. Received May 9, 1894. § 1. That air can be electrified either positively or negatively is obvious from the fact that an isolated spherule of pure watery electrified either positively or negatively, can be wholly evaporated in air.* Thirty-four years ago it was pointed out by one of * This demonstrated an affirmative answer to the question, Can a molecule of a gas be charged with electricity ? ( J. J. Thomson, ' Kecent Eesearches in Electricity and Magnetism,' § 36, p. 53) and shows that the experiments referred to as point- ing to the opposite conclusion are to be explained otherwise. Since this was written, we find, in the ' Electrical Review ' of May 18, on page 571, in a lecture by Elihu Thomson, the following : — " It is known that as we leave the surface of the earth and rise in the air, there is an increase of positive potential 1894.] On the Electrification of Air. 85 us* as probable that in ordinary natural atmospheric conditions, the air for some considerable height above the earth's surface is electri- fied,f and that the incessant variations of electrostatic force which he had observed, minute after minute, during calms and light winds, and often under a cloudless sky, were due to motions of large quantities of positively or negatively electrified air in the immediate neigh- bourhood of the place of observation. § 2. It was proved]; by observations in the Old College of Glasgow University that the air was in general negatively electrified, not only indoors, within the old lecture room§ of Natural Philosophy, but also in the out-of-doors space of the College Court, open to the sky though closed around with high buildings, and between it and the top of the College Tower. The Old College was in a somewhat low situation, surrounded by a densely crowded part of a great city. In the new University buildings, crowning a hill on the western boundary of Glasgow, similar phenomena, though with less general with respect to the ground It is not clearly proven that a pure gas, rarefied or not, can receive and convey a charge. If we imagine a charged drop of water suspended in air and evaporating, it follows that, unless the charge be carried •off in the vapour, the potential of the drop would rise steadily as its surface dimin- ished, and would become infinite as the drop disappeared, unless the charge were dissipated before the complete drying up of the drop by dispersion of the drop itself, or conveyance of electricity by its vapour. The charge would certainly require to pass somewhere, and might leave the air and vapour charged." It is quite clear that " must " ought to be substituted for " might " in this last line. Thus the vagueness and doubts expressed in the first part of the quoted statement are annulled by the last three sentences of it. * " Even in fair weather the intensity of the electric force in the air near the •earth's surface is perpetually fluctuating. The speaker had often observed it, especially during calms or very light breezes from the east, varying from 40 Daniell's elements per foot to three or four times that amount during a few minutes, .and returning again as rapidly to the lower amount. More frequently he had •observed variations from about 30 to about 40, and back again, recurring in uncer- tain periods of perhaps about two minutes. These gradual variations cannot but be produced by electrified masses of air or cloud, floating by the locality of observa- tion."— Lord Kelvin's ' Electrostatics and Magnetism,' art. xvi, § 282. t The out-of-doors air potential, as tested by a portable electrometer in an open place, or even by a water-dropping nozzle outside, two or three feet from the walls of the lecture room, was generally on these occasions positive, and the earth's surface itself therefore, of course, negative — the common fair weather condition — which I am forced to conclude is due to a paramount influence of positive elec- tricity in higher regions of the air, notwithstanding the negative electricity of the air in the lower stratum near the earth's surface. On the two or three occasions when the in-door atmospheric electricity was found positive, and, therefore, the surface of the floor, walls and ceiling negative, the potential outside was certainly positive, and the earth's surface out-of-doors negative, as usual in fine weather." — Ibid., § 300. $ Ibid., Q. 2, § 283. § Ibid., §§ 296—300. 86 Lord Kelvin and Mr. M. Maclean. [May 31r prevalence of negative electricity in the air, have been observed, both indoors, in the large Bute Hall, and in many other smaller rooms, and out-of-doors, in the court, which is somewhat similar to the courts of the Old College, but much larger. It is possible that the negative electricity found thirty years ago in the air of the Old College, may have been due to its situation, surrounded by houses with their fires, and smoking factory chimneys. In the New College much of the prevalence of negative electricity in air within doors has, however, been found to be due to electrification by the burning lamp* used with the quadrant electrometer ; and more recent obser- vations, with electrification by flame absolutely excluded, throw doubt on the old conclusion, that both in town and country negative electrification is the prevailing condition of natural atmospheric air in the lower regions of the atmosphere. § 3. The electric ventilation found in the Old College, and de- scribed in §299 of "Electrostatics and Magnetism," according to which air drawn through a chink, less than ^ in. wide, of a slightly open window or door, into a large room, showed the electrification which it had on the other side of the chink, whether that was the natural electrification of the open air, or positive or negative electri- fication produced by aid of a spirit lamp and electric machine in an adjoining room, has been tried again in the new College with quite corresponding results. It has also been extended to the drawing in of electrified air through a tube to the enclosure represented in fig. 1 of the present paper ; with the result that the water-dropping test indicated in the sketch, amply sufficed to show the electrification, and verify that it was always the same as that of the air outside. When the tube was filled with loosely packed cotton-wool the electri- fication of the entering air was so nearly annulled as to be insensible to the test. § 4. The object proposed for the experiments described in the present communication was to find if a small unchanged portion of air could be electrified sufficiently to show its electrification by ordinary tests, and could keep its electrification for any considerable time ; and to test whether or not dust in the air is essential to what- ever of electrification might be observed in such circumstances, or is much concerned in it. § 5. The arrangement for the experiments is shown in the dia- gram, Fig. 1. AA is a large sheet-iron vat inverted on a large wooden tray BB, lined with lead. By filling the tray with water the air is confined in the vat. There are two holes in the top of the vat ; * ' Electrification of Air by Combustion.' Magnus Maclean, M.A., F.R.S.E., and Makita Goto, Philosophical Society of Glasgow, November 20, 1889 ; ' Electrifica- tion of Air by Water Jet,' Magnus Maclean, M.A., F.R.S.E., and Makita Goto, ' Philosophical Magazine,' August, 18CO. 1894.J On the Electrification of Air. FIG. 1. To electric machine 0 10 £0 30 40 50 60 70 80 90 100 ONE METRE FIG. 2. 88 Lord Kelvin and Mr. M. Maclean. [May 31, one for the water-dropper C, and one for the charging wire D. Both the water-dropper, and the charging wire, ending with a pin- point as sharp as possible, are insulated by solid paraffin, which is surrounded by a metal tube, as shown in half size in Fig. 2. To start with they were supported by pieces of vulcanite embedded in paraffin. But it was found that after the lapse of some days, (possibly on account of ozone generated by the incessant brush dis- charges), the insulation had utterly failed in both of them. The vulcanite pieces were then taken out, and solid paraffin, with the metal guard-tube round it to screen it from electrically influencing the water-dropper, was substituted. This has proved quite satis- factory : the water- dropper, with the flow of water stopped, holds a positive or a negative charge for hours. § 6. A quadrant electrometer E (described in " Electrostatics and Magnetism," §§ 346 — 353) was set up on the top of the vat near the water-dropper, as shown in Fig. 1. It was used with lamp and semi- transparent scale to indicate the difference of potential between the water-dropper and the vat. The sensibility of the electrometer was 21 scale divisions (half -millimetres) per volt, and as the scale was 90 centimetres long, difference of potentials up to 43 volts positive or negative, could be read by adjusting the metallic zero to the middle of the scale. A fractional plate-electric machine was used, and by means of it, in connection with the pin-point, the air inside the vat could be electrified either positively or negatively. § 7. The vat was fixed in position in the Apparatus Boom of the Natural Philosophy Department of the University of Glasgow on the 13th of December, 1893, and for more than three months the air inside was left undisturbed except by discharges from the pin-point through the electrifying wire, and by the spray from the water- dropper. Thus the air was becoming more and more freed of dust day by day. Yet at the end of the four months we found that the air was as easily electrified, either positively or negatively, as it was at the beginning ; and that if we electrify it strongly by turning the machine for half-an-hour, it retains a considerable portion of this electrification for several hours. § 8. Observations were taken almost daily since the 13th Decem- ber; but the following, taken on the 8th of February, the 12th of March, and the 23rd of April, will serve as specimens, the results being shown in each case by a curve. At all these dates the air must have been very free from dust. Both daring the charging and during the observations the case of the electrometer and one pair of quadrants are kept metallically connected to the vat. During the charging the water-dropper and the other pair of quadrants were also kept in connection with the vat. Immediately after the charging was stopped the charging- wire was connected metallically to the out- 1894.] On the Electrification of Air. 89 side of the vat, and left so with its sharp point unchanged in its position inside the vat during all the observations. § 9. Curve 1. February 8, 1894. — The friction-plate machine was turned positive for half-an-hour. Ten minutes after the machine stopped the water-dropper was filled and joined to one pair of quadrants of the electrometer, while the other pair was joined to the case of the instrument. The first reading on the curve was taken four minutes afterwards, that is fourteen minutes after the machine stopped running (18 volts.). A. V f '-6. 8s,. V Cur * /.; ffen 9> •Aaro ino positi vely \ du ring half an h our. \Wa ter p it in to n>a ^er^dr opper afx \ s \ r**\ § \ 1 ^ a ^ Sv^-N •^ l>- CO O5 1 lO O5 CO i— 1 «*-> ^ S ' B| O ** CD CO S n3 . g t-i O B a rh »>, o d d •* lO C3 a rtj'O S §1 • 0 H W CS o 0 C^ O 5? *S *» • I-l o j-j N o C 0 •M- O 0 d S '1 -s '"• '1 rH i-l •* 00 co 10 co -H n B t3 1 1 'Kfe g I-H i-H 1 1 1 1 s bo i d ^g O3 c3 K w ® -S .-§ o 8 ID -a 1 l|^ ^ o ^ o o o o Oi C^ O 00 O O b •< -*a ^.2o °^ TJI CD lO J>- ?H i— i _fi a m •s O H i— 5 a 5 * |^-»7 h o .1 05 Tf (N t- W rf. aj es EH "5 01 w 12 S 'M is a '« bc'S S) "a ^ O J> CD O O rH CM 8§ O £ o t- c3 ^ QJ o o o o o o -°_|_ +3 " O 'Q "8 ^ "^ r>j" «fH T3 g ^"® 0 O ^* 02 -*s § -u 4-1 O 00 JB ^> . b X rQ 1C a ^ Hi 00 03 -2 .SI'S C7i !>• -l^* IO CO ^ •g •*" B 0 O I/si IN •* O IO VO CD t- 1> 'S * 2 ^ a ** _H •*J eg ^ QL 0? ^o a, -2 "bTo g b 0 OJ " ^ o be Q gj * ""&) EH P*-w 0) «r5 — 4j3 ^ O CD -t^* t>* O 2 O c3 O i O CO i— 1 rH O CO -0-5 a'S O SJH ° "£ ft feo »»*<.». 00 ». -" o '-S ^ a' 1 Slog § ^ ° ** § - Q> h-3 "^ E^ ^ "S a c! P— c tc4 bo * C ^ C * ^-^g =3 .2 0 bo 106 Mr. H. Tomlinson. The Effect of Mechanical [May 31 O O § o 05 -= EH 2.2 a.Sg S3 O I I I I I I I "o "bO « '« ° Sr3 ^ r> 1894.] Stress, «M •TS 0 *s j§ 1 ^' § .5 ? 0 .E D o °E ^ J« "IH ^ T3 •id 1 •** C -*J . 3 o a § SO * T x * * oj co p 03 •3 -f- cr1 * CO ^ 1 1 Irr oo oo o co co 10 i— i "a °o 0 » 1 8 1 1 1 1 + .no fsl! 1 2 '•§ g> a o « § 2 iS -M. •J g | i | * -s1 a * i Q> g JD PH g (£ a o 'ffl — O -W ^Tr rt D"* 1 i|l g ' I 1 1 o Is a I-H X * # * ^ rH O *O (N !N CO 111 + § co 1 •ii" .•& s illl'8 ic .« ^173 ° CO r-i rj* O o co 00 N o o ^= ^o ^g .-| g + + + + + + •g a *« a o *? 1—1 ys 6 la -^ ^S a | O "^ bD 55 -15 3 'S 0 a M ^ o 1 a •f o 5 o riff! 7 o <— • 1 £ -2 O &l™ 60 ^ X CO •* - co i— i i— i O eo ^•^ • § flj *53 ® o + + 1 + T + 5 § 13 " r2 "o a 00 ^s ' o Jj trH a £ ^ o E S O-H V J3 p "o • r3 O . w ^ o CO T-i Oj •— i i—* i 03 ,a g g o a *> 0 u 0 ID _a a 3 a. "a — -u>" • • S a a a 3 5 8 •§ OQ O> 'S s 03 O QJ PH O O ' O O § " a 1 p. H i '!! 'I IN IO O 5P IN IN CO g .2 o ft W" fc" S * * *"* 108 Propagation of Magnetisation of Iron. [May 31, V. " Propagation of Magnetisation of Iron as affected by the Electric Currents in the Iron." By J. HOPKINSON, F.R.S., and E. WILSON. Received May 17, 1894. (Abstract.) Consider a solid, cylindrical electromagnet, it is well known that, in reversing the magnetising current, the induction does not instantly reverse, but a certain time elapses before it again attains its full value, that it reverses at a later time at the centre of the core than near its surface, and that the delay in reversal near the centre is due to the electric currents induced in the iron. The object of the pre- sent paper is to investigate these effects. The magnet experimented upon had a diameter of 4 inches, and formed a closed magnetic circuit. Through a part of its length the cylinder of 4 inches diameter was formed of an iron core surrounded by two concentric, closely fitting tubes. Exploring coils of fine copper wire were bedded in the iron between the surfaces of the tubes. The currents induced in these exploring coils were observed when the current in the main coil of the magnet was reversed. These currents in some cases last for over half a minute. Inferences can be drawn from these results as to the behaviour of other diameters than 4 inches. Comparing two cylinders of different diameters, similar events occur, but at times proportional to the squares of the diameters of the cylinders. From this consideration and the experiments, a judgment is formed as to the effects of local currents in the cores of transformers and of the armatures of dynamo machines. VI. " On Rapid Variations of Atmospheric Temperature, espe- cially during Fohn, and the Methods of observing them.'' By J. Y. BUCHANAN, F.R.S. Received May 29, 1894. The variation of the temperature of the air in the course of a day is a matter of familiar observation. It depends in the first instance on the relative positions of the locality and the sun. The temperature is generally highest a short time after the sun has attained its greatest altidude above the horizon, and it is lowest some time after it has attained its greatest depression below the horizon. Observations made at regular intervals over the twenty-four hours show a more or less regular rise of temperature during the early part of the day and a similar fall of temperature during the latter part of the day and the evening. When the interval between the observations is 1894.] On Rapid Variations of Atmospheric Temperature. 109 diminished the regularity of the march of temperature is found to diminish also, but the great variability of the temperature of the air is best shown by the curve drawn by a recording thermometer of sufficient sensibility combined with a clock movement of suitable velocity. Such an instrument draws a sinuous line which is generally smooth during the night and serrated during the day. The shape and the crowdedness of the teeth on the serrated daylight portion of the line have a close connection with, and are to a certain extent an indication of, the character of the existing weather. In general the indented character of the daylight curve is an indication of the disturbing in- fluence of the sun on the equilibrium of the atmosphere which continues just as long as he is above the horizon ; after sunset the atmosphere quickly reverts to a state of greater stability. It is obvious, there- fore, that the indented character of the daylight curve indicates not only changes of temperature in the air but also motions and changes o f motion in it. These motions are generally vertical and too subtle and local to be observed with an anemometer. In the course of frequent observations in the open air and under varying circum- stances, I have many times had occasion to remark these rapid oscillations of temperature and at the same time to deplore the difficulty of accurately measuring them. It is principally with the view of directing attention to this instrumental difficulty that the following observations are put together. At the same time, though few in number, they have to do with a very remarkable species of weather, known by its Alpine name of Fohn. It has been most observed in the valleys stretching in a northerly direction from the main sum- mit line of the chain of the Alps and takes the form of an abnor- mally warm wind blowing from the mountains towards the plain. It has largely occupied the attention of continental meteorologists, and more particularly it has been the subject of exhaustive investi- gations by Hann, who has shown by very strong evidence that its high temperature must be due to its compression in descending from a great altitude. In the descriptions of the Fohn, attention is almost exclusively directed to the high average temperature of the air, and no mention is made of their extraordinary variations, although every observer must have noticed them. They are so great as to be recognised at once by the sensations and at the same time so rapid as to elude almost every other method of estimation or measure- ment. It has also, I believe, not been before remarked that the true Fohn occurs in our own country and with its characteristics quite as well marked as in Switzerland. It is sometimes supposed that a great absolute height of mountain chain is required for its production ; but this is not so. A relative height of 1,000 to 1,200 metres is quite sufficient for its production ; and this is equally available on the west coast of Scotland and on the northern slopes of the Alps. 110 Mr. J. Y. Buchanan. On Rapid Variations of [May 31, The observations were made iu the summer of 1893, which was abnormally warm all over the north of Europe. In the beginning of July I observed the Fohn at Fort William, and in the latter part of August in the upper Engadin, and more particularly in the valley occupied by the Morteratsch glacier. Besides the observation of the varying temperature of the air itself, the investigation of the tem- pei'ature gradient set up between the melting ice surface of the glacier and the hot winds blowing over it presented considerable interest. The curious fact was observed that while the hot wind was blowing over the glacier and melting the surface in abundance, the temperature of the air, as close to the ice as a thermometer could be applied without touching the ice, was never lower than 5*5° C. In the beginning of July at Fort William the weather was very warm, and in the midst of the very warm air still hotter blasts made themselves felt from time to time. The sensation was much the same as is produced when, on the deck of a steamer, the air passing the funnel strikes the face. These hot blasts lasted only for one or two seconds, and repeated themselves every minute or two. Their effect on a thermometer, freely exposed in the shade, was to keep the mercury in a constant state of motion, the temperature rising often more than 1° C. in a minute, and falling again as much. The thermo- meters in the screens were also a good deal affected, though not nearly to the same extent as the freely exposed ones. The recording instru- ments, the clock motion of which was not sufficiently quick to draw the record out into an indented line, showed a broad band which measured the amplitude of the excursions of the instrument, though by no means the amplitude of the oscillations of the temperature of the air. This phenomenon was particularly observed on the 8th July, 1893, when I was employed the greater part of the day in making evaporation experiments. It was very warm, as the following obser- vations of the thermometers in the large observatory screens will show : — 9 A.M. 10 A.M. Noon. 2 P.M. 4 P.M. Dry bulb (C.°) 20-1 22-4 24-9 23 -8 18 '9 Wet bulb (C.°) 177 17'3 18-2 17 '7 16-6 Vapour tension (mm.) Relative humidity .... 13-5 77 11-5 58 11-5 49 11-3 52 12-6 77 It was during the heat of the day, from 10 A.M. to 2 P.M., that the hot puffs made themselves most felt ; but I found it impossible to measure their temperatures, owing to the thermal inertia of the thermometers. The puffs lasted not longer thau one or two seconds, 1894.] Atmospheric Temperature, especially during Fohn. Ill and their temperature, to judge by the sensation, was rather higher than that of the body. The thermometers had only begun to rise when the heating ceased, and they fell back again. From the figures in the above table, it will be seen that the temperature of the air at noon reached 24'9° C., a very high figure for a station in nearly 57° north latitude. Along with the great rise of temperature there is a fall of absolute as well as of relative humidity, indicating that the air has come from a greater altitude. Attempts to measure the actual temperatui-es of the hot puffs gave no satisfactory result. I am much obliged to Mr. Omond and the staff of the Fort William Observatory for their courteous assistance while making these observations. Later in the year, in the middle of August, I visited the upper Engadin, and stayed for some weeks at Pontresina. Here, as else- where the weather was very warm, and I was much struck by observing the same blasts of hot air as I had experienced in Scotland. The general characteristics of the weather were the same, and the temperature of the air in the valley rose nearly as high as it had done at Fort William. On the 18th August I went for an excursion on the Morteratsch glacier with a guide. On my remarking the hot puffs of air, which were much more striking on the ice than on the land, he said it was the Fohn, of which he considered them a characteristic. The sun and the hot wind were causing an enormous amount of surface melting of the ice, and having a thermometer with me, I took the temperature of the air by whirling at a height of about 1 m. from the ice, and found it 12'0° C. ; the wet bulb was 5*0°, so that the vapour tension was 2'3 mm., the relative humidity 22, and the dew point — 8'6° C. The great dryness of the air will be remarked. I then swung the thermometer in a conical path as close to the ice as possible, and the temperature of the air was 10'0° C. Being astonished to find so high a temperature so near the ice, I put the bulb of the thermometer into a crack in the ice, so as to be below the level of the surface of the ice, and its temperature only went down to 7'5° C. All the temperatures were taken with a mercurial thermometer, which was whirled at the end of a string so that its velocity was about 6 m. per second. It was not protected in any way, so that the temperatures observed with it are not free from a certain error due to radiation and reflection, although it was always shaded from the direct sun. These errors are not usually great with a whirled instrument, and most of my observations have to do with differences of temperatures observed with the same instrument and under similar circumstances. On the glacier the thermometer, when whirled, was not apparently affected by radiation or reflection from the ice, and only very slightly by that from the sun. On land I 112 Mr. J. Y. Buchanan. On Rapid Variations of [May 31, remarked that the greatest disturbing effect is produced by sunlight reflected from grass. If the thermometer was whirled in the shade of a north wall with a grass field or hill- side close by, the thermo- meter would be immediately affected to the extent of one to two degrees, according as the sun shone on the grass or was obscured by a cloud. The effect was immediate the moment the sun came out ; sunlight reflected from rocks and light-coloured surfaces did not produce the same effect. On the 19th August I returned to the glacier. At 11 A.M. in the valley below the glacier I found the temperature of the air 22° C., and the wet bulb 12°'f5, whence the vapour tension is 5'0 mm., and the'relative humidity 26. In determining the temperature of the air by whirling the thermometer I found variations of as much as 2°. The hot puffs of air made themselves felt most markedly, and showed that the real variations of the temperature of the air were much greater than the thermometer showed. At 1 P.M., on the hill-side, to the west of the tongue of the glacier, and at a height of about 2,100 m. above the sea, four good observations of the temperature were made, giving 17°'5, 18°'0, 19°*5, and 19°'0; they are all equally trustworthy, and represent the average tempera- tures of the air during the minute, or minute and a half, that the thermometer was whirled. The mean of these values, 18° '5 is taken as the temperature of the air. For determining the temperature of the wet bulb the bulb of the thermometer was wrapped round with one thickness of Swedish filtering paper thoroughly moistened, and the thermometer was whirled as before and until the temperature ceased to fall, it then stood at 9°'5. Still higher up the hill at an altitude of 2,250 m., the temperature of the air at 2 P.M. was 18°'5 C. Having returned to the same spot where the observations were made at 1 P.M. the following air temperatures were observed : — between 2.40 and 2.46 P.M., 17°'5, 180<0, 17°'5, 17°'0, 17°'3, 17°'l ; mean, 17°'4 ; and between 2.50 and 2.54 P.M. 16°'5, 16°'5, 16°7, and 16°'5 ; mean, 16°'55. The mean of the two sets is 17°'06. Again it must be repeated that each of these individual observations is a faithful indication of the average temperature of the air in which the thermometer was whirled, and in so far as its sensibility enabled it to assume the same temperature as the air. From this spot I descended to the glacier and went up it until I got to a position which, judging by the eye, was at the same height as the station just left on the mountain side, and about one kilometre distant from it in a straight line. The weather was rapidly getting colder, the sky being covered with the characteristic Fohn cloud. The wind was fresh down the glacier, which made the exposure of the thermo- meter easy and good. The hot Fohn puffs were also very striking. The thermometer was first swung exposed to sun and wind, showing 1894.] Atmospheric Temperature, especially during Folm. 113 temperatures varying from 10°'5 to 11°'2, the mean being 10°'8 C. Swung in my own shadow, but exposed to the wind, the temperature was 90-8. The wet bulb was 40-7, showing a relative humidity of 37. The thermometer was now exposed, both wet and dry, in a horizontal position with the bulb at a distance of about 2 cm. from the ice, on the top of one of the superficial ridges of the glacier, and fully exposed to the wind, though shaded from the sun. The observed temperatures were : dry, 6°'6 C. ; wet, 3° '7 ; relative humidity, 58°'5. The exposure of the thermometer was as good as could be desired, and, with the fresh breeze blowing, it was thoroughly ventilated. I was again much struck with the highness of the temperature of the air almost in actual contact with the ice. The observations at 1 m. and 2 cm. from the ice were repeated, giving substantially the same results— at 1 m., dry bulb 10°'2, wet 5°'l ; at 2 cm., dry bulb 6°'8, and wet 3°'2. The hot Fohn puffs were more striking on the ice than on the land, owing to the greater difference between their tem- perature and that of the surrounding air. At 4 P.M. I left the ice and returned to the station of 1 o'clock on the hill-side, and took the temperature at 4.35 P.M. — dry bulb 16°'0, wet 80-0, relative humidity 240>5. At the station in the valley below the glacier the temperature was at 5.45 P.M., dry bulb 16°'4, wet 11°'8, and relative humidity 56. These observations, besides showing the remarkable conditions of the air over the glacier, indicate the fineness and warmth of the weather which prevailed. On the 21st August another series of observations was made at the stations on the land and on the ice. The breeze on the ice was not so steady or so strong as on the 19th, and about 5 o'clock in the afternoon there was a heavy squall of rain and thunder. The same hot Fohn puffs made themselves felt as before, without there being any means of measuring their temperature. Their duration at their maximum temperature was never more than a few seconds, during which but little effect was produced on the thermometer. It occurred to me that the only way of gaining a knowledge of the temperature of these puffs of air would be by com- paring the rapidity with which the thermometer moved when exposed to a known difference of temperature, with that observed in th e puffs. A number of observations was made with this view, by warming the thermometer and noting its rate of cooling in air of known tempera- ture. The reverse procedure was also followed on the ice. The thermometer was cooled by being laid close to, but not touching, the ice, it was then quickly raised to a height of 1 metre, and its rate of change of temperature observed. In this way it was found that for an initial difference of 4° the thermometer required 10 seconds to rise 1°, for a difference of 3° 12 seconds, and for a difference of 2°'5 16 seconds. These ratios were observed in the open air, and under the circum- VOL. LTI. I 114 Mr. J. Y. Buchanan. On Rapid Variations of [May 31, stances where the hot puffs are observed. Unfortunately, owing to an accident to the thermometer, very little use could be made of them. Where the rate of change of temperature of the thermo- meter is used to determine the temperature of the air, the movement of the air must be measured or estimated. The observations made on the 19th and 21st August are given in Table I. Table I. — Temperature Observations at Equal Altitudes on the Morteratsch Glacier, and on the Mountain west of it. Thermometer. Diff. Vapour tension. Rel. hum. Dew point. Dry. Wet. 19th, August, 1893. Land station, 2.45 P.M. 4.35 „ Mean Ice station, 3.20P.M. Height 1 metre, 3.55 „ Mean .... Ice station, 3.20 P.M. Height 0-02 m. 3.55 „ Mean .... 17-1 16-0 . 8-6 8-1 8-5 7-9 mm. 3-2 3-2 p. c. 22 24 -5-0 -4-7 16-55 8-35 8-2 3-2 23 -4-85 9-8 10-2 4-7 5-2 5-1 3-26 3-55 36 39 -4-4 -3-5 10-0 4-9 5-1 3-40 37-5 -3-95 6-7 6-6 3-7 3-2 3-0 4-4 4-2 4-0 57 56 -1-4 -3-0 6-65 3-45 3-2 4-1 56-5 -2-2 21st August, 1893. Land station. 1 P.M. „ „ 3.45 „ Mean... .. Ice station, 2.22 P.M. Height 1 metre, 2.54 „ Mean .... Ice station, 2.15 P.M. Close to ice, 2.40 „ Mean .... 14-5 14-3 7-5 8-0 7-0 6-3 3-5 4-2 29 35-0 -3-5 -1-3 14-4 7-75 6-65 3-8 32 -2-4 9-85 11-0 5-6 7-0 4-25 4-0 4-2 5-1 47 52 -1-3 + 1-5 10-43 6-3 4-13 4-6 50 +0-1 7-3 5-5 4-0 3-2 3-3 2-3 4-1 4-2 54 65 -1-5 -0-7 6-4 3-6 2-8 4-2 59 -1-1 For comparison with the temperatures on the ice on the 19th, the mean of the observations on the land station at 2.45 and 4.35 P.M. is taken, and on the ice the mean of the observations at 3.20 and 3.55 P.M. The altitudes of the two stations were as nearly as possible identical, and they were not more than 1 kilometre distant from each 1894.] Atmospheric Temperature, especially during Fohn. 115 other. Considering the temperatures at a height of 1 m. there is a difference of 6°'5 between the, land and the ice. The difference of vapour tension, 0'2 mm., is insignificant, and shows that substantially the air is the same. The dew point in both cases is several degrees below 0°, so that, on coming in contact with the ice, there would be evaporation from it. The evaporating power of the air may be repre- sented by the difference between the tension of saturation and the actual vapour tension. It is very great on land, being 10'75 mm. at 16°'53 C., and it would rapidly evaporate water having that tempera- tare. On coming in contact, however, with ice the air actually in contact, which alone comes under consideration, is first cooled to 0° C., which reduces its saturation tension to 4'6 mm., and the difference is only 1'4 mm. We see, however, that this has been sufficient to increase the absolute humidity of the air in close proximity to the ice. At 1 m. above the ice the air had an average temperature of 10° C. ; at 2 cm. from the ice its temperature was as high as 6°'65 C., and the air in actual contact with the ice must have been at 0° C. Many observations have been made of the temperature of the air at different heights above glaciers, and, as might be expected, considerable dif- ferences have been observed ; but I am not aware that any observations have been made on the air almost but not quite in contact with the ice, as are those which have been made at 2 cm. from the ice. The bulb was perfectly shaded from the sun but freely exposed to the wind, it was also fully exposed to any cold radiations from the ice There is, therefore, no doubt that 6°'65 was the temperature of the air passing the bulb of the thermometer. The vertical distribution of temperature shown by these figures is remarkable. From height of 1 m. to within 2 cm. of the ice there is a gradient of °'4 per metre, in the remaining 2 cm. there is a gradient at the rate )f 33° per metre ; and, from various observations and considerations, it is probable that the moderate gradient is continued to within a lillimetre of the ice, when it becomes precipitous. It is to be noted that the absolute humidity, as shown by the vapour tension of the air, las increased from 3'4 mm., at 1 m., to 4'1 mm., at 2 cm. ; showing lat ice is being evaporated and transferred from the glacier to the itmosphere. The wind was blowing freshly down the glacier, and its velocity was measured by noting the time which pieces of paper allowed to drift took to reach the ice, and then pacing the distance. he mean velocity was found to be from 8 to 10 kiloms. per hour. The observations made on the 21st and on the 22nd confirmed those of the 19th. The same variability of the air temperature at the md stations was noticed. Between 12.55 and 1.6 P.M. the following temperatures were observed by whirling : — 16°'2, 16°'2, 16°'0, 15°'5, L6°-0, 15°-5, 15°-0, 14°-2, 13°'8, 140>0, 13°'5, 13°'5. These are all good ob- servations, and represent real variations of the temperature, or rather i 2 116 Mr. J. Y. Buchanan. On Rapid Variations of [May 31, they indicate real variations of greater amount. Taking the mean of the last five observations, we have the temperature of the air 14°'0. The wet bulb was found at 1.15 P.M. to be 7°'5, giving a difference of 6°'5. On the glacier the air felt closer than on the previous occa- sion. The temperature at 1 m. was 11°'5, and at 2 cm. from the ice 7°'3. The difference 4°'2 is less than on the previous occasion. The wind was much less strong, and yet the temperature close to the ice is higher. The wet bulb, under the same circumstances, showed 4°'0. Five minutes later the dry bulb was observed at 1 m. 10°'2 and 9°'4, mean 9°'85. Another observation of the dry bulb at 2 cm. from the ice gave 6°*6. The interval between the bulb and the ice was now reduced to the smallest possible distance, about 2 mm. The wind fell very light, and the thermometer remained at 8°'0, when the wind returned ib fell to 5°'8. The axis of the thermometer bulb would be about 5 mm. from the ice, and still the air is nearly 6° warmer than the ice. Another observation on the same conditions gave 5°*5. The wet bulb was now exposed, but it had to be kept about 5 mm. off the ice ; it showed 3°'2. At 2.43 P.M. a great volume of warm air came down, and the wet bulb ran up to 4°'5 in three or four seconds. With the return of the breeze the wet bulb went back to 3°'0. The Fohn puffs were now very troublesome. At 2.52 P.M. the wet bulb at 1 m. was 7°'0 ; the dry bulb showed — at 2.54 P.M., 11°-0 ; at 2.55 P.M., 13°'5 ; and at 2.57 P.M., 14°-5. In one puff the thermometer was observed to rise one degree in eight seconds, which would make the true temperature of the air at the moment about 6°'0 higher, or 19°'5. At 3.30 P.M. I returned to the land stations, and again found the same variable temperatures. Between 3.35 and 3.45 P.M. the tem- perature varied between 16°'0 and 13°'5. The following averages- were taken : — 3.45 P.M., dry, 14°'3; wet, 8°'0; relative humidity, 35°. 4.0 „ „ 14°-0; „ 8°-5; „ „ 42°-5. Taking the first of these and the observations at 1 o'clock, we have for the mean temperature of the air 14°'15, and the wet bulb 7°'75r On the ice we have — At 1 m., dry bulb, 9°'85 ; wet, 5°'6, and At 2 cm., „ 7°-3 ; „ 4°'0. The difference in the temperature of the air at 1 m. is only 4°'3, and that between 1 m. and 2 cm. above the ice is only 2°'55, while- the air at 2 cm. is 7°'3 warmer than the ice. On the 22nd August, the observations on the ice were repeated with very much the same results. The temperature of the air ranged from 9°'0 to 9°'5 at I m., and was 58/5 at 1 cm. from the ice. 1894.] Atmospheric Temperature, especially during Fohn. 117 The result of the few observations here quoted is to show that the air, which over land has a temperature of 15° to 20° or higher, in pass- ing over a glacier is cooled to a comparatively slight degree. Although the air appears to be thoroughly- mixed by its own motion, very sharp gradients of temperature are produced and maintained. The great and abnormal temperature of the air of the valley is kept up by the heat liberated by the compression accompanying the descent of local streams or striae of air from high levels. These keep up an extra supply of heat over and above what is supplied by the direct radiation of the sun. The result is that the melting of the glacier in Fohn weather greatly exceeds that of even the hottest day of ordinary weather. In order to convey a general idea of the climate in the neighbour- hood during the period when my observations were made, I subjoin a table of the air temperatures observed at the Pfarrhaus in Pontresina three times daily, and obligingly supplied to me by Herrn Pfarrer Falliopi. Table II. — Temperature of the Air at Pontresina. Temperature of the air observed at Date. 7 A.M. 1 P.M. 9P.M. Temp. Diff. from mean. Temp. Diff. from mean. Temp. Diff. from mean. 1893 C°. C°. C°. August 15 . 4-7 -2'92 19-2 -1-26 10-0 -1-36 16. 5-9 -1-72 20-0 -0-46 10-8 -0-56 17. 7-2 -0-42 20-8 •fO'34 11-8 + 0-44 18. 8-2 + 0-58 21-8 + 1-34 12-8 + 1-44 19. 8-6 + 0-98 21-2 + 0-74 12-8 + 1-44 20. 10-0 + 2-38 19-8 -0-66 12-6 + 1-24 21. 7-6 -0-02 22-2 + 1-74 10-2 -1-16 22. 8-2 + 0-58 20-2 -0-26 10-2 -1-16 23. 6-9 -0-72 19-2 -1-26 12-8 + 1-44 24. 8-9 + 1-28 20-2 -0-26 9-6 -1-76 Mean 7-62 •• 20-46 •• 11-36 •• In this table the very high temperature on the 18th, 19th, 20th, id 21st is very apparent. The Fohn prevailed during all these days. On the 23rd August, which was a very warm day, I made a series :>f observations between Pontresina and the top of the Piz Languard, rhich is the highest peak on the ridge immediately behind Pontre- 118 Mr. J. Y. Bucliauau. On Rapid Variations of [May 31, sina, and is very easily accessible. It had been raining heavily in the night, so that in the early morning the air was rather cool; but the following observations made before starting up the mountain will show how rapidly the temperature was beginning to rise. 8.0 A.M Dry bulb, 10°'4 ; wet, 90<2. 9.10 „ „ 14°-8; „ ll°-4. 10.0 „ „ 17°-0. At 10 A.M. I started up the mountain, following the excellent path which leads to the summit. In the following table the temperatures observed at various stations are entered along with corresponding ones observed in the porch of the Hotel Reseg at Pontresina. Temperature. Height above sea. Time. On Difference. \JH At hotel. mountain. m. o o Pontresina . . . 1800 10.0 17-0 2100 10.50 16-5 19-5 3-0 2250 11.5 16-5 20-0 3-5 2370 11.35 16-5 20-5 4-0 2670 32.0 14-5 20-75 6-25 2790 12.80 13-3 21-0 7-7 2970 1.0 14-0 21-5 7-5 3180 1.30 13-1 22-0 8-9 3266 2.10 11-0 2.40 10-5 22-0 11-25 Excepting in the first interval the rate of fall of temperature between Pontresina and the station on the mountain is less than 1° per hundred metres. At the summit the mean temperature of the dry bulb was 10°'75, and of the wet bulb 6°'45, whence we have the vapour tension 4'5 mm. and the relative humidity 47. The weather was of the same kind as in the valley, abnormally warm, and the air very dry. Eeceived May 31, 1894. The thermometer which was used in these observations was not very sensitive, and when it was broken I could only replace it by one •which was considerably less so. They were, therefore, of no use for determining rapidly varying temperatures. The method indicated 1894.] Atmospheric Temperature, especially during Fohn. 119 above, whereby the temperature of the air is inferred from the velocity with which the thermometer rises or falls when immersed in it, either at rest or moving with a known speed, is in itself quite satis- factory. The difficulty in applying it is to ascertain the rate of motion of the air, because, other conditions being the same, the thermometer changes its temperature in proportion to the velocity of the air passing it. When the air has a horizontal motion it is called wind, and there are many instruments for its measurement ; but there is probably nearly as much vertical as horizontal motion in the atmosphere, but it is seldom observed and not easily measured. In fact, a very good way of detecting these movements in air which, to the senses, appears to be motionless, is to observe the rate of cooling or heating of a thermometer in it. A thermometer similar to the one used in these investigations was carefully tested as to its rate of cooling, in connexion with a series of observations made in the winter. Its rate of cooling was repeatedly determined in a room of constant temperature, and in the open air, when it was, to all appearance, motionless. Sometimes the rate of cooling in the open air was very nearly the same as in the room, but at other times it was much greater. It was never less. In four experiments, taking the same excess of temperature above the air, namely, 5°'5 C., the temperature of the thermometer fell by half that amount, 2°'75 C., in the room in 125 sees., and in the open air, which was apparently still, in 100, 70, and in 55 sees. The volume of the bulb of this thermometer, which was cylindrical, was 0'92 c.c. ; it was rather sluggish. Applying Leslie's rule for finding the " range " of the thermometer from the time it takes to cool to half the extent of the difference between its initial temperature and that of the air, we multiply it by 101/70. Leslie* defines the " range " of a thermometer or other body cooling to be the reciprocal of the fraction of the whole initial difference of temperature between the thermometer and the air, by which it cools in the first interval of time ; or it is the time in which the thermometer would fall to the temperature of the' medium, if, in each successive interval of time, its temperature had fallen by the same amount as in the first interval of time. The " ranges " of our thermometer cooling in the above conditions are found to be 180, 143, 100, and 80 sees, respectively. Having recognised that, in the conditions under which he experi- mented, the refrigerant power of a stream of air is exactly propor- tional to its velocity, he givest a formula for finding the velocity of the wind from the rate of cooling of a thermometer, or other similar * ' An Experimental Inquiry into the Nature and Propagation of Heat,' by John Leslie, Edinburgh, 1804, p. 264. t Page 283. 120 Mr. J. Y. Buchanan. On Rapid Variations of [May 31, vessel, in it. If T be the "range" in still air, and t the observed " range," then the velocity of the wind is 20 T— t • c v = — . in feet per second, or 3 t y £ v = X 4^ in miles per hour. t Converting into metrical units we have T-t v = 2'032 -— — m metres per second. t If in our experiments we ascribe the whole difference in the rate of cooling in the room and in the open air to motion of the air, and apply Leslie's formula, we find that the air must have been passing the thermometer at the rate of 0'5, 1'6, and 2'5 m. per second respec- tively. On each occasion there was no perceptible horizontal motion of the air, and the differences in the rates of cooling observed may, in the absence of a better explanation, be held to indicate the presence of ascending or descending currents of probably very local character. In the winter of this year I revisited the Engadin, and stayed for a fortnight at St. Moritz. As the room which I occupied faced due north the window of it was convenient for making observations of the temperature of the air. From the 24th February to the 3rd March I made every morning a series of observations of the temperature of the air, beginning when there was just light enough to read the thermometer, and continuing till between 8 and 9 o'clock in the morning. At first I took the temperature every minute, but finding the oscillations of temperature very great, I reduced the intervals to twenty seconds, and sometimes to fifteen seconds. The thermometer used was the one whose " ranges " in the still air of a room and outside have been given above. As before remarked, it is a sluggish instrument, yet the variations which it indicated in these short intervals of time were much greater than I could have anticipated. To print the observations in extenso would occupy too much space, but the striking features can be easily sum- marised. They are given in Table III. Excepting on the 26th February, when it was snowing all the morning, the observations embrace the interval of an hour or an hour and a half after sunrise. The time was devoted entirely to this object, and observations were made at as close dates as possible. Working alone, an interval of twenty seconds is quite convenient ; shorter intervals cause hurry. The time immediately following sunrise is when one would expect the tem- perature of the air to rise continuously, if not regularly ; but we see that so far from rising continuously and regularly the thermometer 1894.] Atmospheric Temperature, especially during Fohn. 121 :=! . 03 — go II PR o o o o o o o ** o S co i> in m co oo • rH CO (M 5*1 N CM 1 M fe O O O O O O • |j 11 i> co oo m i— i i> m CD (M CO 00 ^P O OO 1 r-l i-l i-l IM -^1 CM a t , — — -^ £g 11 co i-i m m o IH t> rfi ^ 1 "" i-i •J| rt| 00 O |I 1 CQ CO CO l> 00 00 Od IN CO rfi CO CO i-l 00 I-l o § • 3 p ^ 1 o> P< •2 a O TO co co oo o o in O So o o m co o N CM (M i-l iH CSJ S3 iH i-l CO 00 t» 00 CO t* CO 00 CD 00 I-l r- d I ^ - - 4 1 i|S P fe F^ >.O CD £•• 00 »H Cl CO CM CM (M C\l 122 Mr. J. Y. Buchanan. On Rapid Variations of [May 31, rises, falls, and remains stationary quite irregularly. On some days, as on the 28th February, these irregularities are comparatively few ; on others, as on the 1st and 2nd of March, they are numerous. The largest rise or fall in twenty seconds is 0°'5 C. From experiments in t;alm air outside and in still air in a room we find that for this thermometer to rise or fall 0°'5 C. in twenty seconds the temperature of the air around it must be from 2°'25 C. to 4°'65 C. hotter or colder than the thermometer. Taking even the lowest of these values, we see how great the possible error is in measuiing the actual temperature of the air at any moment with a thermometer, and the error is the greater the more sluggish the instrument is. In Table IV the detailed observations are given for a few minutes on the 26th February, when the temperature was changing very rapidly. In the third and fourth columns the rise or fall of the Table IV. — Temperature of the air at St. Moritz, observed at intervals of twenty seconds. Date, 26 February, Observed tempera- ture. Difference. Correspond- ing difference of tempera- ture of air. Amended tempera- ture of air. Differences of amended temperatures. Fall. Else. 1894. T. C°. - + -t. + t. T' = T + 1. Fall. Eise. A.M. h. ID. s. 11 18 45 5-88 ,. ^ B J >t 6-48 19 5 6-00 0-12 0-60 6-60 ^ t 0-12 25 6-12 0-12 . 0-60 6-72 „ . 0-12 45 6-25 . 0-13 § 0-60 6-25 0-47 20 5 6-25 g B ^ g 6-25 25 6-25 . . . t ]* 5-25 1-00 45 6-00 0-25 . . 1-00 4-30 0-95 21 5 5-62 0-38 , , 1-70 3-37 0-93 25 5-12 0-50 . . 2-25 — 4-12 . . 0-75 45 4-88 0-24 ^ i 1-00 2-63 1-49 22 5 4-38 0-50 ( t- • 2-25 2-13 0-50 25 3-88 0-50 , . 2-25 '.'. 3-88 t „ 0-75 ; 45 3-88 . . . . ^ B 3-28 0-50 23 5 3-75 0-13 . . 0-60 3-75 0-47 25 3-75 . . . . ^ ^ 4-37 t ^ 0-62 45 3-88 . . 0-13 , , 0-60 3-88 0-49 24 5 3-88 . . . 9 , ^ 3-28 0-50 25 3-75 0-13 m 9 0-60 3-15 0-13 45 3-62 0-13 9 % 0-60 3-12 0-03 25 5 3-50 0-12 9 ^ 0-50 3-00 0-12 25 3-62 . . 0-12 . . o'-50 4-00 , . 1-00 45 3-50 0-12 . . 0-50 3-12 0-88 26 5 3-50 . . . , . , ^ t 3-50 0-38 25 3-50 1894.] Atmospheric Temperature, especially during Fb'hn. 123 observed temperature is given. In tbe fiftb and sixth columns the corresponding differences between the temperature of the air and that of the thermometer which would cause the observed rate of change of temperature are given ; with these and the observed tem- peratures we obtain the amended temperatures of the seventh column. Although it was snowing on the 26th the air was perfectly still, and the rate of cooling corresponding to the " range " 80 sees, has been applied. Had the rate of cooling of the thermometer in tbe still air of a room been taken the difference between amended and observed temperatures would have been nearly twice as great. It was interesting to know what could be obtained with a record- ing thermometer of ordinary type, and in Table V the results of some observations made in Cambridge with a Richard's recorder are given. Table V, giving the Time in Seconds required by a Richard's Record- ing Thermometer to change its Temperature by 1° C. for a given Difference of Temperature between it and the Air. Difference of tempera- ture between thermometer and air 12°. 11°. 10°. 9°. 8°. 7°. 6°. 5°. 4°. 3°. 2°. at beginning of exposure. gO f In the [ 20" 20" 25" 25" 30" 30" 30" 65" 90" 90" 240" ^M open air I and fresh j .. • • .. .. 35 45 120 130 150 300 ,£3 O breezes. .. .. .. .. 9 9 20 35 40 45 80 240 C 'i.r -\ '3 0 K from > 20 22 24 26 28 30 35 52 84 140 250 oa^ | 60 70 110 130 210 §^1 In still 1 4> b ° air in ^ 90 100 300 450 -u O e .c a room. | •a. I * 0 2 a I 120 160 300 H 60 80 110 180 320 The figures in this table are taken from the curves drawn by the instrument on a drum revolving once in forty-eight minutes. The instrument was allowed to take the temperature of the room, then exposed in the shade in the open air when a fresh breeze was blowing and allowed to remain there until it had taken the temperature of the air. It was then transferred to the room, and allowed to rise until it attained its temperature. 124 Mr. J. Y. Buchanan. On Rapid Variations of [May 31, In this way two sets of curves were obtained, consisting of three curves in still air and three in a fresh breeze. The results are not very concordant, for, although the scale of time is very open — 1 min. occupying 5 mm. — the temperature scale was very close, 1° occupy- ing only 1 mm. The object, however, of the table is to show what can be expected from an instrument of the kind in the measurement of changes of temperature. The results obtained in the open air would necessarily vary somewhat, because, although a fresh breeze was blowing all the time, a fresh breeze varies in velocity. In order to obtain the best results from a thermometer it should be exposed to uniform ventilation. This can only be effected by artificial means, and they necessarily tend to efface sharp variations of temperature. The arrangement adopted by Professor Assmann in his psychrometer for ventilating and exposing his thermometers ought to be suitable for this purpose. The current of air produced must be uniform, and the behaviour of the thermometer as regards rate of change of temperatures in the current produced must be accurately determined. In Assmann's arrangement the thermometer is enclosed in a metal tube, consequently the diameter of the bulb, on which the sensitive- ness depends, can be made smaller and its length greater than would be safe with a a unprotected instrument. A mercurial thermometer, therefore, ventilated on Assmann's system, ought to be efficient for the measurement of temperatures changing with considerable rapidity. Departing from the mercurial thermometer I have found the simple air thermometer very good for indicating and measuring quick varia- tions of temperature. It has the advantage of lightness and cheap- ness. The form which I use is a glass bulb, of about 3 cm. diameter on a straight stem of about 10 cm. length. This can be attached to a (J-tube of greater or less diameter, according as the differences of temperature to be observed are great or small. The \J -tube has some coloured Avater as indicator, and the indications of the instrument are compared with those of a thermometer. As the instrument is only put together when it is wanted, the variations of barometric pressure do not affect it. It has the great advantage that it can be connected with a tambour, and thus be made to record. The sensitive- ness of the glass air thermometer is about the same as that of a very fine mercurial thermometer made for me by Messrs. Hicks. The air thermometer, however, would be very much more sensitive if the ball were made of thin metal instead of glass. There is a limit to the sensitiveness of all thermometers depending on the dilatation of a fluid, and I do not think that any such thermo- meter can be constructed which would give directly the true tem- perature of the air in the puffs of Fohn wind which we have been 1894.] Atmospheric Temperature, especially during Fohn. 125 discussing; by taking account of the rapidity of their movement they can be constructed to give the temperature inferentially. The only probable method of observing directly such rapid changes of temperature is by electric or thermoelectric methods. A thermo- electric junction is made of metals which conduct the heat rapidly, and as their mass can be made very small and their specific heat is low they can be made to follow the temperature of the medium in which they are immersed more closely than any other form of ther- mometric apparatus. The galvanometer necessary for measuring the currents produced is the inconvenient part of the apparatus, but I am informed by those familiar with such apparatus that a suitable instru- ment for use in the field could be constructed without difficulty. Thermometers as Calorimeters. — If we know not only the rate of cooling of a thermometer, if we have the figure which, in Leslie's language, is called the "range," and if in addition we know the thermal mass of the bulb which is generally expressed by its " water value," the thermometer becomes an efficient calorimeter. It is a familiar observation that the thermometer and the senses frequently disagree about the warmth or coldness of the weather. This is because they measure different things. The thermometer measures the temperature of the air, the senses measure the heating or cooling power of the atmosphere, or the rate at which the body is called upon to receive or supply heat. The body is a calorimeter and not a nere thermometer. But with a knowledge of the constants above men- tioned, the thermometer becomes also a calorimeter. In connection with the melting of ice by the hot wind in the Engadin, and the corresponding abstraction of heat from the air, I made a number of experiments by whirling thermometers at various speeds in air of definite temperature, having previously warmed the thermometer to a higher temperature. In order to give calorimetric expression to the result, and to express the heat exchange which had taken place, it was necessary to know the water value or thermal mass of the thermometer bulb. In similar experiments made by Leslie, he used a tin sphere 4 in. in diameter filled with water, of which it contained more than half a litre, and there was no difficulty in finding the ^thermal mass, as that of the thermometer was an insignificant fraction of it. With a mercurial thermometer, however, of ordinary type the glass envelope of the bulb is as important from a calorimetric point of view as the mercury contained in it ; and it is impossible to know the proportions in which the two substances are present, except by weighing them in process either of construction or of destruction. The former of these processes was excluded, and I hesitated to adopt the latter before some more use had been got out of the thermometer. Meantime I endeavoured to estimate the probable thermal mass of the bulb by 126 Mr. J. Y. Buchanan. On Rapid Variations of [May 31, carefully measuring it, and assuming a probable thickness of the glass. In dealing with problems of this sort it is necessary to express the specific heat in terms of the volume, and for this purpose the ordinary numbers which express the capacity for heat of unit weight have to be multiplied by the density, which expresses the weight of 1 c.c. of the substance. The density of mercury is 13'596, and that of ordinary glass is 2-45 ; their specific heats per unit weight are 0-033 and 0*19 respectively; whence the capacity for heat of 1 c.c. of mercury is 0*449, and of glass 0'466. If their specific heats are taken as identical and equal to 0*457, the error made will not be more than 2 per cent., in the extreme case where the bulb is all glass or all mercury. Hence it appeared that there was no necessity for knowing the thickness of the glass of the bulb or the weight of mercury in it. For calorimetric purposes, a knowledge of the volume of the bulb suffices, and it is immaterial in what proportion the two substances are present. The figures on which this calculation are based are for ordinary soda or potash glass, which was no doubt nsed in the con- struction of the German thermometers which I was using. Using the value 0*457 for the specific heat per unit volume of the bulb, and whirling the thermometer at the uniform rate of 6 m. per second, twelve observations were made of the thickness of the film of air heated to the full amount, corresponding to the fall of tempera- ture of the thermometer. The difference between the initial tem- perature of the thermometer and that of the air varied from 18° C. to 2° C., and the resulting computed thicknesses of the film of air heated varied from 0*209 to 0*267 mm. ; the mean value was 0*237 mm. The measurement of the volume of the bulb requires some atten- tion. The most convenient form of the bulb is the cylindrical, and it is also the most common. But the bulbs are very rarely truly cylindrical, they are often considerably tapered. It is not sufficient to measure the diameter of the bulb with callipers, it is necessary to measure the circumference at various parts of the bulb. One simple way is to envelop the bulb with a wrapper of tissue paper, like a cigarette, to blacken the edge of the paper which is laid inside. When the paper is neatly and smoothly laid on, pressure with the finger along the line of the inner edge of the paper produces a sharp impres- sion of the edge on the overlapping paper. On unrolling the paper the exact envelope of the bulb lies between the blackened edge of the paper and the impression which it has made on the paper underlying it. The length of the bulb is very easily measured, and when the paper envelope has been, to begin with, given the proper length, it measures the outer surface of the bulb, less the surface of the end. This is assumed to be hemispherical, and is added accordingly. The upper end of the bulb, where the stem joins on, is neglected, as in 1894.] Atmospheric Temperature, especially during Fohn. 127 thermometers of German pattern, it takes little part in the exchange of heat with the outside. Another method of obtaining the exact circumference of the bulb, which is a little easier and perhaps more exact, is to wind fine thread round it, each turn touching its neigh- bour closely until, say, ten turns have been taken. The thread is then unwound and measured. The tenth part of the length is the circumference of the bulb. By measuring the axial space occupied by the ten turns, the correction for " pitch " can be ascertained, but if anything but very coarse thread is used it is negligible. The active superficial area of the bulb is given by adding to the hemi- spherical end surface the product of the mean circumference into the total length of the cylindrical part of the bulb. In like manner, the volume of the bulb is obtained by adding to the hemispherical volume of the end the product of the mean circular area into the length of the cylindrical part of the bulb. The volume, multiplied by 0'457, gives the thermally equivalent volume or weight of water. Air thermometers of the simple kind described above, are very easily made so as to give calorimetricai results. It is only necessary to weigh and measure the piece of glass tube before blowing the bulb. The shortening of the straight part of the tube after blowing gives the length of it which has been expanded into a ball, and from the known length and weight of the original piece of tube, the weight of the bulb is found. By carefully gauging the volume of the ball its volume can be obtained, and from that the thickness of the glass. When the specific heat of the glass is known, the water value of the bulb is given ; if the air contained is taken into account, the value is increased by from 1 to 2 per cent. The surface of the ball divided by TableVI. — Particulars of Calorimetric Air Thermometers made of Lead Glass. Number of Instrument. 1. 2. 3. 4. 5. Original weight of tube (grm.) ,, length of tube (mm.) Ditto after blowing 17 -724 225-7 197 -'0 18 -508 193-0 144-0 18 -4186 192-1 137-0 18-8136 196-4 126 -0 18 -6169 194-25 104-0 Difference 28-7 49 -0 55-1 7o-4 90-25 Weight of 10 mm. tube (grm.) Weight of bulb (grm.) Diameter of bulb (mm.) .... Volume of ditto (c.c.) 0 -7853 2 -2538 24 7-238 0 -9590 4 -6991 32 17-157 0-9590 5 -2841 38 28 '731 0-9580 6-7443 45 47 -713 0 -9580 8 -6550 51 69 -456 Surface of buib (sq. cm.) . . . Volume of glass at sp. gr. = 3-0 18 -095 G'7513 32 -170 1-5664 45-364 1 -7614 63 -617 2 -2481 81 -713 2-8850 Thickness of glass (mm.) . . . Water value of bulb, sp. heat = 0'57 0-415 0 -4282 0-487 0-8928 0-388 1-0040 0-353 1 -2814 0-353 1-6445 Surface -f- water value 42-26 36-03 45 -18 37-25 42 -24. 128 The Root of Lyginodendron Oldhamium, Will. [May 31, the water value gives an expression for the sensitiveness of the instru- ment. In Table VI the particulars of several air thermometers which I have had made are given. As they are made of lead glass, both the density and the capacity for heat are higher than in the case of German glass. VII. " The Root of Lyginodendron Oldhamium, Will." By W. C. WILLIAMSON, LL.D., F.R.S., and D. H. SCOTT, M.A., Ph.D., F.L.S., F.G.S. Received March 14, 1894. During a re-investigation of the structure of Lyginodendron* the results of which we hope to lay before the Royal Society on a future occasion, an important fact has come to light, which we desire to place on record without delay. A carboniferous fossil, with the structure perfectly preserved, has been described in previous memoirs, under the name of Kaloxylon HooJceri, Will.f We have now established the fact that Kaloxylon was not an independent plant, but was the root of Lyginodendron Oldhamium. Specimens, presenting in every respect the typical Kaloxylon struc- ture, have been found in actual continuity with the stem of Lygino- dendron^ arising from it as lateral appendages. Their structure and mode of origin prove that they were adventitious roots. These organs branched freely, and we have roots and rootlets of all sizes, and at all stages of development. This discovery enables us to give a complete account of the vege- tative organs of Lyginodendron, as we are now fully acquainted with the structure, not only of the stem and foliage, but also of the adven- titious roots. Presents, May 31, 1894. Transactions. London : — Camera Club. Journal. Vol. VIII. No. 96. 8vo. London 1894. The Club. Entomological Society. Transactions. 1894. Part 1. 8vo. London. The Society. Royal United Service Institution. Journal. Vol. XXXVIII. No. 195. 8vo. London 1894. The Institution. * Cf. Williamson, " On the Organisation of the Fossil Plants of the Coal Measures," Part IV, ' Phil. Trans.,' 1873, p. 377; Part XVII, ' Phil. Trans.,' 1890, B., p. 89. t Cf. " On the Organisation of the Fossil Plants of the Coal Measures," Part VII, ' Phil. Trans.,' 1876, Part 1, p. 1 ; Part XIII, ' Phil. Trans.,' 1887, B., p. 289. 1894.] Presents. 129 Transactions (continued). Munich : — K. B. Akademie der Wissenschaften. Sitzungsberichte (Math.-phys. Classe). 1894. Heft 1. 8vo. Munchen 1894. The Academy. New York: — American Museum of Natural History. Bulletin. Vol. VI. Pages 97—128. 8vo. [New York] 1894. The Museum. Palermo : — Circolo Matematico. Bendiconti. Tomo VIII. Fasc. 1—3. 8vo. Palermo 1894. The Society. Rome: — B. Comifcato Geologico d'ltalia. Bollettino. Anno 1894. No. 1. 8vo. Roma. The Committee. Turin : — B. Aocademia delle Scienze. Atti. Vol. XXIX. Disp. 5—10. 8vo. Torino 1894. The Academy. Vienna : — Anthropologische Gesellschaft. Mittheilungen. Bd. XXIV. Heft 2. 4to. Wien 1894. The Society. Observations and Beports. Calcutta : — Meteorological Department, Government of India. Meteorological Observations recorded at Seven Stations in India. December, 1893. 4to ; Monthly Weather Beview. December, 1893. 4to. Calcutta. The Department. London : — Army Medical Department. Beport. 1892. 8vo. London 1894. The Department. Potsdam : — Astrophysikalisches Observatorium. Publicationen. Bd. IX. 4to. Potsdam 1894. The Observatory. Journals. Archives Cliniques de Bordeaux. Annee III. No. 4. 8vo. Bor- deaux 1894. The Editors. Burdett's Hospital and Charities Annual. 1894. 8vo. London. Mr. H. C. Burdett. Firket (Ad.) Sur Quelques Boches Combustibles Beiges assimilees ou assimilables au Cannel-coal Anglais. 8vo. Liege 1893 ; L' Origin e et le Mode de Formation de la Houille. 8vo. Liege 1894. The Author. Hinrichs (G.) Centenary Commemoration of Antoine-Laurent Lavoisier. 1794—1894. 4to. St. Louis 1894 ; Contributions to Atom-Mechanics, published in the Comptes Bendus of the Academy of Sciences of Paris, and in other serials. 8vo. St. Louis 1894. The Author. Hjman (C. P.) An Account of the Coins, Coinages, and Currency of Australia. 8vo. Sydney 1893. The Author. Beade (T. M.) A Cooling and Shrinking Globe and the Origin of Mountain Banges. 8vo. London 1894; Continental Growth and Geological Periods. 8vo. London 1894. The Author. VOL. LVI. K OBITUARY NOTICES OF FELLOWS DECEASED. FREDERICK LE GROS CLARK, F.R.S., F.R.C.S.Eng., who died on the 19th July, 1892, after a brief illness, at the ripe age of 81, was born on the 7th February, 1811, in Mincing Lane, the youngest of nine children of a city merchant. His early years were spent in the city. In 1822 he went to reside as a pupil with the Rev. Ford Richardson, at Iron Acton, in Gloucestershire, where he remained four years. Here he received a very excellent education. He had always expressed a desire to become a Surgeon ; but his father, before deciding, consulted his friend, Mr. Benjamin Travers, then the dis- tinguished Senior Surgeon to St. Thomas's Hospital, who allowed the son to have the run of the hospital for a couple of months, at the end of which time his father gave him the choice of entering his own counting house or of being apprenticed to Mr. Travers. In February, 1827, at the age of 16, he was apprenticed, and at once began his hospital career. He appears to have been an industrious and dis- tinguished pupil, for in 1830 he obtained the Cheselden Medal, and in the same year was appointed Assistant Demonstrator of Anatomy, under Mr. Tyrrell. He spent the summer session of this year in Dublin. In 1833 he passed his examination at the Royal College of Surgeons. The summer session of that year he spent in Paris, that of 1834 in Edinburgh, that of 1835 in Berlin ; and in 1836 he passed three months at Gottingen. In 1837 he took rooms near the hospital and started in practice. He continued working at the hospital, and teaching anatomy ; and in 1842, when the Hospital Medical School was remodelled, he was elected Lecturer on Descriptive and Surgical Anatomy ; and in 1843, on Mr. Tyrrell's ^eath, he was appointed Assistant Surgeon. He then removed to Finsbury Square. In 1847 he was elected Surgical Secretary to the Medico-Chirurgical Society, and in the following year moved to Spring Gardens. In 1853 he was appointed full Surgeon to the hospital ; and increasing engagements compelled him to retire from the Chair of Anatomy in 1854, though still retaining the lectures on Regional and Surgical Anatomy. In 1858 he removed to St. Thomas's Street, at the request of the Governors, and in 1860 became Lecturer on Surgery, an appoint- ment which he held down to his retirement from the hospital. Iii 1864 he was appointed Examiner in Surgery to the Royal College of ' 11 Physicians for two years ; and in July of the same year was elected a Member of the Council of the Royal College of Surgeons, and was at the head of the poll. In 1866 he was appointed Examiner in Surgery to the University of London for a period of five years ; in 1867 Professor of Human Anatomy and Surgery to the College of Sur- geons ; in 1868 Hunterian Professor of Surgery and Pathology ; and Examiner at the College in 1870. In 1872 he was appointed Vice- President of the College, and in 1874 President, giving the Hunterian Oration on the 13th February, 1875, the forty-eighth anniversary of his apprenticeship to the college. In 1872 he was elected a Fellow of the Royal Society. In 1873 he retired from the hospital, having retained office for a year or two at the special request of the Gover- nors. In 1877 he gave up practice, and took up his residence per- manently at Sevenoaks ; and in 1879 he retired from the Council of the Royal College of Surgeons. But even after his retirement he remained a busy and active man. He still continued Consulting Surgeon to the South Eastern Railway Company; he was always ready to give professional assistance to his neighbours ; he was Consulting Surgeon to and an active Governor of the Hospital with which he had been so long connected, took great interest in the welfare and progress of the Medical School, attended all anniversary and other important or interesting meetings, taking part in their proceedings ; and he retained his connexion with the Salters Com- pany, of which he had been twice master, at twenty years' interval. It should be added, that in addition to other duties he was for some years Surgeon to the Magdalen Hospital and to the London Female Penitentiary, and Consulting Surgeon to the Surrey County and Great Northern Hospitals. Mr. Le Gros Clark was in many respects a remarkable man. In the first place he had striking physical endowments ; he was tall and well made, spare but very muscular, singularly handsome, with dark curly hair and whiskers, dark grey eyes and bushy eyebrows, and well- formed features ; and was remarkably dignified and gentlemanly in appearance and demeanour. As a young man he was a great athlete, excelling especially in rowing, boxing, and riding, and he retained this activity of body to the last. He was hardly what one would term a genial man ; but he was a man of the highest character, he was absolutely unselfish and unself seekiDg ; whatever he undertook to do he did with all his might, he was perfectly truthful and trust- worthy, and always kind and considerate for others, and a warm and appreciative friend. He was a devout Christian, and member of the Church of England. As a surgeon and a teacher he was excellent. He was a thorough anatomist,, and an admirable lecturer on anatomy. He had had a wide experience as a surgeon, and was admirably well up in the subject ; he was conscientiously attentive to his patients, Ill and neglected nothing for their benefit, and he was a faultless operator. As a clinical teacher he was admirable. He was not, and did not aim at being, a speaker of commanding eloquence ; nevertheless he was a most excellent and ready lecturer and speaker. His manner was always quiet and gentlemanly; his language was always simple and well chosen ; and his matter was always appropriate to the occasion. He was consequently not only a clear and attractive lecturer, but he was a clear and attractive speaker on all festive and other occasions. Although he was a scientific surgeon and anatomist he did not do much original scientific work. He wrote many papers on points in anatomy and. surgery that interested him ; he lectured (as before stated) at the Royal College of Surgeons. He delivered three introductory addresses at St. Thomas's Hospital. While president he delivered the Hunteriaii Oration at the College of Surgeons, an address which was philosophical and full of thought. He contributed a paper on ' The Mechanism of Respiration ' to the ' Proceedings of the Royal Society ' (vol. 20, 1872). In 1836 he published a work on the anatomy and physiology of the nervous system ; in 1847 and 1853 he translated two volumes of Dupuytren's. ' Le9ons Orales ' for the Sydenham Society ; he published ' Lectures on Surgical Diagnosis of Visceral Lesions ' in 1870 ; and ' Outlines of Surgery,' of which a second edition appeared in 1872 ; Paley's ' Natural Theology,' edited for the S.P.C.K. in 1875 ; a little manual of physiology for the same society in 1883 ; several articles on anatomy and physiology in the ' Encyclopaedia Metropolitana ' about 1840 ; several papers in the ' Medico- Chirurgical Transactions ; ' critical articles in the ' British and Foreign Quai-terly,' and he made a few contributions of cases to the medical papers. Lastly, he pub- lished a collection of 'Miscellaneous Essays,' which had already appeared in various periodicals, &c. J. S. B. 1894.] Presents. 129 Transactions (continued). Munich : — K. B. Akademie der Wissenschaften. Sitzangsberichte (Math.-phys. Classe). 1894. Heft 1. 8vo. Munchen 1894. The Academy. "New York : — American Museum of Natural History. Bulletin. Vol. VI. Pages 97—128. 8vo. INew York] 1894. The Museum. Palermo : — Circolo Matematico. Rendiconti. Tomo VIII. Fasc. 1—3. 8vo. Palermo 1894. The Society. Rome: — R. Comitato Geologico d'ltalia. Bollettino. Anno 1894. No. 1. 8vo. Roma. The Committee. Turin : — R. Accademia delle Scienze. Atti. Vol. XXIX. Disp. 5—10. 8vo. Torino 1894. The Academy. Vienna: — Anthropologische Gesellschaft. Mittheilungen. Bd. XXIV. Heft 2. 4to. Wien 1894. The Society. Observations and Reports. Calcutta : — Meteorological Department, Government of India. Meteorological Observations recorded at Seven Stations in India. December, 1893. 4to; Monthly Weather Review. December, 1893. 4to. Calcutta. The Department. London : — Army Medical Department. Report. 1892. 8vo. London 1894. The Department. Potsdam : — Astrophysikalisches Observatorium. Publicationen. Bd. IX. 4to. Potsdam 1894. The Observatory. Journals. Archives C Uniques de Bordeaux. Annee III. deaux 1894. Burdett's Hospital and Charities Annual. No. 4. 8vo. Bor- The Editors. 1894. 8vo. London. Mr. H. C. Burdett. Firket (Ad.) Sur Quelques Roches Combustibles Beiges assimilees ou assimilables au Cannel-coal Anglais. 8vo. Liege 1893 ; L'Origine et le Mode de Formation de la Houille. 8vo. Liege 1894. The Author. Hinrichs (G.) Centenary Commemoration of Antoine-Laurent Lavoisier. 1794 — 1894. 4to. St. Louis 1894 ; Contributions to Atom-Mechanics, published in the Comptes Rendus of the Academy of Sciences of Paris, and in other serials. 8vo. St. Louis 1894. The Author. lljman (C. P.) An Account of the Coins, Coinages, and Currency of Australia. 8vo. Sydney 1893. The Author. Reade (T. M.) A Cooling and Shrinking Globe and the Origin of Mountain Ranges. 8vo. London 1894 ; Continental Growth and Geological Periods. 8vo. London 1894. The Author. VOL. LVI. K 130 Election of Fellows. [June 7, June 7, 1894. The Annual Meeting for the Election of Fellows was held this day. The LORD KELVIN, D.C.L., LL.D., President, in the Chair. The Statutes relating to the election of Fellows having been read, Sir Erasmus Ommanney and Mr. Scott were, with the consent of the Society, nominated Scrutators to assist the Secretaries in examining the lists. The votes of the Fellows present were then collected, and the fol- lowing candidates were declared duly elected into the Society : — Bateson, William, M.A. Boulenger, George Albert. Bradford, John Rose, M.D. Callendar, Professor Hugh Long- bourne. Cheyne, Professor William Wat- son, M.B., F.R.C.S. Froude, Robert Edmund. Hill, Professor M. J. M., M.A., D.Sc. Jones, Professor John Viriamu, M.A., B.Sc. Love, Augustus Edward Houghr M.A. Lydekker, Richard, B.A. Penrose, Francis Cranmer, M.A., F.R.A.S. Scott, Dukinfield Henry, M.A.,. F.L.S. Smith, Rev. Frederick John, M.A. Swan, Joseph Wilson, M.A., F.I.C. Veley, Victor Herbert, M.A.. F.C.S. Thanks were given to the Scrutators. 1894.] On the Newtonian Constant of Gravitation. 131 June 7, 1894. The LORD KELVIN", D.C.L., LL.D., President, in the Chair. A List of the Presents received was laid on the table, and thanks ordered for them. The following Papers were read : — I. " On the Newtonian Constant of Gravitation." By C. V. BOYS, F.R.S., A.R.S.M., Assistant Professor of Physics, Royal College of Science, South Kensington. Received May 31, 1894. (Abstract.) The Newtonian constant of gravitation G, i.e., the force in dynes between 2 grams of matter 1 cm. apart, has been determined with a very accurately constructed piece of apparatus, designed on the lines which I laid down in my paper on the Cavendish experiment (' Roy. Soc. Proc.', vol. 46, p. 293). The important dimensions are approxi- mately— Distance between centres of lead balls in plan . . 6 in. „ „ gold „ . . 0-9 in. Diameter of lead balls 4^ in. gold „ 0-2 and O25 in. Difference of level between right and left sides . 6 in. The lead balls were hung by phosphor bronze wires from pillars in the lid of the apparatus, and the gold balls by quartz fibres from the ends of the " beam mirror." The beam mirror was supported by a quartz fibre, 17 in. from a torsion head. An elaborate system of screens protected the apparatus from temperature variations. An " optical compass " of extreme precision was employed in measuring the horizontal distances between the fibres and between the wires, which alone among the geometrical magnitudes need be known with a very high degree of precision. The scale was 9 ft. long, divided into 50ths of an inch. It was placed at a distance equal to 14,000 divisions. It could be read with certainty to 1/10 division. The deflections varied according to the circumstances of each experiment from 351 to 577 divisions, and the squares of the periods from 35,431 to 58,519 sees.2 K 2 132 Mr. S. Bidwell. On the Recurrent [June 7, The experiments were carried out by permission of Professor Clifton, under the Clarendon Laboratory, at Oxford. The result is for G, the Newtonian constant of gravitation. . . . 6'6576 X 10~* A, the mean density of the earth 5 '52 70. II. " On the Recurrent Images following Visual Impressions." By SHELFORD BIDWELL, M.A., LL.B., F.R.S. Received March 27, 1894. The earliest recorded observation which I have been able to find of a certain curious phenomenon associated with optical after-images is that of Professor C. A. Young, who published a note on the subject in the year 1872, and proposed that the phenomena should be called "recurrent vision."* He noticed that when a powerful Leyden jar discharge took place in a darkened room, any conspicuous object was seen twice at least, with an interval of a little less than a quarter of a second ; often it was seen a third time and sometimes even a fourth. He thought that the phenomenon suggested the idea of a reflection of the nervous impulse at the nerve extremities, as if the intense impression upon the retina, after being the first time propagated to the brain, was 7'eflected back to the retina and thence again to the brain, thus renewing the sensation of vision. A few months later an account of two experiments on the same subject was published by Mr. A. S. Davis. t In the first, a piece of charcoal, one end of which was red-hot, was waved about so as to describe an ellipse or circle a few inches in diameter. A blue image of the burning end was seen following the charcoal at a short distance behind it, the space between the charcoal and its image being ab- solutely dark. The interval of time after which the sensation of blue light succeeded the primary sensation was estimated to be about a fifth of a second. The other experiment was made with a piece of apparatus resembling a photographic instantaneous shutter. The shutter was interposed between the observer's eye and the sky and was covered with pieces of coloured glass, through which momentary flashes of light were allowed to pass. It was found that each flash was, after a short interval, generally succeeded by a recurrent image, the colour of which was quite different from that of the glass. The results of Mr. Davis's observations are summarised below. Mr. Davis remarks that except as regards the red glass, the re- current colour does not differ much from the complementary colour, * ' Phil. Mag ,' vol. 43 (1872), p. 343. t Ibid., vol. 44 (1872), p. 526. 1894.] Images following Visual Impressions Table of Mr. Davis's Observations. 133 Light transmitted. Complementary colour. Recurrent colour. Deep blue Yellow Greenish- yellow Blue-red Reddish-blue Yellow Blue Orange-red Blue-green Red-blue No image and he concludes that when any one of the three kinds of Young- Helmholtz nerve 6bres is excited, an excitation is induced in the nerve fibres of the other kinds, the process being analogous to the in- duction of electric currents. In 1885 I called attention to a very simple and effective method of exhibiting a recurrent image.* If an ordinary vacuum tube, illu- minated by an induction coil discharge, is made to rotate slowly upon a horizontal axis fixed at right angles to the middle of the tube, the tube is seen to be foJ lowed at a distance of a few degrees by a ghost-like image of itself, the ghost exactly imitating the original in form, but having a uniform steel-grey colour. In the same paper the following observation is noted : — " The vacuum tube being at rest in a feebly lighted room, I concentrated my gaze upon a certain small portion of it while the discharge was passing. The current was then interrupted and the luminous image was almost instantly replaced by a corresponding image which appeared to be intensely black upon a less dark back-ground. After a period which I estimated at from a quarter to half a second the black image again became luminous ; this luminous impression lasted but for a small fraction of a second and the series of phenomena terminated with its disappearance It was also found desirable to make the preliminary illumination as short as possible, a single flash being generally sufficient to produce the phenomena." The following comment was added : — " The series of phenomena seem to be due to an affection of the optic nerve which is of an oscillatory character. Abnormal dark- ness follows as a reaction after the luminosity, and again after ab- normal darkness there is a rebound into feebler luminosity." The subject has recently attracted much attention in connection, with the experiments of M. Aug. Charpentier. The account of them given by M. Charpentier in a paper on " Retinal Oscillations"! is briefly as follows : — If a black disk having a white sector is illu- minated by a strong light, and slowly turned round while the * ' Nature,' vol. 32 (1885), p. 30. t "Oscillations retiniennes," ' Comptes Rendus,' vol. 113 (1891), p. 1.47. See also " Reaction oscillatoire de la Retine," ' Arch, de Physiologic,' 1892, p. 541. 134 Mr. S. Bidwell. On Ike Recurrent [June 7, observer's eye is fixed upon its centre, there appears upon the white sector, near to its leading edge, a well-defined dark band, which is separated from the black ground of the disk by a similar white band. The angular extension of the dark band increases with the speed of rotation, so that it always takes the same time to pass over a fixed point on the retina ; it begins about one-sixty-fifth or one-seventieth of a second after the first passage of the white, and lasts sensibly the same time. He goes on : — " The dark band is in fact only a kind of reaction of the retina after the luminous excitation, a reaction which can be demonstrated in a totally different manner. I have found that if an instantaneous luminous excitation is produced in complete darkness the sensation appears to be reduplicated ; shortly after its first generation it seems to disappear and then manifest itself again. This is the case, for example, when a single discharge from a Ruhmkorff coil is passed through a Crookes or Geissler vacuum tube, or simply, but less obviously, through the air There is, then, in this last experiment, as in the first, a negative reaction of the retina under the influence of excitation It would be difficult, and in any case premature, to indicate the cause of this phenomenon, but it may fairly be characterised as the result of a retinal oscillation set up under the influence of the beginning of the luminous excitation." I think it clearly appears from the above extract that M. Charpentier was unacquainted with the earlier observations of myself and others. In consequence of the importance which seemed to be attached by physiologists to the phenomena of visual reaction, as evidenced by Professor Burdon Sanderson's recent Presidential Address to the British Association,* I was induced to undertake the farther experi- mental investigation, of which an account is given in the present paper. This deals partly with the colours of recurrent images under different conditions, and partly with the reaction attending the earl}* stages of a luminous impression as noticed by Charpentier. In the observation of the recurrent images set up by the action of light of different colours I began, like Mr. Davis, by using coloured glasses. A metal disk, about 8 cm. in diameter, was arranged so as to rotate slowly and steadily about its centre in front of the condenser of a projection lantern. Near the edge of the disk was a circular aperture about 0'5 cm. in diameter, the image of which was focussed upon a distant screen. A plate of coloured glass was placed before the pro- jecting lens, and thus was obtained a small, coloured disk of light, which described a circular path upon the screen. The coloured disk was, in most cases, seen to be followed at 8,n interval of a few degrees by a ghost of the same size and shape, but of feebler luminosity, and of a hue which varied more or less with the colour of the glass * ' Brit. Assoc. Rep.,' 1893. ' Nature,' vol. 48, p. 468. 1894] Images following Visual Impressions. 135 •employed. With white electric light the colour of the ghost was violet. This method of experimenting was, however, found to be unsuited for the purpose in view, and I mention it only on account of the facility which it affords for exhibiting the phenomenon to a large number of persons. To obtain results of any value, it was necessary to employ the simple colours of the spectrum, and the arrangement finally adopted for this purpose is indicated in fig. 1. L is a lantern con- Fia. 1. -/•38m. taining a high-pressure oxy hydro gen light, which is better adapted for the experiment than an arc lamp, the intensity of the light being •easily varied. S is an adjustable slit, M a projection lens, P a bi- sulphide of carbon prism, D a metal plate, in the middle of which is a circular aperture 2 mm. in diameter. A spectrum, 6 or 7 cm. in length, can be projected upon D, a small selected portion of it passing through the aperture and falling upon the mirror Q, which is 8 cm. in diameter. To the back of the mirror is attached a horizontal arm, which is not quite perpendicular to the mirror, its inclination being capable of adjustment. The arm is rotated by clock-work, and turns once in 1| sees. It was at first attempted to study the phenomenon by direct eye observations of the reflected image of the aperture in the rotating mirror, the aperture being covered by a piece of finely-ground glass ; but, for pretty obvious reasons, no satisfactory results could be thus obtained. A telescope was then employed, having a power of 12, and an eye- 130 Mr. S. Bidwell. On the Recurrent [June 7y piece with a large field. I believe that, after sufficient practice, this would he found the best possible method of observation ; but it is exceedingly difficult to keep the eye absolutely steady, and untrained observers never succeeded in seeing the looked-for phenomena at all.* Since it seemed desirable that my own observations should be confirmed by others, I abandoned the telescope and the ground glass, and by means of the lens N" focussed the reflected image of the aperture upon a white screen, R. The diameter of the projected disk of coloured light was about 1'5 cm., and that of the approxi- mately circular path which it described, 30 cm. To aid in steadying the eye, a spot of luminous paint, upon which the gaze might be directed, was applied at the centre of the circle. With this arrange- ment, almost any one can see the ghosts without the smallest difficulty. When the mirror turns once in 1^ sees., the ghost or recurrent image appears about 50° behind the coloured disk, the corresponding time interval being one-fifth of a second. Exact measurement is, however, not easy, and it is probable that the interval is not quite the same with light derived from different portions of the spectrum. The ghost appears to be circular in form, its diameter being generally rather less than that of the original. The colours of the recurrent images, as specified below, have all been observed by several personsr and, except as to those at the extreme limits of visibility, all the observations were in agreement. Experiment 1. Spectrum colours. Recurrent colours. Extreme violet No perceptible image. Middle violet A pale image, variously described as grey, yellow, and greenish-yellow. Dark blue Feeble violet. Light blue Brighter violet. Middle green Bright violet. The image is more conspicuous with green light than with any other. Greenish-yellow Blue. Orange-yellow Bluish-green, Orange Dark bluish-green. Orange-red., . , Very dark bluish-green. Red No image at all, however bright the red was made. The violets all appear to my own vision slightly redder than the violet of the spectrum. The following experiment was then made. * If a telescope is used, the mirror must be silvered on its outer surface, and in the air of a laboratory is quickly tarnished. 1894.] Images following Visual Impressions. 137 Experiment 2. For the screen with the aperture at D, tig. 1, another was substi- tuted, having a horizontal slit 7 cm. long and 2 mm. wide, the image of which was projected upon the screen B after reflection from the rotating mirror. Thus a small spectrum was produced, which re- volved parallel to itself, in a circle about 1 metre in diameter.* The eyes were directed upon a fixed spot near one end of the horizontal diameter of the circle. The spectrum was followed by a ghost of the form rather roughly indicated in fig. 2. It extended from the orange of the spectrum to the beginning of the violet, terminating somewhat abruptly at the orange end, and fading away gradually at the other. The image was distorted, as shown in the figure, approaching nearest to the spectrum at about the middle of the green, a little on the more refrangible side of the most luminous portion. The distance sepa- rating the spectrum from the image increased more rapidly towards the red end of the spectrum than towards the violet end, and the image was widened out considerably at the violet end; but neither the moving spectrum itself nor its recurrent image was so sharply defined as appears in the diagram. It was remarkable that the whole of the recurrent image of the spectrum was of a violet hue, being brightest where the distance irom the spectrum was least. No trace whatever of yellow or greenish-yellow could be detected at the more refrangible end, nor of blue or bluish-green at the other. The apparent absence of any colour except violet in the. recurrent image of the complete spectrum is capable of two possible explana- tions. The greenish-yellow seen at one end, and the blue and bluish- green seen at the other, when the spectrum colours are tested sepa- rately, may be due merely to an effect of contrast, the true colour of the image being in both cases a weak violet. Or, on the other hand, these colours may really be present at the ends of the image of the whole spectrum, being, however, of too weak an intensity to be distinguishable when in proximity to the more luminous por- tions of the spectrum itself and of its image. Two experiments were made in the hope of settling this question. Experiment 3. The slit at D was removed, and in its place was put a zinc plate having two small apertures close together. A second lantern and prism were set up, and two spectra were projected upon the zinc plate. By the help of screens, things were so arranged that a violet * Helmholtz observed the after-images of a spectrum seen for an instant, but failed to notice the dark interval which preceded their appearance (' Phys. Opt.,' p. 376). 138 Mr. S. Bidwell. On the Recurrent FIG. 2. [June 7, ray from one spectrum passed through, one of the apertures, while a violet ray from the other spectrum passed through the other. These were reflected from the mirror (which was not rotated), and formed two violet disks side by side upon the screen. By adjusting the widths of the slits and the intensities of the limelights, one of 1894.] Images following Visual Impressions. 139 these disks was made as bright as possible, and the other very feeble. By the side of the bright violet disk the feeble one often seemed to be of a greenish-yellow hue, though, when seen alone, it was distinctly violet. Experiment 4. In a similar manner a feeble violet and brighter greenish-yellow were placed side by side, but, however much the intensity of the former was diminished, it could never be made to assume a blue colour at all comparable to that possessed by the recurrent image of the greenish-yellow. Nor did it appear bluish-green beside orange- yellow or orange-red. While, therefore, the result of Experiment 3 is consistent with the contrast hypothesis, that of Experiment 4 appears to be opposed to it ; but there is so great a difference in the circumstances of the two kinds of observation, the one involving a deliberate comparison of the colours of two stationary luminous disks, and the other an estimate formed while a disk and its recurrent image were in rapid motion, that the opposing evidence cannot be regarded as conclusive. Another experiment was therefore devised. Experiment 5. The original screen with one small aperture was placed at D, and two spectra were projected upon it in such a manner that a green ray from one spectrum, and a red ray from the other, passed through the aperture, forming red and green images which were exactly superposed upon the screen B. The colour of the single image thus formed could, by suitable regulation of the limelights, be made greenish -yellow, yellow, or orange-yellow, these colours being, of course, not simple ones, but compounds of red and green. Now, red by itself gives no recurrent image whatever (this was verified before proceeding further by shutting off the green ray), while green by itself gives a violet recurrent image. The question to be decided was whether the green, when accompanied by the inert red, would give a violet recurrent image as if it were alone, or whether the com- pound colour formed by the combination — greenish-yellow, for example — would be attended by a blue or bluish-green recurrent image, just as if the compound were a simple spectrum colour. The latter was found to be the case. The same hue of greenish- yellow, Avhether a simple spectrum colour or a compound of red and green, was always attended by a blue ghost. When the red ray of the compound was shut off by a screen, the ghost instantly became violet : when the screen was removed it at once resumed its blue colour. 140 Mr. S. Bidweli. On the Recurrent [June 7, This experiment, though not conclusive, is clearly in favour of the probability that the blue and bluish-green recurrent colours apparently observed when the yellow and orange portions of the spectrum are tested separately are due merely to an effect of mental judgment, and not to any cause of a physiological nature. There are, therefore, four independent facts which are consistent with the conclusion that luminous recurrent images are due to a reaction of the violet nerve fibres only. (a.) With white light the recurrent colour is violet. (6.) In the recurrent image of the complete spectrum no colour but violet can be detected. (c.) A pure red light, however intense, gives no recurrent image. It is generally supposed by the supporters of the Young- Helmholtz theory that red light has no action upon the violet nerve-fibres. (d.) The apparently blue colour of the ghost of simple spectrum yellow is just as well produced by a compound yellow con- sisting of green and red, the latter of which is inert when tested separately. The path of the revolving spot of light is generally marked by a phosphorescent track, which, when the rate of revolution is not less than one turn in 1^ sees., often forms a complete circle. The bril- liancy of this luminous trail seems to vary with different observers, in some cases appai'ently being so intense that the recurrent image cannot be distinguished from it at all. The trail is due to the usually feeble continuation of the after-image, of which the bright initial stage constitutes the recurrent image. A spot of red light, although it is never followed by a ghost, is always considerably elongated daring its revolution, and its colour ceases to be uniform, the rear portion assuming a light bluish-pink tinge. However small the spot of light is made, and however high the speed of revolution, no complete separation of the spot into red and pink portions has ever been effected. In the experiment next to be described the Charpentier effect and the recurrent image are made to exhibit themselves simultaneously, Experiment 6. Two blackened zinc disks, 15 cm. in diameter, from each of which two opposite quadrants were cut out.* were mounted in contact with each other on a horizontal axis, driven by clockwork and making one turn in 1| sees. By slipping the disks over one another round their * It was found necessary to cut out two quadrants instead of only one, in order to balance the disks and secure uniform rotation. 1894.] Images following Visual Impressions. 141 centres, opposite open sectors might be obtained, of any aperture from 0° to 90°. The apparatus was set up opposite a box containing a 32-candle power incandescent lamp, with a variable resistance in the circuit, the side of the box between the lamp and the disks being covered with a sheet of ground glass. The sectors being in the first place opened as widely as possible, I fixed my eye upon the centre of the double disk, and at once saw Charpen tier's dark band upon the illuminated, background. The sectors were then gradually closed up, until the posterior edge of the dark band approximately coincided with that of the sector.* When this was accomplished it was found that the arc of the open sector was equal to about -Jy part of the whole circumference. The dark reaction, therefore, ceased in (T\ of 1^ sees. =) -^ sec- after the first impact of the light upon the eye. For more readily demonstrating the succeeding phenomena, it was found convenient to again open the sectors a little, so that they covered an angle of about 10° or 12°. Resuming the observation, it was seen that the posterior edge of the open sector was bordered by a luminous fringe due to persistence. A little beyond the termina- tion of the fringe there appeared an intensely black radial band, estimated to cover a space of from 3° to 4°, and easily visible even upon the black ground of the metal disk, though it is shown far more conspicuously upon a translucent disk made of stout writing- paper, with a sector cut out. Lastly, after another interval of, perhaps, 35° or 40°, came the luminous recurrent image,f which, with the yellowish light of the incandescent lamp, appeared to be of a blue colour. By varying the angular aperture of the sector, it was ascertained that the recurrent image appeared at a fixed interval after the light was cut off, and not after its first impact. This method of observation revealed one other point of interest, which seems hitherto to have escaped notice, though it is evident enough with a Charpentier disk, when once attention has been directed to it. The average illumination of the bright band inter- vening between the dark band and the leading edge of the sector is much more intense than that of the other portion of the sector. Moreover, it is not uniform, but increases, gradually at first, and very rapidly at last, from the leading edge up to the dark bnnd. In fact when the light used is not strong, the luminous margin of the bright band is a far more conspicuous object than the dark band itself : it appears to glow almost like a white-hot wire. Charpentier states that, under favourable conditions, he has been * This was not a very easy operation, because the luminous sector was slightly widened by persistence, especially near the circumference. f This, of course, cannot be seen upon a translucent paper disk being over- powered by the transmitted light. 142 Mr. S. Bidwell. On the Recurrent [June 7, able to detect the existence of a second, and even of a third, dark band of greatly diminished intensity, though he adds that the obser- vation is a very difficult one.* What is probably the same effect in a- different form can, however, be shown quite easily in the following manner. Experiment 7. In a blackened zinc disk 15 cm. in diameter, there were cut two opposite radial slits, about 0'5 mm. in width. The disk was rotated at the rate of one turn per second in front of a sheet of ground glass, behind which was an incandescent lamp. The glass was covered with opaque paper, in which a circular opening was made of slightly less diameter than the disk. The disk was placed opposite this opening, and no light reached the eye except such as passed through the two slits. When the disk was observed from a distance of about 1^ metres, the eye being fixed upon its centre, each slit 'appeared to give four (or possibly five) luminous images, arranged like the ribs of a partly opened fan. The images were distinctly separated by dark intervals near the circumference, but overlapped one another towards the centre. The leading image was naturally the brightest, each consecutive image being considerably weaker than its precursor. All had the same tone of colour, namely that of the yellowish-light given by the electric lamp. The usual blue recurrent image could also be seen following the images of the radial slits, at an angle of about 80°. It appears, then, that when the retina is exposed, to the action of light for a limited time, the complete order of visual phenomena is as follows : — (1) Immediately upon the impact of the light there is experienced a sensation of luminosity, the intensity of which increases for about one-sixtieth of a second : more rapidly towards the end of that period than at first. (2) Then ensues a sudden reaction, lasting also for about one- sixtieth of a second, in virtue of which the retina becomes partially insensible to renewed or continued luminous impressions. These two effects may be repeated in a diminished degree, as often as three or four times. (3) The stage of fluctuation is succeeded by a sensation of steady luminosity, the intensity of which is, however, considerably below the mean of that experienced during the first one-sixtieth of a second. * I Lave noticed that the intensity of the dark band always appears to fluctuate very rapidly, perhaps twenty or thirty times in a second. The rate of fluctuation is quite regular, and independent of the rate of rotation. 1 894] Images following Visual Impressions. Fia. 3. 143 (4) After the external light has been shut off, a sensation of diminishing luminosity continues for a short time, and is suc- ceeded by a brief interval of darkness. (5) Then follows a sudden and clearly-defined sensation of what may be called abnormal darkness — darker than common dack- ness — which lasts for about one-sixtieth of a second, and is followed by another interval of ordinary darkness. (6) Finally, in about a fifth of a second after the extinction of the external light, there occurs another transient impression of luminosity, generally violet coloured, after which the uni- formity of the darkness remains undisturbed. An attempt has been made in fig. 3 to give a rough diagrammatic representation of the above-described chain of sensations. No account has been taken of the comparatively feeble after-image, to which the phosphorescent trail before referred to is due, and which may last for two seconds or more. In conclusion, it may not be unnecessary to add a warning that, though all the effects here described have been witnessed without much difficulty by several persons besides myself, it is hardly probable that any one, who is quite unaccustomed to observations of the kind, will be entirely successful in a first attempt at repeating the ex- periments. Addendum. May 24th. Since the above was written, there has been brought to my notice an important paper by Dr. Carl Hess, " On the After-images follow- ing luminous Impressions of short Duration."* In his principal experiments momentary illumina'tion was produced by means of an instantaneous shutter, giving an exposure of 1/100 or 1/200 second. * Pfluger's ' Archly fur Physiologic,' vol. 49 (1891), p. 190. 144 Recurrent Images following Visual Impressions. [June 7, Observations were made of the effects following excitation by white light, by coloured light derived from different portions of the spec- trum, and by the whole spectrum at once. In all his experiments, Dr. Hess noticed the occurrence of what he speaks of as a negative after-image of very short duration (corresponding to what I have called an interval of darkness) which followed almost immediately upon the termination of the illumination, and preceded what is " commonly known as the positive after-image." He states that this negative after-image which, according to his measurements, lasted for about one-third or one-half second, was overlooked by Helmholtz, Aubert, Fich and others. Hering, however, had reasons for suspect- ing its existence, and it was to test this point that the experiments in which Hering himself co-operated, were undertaken. The negative after-images are stated not to have been represented in all cases by mere darkness. Under favourable conditions, the " dark after-image " succeeding a momentary excitation by coloured light, was tinted with a colour complementary to the original one ; and when the stimulus was generated by the complete spectrum, all the complementary colours were seen for a short time after its disap- pearance. No such complementary tints have ever revealed them- selves in my own experiments, the space between the primai'y luminous image and its ghost always appearing as simply dark. The colours assigned by Dr. Hess to the " positive after-images " also differ from those observed by myself. In most cases he describes the positive after-image as either having a feeble colour of the same hue as that of the light employed for the stimulus, or as being colourless. Dr. Hess considers, as I do, that the brightest portion of the posi- tive after-image of the spectrum corresponds with the green, and remarks that the brightness decreases gradually towards the more refrangible end of the spectrum, and much more quickly towards the less refrangible end. Such discrepancies as seem to exist between Dr. Hess's results and my own may perhaps be accounted for by the very different methods of observation which we employed. A stationary stimulus would, no doubt, be better adapted than a moving one for developing the feeble tints of the dark negative after-images, as well as those exhibited by the bright positive after-images during by far the greater part of their continuance, which, according to Dr. Hess's estimate, is as long as from four to eight seconds. On the other hand, the method adopted by myself discloses the important fact, of which Dr. Hess makes no mention whatever, that the positive images are immensely brighter for a very brief initial period — not more than one-tenth of a second — than during their subsequent existence. While this phase of transient brilliancy altogether failed to attract Dr. Hess's notice, 1894.] Niagara Falls as a Chronometer of Geological Time. 145 it constituted iu my own experiments the chief and most striking phenomenon : and it was to the colours which appeared during the bright phase that my attention was exclusively directed, the tints of the relatively insignificant " luminous trails " being too faint to be distinguish able . It is clear that the momentary excessive brightness of the positive image is no less essential than the dark interval (or negative after- image) for the generation of the phenomenon of recurrent vision which forms the subject of the present paper. III. " Niagara Falls as a Chronometer of Geological Time." By J. W. SPENCER, Ph.D. Communicated by Professor T. G. BONNET, F.R.S. Received March 16, 1894. (Abstract.) 1. Conjectures as to the Age of Niagara Falls. — Prior to the writing of the present paper, most of the conjectures as to the age of the Falls have been based simply upon the supposed uniform rate of recession. Thus, in 1790, Andrew Ellicott assigned 55,000 years as the age of the Falls. In 1841, Sir Charles Lyell allowed 35,000 years ; in 1886, Professor B. S. Woodward, after three surveys had been made, calculated the age as 12,000 years ; and later, Mr. G. K. Gilbert, supposing the recession to progress at' the maximum axial retreat lone, reduced the age of the Falls 6,000 years. This latter was not itended as an estimate, as he fully recognised that such a time must lave been greatly lengthened by many changing conditions. The rate iopted by the first two writers was only conjectural, as no surveys id then been made. Three surveys had been completed before the writings of the latter two writers, and I have had the benefit of a fourth. Woodward's calculation was upon the mean mathematical ilargement of the Horseshoe gulf at the end of the chasm, which ite was less than the geological rate of retreat. The author's method differs from the others in that it takes into consideration the rate of recession throughout the changing episodes of the river, which have been entirely discovered by Gilbert or himself. His com- putations make the age surprisingly near to the conjecture of Lyell. 2. Modern Topography. — This section of the paper gives such etails as bear upon the subject, some of which do not appear else- where. 3. Geology of the District. — Besides what may be found in other forks, there are several measured sections and descriptions showing le amount of work the river had to do. Several figures illustrate the varying conditions. VOL. LVI. L 146 Dr. J. W. Spencer. Niagara Falls [June 7, 4. Ancient Topography. — The Niagara is a modern river. It crosses a broad ancient valley nearly 100 ft. deep, in the vicinity of the Falls. This depression has largely escaped the attention of even geologists, and entirely in its bearing upon the history of the Falls. The pecu- liar extension of the chasm at the Whirlpool, and the buried valley of St. David's, have been considered by many as part of a preglacial Niagara river. This is now found to be a branch of a buried valley outside the Niagara canon, and hundreds of feet shallower, with ancient sloping \/-shaped walls, whilst those of the gorge are vertical. It is only an incident that the modern river touched this drift-filled valley, but it has given rise to the elongation of the chasm at the Whirlpool. The drainage of the tableland in ancient times was across the direction of the Niagara river, and was strongly marked by bold limestone ridges, which have only been penetrated by the Falls in modern times. Even the Erie basin emptied by a route several miles west of the Niagara. 5. Basement of the River. — In order to explain the work done by the river, this feature is described, part of the banks of the original course, before sinking into the chasm, being on hard rocks, and part on local deposits of drift. Even the deserted river banks carved out of such accumulations are still well preserved. 6. Discharge of the Niagara River. — This is only important in order to learn what is the discharge of the Erie basin alone ; for during a considerable portion of the life of the Niagara only the Erie waters fell over the falls. The drainage of the Erie basin is 3/11 of that of the four great upper lakes. 7. Modern Recession of the Falls. — From four surveys, extending over a period of forty-eight years, the mean modern rate of recession of the Falls is found to be 4'175 ft. a year. Its rate is variable with secular episodes of rapid medial recession, followed by its cessation along the axis, but with increased lateral retreat. This cycle appears to take about fifty years. But the detailed figures are given with a map. This rate is, however, excessive, on account of the geological con- ditions favouring the rapid modern recession, but the rate taken for the mean recession under the conditions of the modern descent of the river with the present discharge is 3' 75 ft. a year. 8. Sketch of the Lake History and the Nativity of Niagara River. — At one time a great proportion of the lake region was covered by a single sheet, or the Warren Water. Upon its dismemberment — in part, at least, by the rise of the land — one large lake was formed occupying the basins of Huron, Michigan, and Superior ; and another a portion of the Erie extending into the Ontario basin. The waters in these two basins were subsequently lowered, so that they fell to their rocky eastern rims, and the three upper lakes discharged by way of Lake Nipissing and the Ottawa river, and the Niagara had its birth, 1894.] as a Chronometer of Geological Time. 147 draining only the Erie basin. Then the Niagara river descended 200 ft. In course of time tbe waters subsided 220 ft. more, but eventually they were raised again 80 ft. at the mouth of the Niagara, thus reducing the descent of the river, from the head of the rapids above the falls to the foot of the last rapids in its course to the lake, to 320ft. During the lowest stage, Ontario lake receded twelve miles from the end of Niagara gorge, where the falls had been located at their nativity. 9. Laws of Erosion. — Theoretically the erosion varies as the height of the falls and the volume of the water, but some of the work is converted into heat. The recession is largely due to the work being expended in the undermining of the hard capping rocks, by the removal of the underlying shales. The rate of the modern recession has been determined under the changing conditions of erosion, so that the theoretical variations of other portions of the river's work includes their modification. 10 and 11. Episodes of the River and the amount of Recession in each. Duration of each Episode. — First episode : Water falling 200 ft., in volume 3/11 of modern discharge ; gorge, 11,000 ft. long ; duration, 17,200 years. Second episode : river descending 420 ft., in three cascades ; first stage, only the discharge of the Erie waters ; length of chasm, 3,000 ft. ; duration, 6,000 years ; second stage, drainage of all the upper lake ; length of chasm, 7,000 ft. ; duration, 4,000 years. Third episode : same volume and descent as in last, but the three falls united into one fall ; length of chasm, 4,000 feet ; duration, 800 years. Fourth episode : volume of water as at present, the level of lower lake as to-day ; first stage, a local rapid making the descent of 365 ft. ; work particularly hard ; length of gorge, 5,500 ft. ; dura- tion about 1,500 years ; the second stage as at present ; work easy ; length of canon, 6,000 feet ; descent of water, 320 ft. ; rate of reces- sion here taken as the full measured amount of 4'175 ft. a year ; duration, 1,500 years. Thus the age of the falls is computed to be 31,000 years, with another 1,000 years as the age of the river before the nativity of the Falls. The turning of the Huron waters into the Niagara was about 8,000 years ago, A difficult question was the amount of work done in each episode. This was in part determined by the position of the remaining terraces corresponding to different stages of the river, and by the changing effects of erosion. 12. Relations betiveen the Terrestrial or Epeirogenic Movements and the Falls. — The deserted beaches in the lake region have been deformed by unequal terrestrial elevation, and this movement has caused the changing conditions of the river in a large part, such as the turning of the Huron waters from the Ottawa valley to the Erie basin. This deformation affecting the Niagara district, since the commencement of the river epoch, amounts to 2'5 ft. per mile ; east of Lake Huron, L 2 148 Dr. V. Harley. The Influence of Intra- Venous [June 7, 4 ft. per mile ; and at the outlet of Lake Ontario, 5 ft. per mile ; all in a north-eastward direction. Taking the amount of movement in each district as representing also the proportional measure of time, then calculations can be made upon several of the beaches, and in terms of the age of Niagara their antiquity can be inferred. The importance of the computations in this paper is that they support the correctness of the calculated age of the Falls. In the application of these results it appears that the rate of terrestrial uplift in the Niagara district is about 1'25 ft. a century ; 2 ft. east of Lake Huron, and 2*5 ft. at the outlet of Lake Ontario. Here was found the first long looked-for indication of the rate of uplift. 13. The Relation of Niagara Falls to Geological Time. — From the study of the deserted beaches, it appears that the commencement of the lake epoch was as long before the birth of Niagara Falls as the Falls are old, so that the beginning of the lake age was probably 64,000 years ago, or perhaps even 80,000 years. Against this con- jecture we have as yet no proof. On the other hand, some suppose the lakes to have been held in by glacial dams, continuing for long episodes at the same level, and by the withdrawal of the glaciers the waters were lowered in addition to the terrestrial deformation. With this assumption, the retreating ice continued until the end of the Iroquois episode, or from our computations until 14,000 years ago. But here we need much more investigation. The present paper is merely a contribution in a field of work in America, in which only a few workers have so far contributed the detailed labours upon which this study is built. 14. The End of the Falls. — From the rate of terrestrial elevation and the rate of recession of the Falls, it appears that if the move- ments continue as they have been progressing, then before the Falls shall have retreated to Lake Erie, the Niagara outlet will have been deserted, and the waters of the upper lakes will discharge by way of Chicago into the Mississippi drainage, a change analogous to the turning of the Huron waters into the Erie valley from the Ottawa outlet. This change might be expected 7,000 — 8,000 years hence. IV. " The Influence of Intra- Venous Injection of Sugar on the Gases of the Blood." By VAUGHAN HARLEY, M.D., Teacher of Chemical Pathology, University College, London, Grocer Research Scholar. Communicated by GEORGE HARLEY, M.D., F.R.S. Received May 9, 1894. In a paper on "The Effects and Chemical Changes of Sugar injected into a Vein "* I showed that when grape sugar is injected- * ' Koy. Soc. PPOC.,' 1893. 1894.] Injection of Sugar on the Gases of the Blood. 149 into the jugular vein of a dog it causes an augmentation in the quantity of lactic acid in the circulation, the quantity of the acid steadily increasing until it reaches its maximum in about three hours after the injection. It then gradually, hour by hour, decreases, until in about six hours it returns to the normal amount. The question as to the base with which the lactic acid combines to form a lactate is, however, still unsettled. The results of Walter's* experiments, in conjunction with the often-noticed fact that ammonia is increased in the urine of diabetes, led me to imagine that the lactic acid combined with ammonia, until I found that the breaking up of sugar in the organism has no influ- ence whatsoever on the amount of ammonia in the blood, and conse- quently it cannot be the base. It then appeared to me probable that the lactic acid had combined with the bases of carbonates in the blood, having driven out the carbonic acid from its compound. In order to try and settle this point, I estimated the quantity of carbonic acid in the blood under different conditions. The series of experiments I am now about to record were performed in the Physiological Institute at Leipzig, and I wish to express my gratitude to Professor C. Ludwig for the kind assistance he gave me in the matter. The experiments, which were made on dogs, were conducted in the same manner as in my previous researches, above alluded to, except that blood was withdrawn only three times from each dog. In order to obtain a normal standard, the first specimen of blood was taken before the sugar was injected, the second was withdrawn in an hour, and the third in from three to five hours after the intra-venous injection of the sugar. In order that the composition of the blood might be altered as little as possible by the bleeding, only 30 c.c. of blood was collected each time. In all cases the blood was collected under mercury from the carotid artery. The gases were pumped from the blood by means of a Ludwig mercurial pump, and analysed by Bunsen's method. The quantities of gases found were calculated at 0° C. and 760 mm. of mercury, and are expressed in volumes per cent. Before alluding to the changes found in the blood gases, I will briefly give, in a tabular form, the results obtained from each experiment : — * Walter, 'Arch. Exper. Path. u. Pharm.,' TO!. 7, p. 158, 1877. Dr. V. Harley. The Influence of Intra-Venous [June 7r Experiment 1. Weight of animal. Condition. Quantity injected of Volumes per cent, at 0° C. and 760 mm. Hg. Sugar, in grams. NaCl solution. Carbonic acid. Oxygen. kilos. 7 Before sugar injection . . . 1 hour after . . 5 hours after. . total. 60 per kilo. 8-56 c.c. 120 37 '380 27 '006 34-357 22-280 17 -071 14 -886 The only nerve symptoms after the sugar injection were manifested in vomiting and muscular tremors. These were not accompanied by coma or any other symptom. The quantity of carbonic acid found in the standard specimen of blood was 37'380 per cent., whereas, in that taken an hour after the sugar injection it was only 27*006 per cent., that is to say, a diminu- tion of 10'374 per cent, in the amount of carbonic acid followed upon the intra-venous injection of the sugar, while the blood withdrawn five hours later contained 34'357 per cent, of carbonic acid, this being only 3*023 per cent, less than that in the standard blood. The carbonic acid, thus tending to return to the normal amount, showed that the influence of the sugar on the carbonic acid in the blood is merely temporary. The quantity of oxygen in the standard blood was found to be 22 '280 per cent. An hour after the sugar injection it had fallen to 17'071 per cent., thus giving a diminution of 5'209 per cent. Five houi's later it was still further decreased, being only 14'886 per cent. Consequently, in this respect the effect of the sugar on the oxygen is different from that upon the carbonic acid. Experiment 2. Weight of animal. Condition. Quantity injected of Volumes per cent, in 0° C. and 760 mm. Hg. Sugar, in grams. NaCl solution. Carbonic acid. 4 Oxygen. kilos. 5 Before sugar injection . . . 1 hour after . . 4| hours after. total. 50 per kilo. " c.c. 100 38-541 28-042 28 '92(5 19 '902 7-220 13 -968 1894.] Injection of Sugar on the Gases of the Shod. The nerve symptoms following the injection of the sugar were greater in this case. The vomiting and tremors of the limbs were followed by well-marked epileptic fits, which, an hour later, were succeeded by a semi-comatose condition. Although the animal could still be roused, it remained in a sleepy condition up to the third bleeding, when it was killed. The percentage of carbonic acid fell in the first hour after the sugar injection, while the animal was in a drowsy condition, from 38'541 to 28'042. This gives a diminution of carbonic acid of 10'499 per cent. Four and a half hours after the injection the carbonic acid had risen to 28'926 per cent. That is to say, it was 9'615 per cent, less than the quantity found in the standard blood. The oxygen which originally stood at 19*902 per cent, fell, in the first hour, to 7'220 per cent. ; therefore it was 12-682 per cent, less than the normal amount. In four and a half hours after the sugar injection it increased to 13*968 per cent., which is only 5'934 per cent, less than the original quantity found. Thus it appears in this case there was a greater diminution in both the carbonic acid and oxygen of the blood during the first hour than in the former experiment ; a result corresponding with the far greater nerve disturbances, and no doubt due, as stated in my former paper, to a larger percentage of sugar to bodily weight having been injected iuto the circulation. It was found in this case that the carbonic acid was, four and a half hours after the injection of the sugar, while the animal was in a semi-comatose state, almost as low as during the first hour. The oxygen had by this time, on the other hand, markedly increased in quantity. These results having been obtained, it was decided to withdraw the third portion of blood somewhat earlier after the sugar injection than in the foregoing cases. Experiment 3. Weight of animal. Condition. Quantity injected of Volumes per cent, at 0°C.and760mni.Hg. Sugar, in grams. NaCl solution. Carbonic acid. Oxygen. kilos. 23 Before sugar injection. .. 1 hour after . . total. 240 per kilo. 10 -435 c.c. 480 42 -260 33-075 10-217 3 hours after. . •• •• •• 38-000 14-569 Here the nervous symptoms which showed themselves in the form of vomiting, trembling, and two epileptic attacks were followed by 152 Dr. V. Harley. The Influence of Intra-Venous [June 7, drowsiness, which lasted until after the second bleeding. The drow- siness in this case passed off before the third bleeding. In the standard specimen of blood the carbonic acid was 42 '260 per cent., and an hour after the intra- venous injection of the sugar it fell to 33'075 per cent., being a decrease of 9'185 per cent. The third portion of blood taken in three hours, that is to say after the drowsiness had passed off, was found to contain 38'000 per cent, of carbonic acid, a decrease from the normal of 4'260 per cent. The specimen of oxygen from the normal blood was lost. The quantity found an hour after was 1O217 per cent., and three hours after the sugar injection it had increased to 14*569 per cent. The results of this experiment, as far as they go, correspond very closely with those of Experiment 1 ; in which there was likewise only a very slight nervous disturbance. Experiment 4. Weight of animal. Condition. Quantity injected of Volumes per cent, at 0° C. and 760 mm. Hg. Sugar, in grams. NaCl solution. Carbonic acid. Oxygen. kilos. 20-5 Before sugar injection. . . . 1 hour after . . total. 230 per kilo. 11.2 c.c. 460 39 -520 32-140 16 -025 15-561 3 hours after. . •• •• 24-725 17 '767 Although this dog had vomiting and marked tremor of the limbs there were no epileptic seizures. Sleepiness, however, came on later, and was marked at the time of the third bleeding. In the first specimen of blood the quantity of carbonic acid was 39'520 per cent., and it diminished during the first hour after the sugar injection to 32'140 per cent., being a decrease of 7*380 per cent. Three hours after the injection of the sugar, the carbonic acid fell still further, it being then only 27' 725 per cent., that is to say 14' 795 per cent, less than the original amount. The oxygen, which at the beginning was 16*025 per cent., de- creased during the first hour to 15'561 per cent., being a loss of 0*464 per cent. By the third hour it again rose to 17*767 per cent., that being 1'742 per cent, more than was found in the normal blood. As in Experiment 2, this dog had become semi-comatose by the third bleeding, the carbonic acid being then even less than what it was during the first hour. 1894.] Injection of Sugar on the Gases of the Blood. 15B Having now briefly given the results met with in each separate experiment, I will now consider the results as a whole. In the first place, we see there was a decrease in the quantity of carbonic acid in all the different specimens of blood during the first hour after the sugar was injected, the diminutions being 10*374, 10-499, 9-185, and 7*380 per cent. In the second place, the blood, taken five hours after the sugar was injected, showed a decrease of 3-023 and 9'615 per cent, (in Experi- ments 1 and 2) ; while after three hours there was a decrease of 4'260 per cent, (in Experiment 3). In all the three cases it had therefore shown, during the later hours, a more or less marked tendency to return to the normal amount. In Experiment 4 the blood at the third hour contained 14' 795 per cent, less carbonic acid than the normal blood, and 4*415 per cent, less than what it con- tained at the first hour. This discrepancy may be due to the fact that a greater percentage of sugar was injected, and the dog was in consequence rendered more comatose. This view seems the more likely, as in Experiment 2, when the dog was semi-comatose, the carbonic acid was markedly diminished at the fifth hour.* These united results support the view that the lactic acid derived from the splitting up of the sugar in the animal body drives off the carbonic acid from the sodium salts and replaces it. This view is still further supported by the fact that the quantity of carbonic acid in the blood withdrawn at the different periods after the sugar injection, varied in the same manner as the quantity of lactic acid had been found to do. In both cases during the first hour after the sugar injection, one finds larger quantities than during the later hours. Whether the percentage decrease in the amount of the carbonic acid hinders its elimination by the lungs or not will depend upon how much power the combined lactic acid has of hindering the blood from taking up the carbonic acid from the tissues and the tension of the existing gas. This point would be ascertained by estimating the quantity of carbonic acid expired after the sugar injection. The experiments I have already published! on this point show that there is no decrease the amount of carbonic acid expired from an animal immediately Eter sugar has been injected into its circulation. In fact there was an actual increase of carbonic acid in all but one casej during the irst hour. " Small dogs," as I stated in my former paper, " are relatively much more sus- ceptible than large ones to the effect of sugar injection." t Vaughan Harley, " Influence of Sugar in the Circulation on the Respiratory Oases," ' Journal of Physiol.,' vol. 15, p. 139, 1893. Ibid., Exp. 9, p. 147. 154 The Influence of Intra- Venous Injection of Sugar. [June 7, There was a marked decrease of carbonic acid during the later hours in those cases which suffered from coma. It would thus seem that in those cases when there are no nervous symptoms caused by the intra-venous injection of sugar, while the quantity of lactic acid is at its highest and the quantity of carbonic acid in the blood is at its lowest, more carbonic acid is expired than before the injection of the sugar. An explanation to this fact, if it really exists, is at the present moment impossible. In order to settle this point it would be neces- sary in the same animal to make all the analyses at the same time ; which would be impossible without an exceptionally large dog, as the quantity of blood needed would cause of itself changes in the metabolism. In the next place the changes met with in the quantity of oxygen in the blood are still more surprising, as there is no known reason why haemoglobin should not take up the usual amount of oxygen as. it does in health after the intra-venous injection of sugar. In all of the experiments the quantity of oxygen is seen to have been markedly diminished during the first hour after the sugar in- jection. In three of them it fell 0'464, 3'209, and 12-682 per cent, below the- normal standard during the first hour. This result can be partially explained by the influence of the endosmotic flow of the juices of the tissues into the circulating blood, which is known to occur when the quantity of sugar in the blood is increased. For in a series of similar experiments Brasol* found that during the first five minutes after the injection of sugar the proteids of the serum were reduced to even below one half of their previous amount, while one or two hours after the injection the proteids had, as a rule, returned to the normal amount. During the third and fifth hours after the sugar injection it will be noticed that the quantity of oxygen in the arterial blood was 14'886> 13'968, 14*569, and 17'767 per cent. ; that is to say the quantity that is usually found in venous blood. The diminution in the quantity of oxygen, even from three to five- hours after the sugar injection, cannot therefore be explained on a dilution theory. * Brasol. Du Bois-Beymond's ' Archir.,' 1884. 1894.] Contributions to the Life-History of the Foraminifera. 155 V. " Contributions to the Life-History of the Foraminifera.'* By J. J. LISTER, M.A., St. John's College, Cambridge. Communicated by Professor ALFRED NEWTON, F.R.S. Re- ceived May 7, 1894. (Abstract.) The phenomenon of dimorphism is now known to be presented by many different species of Foraminifera. The individuals of a species fall into two groups. In one the central chamber (the Megasphere of Munier-Chalmas and Schlum- berger) is of considerable size, while in the other it is small (Micro- sphere). These two forms of a species may be distinguished as the Megalospheric and Microspheric forms. They have been shown to differ, not only in the size of the central chamber, but. in some instances (Miliolidse), in the plan on which the chambers are arranged, in the size attained by the full-grown shell, and also in the frequency of their occurrence, the megalospheric form being much the more abundant. It has been suggested that the different conditions under which Orbulina universa is found represent the megalospheric and micro- spheric forms, but the reasons urged in favour of this view appear inconclusive. Polystomella crispa (Linn.). With the hope of throwing light on the life history of the Foraminifera, a large number of specimens of this species have been examined. Like so many others, it is dimorphic. Though the two forms are indistinguishable when the shell is complete, on examining decalcified and stained specimens they may be at once referred to one form OF the other. The central chamber of the megalospheric form is gener- ally about 80 ft in diameter, while that of the microspheric form is about 10^. Associated with the difference in the size of the central chambers there is a marked difference in the nuclei of the two forms. The relative frequency of the megalospheric form to the microspheric- in 1812 examples, is as 34 to 1. In the Microspheric form numbers of small nuclei are present, scattered through the protoplasm, but not extending into the terminal chambers. Those in the inner chamber are smaller than those situated further on. The nuclei contain nucleoli of different sizes, lying in an apparently homogeneous internucleolar substance. It is shown that the nuclei increase in number by simple division, and it appears probable that they are so derived from a single nucleus.. 156 Mr. J. J. Lister. Contributions to [June 7, After maintaining their rounded form for a certain time, the nuclei .give off portions of their substance into the surrounding protoplasm. This process appears to begin in the innermost chambers, but it ex- tends to the nuclei in the outermost chambers, and ultimately the whole of the nuclear material is distributed through the protoplasm in the form, in preserved specimens, of irregularly branched and deeply staining strands. Of the further history of the microspheric form I have no clear evidence. The Megalospheric form during the vegetative period of its life has a single large nucleus, which grows in size with the growth of the protoplasm, and passes on from chamber to chamber, moving to- wards the centre of the protoplasm contained in the series of •chambers, though lagging some distance short of it. It consists of a nuclear reticulum, nucleoli which occupy the nodes of the reticu- lum, and of a substance occupying the meshes. The nucleoli appear to increase in number and diminish in size with the advance of the organism. There is reason to believe that as the nucleus moves on through the chambers portions of its substance are given off into the protoplasm. It appears that this may occur either by the separation of considerable portions, sometimes containing several nucleoli, which lie strewn along the track of the nucleus, or by the dispersal of minute fragments into the surrounding protoplasm, causing in stained specimens a flush in the neighbourhood of the nucleus. In some specimens the nucleus has lost its rounded form, and sends irr.egular processes into the protoplasm. Its staining pi'operties are at the same time diminished. It appears probable that these nuclei are such as have given off a large part of their substances as above described, and are now in process of dissolution. In the reproductive phase no large nucleus is present, but hosts of minute nuclei (1 — 2/t in diameter) are found scattered through the protoplasm. At the same time broad channels of communication have become opened up, setting the inner chambers in direct com- munication with the outer. At first the small nuclei are most abundant in the terminal cham- bers, but ultimately they become uniformly distributed through the protoplasm. They then divide by karyokinesis, the protoplasm being aggregated about them in spherical masses, 3'5/t in diameter, each of which contains a dividing nucleus. At a later stage each nucleus, presumably the daughter-nuclei of this division, becomes the centre of a flagellated spore. These spores «re all of approximately equal size, in other words, they are isospores. In one instance spores of a different character were observed escaping. These were anisospores. They consisted of macrospores, globular bodies having a diameter of 11 — 10/t, and with indica- tions of a flagellum, and microstores of a globular or oval shape, from 1894.] the Life-History of the Foraminifera. 157 6 — 1 /* in diameter and provided with two flagella, one longer thanthe- other, rising close together from the body of the spore. I am. unable to say whether the parent of these spores was megalospheric or micro- spheric, but as the isospores are produced by the individuals of the former type it is possible that the anisospores belong to those of the latter. Orbitolites complanata, LamJc. In the Microspheric form, the centre of the disc is occupied by small chambers. Numbers of rounded nuclei are distributed through the protoplasm, often in pairs, and in some cases they may be seen to be united by a constricted band, as though in process of simple division. Larger solitary nuclei with a well marked reticulum are also present. In the later stages of growth large brood chambers are formed at the periphery of the disc, which Brady found to be crowded with young (" primitive discs ") of the megalospheric form. Examination of specimens preserved in spirit in which the young are present in the brood chambers, shows that the inner part of the shell is empty, its contents being represented only by the young. A large nucleus is present in the " primordial chambers " of the young discs. The centre of the Megalospheric form is occupied by the " primitive disc." This consists of a large " primordial chamber " (the megalo- sphere), which is usually pyriform, and measures about 100/t in length, surrounded by the very large " circumambient chamber." The small chambers of the remainder of the disc are arranged about the primitive disc in rings. The nucleus which, as has been said, occupies the primordial chamber in the young form, maintains that position during a large part of the growth of the shell. Ultimately it appears to break up into irregular fragments, which become dispersed through the adjoin- ing chambers. The specimens of this form from Celebes have all attained a larger size than those from Tonga and Fiji. In three cases (out of 114) the protoplasm has left the central region of the disc, and is massed in brood chambers at the periphery in the form of megalospheric young, exactly resembling in shape and size those borne by the microspheric- form. It is thus established that both the megalospheric and micro- spheric forms of Orbitolites under certain circumstances, produce young of the megalospheric type. An examination of specimens of Botalia beccarii (Linn.), Truncat- ulina lobatula, Walker and Jacob, Calcarina hispida, Brady, and Cyclo- clypeus has furnished evidence of the relation of nuclear characters to the two forms of a species analogous to that obtained in Polystomella. 158 Mr. J. J. Lister. Contributions to [June 7, Summai'y and Conclusions. The following statements relating to the life-history of the For- aminifera appear to be justified : — 1. The species are in a great number of cases dimorphic. The dimorphism has been stated to exist in twenty-three genera, belong- ing to four out of the ten families into which Brady divided the •group. 2. The two forms differ from one another — (a) In the size of the central chamber. Their difference in this respect is in many cases very marked but may be slight ( Truncatulina) . (fe) In the shape and mode of growth of the chambers succeeding the megalosphere and microsphere. (c) In the character of their nuclei. In this paper it is shown that in several species the microspheric form has many com- paratively small nuclei, while the megalospheric form has a single large nucleus. 3. The megalospheric form of a species is much more numerous than the microspheric. 4. The megalospheric form has been seen to arise in some cases (at least seven genera) as a young individual already invested by a shell, produced in the terminal or peripheral chambers of the parent. While in some cases (Orbitolites) the parent of such megalospheric young was microspheric, in others (Peneroplis, Ch-bitolites) it was megalospheric. 5. Foraminifera, in certain conditions, give rise to active swarm cells. These have been previously recorded in Gromia and Cymbalopora. In Polystomella the protoplasm of a megalospheric form was found broken up into swarm cells of uniform size (isospores), and similar bodies in a flagellated condition have been seen escaping. The production of anisospores has been recorded in Miliola (Schneider), and it occurs also in Polystomella as stated above. The question has arisen : are the two forms of the Foraminifera distinct from their origin, or is one a modification of the other ? The following reasons may be urged for rejecting the latter hypothesis : — Among the Miliolidae the plan of growth is often entirely different in the two forms. The hypothesis of modification would in this case require a remodelling of the whole shell. If such modification were to occur, various stages in the replace- ment of the megalosphere by small chambers should be found. So far as I am aware such stages have not been found. 1894.] the Life-History of the Foraminifera. 159 While the megalospheric form is not found in process of transition into the microspheric, it is found, either with the protoplasm broken up into swarm cells (Polystomella), or containing megalospheric young in the peripheral chambers, while the central chambers are empty (Orbitolites) . In both cases the megalosphere remained un- absorbed at the centre of the shell. The microspheric form is found in the young condition. The nuclear characters of the two forms are, at any rate, in the species which I have examined, quite distinct. It appears then that it may safely be concluded that the microspheric and megalospheric forms are distinct from their origin. What then is their relationship ? When two forms of a species are met with in animals or plants they generally either represent different sexes, or they are members of a recurring cycle of generations. The hypothesis that the two forms of the Foraminifera represent the two sexes appears to be disproved by the fact that in Orbitolites complanata, both megalospheric and microspheric forms are found with the young of the megalospheric form (primitive discs) in their brood chambers. Other genera furnish analogous, though less com- plete evidence. Hence it is impossible to regard either form as male. We turn then to the other hypothesis that the two forms are members of a recurring cycle of generations. On this view it is necessary to suppose, from the evidence afforded by Orbitolites com- planata, in which both microspheric and megalospheric forms have been found with the young of the megalospheric form in their brood chambers, that the megalospheric form may, at any rate in some genera, be repeated for one or more generations, before the microspheric form recurs. No evidence of such a repetition has, however, been furnished by the examination of Polystomella. The view that the life-history of the Foraminifera comprises more than one generation is in harmony with the fact that the nuclear history of the two forms in Polystomella, so far as it has been ob- served, presents resemblances to that which Brandt has recently described in Thalassicola amo'ng the Badiolaria. In this group, as is well known, the individuals of a species fall into two sets, those pro- ducing isospores and those producing anisospores, which are regarded as an asexual generation alternating with a sexual. The simultaneous division of nuclei by karyokinesis immediately before the formation of the reproductive elements which was ob- served in the megalospheric form of Polystomella is a phenomenon of very general occurrence. A similar division has been shown to occur 160 Presents. [June 7, in several genera of the Mycetozoa immediately before the formation of the spores, and it appears probable that the phenomenon is akin to the division of the micro-nucleus which precedes conjugation in the Infusoria, and to the division of nuclei which occurs in the maturation of the reproductive elements in the higher forms of ani- mals and plants. Presents, June 7, 1894. Transactions. Bnda-Pesth : — Ungarische Geologische Gesellschaft. Foldtani Kazlony. Kotet XXIV. Fiizet 1—5. 8vo. Budapest 1894. The Society. Cracow : — Academic des Sciences. Bulletin International. Avril, 1894. 8vo. Cracovie. The Academy. Kharkoff : — Societe des Sciences Experimentales. Travaux de la Section Medicale. 1893. [Russian.'] 8vo. KharJco/ 1894. The Society. Lausanne : — Societe Vaudoise des Sciences Natiirelles. Bulletin. Vol. XXX. No. 114. 8vo. Lausanne 1894. The Society. Leipsic : — Konigl. Sachs. Gesellschaft der Wissenschaften. Ab- handlungen (Math.-phys. Classe). Bd. XXI. No. 1. 8vo. Leipzig 1894; Berichte iiber die Verhandlungen (Math.- phys. Classe). 1894. No. 1. 8vo. Leipzig 1894. The Society. London : — British Astronomical Association. Journal. Vol. IV. No. 6. 8vo. London 1894. The Association. Geologists' Association. Proceedings. Vol. XIII. No. 7. 8vo. London 1894. The Association. Photographic Society of Great Britain. Journal and Transac- tions. Vol. XVIII. No. 9. 8vo. London 1894. The Society. New York : — American Museum of Natural History. Bulletin. Vol. VI. Pages 139—160. 8vo. [New York] 1894. The Museum. Paris : — ficole Normale Superieure. Annales Scientifiques. Tome XL No. 5. 4to. Paris 1894. The School. Societe de Geographic. Bulletin. Tome XIV. Trim. 4. 8vo. Paris 1894. The Society. Societe Mathematique. Bulletin. Tome XXII. No. 4. 8vo. Pans [1894]. The Society. Santiago : — Sociedad Nacional de Mineria. Boletin. Ano XI. No. 65. 4to. Santiago de Chile 1894. The Society. 1894] Presents. 161 Journals. Archives Neerlandaises des Sciences Exactes et Naturelles. Tome XXVII. Livr. 4—5. Tome XXVIII. Livr. 1. 8vo. Harlem 1894. Societe Hollandaise des Sciences, Harlem. Boletin de Minas Industria y Construcciones. Afio X. Num. 2 — 3. 4to. Lima 1894. Escuela Especial de Ingenieros, Lima. Giurisprudenza Internazionale. Anno I. Fasc. 1 — 12. 8vo. Napoli 1893. The Editor. Alcock (A.). Natural History Notes from H.M. Indian Marine Survey Steamer " Investigator." Series II. No. 1. On the Results of the Deep-Sea Dredging during the Season 1890-91 (cont.). 8vo. [London] 1894. The Author. Caruel (T.). Epitome Florae Europae terrarumque affinum. Fasc. 2. 8vo. Floreniiae 1894. The Author. Charleton (A. G.). The Choice of Coarse and Fine-crushing Machinery and Processes of Ore Treatment. 7 Parts. 8vo. Newcastle-upon-Tyne 1892-94. The Author. Collins (F. H.). Twelve Charts of the Tidal Streams on the West Coast of Scotland. Folio. London 1894. The Author. Espin (Rev. T. E.). The Distribution of Stars of Type III and of Stellar Spectra in Space. 8vo. Crook [1894]. The Author. Galilei (Galileo). Opere : Edizione Nazionale. Vol. IV. 4to. Firenze 1894. R. Ministero della Istruzione Pubblica, Rome. Helmholtz (H. von), For. Mem. R.S. Uber den Ursprung der richtigen Deutung unserer Sinneseindriicke. 8vo. Hambiirg [1894]. The Author. Hull (E.), F.R.S. Artesian Boring at New Lodge, near Windsor Forest (Berks.). 8vo. [London'] 1894. The Author. Laws (F. A.). On an Apparatus for the Measurement of Coefficients of Self-induction, and the Investigation of the Phenomena of Alternating Currents. 8vo. [Boston 1894] ; Preliminary Note on a Method for the Harmonic Analysis of Alternating Currents. 8vo. [Boston 1894]. The Author. Meyer (0. E.). Ueber electrische Eisenbahnen. 8vo. Breslau [1894]. The Author. Meyer (0. E.) und K. Miitzel. Ueber die Stb'rungen des Fern- sprechverkehrs durch electrische Strassenbahnen. 8vo. Berlin 1894. The Authors. VOL. LVI. 162 Prof. W. Ramsav aud Miss E. Aston. [June 14, June 14, 1894. The LORD KELVIN, D.C.L., LL.D., President, in the Chair. Mr. William Bateson, Mr. George Albert Boulenger, Professor Hugh Longbourne Callendar, Professor William Watson Cheyne, Mr. Robert Edmund Proude, Mr. Augustus Edward Hough Love, Mr. Francis Cranmer Penrose, Dr. Dukinfield Henry Scott, the Rev. Frederick John Smith, Mr. Joseph Wilson Swan, and Mr. Victor Herbert Veley were admitted into the Society. A List of the Presents received was laid on the table, and thanks ordered for them. The following Papers were read : — I. " The Molecular Surface-energy of the Esters, showing its Variation with Chemical Constitution." By Professor W. RAMSAY, Ph.D., F.R.S., and Miss EMILY ASTON, B.Sc. Received April 26, 1894. The investigation of the thermal relations of a series of esters by Professor Young* has made it possible to determine their molecular surface-energies between ordinary temperature and their critical points ; for the two important constants required for the calculation of these properties, the densities of the liquids and of their vapours in the saturated state (their orthobaric volumes) have been carefully determined by him. Professor Young has had the kindness to place his specimens at our disposal ; their purity is sufficiently guaranteed by the proofs afforded in his paper. Before using them they were tested for acidity, to ensure that no hydrolysis had occurred during accidental exposure ; but in no case was the reaction acid. The chief question to which an answer was sought was : Do these bodies confirm the general law of which experimental proof was fur- nished by one of the authors in conjunction with Dr. Shields, which may be thus stated — At approximately equal intervals of temperature below their critical temperatures all normal liquids possess equal molecular surface-energy ? The analogy of this law with that of Boyle is very striking; the latter may be expressed in almost identical terms — * ' Trans. Cliem. Soc.,' vol. 63, p. 1191. 1894.] The Molecular Surface-energy of the Esters, Sfc. 163 At equal intervals of temperature above absolute zero all normal gases possess equal molecular volume-energy. By " molecular volume-energy " is understood the product of pres- sure into molecular volume, that is, into the volume occupied by the molecular weight of the gas taken in grams ; while molecular surface- energy signifies the product of surface-tension and molecular-surface, that is, the surface on which equal numbers of molecules are supposed to be uniformly distributed, equal to the two-thirds power of the molecular volume of the liquid. The apparatus employed for low temperatures was that figured in the * Transactions of the Chemical Society,' vol. 63, p. 1094. A double set of observations was made, each set with a different capillary tube. One of the tubes was accidentally broken during the experiments, and was replaced by one of approximately the same radius. The radii, as described in a previous paper, were measured by means of a micro- scope with micrometer eye-piece ; tube A had a radius of O01843 mm. ; tube B of 0'01708 mm. ; and tube C of 0'01046 mm. These measure- ments were confirmed by determining the ascent of pure benzene in the tubes at known temperatures, and this is, on the whole, the easiest and most accurate method of determining their diameters. For higher temperatures, the apparatus, described in the ' Philo- sophical Transactions,' 1893, A, p. 662, was employed. In order to apply a correction for the capillary rise in the barometer-tube in which the capillary tube D was confined, a determination was made with each ester at some temperature (usually the boiling point of ilcohol under atmospheric pressure) at which capillary rise had been letermined in a wide tube with tube A, B, or C, where correction was innecessary, the ascent being taken, as customary, in inverse propor- tion to the radii of the tubes. As the variation of capillary rise with jmperature is approximately a linear one, a sufficiently accurate arrection may be obtained by assuming a rectilinear relation. Thus, for example, if at 78° the rise in the wide tube was 80 mm., and in le narrow tube 27 mm., it was necessary to add 3 mm. to the rise in ic narrow tube at that temperature. At the critical temperature the correction is, of course, zero, since at that temperature there is no capillary ascent in any tube. It was held that this difference decreased in the barometer tube proportionately with rise of tempera- ture, so that, for example, if the critical temperature were 278°, at the temperature 178° the correction applied amounted, to 1*5 mm. We regarded it as unnecessary to increase labour by taking ob- servations at each 10° rise of temperature, since a few points on the curve serve to show whether the rectilinear relation hol'ds. The plan of experiment was as follows : — The pressure tube containing the capillary tube was heated ill the vapour of chlorobenzene, boiling under atmospheric pressure about 132°, the exact temperature li 2 164 Prof. W. Ramsay and Miss E. Aston. [June 14, naturally depending on tlie barometric pressure of the day. As none of the esters boiled much above 100° at atmospheric pressure, it was possible by lowering pressure to cause them to boil at 132°, especially as in filling the tube a trace of air was purposely left in the liquid. It is not necessary that this air should be visible as a bubble, but it is sufficient if the liquid is not thoroughly boiled in vacuo. It may appear strange that such a course was followed, but repeated expe- rience has shown that if a liquid is wholly deprived of dissolved gas by boiling it in vacuo, it is impossible to cause it to boil, even at atmospheric pressure, although heated to 100U above its normal boiling point. Having determined the capillary rise at 132°, the pressure in the jacket was lowered, so as to cause the chlorobenzene to boil at 78° or 80°, care being taken not to allow the gas present in the upper part of the tube to condense wholly. The rise was again noted. The tube was then jacketed with quinoline vapour at about 185°, as well as at higher temperatures, and readings were again taken. Some six or seven points on the curve were thus determined, a sufficient 'number to characterise it. It will conduce to clearness to give the essential data at this stage, reserving details of experiment to the Appendix, where they are tabulated. For completeness' sake, the results previously published in the ' Philosophical Transactions ' for methyl formate and for ethyl acetate are here included. As the molecular surf ace- energy of a liquid, provided it does not dissociate with rise of temperature, may be calculated by means of the equation 7(Mt>)» = k(-r-d) (where Jc is a constant characteristic of each liquid but varying only slightly from 2'1, ? is the temperature measured from the critical point downwards, and d is a constant), the liquid is sufficiently characterised by giving the values of &, the critical temperature, and d. They are as follows : — Table I. Critical temperature, Ester. C.° Jc. d. Methyl formate 214'0 2'042 5'9 Methyl acetate 233'7 2'109 4'5 Methyl propionate 257'4 2'182 5'3 Methyl butyrate 281'25 2'220 3'75 Methyl isobutyrate 267-55 2'248 5'25 1894.] The Molecular Surface-energy of the Esters, $c. 165 Table I — continued. Critical temperature. Ester. C.° Tc. d. Ethyl formate 235'4 2-020 4'5 Ethyl acetate 251-0 2'226 6'7 Ethyl propionate 272'9 2'240 4'9 Propyl formate 264'85 2'110 4'85 Propyl acetate 276'2 2'227 5'0 It is evident at the first glance that it is the acid radical which determines the value of Tc, for it increases progressively with the pro- gressive increase of its molecular weight. Young has noticed a similar relation to hold with the ratios of absolute temperatures at corresponding pressures to absolute critical temperatures, but in other relations which he has investigated, there does not appear to be any analogous regularity. As regards the values of d, they appear to fluctuate as the series is ascended in the order of complexity of acid radical, but too great dependence must not be placed on the values given. A very small change in Jc would make a considerable difference in the value to be assigned to d, An attempt has been made to ascertain whether molecular volumes admit of more regular comparison at temperatures at which mole- cular surface-energies are equal. This appears, however, not to be the case. Thus, at the value 390 ergs, the group of four isomerides gives the following numbers : — Ester. Molecular volume. Methyl butyrate 126'17 Methyl isobutyrate 125'93 Ethyl propionate 126'80 Propyl acetate 127'16 The agreement is no better than at their boiling points under normal pressure. An attempt has also been made to find whether the boiling points at corresponding pressures bear a constant ratio to the temperatures of equal molecular surface-energy. Taking that ratio of pressure to critical pressure given in the fifth line of the table on p. 1245 of the iper in the ' Transactions,' and dividing the corresponding tem- perature for each liquid by the temperature of equal molecular surface-energy, the following ratios are obtained : — 166 Prof. W. Ramsay and Miss E. Aston. [June 14, Katio. Methyl formate 295-6/280' 1 = 1-055 Ethyl formate 3ir4/310'9 = 1-002 Methyl acetate 31375/317-5 = 0'9879 Propyl formate 332-5/348' 2 = 0-9545 Ethyl acetate 327'7/342'l = 0'9625 Methyl propionate 331-1/346-3 = 0'9559 Propyl acetate 34675/369-2 = 0-9393 Ethyl propionate 344-6/367-0 = 0'9391 Methyl butyrate 348-3/375-0 = 0'9290 Methyl isobntyrate 338'75/362'0 = 0'9356 These numbers may be roughly arranged into four groups : methyl formate, the isomerides of the formula C3H602, those of the formula CiHgC^, and those of the formula C5Hi002. They suffice to show that the molecular surface-energies are not comparable for non-isomeric bodies at corresponding pressures. It must, therefore, be concluded that, although a certain rough analogy exists between the corresponding temperatures and pressures of the esters and their molecular surface-energy, yet the causes which determine deviation from the deductions from the equations of condition for fluids, are still more operative in causing deviations when surface forces are under consideration. These experiments add eight more compounds to the list of six given in the ' Phil. Trans.,' 1893, A, p. 662, showing that within wide limits of temperature the molecular surf ace- energy of non-associating compounds is a linear function of the temperature ; and as the law has been found to hold between more restricted limits of temperature for other thirty (' Trans. Chem. Soc.', vol. 63, p. 1191), it may be taken as placed on a firm basis. A certain number of substances, among which are to be found the alcohols and the acids, show deviation from this law. Reserving to another occasion the grounds for inferring that this deviation points to molecular complexity, it is advisable to inquire here whether it is legi- timate to assume for compounds which follow the law that their mole- cular weight in the state of liquid is the same as that of their gases. Strictly speaking, the conclusion does not follow. The similar form of the surface-energy equation to that expressing volume-energy is a mere analogy ; there is no physical connexion as yet manifest between the two.* There is no positive evidence to show that the molecules of such liquids as follow the law do not associate in twos, or threes, on assuming the liquid state. But one thing is certain, if they do, all associate to an equal extent, and the degree of associa- tion is not altered by rise of temperature. * This conclusion must be modified in view of the recent memoir by ran der Waals (seep. 181). 1894,] The Molecular Surface-energy of the Esters, fyc. 167 These two assertions are probably not true ; it is unlikely that mere liquefaction should produce in all cases equal association ; and it is unlikely that a rise of temperature should not cause the dissocia- tion of an associated body. Change from the gaseous to the liquid state may be regarded as essentially equivalent to increase of pres- sure, since each produces approach between the molecules, diminish- ing intermolecular distance, and bringing so-called chemical forces into play. Now it is well known that equal rise of pressure does not always produce equal increment of association ; hence it is unlikely that association to an equal extent should be caused by the reduction of the volumes of compounds until they are approximately equal. This kind of proof is not unknown to chemists ; it is employed, tacitly perhaps, in the fundamental statement that the molecular formulae of hydrogen, oxygen, nitrogen, &c., are H2, 02, and N2. On this basis rests the usually accepted molecular formulae of all com- pounds, and they are accepted because they are the simplest expres- sions which admit of equations of chemical interchange being written. It is true that subsequently to the adoption of such a standard its justice was confirmed by Kundt and Warburg's deter- mination of the ratio between the specific heats of mercury gas at constant pressure and at constant volume, thereby rendering it ex- tremely probable that the molecular formula of mercury is Hg!, and consequently that of hydrogen H2 ; and by the discovery by Victor Meyer that the molecular formula of iodine at high temperatures cor- responds with Ii. But such confirmations merely supported the generally received assumption (for assumption it was then) that the molecular formulae of most gases are directly comparable with that of hydrogen as H2. Even at this present date the doctrine of the uniform expansion of gases at high temperatures rests on a similar basis. It has been shown by Victor Meyer that at the highest temperature attainable in a gas-furnace — some 1700° — hydrogen, oxygen, and nitrogen main- tain the same ratio of expansion towards each other. One of two conclusions follows : — either that the expansion of all three gases is uniform with increase of temperature, or that all three gases dis- sociate equally with equal rise of temperature. Needless to say that the first alternative is universally adopted. We have thought it well to state in full the reasons for adopting the assumption that the molecular weights of such liquids as the esters are not changed on their assuming the liquid state. It is now evident that such a statement is an assumption, a hypothesis ; but it is one for which there is a great deal of probability, probability of the same kind as that which led to the adoption of the usually received molecular formulae for gases. IliS Prof. W. Ramsay and Miss E. Aston. [June 14, 4 a 0 J o S t*f}* O IO r-l O5 (N O T? I-l IN ff7 ra oo IN IN (N 00 O CO to lO Ti a o IN r-l 00 (N t> CO CO C3 >O O -# CO O r-l •«* O t> •* 1-1 oo co »O OS (N 00 -* rH 00 00 • O IO rH d S 10 T? CO IO 'S1 CO l*^'* XO T? -* o-^ ?S »O vo O (N CO O (N » o co oo S38 O CO 00 l- O 7-1 00 05 IN CD CO O co co oo IN O O | •^i co co IN CO 00 CD 00 — ' •^ co co CO CO t^ (N •* 00 o •* co Tf ^»» 00 lO TJ4 CO a o O CO rH IO to •* 1 a o 5? oo o GO O 10 CD (N 'N OC W CO IN IN r>- (N O CD to § co xra -J i^ (M 00 rH r-l oo o co 10 S 00 OO CO rH 00 1— 1 o o ^ Oi *O C^l i— ( rH to o co IN IN r-l »3 O t> (N IN rH IO O t- • IN IN r-l o II • • IN N rH o II II i-* O * GO ^? O O IN r-l J> O (N O3 10 r-l i— I i— 1 O O CO CO oo co o .S ^T* rH CO '"O (N 00 •* c3 rH CD 00 .S Oi O3 rH ^ (N 00 IO =3 ^ M INININ IN (N (N CO 1 CO . O CD 00 •*> O rH TH O O O5 co co o 1 'S co oo t^ OS (M 1O •* CD Oi rH HH 1O IO ^ CO 1> 0 CC 00 rH o 8 1> 00 Tfl 10 >a IN co o o CO IO rH rH O O rH O 6 ii * N§^i 1C rH 00 ••# O CD IN (M rH .2 CO CD rH 1C 10 O (M 00 1C 1C CD O rH CO rH vc co co OS O I dy (M»)l J and it was on this assumption that the results given in the papers referred to for the alcohols, the acids, and water were calculated. The numbers obtained were, however, as will be shown, much in excess of the truth. An attempt was made to approximate to the true value of x, by calculating it by means of equation (4) approximately, and using the results obtained to correct equation (3), by inserting the neglected second term. This was found to be impossible, and to lead to absurd results; hence it was inferred that the variation of x with tem- perature was such as to make it imperative that attention should be paid to the second term of equation (3). At the same time it was noticed in mapping x that its alteration with temperature was 'approximately linear ; and this fact greatly simplified the problem. Mr. J. Rose-Innes, who has taken much interest in this work, and lias on several occasions given valuable assistance, was kind enough to endeavour to find an expression which would satisfy these condi - tions. A formula of the form (5) agrees admirably with the experimental values of molecular surface- energy for methyl and ethyl alcohols, water, and acetic acid between low temperatures and some 30° below their critical points. Even at — 89'8°, it will be noticed, the agreement for methyl and ethyl alcohols is reasonably good. The constants for these substances are : — Jc. Methyl alcohol ...... 1-489 Ethyl alcohol ....... 2-170 Water ............. 2'631 Acetic acid ......... 1-910 A comparison between the calculated and found values of •given in the following table : — d. -4-22 /*• 0-00104 Critical temperature. 240-0° 4-8 0-00193 243-1 19-5 0-00218 358-1 11-9 0-00163 321-5 1804.] Dissociation of the Molecules of Liquids. 177 TJ r-lOi-l(Mi>OSeOi-IO5 ^ 0 3 iHCO-.COSusobf-ioii^.iO(NOt-^r-iaoxn COCOCOCOiM(M(MlMiMr-lr-li-C s n3 CJ CO'* g- W).o kg | I* o .^i C>JIO?DCDIOCOOJ> OOOcC)'*t'ir'OOOlO Ifc !> t OOOOOQOGOXQOOOVD COWi— IOit^"Oj>oco ^ 15 3 0 COr-iaSUS!H»Oi>*>«O(NCOt>OOt- eoeooooccwot^-^i-it^kflcoiH •*cdCO|MNfH(MOO "o 1 ?- 1 COrHl^^rH»CQOOOt-lMWSt-O5G5 co«OQO-*i-ij>ioe<3r-i TjiCOCO(MNS<150THr-lr-l K w » • OOsICO WCOCOCOCOCOS5 05COCOMCO co-iJ-3l«DCOO!-l-(NO5I>U9lMOOTI>CO-*CO 's1 O s T3 00"**5COJ>OO05COOi-ICD-*^li-H"*i-ICDQU50i ^)t-i-(O5I>OCO OO51-^!O'^I(M CONINi-l'-lr-lrHr-l ^ i I oo n05OOOOOOOOOOOO (N0^t->racoi— lOit^coiOTficoiM CO N i-H iH i— 1 i-( e H 00 ncnoooooooooooo QONf>CJ>-iCOiOt-COC5Oi-((N 1 + VOL. T.VI. 178 Prof. W. Ramsay. The Complexity and the [June 14,. It may be remarked that at temperatures within 20 or 30 degrees of the critical point the former no longer accurately expresses the results. This is not peculiar to associating compounds, as has already been shown in the ' Phil. Trans.,' loc. cit., p. 657. Should it be desired to secure more accurate correspondence between the found data near the critical temperatures and those calculated, the last term may be modified. The equation then becomes For ethyl alcohol, the value of X is 0'044, and, on introducing this correction, the calculated values near the critical point, above 180°, are as follow : — T yU,, T ,M., Found. Calctd. Found. Calctd. 1 o O 240 3-1 3-7 3-9 200 43-1 75-7 76-9 236 7-1 9-9 9-9 190 53-1 94-9 95-1 234 9-1 13-3 13-2 180 63-1 112-6 112-8 230 13-1 19-8 20-3 178 73-1 130-1 129-9 220 23-1 39-2 39-0 160 83-1 147-2 146-4 210 33-1 57-1 58-1 150 93-1 163-0 162-5 Similar corrections could be introduced for methyl alcohol, acetie acid, and water, which would have the effect of reproducing the experimental numbers at low values of T. The following considerations show how it is possible to calculate the degree of association of such compounds at any desired tem- perature. Neglecting for the present the " X " term, which is intro- duced to secure concordance at temperatures near the critical point, let us consider equation (1), where k has the value 2' 121 for unasso- ciating liquids, viz., 7 (Mi/)* = 2-121 (T— d). Supposing that the liquid is composed partly of complex molecules,, and that a; is a measure of the complexity, we should have 7 (xM.v)% = 2-121 (T— d), or 7 (Mt>«) = 2-121 X -= X (7— l 1894.] Dissociation of the Molecules of Liquids. 179 Comparing this with equation (5), which reproduces the results for associating liquids with fair accuracy, it is evident that x corresponds to the expression >}'• It is, of course, easy to include the " X " term, when x follows, as before. There can, I think, be no doubt that this method gives a correct value to the factor of association, within certain limits. These limits are conditioned by the fact that the number chosen for &, viz., 2*121, is not absolutely constant, but varies with the nature of the com- pound. The extreme variation found for the fourteen substances which have been most carefully investigated is between 2'020 for ethyl formate, and 2'248 for methyl isobutyrate. On the assumption that this is the extreme divergence, there may be an error of 5 per cent, in a negative or positive direction caused by assuming the mean value 2-121. But there is another assumption involved in such calculations. It is that a mixture of two liquids possesses such a molecular surface- energy that the mean molecular weight of the mixture, calculated from the proportion in which they are present in the mixture, shall be deducible from the molecular surface-energy. It is conceivable that the surface of such a mixture should not exhibit the same dis- tribution of molecules as the interior, and evidence is required to show that the assumption that it does is correct. This evidence is riven in another communication, and it appears therefore that the ssumption is justified. With these premises, therefore, I proceed to give the molecular ssociation of methyl and ethyl alcohols, water, and acetic acid. N 2 180 Prof. W. Ramsay. The Complexity and the [June 14, fi co i-H COOSNCOOSINCOCSCO t^. i-H >O O Tp ^i co co IN (N *o fc US US US 1O US US US USIQ us us us us XCD-* o ^ w co co us co os H £• COUSUSrfi^eON . ^' e 1 l! X US co X X X X X X X co 1—1 os x> us co F—I CO CO (N IN IN IN (N . H b O O O O Q O O (N ^ CD X O IN -^t 1-1 l-l i-H : : : : H 8 lOOSINCOOSCOt^r-IUS COUSUS-^COCOININi-l OS CO CO Q 0 0 O O r-ICDO"#OSCOCO COi-H^IOOOJXXX i-l X »O X J> t^- (N 'o o 1^> o * °x IN CO ooooooooo INt—1-OCOi— lOSt— COU5 N i— 1 rH i— 1 i— 1 o o o -* CO IN H x .i? OOOOOOOOO £J i>» OS i— 1 CO US 1> X CS i— 1 i— 1 i— 1 rH i— 1 i— 1 8 2 i N N IN 1894.] Dissociation of the Molecules of Liquids. 181 Addendum. Since the foregoing pages were written, Professor van der Waals has published a long memoir on the " Thermodynamic Theory of Capillarity on the Assumption of Continuous Change of Density " (' Zeitschrif t fur physikalische Chemie,' vol. 13, pp. 657 — 725). On the main part of his work I have no criticism to offer; but on p. 714, he states some objections to the method previously employed by Ramsay and Shields in calculating the factor of association x. These remarks are fully justified, as will have been seen from the preceding pages of this paper ; but in the formula which he suggests to replace it, he makes an assumption which, at first sight, is no less untenable than our assumption that the factor of association, x, does not vary with temperature. In placing the factor of association as equal to unity at temperatures near the critical temperature, he obtains the formula . _ 7 (Mi;)1 and, inasmuch as this assumption is apparently very nearly true for methyl and ethyl alcohols and for acetic acid, the numbers he gives are nearly identical with those in the last table of this paper. But they differ in the case of water, and x, according to him, is equal to 1-9, instead of to 1707. The formula given by him on p. 716 to calculate 7 yields remark- ably good results. In fact, if 7 be calculated for ether at — 89'8°, a result is obtained identical with that found. This result was unfore- seen by van der Waals, for the value for 7, 30*65, was not given by us at that low temperature in our previous paper. Professor van der Waals, however, makes two criticisms which appear to me to be hardly justified. The first refers to a correction applied by us in order to allow for the capillarity in the wider, yet still narrow, tube in which the capillary tube stood. He thinks that this correction would be affected by the curvature of the meniscus not being the same at high as at low temperatures. The remark is certainly true ; but as the alteration in height due to altered curva- ture of meniscus would be well within the range of experimental error, it is negligible. The second criticism deals, with the capil- larity near the critical point, and van der Waals states that the simple formula, applicable to narrow tubes, no longer holds when the capillary rise is only a few times greater than the radius of the tube. This objection would be justified were it not that the capillary heights are nearly a linear function of the temperature ; and with non-dissociating liquids, which he is here considering, 182 Prof. W. Ramsay and Miss E. Aston. [June 14, it is quite unnecessary to take measurements at temperatures very close to the critical temperature, because, if that temperature is known, there can be only one curve joining the points experiment- ally found at somewhat lower temperatures and the critical tempera- ture. The form of the curve is such that no doubt can exist as to its course. Indeed, with chlorobenzene, measurements were not carried out at all in the immediate neighbourhood of the critical point, but only at much lower temperatures, and yet there could be no question as to the course of the curve, when it was mapped. But these are minor points ; and it is very gratifying to find that the material provided by Dr. Shields and myself affords such a remarkable con- firmation of the justice of Professor van der Waals' views. — 18th May, 1894. III. " The Molecular Surface-Energy of Mixtures of non- associating Liquids." By Professor WILLIAM RAMSAY, Ph.D., F.R.S., and Miss EMILY ASTON, B.Sc. Received April 26, 1894. It has been shown in the previous paper that it is possible to cal- culate the degree of association of an associating liquid such as alcohol, on the assumption that molecules of less complexity remain uniformly distributed along with molecules of greater complexity throughout the liquid, and that no one kind of molecule tends to congregate on the surface to the exclusion of the other. It is neces- sary, however, to justify this assumption ; and for this reason experi- ments have been made on mixtures of liquids the molecules of which do not unite to form complex groups ; such are most of the liquids investigated by Ramsay and Shields (' Trans. Chem. Soc.,' vol. 63, p. 1099, etseq.). The experiments of which an account is here given, show that while the height to which a mixture of two liquids ascends in a capil- lary tube is not the mean of the heights to which each singly would ascend at the same temperature, while the surface-tensions and the surface-energies are not necessarily the mean of those possessed by the liquids unmixed with each other, regard being paid to their rela- tive proportion in the mixture, yet the coefficient of decrease of mole- cular surface-energy, and consequently the calculated molecular weights, are true means of those of the two liquids. The substances used in these experiments were chosen in pairs, and as it was necessary in closing the tubes to evaporate some of the con- tained liquid in order to ensure the expulsion of air, mixtures of such liquids were taken as possess approximately equal boiling points, so 1894.] Molecular Surface-energy of Mixtures of Liquids. 183 that each should evaporate to nearly the same extent. For this reason the following liquids were chosen: — f Toluene U0°'6 at 761'2 mm. pressure. I Piperidine 105 — 1060<2 at 769 mm. pressure. * f Benzene Constant, about 80°. 1 Carbon tetrachloride Constant „ 77°. f Chlorobenzene Constant „ 132°. I Ethylene dibromide . Within O5°, about 131°. f Carbon disulphide . . . Constant, about 46'2°. I Chloroform Constant, about 62°. The amount of toluene distilled was 750 c.c. The thermometer did not vary during the distillation by the fiftieth part of a degree. The amount of piperidine was much less, about 75 c.c. The alteration of boiling point appears to be due, in part at least, to its eager absorp- tion of carbon dioxide. The benzene was part of a large stock which had been repeatedly frozen and thawed. It was free from thiophene, and had an absolutely constant boiling point. The carbon tetra- chloride boiled constantly while 400 c.c. passed over. The chloroben- zeue was part of a stock used for securing constant temperatures, and had been repeatedly fractionated ; it boiled with absolute con- stancy while 750 c.c. passed over. The ethylene dibromide was not quite so pure ; the rise of 0'5°, however, was spread over 200 c.c. ; while the purity of the carbon disulphide and the chloroform was guaranteed by the constancy of boiling point while large quantities distilled. The molecular surface-energies of the pure substances were first determined. They are given in the tables which follow : — t = temperature. h = rises in centimetres in capillary tube. p = density of liquid. 7 = surface-tension calculated by the equation 7 = \rgln>. r = radius of tube. 7(Mw)* = molecular surface-energy, where M = molecular weight, and v = volume of one gram. * The constancy of the boiling-point of a liquid is the guarantee of its purity, provided a considerable quantity boils at a constant temperature. The determina- tion of the actual temperature involves the accuracy of the thermometer. 184 Prof. W. Ramsay and Miss E. Aston. [June 14, S o co "* JO t— I p o oo oo 10 >*t g 3 9 IN IN (N ^ co ^ 00 co | 10 CO CD 0 I-l § 1 0 I-H (N 3 IN >>• os cc IN co cq IN *& o ^ co CO CO IN IN CO »O •it °10 o oo CO 00 0 00 IN co CO IN IN IN co *j °«3 *9 CO IN eo i-i 1O N (N *>» fW i— 1 0 IN IN T ^ CO CO S( U5 IN CO ^— ^ co CS (N ^ 10 o rs 'C eo 9 O CD OS eo CO IN J> CM rH 0 co 10 * rH X> os m oo •* co co CO 1> O eo CO CM - CO OS i> oo 00 iH rH rH CM CM CM «> S 3 O OS CM rH . CD 00 00 OS 00 CO co co N eq CM OS IN CO °3 1 t* eo rH o ^ CM O eq rH OS o CM co rH OS 00 CO * rH 0 ^~ CO eo os CO CM co eo co os m os (N rH cL CM in oo os rH' J^ rH O "^ CO O OS rH rH rH O vn o 00 CM 0 0 gs in i>» OS in ta ^ CO \a cp oo co 18 t- eo rH - CM 00 8 £ rH CM * "J* CM OO CO 4 ~ O5 H< eo OS CO O co oo w i> co co ^ ** CM t- m CO CM CM CO *r-l a o eo •* o 1 cL t- CM OS CM CM rH 0-01708. O rH rH rH i .t~ O m in o CM O i>- W CO CM CM II J> 0 0 i> °OS CO rH ^ CO CO O «o f*e o o o rH 0 o CM CM M>* II ^ t*» to C? i. § 1 § S CM CO 00 CO O OS Chloroform £•« co os CM CM rH «L i> in co O CO t- rH rH rH O O Q t~ (M 0 IO rH 1> CO CO 186 Prof. W. Ramsay and Miss E. Astoii. [June 14, These results call for no special remark, except that, contrary to the experiments of Ramsay and Shields, carbon disnlphide appears to associate somewhat at low temperatures. Further experiments will be made on this matter at still lower temperatures. The result given here may be taken as reliable, for it was carefully repeated several times, special precautions being taken to ensure the absolute purity of the bisulphide, and using a different capillary tube. The densities were taken from the following sources : — Toluene, Nasini and Pagliani, ' Jahresb.,' 1862, p. 63. Piperidine, Beilstein, ' Organische Chemie,' vol. 3, p. 616. Benzene, Kopp, 'Annalen,' vol. 64, p. 215. Carbon tetrachloride, Thorpe, ' Trans. Chem. Soc.,' vol. 37, p. 200. Chlorobenzene, determined by ourselves at the temperatures chosen. Ethylene dibromide, Thorpe, ibid., p. 197. Chloroform, ibid., p. 197. Carbon disulphide, ibid., p. 364. DETERMINATION OF THE CAPILLARY RISE OF MIXTURES. I. Toluene and Piperidine. (a.) 5C6H5.CH3 to 1C6H10:NH. In filling the tube with this mixture O'llS gram was lost out of a total of 4 grams, or a little over 2 per cent. It may be assumed that, owing to these liquids having so nearly the same boiling point, no material alteration of their ratio is due to this cause. The density of the mixture was assumed to be the mean of those of the constituents, taken in the proportion in which they were present. As will be after- wards shown, no appreciable error is involved in this assumption. The values of h are the mean of four observations in each case. In the columns with the heading " calculated " the mean height, surface-tension, and surface-energy have been inserted, together with the mean values of k. o IC«HIO:NH. Found. Calculated. t. h. p(calctd.). 7- 7(Mv)i. k. h. 7- 7(Mt;)l. k. 14-5 3-647 0 -8684 28-63 635-5 2 -191 3-622 28-63 631-8 2-079 46-6 3-285 0 -8377 24-88 565-6 2-032 3-283 24-53 565-0 2-277 78-4 2-945 0-8077" 21-54 501-0 2-123 2-904 21-19 492-6 2-013 132-5 2-323 0-7535 15-82 386-1 2-309 1573 383-7 1894.] Molecular Surface-energy of Mixtures of Liquids. 187 The mean value of K, calculated over the whole range of tempera- tare, is 2*115 ; the mean value found is 2*123 ; hence it may be con- cluded that these liquids are without influence on one another. It will also be noticed that the found and calculated values of 55 Silver (pure) . ivity between Temper i i 1 atures 10° C. Thermal conductivity n C.G.S. units. 0-9594 0-88838 0-8612 0-3497 0-3198 0-9628 0-7464 0-1861 and 97° C. Diameter. 2-00 mm. 2-11 „ 3-09 „ 2-04 „ 2-04 „ 2-02 „ 2-00 „ 2' 00 2 3 4 (very impure) 5 Gold Platinum Experiments to find out if there is any relation between the elec- trical and thermal conductivities confirmed what has been found by previous investigators, that if one metal is a better conductor for heat it is also a better conductor for electricity. The results did not, however, prove that the ratios were always the same, although in some cases they agreed very closely. For example — Conductivity of Specimen 2 in above table 0 «0 ,. T ^ . J •=-. — r- =2'78 for heat. Conductivity of Specimen 5 in above table = 2'86 for electricity. Conductivity of Specimen ^ _ 0.54 f v, f Conductivity of Specimen 4 = 2'56 for electricity. Conductivity of Specimen 1 , A0 £ -, — *— - = r08 for heat. Conductivity or Specimen 2 = T066 for electricity. 1894.] Presents. 203 While, however, these numbers agree very closely, other wires were tested in which the numbers varied considerably. It is intended to go on with tests of alloys, such as platinoid and German silver; also, by using liquids other than water, to obtain values of the variation of conductivity with temperature. Presents, June 14, 1894. Transactions. Berlin : — Gesellschaft fiir Erdkunde. Verhandlungen. Bd. XXI. No. 5. 8vo. Berlin 1894. The Society. Bucharest: — Societatea de Sciin^e Fizice. Buletinul. Anul III. No. 1 — 2. 8vo. Bucurescl 1894. The Society. Calcutta: — Indian Museum. Notes. Vol. III. No. 3. 8vo. Calcutta 1894. The Museum. Kew: — Royal Gardens. Bulletin of Miscellaneous Information. 1894. Appendix 2. 8vo. London. The Director. Liege : — Societe Geologique de Belgique. Annales. Tome XXI. Livr. 2. 8vo. Liege 1893-94. The Society. London : — Anthropological Institute. Journal. Vol. XXIII. No. 4. 8vo. London 1894. The Institute. East India Association. Journal. Vol. XXVI. No. 4. 8vo. London 1894. The Association. Entomological Society. Transactions. 1894. Part 2. 8vo. London. The Society. Manchester : — Geological Society. Transactions. Vol. XXII. Parts 16—18. 8vo. Manchester 1894. The Society. Meriden : — Meriden Scientific Association. Transactions. Vol. V. 8vo. Meriden, Conn. 1894. The Association. Milan: — Societa Italiana di Scienze Naturali. Atti. Vol. XXXIV. Fasc. 4. 8vo. Milano 1894. The Society. Moscow : — Societe Imperiale des Naturalistes. Bulletin. Annee 1893. No. 4. 8vo. Moscou 1894. The Society. Philadelphia : — American Philosophical Society. Proceedings. Vol. XXXI. No. 142. 8vo. Philadelphia 1893. The Society. Rio de Janeiro: — Museu Nacional. Archives. Vol. VIII. 4to. Rio de Janeiro 1892. The Museum. Washington : — Smithsonian Institution. Contributions to Know- ledge. Vol. XXVII. No. 884. 4to. Washington 1893. The Institution. United States National Museum. Bulletin. Nos. 44 — 46. 8vo. Washington 1893 ; Proceedings. Vol. XV. 8vo. Washington 1893; Report for the year ending June 30, 1891. 8vo. Wash- ington 18y2. The Museum. 204 Presents. Observations and Reports. India : — Archaeological Survey of India. The Bower Manuscript. Edited by A. F. R. Hoernle. Part 2. Fasc. 1. 4to. Calcutta 1894. The Survey. Madison : — Washburn Observatory of the University of Wisconsin. Publications. Vol. VIII. 8vo. Madison, Wis. 1893. The Observatory. Mauritius : — Royal Alfred Observatory. Annual Report of the Director. 1892. Folio. [Mauritius 1894.] The Observatory. Sydney : — Department of Public Works. Report. 1892. Folio. Sydney 1893. The Department. Washington: — United States Commission of Fish and Fisheries. Report of the Commissioner. Part XVII. 8vo. Washington 1893. The Commission. United States Department of Agriculture. Report of the Chief of the Weather Bureau. 1891-92. 4to. Washington 1893. The Department. Journals. Horological Journal. Vol. XXXVI. No. 430. 8vo. London 1894. British Horological Institute. Medico-Legal Journal. Vol. XI. No. 3. 8vo. New York 1893. The Editor. Nature Notes. Vol. V. No. 54. 8vo. London 1894. Selborne Society. Scientific Memoirs by Medical Officers of the Army of India. Part 8. 4to. Calcutta 1894. India Office. Debus (H.), F.R.S. Ueber einige Fundamental- Satze der Chemie insbesondere das Dalton-Avogadro'sche Gresetz. 8vo. Cassel 1894. The Author. Garcia de la Cruz (V.). Lois Mecaniques des Liquides Troubles et des Graz Nebuleux. 8vo. Barcelone 1894. The Author. Researches on Explosives. 205 June 21, 1894. The LORD KELVIN", D.C.L., LL.D., President, in the Chair. Dr. John Rose Bradford and Professor M. J. M.'Hill were admitted into the Society. A List of the Presents received was laid on the table, and thanks ordered for them. The following Papers were read : — I. " Researches on Explosives. Preliminary Note." By Captain Sir A. NOBLE, K.C.B., F.R.S., M.I.C.E., &c. Received June 13, 1894. The researches on which I, in conjunction with Sir F. Abel, have been engaged for very many years, have had their scope so altered and extended by the rapid advances which have been made in the science of explosives, that we have been unable to lay before the Society the results of the many hundreds of experiments under varied conditions which I have carried out. We are desirous also of clearing up some difficulties which have presented themselves with certain modern explosives when dealing with high densities and pressures, but the necessary investigations have occupied so much time that I am induced to lay a few of our results before the Society, trusting, however, that before long we may be able to submit a more complete memoir. A portion of our researches includes investigations into the trans- formation and ballistic properties of powders varying greatly in composition, but of which potassium nitrate is the chief constituent. In this preliminary note I propose to refer to powders of this descrip- tion chiefly for purposes of comparison, and shall devote my attention principally to gun-cotton and to those modern explosives of which gun-cotton forms a principal ingredient. In determining the transformation experienced during explosion, the same arrangements for firing the explosive and collecting the gases were followed as are described in our earlier researches,* and the gases themselves were, after being sealed, analysed either under the personal superintendence of Sir F. Abel, or of Professor Dewar, * ( Phil. Trans.,' rol. 165, p. 61. 206 Capt. Sir A. Noble. [June 21, and to Professor Dewar's advice and assistance I am indebted, I can hardly say to what extent. The heat developed by explosion, and the quantity of permanent gases generated were also determined as described in our researches, but the amount of water formed plays so important a part in the transformation that special means were adopted in order to obtain this product with exactness. The arrangement employed was as follows : — After explosion the gases formed were allowed to escape through two U -tubes filled with pumice stone and concentrated sulphuric acid ; when the gases had all escaped the explosion cylinder was opened, and the water deposited at the bottom of the cylinder was collected in a sponge, placed in a closed glass vessel and weighed. The cylinder was then nearly closed and heated, and a measured quantity of air was, by means of an aspirator, drawn slowly through the U -tubes till the cylinder was perfectly dry. This was easily ascertained by observing when moisture was no longer deposited on a cooled glass tube through which the air passed. The (J -tubes were then carefully weighed, the amount of moisture absorbed determined, and added to the quantity of water directly collected. The aqueous vapour in the air employed for drying was, for each experiment, determined and deducted from the gross amount. Numerous experiments were made to ascertain the relation of the tension of the various explosives employed, to the gravimetric density of the charge when fired in a close vessel, but I do not pro- pose here to pursue this part of our enquiry, both because the sub- ject is too large to be treated of in a preliminary note and because approximate values have already been published* for several of the explosives with which we have experimented. With certain explosives, the possibility or probability of detona - tion was very carefully investigated. In some cases the explosive was merely placed in the explosion vessel in close proximity to a charge of mercuric fulminate by which it was fired, but I found that the most satisfactory method of experiment was to place the charge to be experimented with in a small shell packed as tightly as possible, the shell then being placed in a large explosion vessel and fired by means of mercuric fulminate. The tension in the small shell at the moment of fracture and the tension in the large explosion vessel were in each experiment, carefully measured. It may be desirable here to explain that I do not consider the presence of a high pressure with any explosive as necessarily denot- ing detonation. With both cordite and gun-cotton I have developed enormous pressures, close upon 100 tons per square inch (about * Noble, ' Internal Ballistics/ 1892, p. 33 ; ' Eoy. Soc. Proc.,' vol. 52, p. 128. 1894.] Researches on Explosives. 207 15,000 atmospheres), but the former explosive I have not succeeded in. detonating, while gun-cotton can be detonated with the utmost ease. It is obvious that if we suppose a small charge fired in a vessel impervious to heat, the rapidity or slowness of combustion will make no difference in the developed pressure, and that pressure will be the highest of which the explosive is capable, regard being of course had to the density of the charge. I say a small charge, be- cause, if a large charge were in question and explosion took place with extreme rapidity, the nascent gases may give rise to such whirlwinds of pressure, if I may use the term, that any means we may have of registering the tension will show pressures very much higher than would be registered were the gases, at the same temper- ature, in a state of quiescence. I have had innumerable proofs of this action, but it is evident that in a very small charge the nascent gases will have much less energy than in the case of a large charge occupying a considerable space. The great increase in the magnitude of the charges fired from modern guns has rendered the question of erosion one of great im- portance. Few, who have not had actual experience, have any idea how rapidly with very large charges the surface of the bore is re- moved. Great attention has therefore been paid to this point, both in regard to the erosive power of different explosives and in regard to the capacity of different materials (chiefly different natures of steel) to resist the erosive action. The method I adopted for this purpose consisted in allowing large charges to escape through a small vent. The amount of the metal removed by the passage of the products of explosion, which amount was determined by calibration, was taken as a measure of the erosive power of the explosive. Experiments have also been made to determine the rate at which the products of explosion part with their heat to the surrounding envelope, the products of explosion being altogether confined. I shall only briefly allude to these experiments, as. although highly interesting, they have not been carried far enough to entitle me to speak with confidence as to final conclusions. Turning now to ballistic results. The energies which the new ex- plosives are capable of developing, and the high pressures at which the resulting gases are discharged from the muzzle of the gun, render- length of bore of increased importance. With the object of ascer- taining with more precision the advantages to be gained by length, the firm to which I belong has experimented with a 6-inch gun of 100 calibres in length. In the particular experiments to which I refer, the velocity and energy generated has not only been measured at the muzzle, but the velocity and the pressure producing this velocity have been obtained for every point of the bore, consequently 208 Capt. Sir A. Noble. [June 21, the loss of velocity and energy due to any particular shortening of the bore can be at once deduced. These results have been obtained by measuring the velocities every round at sixteen points in the bore and at the muzzle. These data enable a velocity curve to be laid down, while from this curve the corresponding pressure curve can be calculated. The maximum chamber pressure obtained by these means is corroborated by simul- taneous observations taken with crusher gauges, and the internal ballistics of various explosives have thus been completely de- termined. Commencing with gun-cotton, with which a very large number of analyses were made, with the view of determining whether there was any material difference in the decomposition dependent upon the pressure under which it was exploded, two descriptions were em- ployed : one in the form of hank or strand, and the other in the form of compressed pellets. Both natures were approximately of the same composition, of Waltham Abbey manufacture, containing in a dried sample about 4'4 per cent, of soluble cotton and 95'6 per cent, of insoluble. As used, it contained about 2'25 per cent, of moisture. The following were the results of the analyses of the permanent gases. They are placed in five series, viz. : — First. Analyses showing the decomposition of the strand or hank gun-cotton. Second. Analyses showing the decomposition of pellet gun-cotton. In both these series the analyses are arranged in the order of the ascending pressures under which the decomposition took place. Third and fourth. Examples of the decomposition of strand and pellet gun-cotton when exploded by means of mercuric fulminate ; and, fifth, a series showing the decomposition experienced by pellet gun-cotton saturated with from 25 to 30 per cent, of water, and deto- nated by means of a primer of dry gun-cotton and mercuric fulminate. I leave these results for discussion in the memoir which Sir F. Abel and I hope before long to submit, and will only remark that, in Tables I and II, the same peculiarity we have before remarked upon in reference to gunpowder, is again exhibited ; I mean the marked manner in which the carbonic anhydride increases with the pressure. ' It will be noted that in Table I the volumes of carbonic anhydride and carbonic oxide are nearly exactly reversed ; again, considering that the composition of the pellet and strand gun-cotton is practically the same, the distinct difference between the proportions of these pro- ducts in the two series is sufficiently remarkable. It not improbably is connected with the rapidity of combustion of the two samples. Another striking peculiarity is the manner in which the C02 is in- creased (as exhibited in Table V) when saturated pellet cotton is ' detonated. 1894] Researches on Explosives. 209 o w. a> rt be ? a . s o> a I > a a.. O oo t>. co 10 •* i-l IO t- rH CO O lO CD i> co co co CO o eo CO 0 >o oo co »o "* •* 00 CO CD CO CO IN r-l i-l o o IM r-l (N IO O N O CO IN CO 00 co o oo co 1-1 CO CO i-< rH O eo O os eo ic «O - oo os w> oo i-l O O oo Oi iM i-H (N •* X 1' O (M OS t>O CO CO r-t rH Ift (N CO CO 00 TP CO O i-l rH O w O5 IO i> 00 O (N CO r-l r-l o os CD GO >ra eq T}( co co oo co i-l CO CO C5 CO O (N CO i— 1 r-l •jj £ 1 1 • X > r5 „ O S S R S t> 1 T E A 1 o"o ' ' w QOMr^O o i 00 -* O 00 00 CO >p O5 oo to o -* N CO (N rH 0 >o N ^ 'Jl O 05 CO p Oi 00 •* O W> O C 00 CD to >o •<* o co l>- rH 00 IO 00 CD iH »1 CO iH T-H 1 0 VO O rH CO 00 J> O t» N CM p • E, »ft rH 10 00 OS »O i-H H CO rH -H Ions per sc t o •* 1-1 T-. CO J> W5 •* -* (N CO CO CO 1>1> Gi O O IOJ rH U5 CD CO W O5 OO CO CO Oi VO O IM CO rH i-H \n rH r-l O t- rH !O lO 00 O5 r-l co US OJ 00 Iffl O N CO rH i-l LQ CO IO Q X •. 00 »O CO rH rH O5 (M 1C O rj co «q I-H C4- 1 1 > a • 2 1 ,S) : : . . . 1 < t 1— 1 t ooHSo YOL. LTI. 210 Capt. Sir A. Noble. [June 21, III. Results of the Analyses of Strand Gun-cotton when fired in a Close Vessel by Detonation. Pressure* per sq. inch. A^ ^7 N 1 ton. 3 tons. C02(vols.) 19-21 29-08 CO „ 41-25 32-88 H „ 23-07 20-14 N „ 16-21 17-50 CH4 „ 0-26 0-75 IV. Similar Results for Pellet Gun-cotton. Pressure per sq. inch. 3 tons. 10 tons. C02 (vols.) 25-76 26-50 CO „ 39-34 37-48 H „ 18-71 20-97 N „ 16-19 15-05 CH4 „ Nil Nil V. Results of Analyses of Saturated Pellet Gun-cotton fired in a Close Vessel by Detonation. Pressure per square inch Under 10 tons. 10 '5 tons. 16 tons. 16 "5 tons. C02 (vols.) 32-14 33-25 32-93 35-60 CO „ .... 27-04 25-90 27'25 23-43 H „ .... 26-80 26-53 25-76 24-22 N , 13-83 14-32 14-06 15-25 CH4 „ .... 0-19 Nil Nil 1-50 Such are the average analyses of the permanent gases generated by the decomposition of gun-cotton under the various conditions I have described, and it will be evident from these analyses that the volumes of the permanent gases may be expected to differ to some very appreciable extent, depending both upon the density under which it is exploded, and also upon the mode of explosion. I have found it most convenient to explode the charges, the permanent gases from which were to be measured, under a pressure of about 10 tons per square inch (1,524 atmospheres), and, under these circumstances, the average of several very accordant determinations gave, at 0° C. and 760 mm. of mercury, 689 c.c. per gram of strand gun-cotton and 725 c.c. per gram of pellet gun-cotton. * The pressures given are those due to the gravimetric density of the charge. 1894.] Researches on Explosives. 211 At the temperature of explosion the whole of the water formed is in the gaseous state. It is therefore necessary, in order to obtain the total gaseous volume, to add to the above volumes of permanent gases the equivalent volume of aqueous vapour at the temperature and pressure stated. Now the quantity of water formed by the explosion of 129*6 grams of gun-cotton was found to be 16'985 grams ; hence 1 gram of gun-cotton generated 0'1311 gram of water, equivalent to 162'6 c.c. of aqueous vapour, and the total volume of gaseous matter at the temperature and pressure stated is for strand gun-cotton 852'2 c.c. per gram, for pellet 887'6 c.c. The heat measured reached, with strand gun-cotton, 1068 gram- units water fluid, or 988 gram-units water gaseous, while with pellet gun-cotton these figures were 1037 or 957 gram-units respec- tively. Pellet gun-cotton made at Stowmarket generated 738 c.c. of permanent gas and 994 units of heat per gram, while dinitro-cellulose containing 12'8 per cent, of nitrogen generated 748 c.c. of gas and 977 units of heat, the water in both cases being fluid. Gun-cotton, both pellet and strand, I have detonated by means of mercuric fulminate with ease and certainty. The effect of employing this means of ignition in a close vessel is very striking, and the indications of intense heat are much more apparent than when the charge is fired in the ordinary way. This effect is probably partly due to an actual higher temperature, caused by the greater rapidity of combustion. I allude elsewhere to the extreme rapidity with which the gases part with their heat, but this higher heat is, I think, clearly indicated by the surfaces of the internal crusher gauges becoming covered with innumerable small cracks and by thin laminae occasionally flaking off exposed surfaces; but perhaps the most striking proof of the violence of this detonation is shown by its action on a cast-iron shell fired as I have described ; where no detonation takes place the shell is broken into fragments of various sizes, such as are familiar to all acquainted with the bursting of shell ; but when detonation, with gun-cotton, for example, takes place, the whole shell is reduced to very minute fragments, and, what is more remark- able, two-thirds of the total weight are generally in the form of small peas and of the finest dust. The ease with which gun-cotton can be detonated renders it unsuit- able for use as a propulsive agent unless this property be in some way neutralised. I have, therefore, made but few experiments in this direction, and shall not further allude to them in this note, as more suitable explosives, explosives also of which gun-cotton is a principal component, have been elaborated, and these not only possess to the full the high ballistic properties of gun-cotton, but are more or less free from the tendency to detonate, which, however useful it may be p 2 212 Capt. Sir A. Noble. [June 21, in other directions, is a fatal objection to the employment of gun- cotton for propelling purposes. Turning now to cordite ; cordite consists, as is well known, of nitro-glycerine and gun-cotton as its main ingredients. As now made it contains 37 per cent, of gun-cotton (trinitro-cellulose with a small proportion of soluble gun-cotton), 58 percent, of nitro-glycerine, and 5 per cent, of a hydrocarbon known as vaselin. On account of the importance of this explosive, I have made numerous experiments, both with large and small charges, to determine the relation of the tension to the density of the charge. Up to densities of 0'55 the relation may be considered to be very approximately determined ; above that density, although many determinations have been made, these determinations have shown such wide variations that they cannot, until certain discrepancies are explained, be assumed as at all accurate. The average results of some of the analyses of the permanent gases are given below : — The first four analyses were made from experiments with the earlier samples of cordite when tannin formed an ingredient of cordite. They are not, therefore, strictly comparable with the later analyses. There appears also to be a difference in the transformation, slight but decided, which the same cordite experiences, dependent upon the diameter of the cord, and this difference is shown at once in the analyses, in the volume of permanent gases, in the heat developed, and, I think, in the amount of aqueous vapour formed. The following are some of the analyses : — VI. Pressure per square inch. / " > 0-048 Cordite. 0'255 Cordite. c 2 '5 tons. 6 tons. 10 tons. — \ 14 tons. 10 tons. 12 tons. 11 tons. — \ 14 tons. C02 CO H N CH4 .... 29-9 28-3 30-4 30-7 20-0 18-9 32-0 32-9 18-0 17-1 31-6 27-0 32-1 34-2 21-6 26-9 14-8 12-0 traces. 28-4 33-8 24-4 13-4 23-9 37-2 28-4 10-4 26-3 35-8 26-1 n. 19-3 . 22-5 In the whole of these analyses the water formed by the explosion smelt strongly of ammonia. The quantity of permanent gases measured, under the same con- ditions as in the case of gun-cotton, was found to be — For the earlier cordite, 655 vols. For the present service cordite, O255 in. in diameter, 692 vols., and for that 0'048 in. in diameter, 698 vols. In the two latter samples the aqueous vapour was determined, and was found to 1894.] Researches on Explosives. 213 amount to 20'257 grams for the 0'255-in. cordite, and to 20*126 grams for the 0'048-in. cordite ; or, stating the result per gram, these figures are respectively equivalent to 0'1563 gram, or 194 c.c. aqueous vapour, and to O1553 gram, or 192'5 c.c. per gram of cordite. Hence the total gaseous products generated by the explosion of cordite amount per gram to 886 c.c. for the 0'255-in. cordite, and to 890'5 c.c. for the 0'048-in. cordite, the volumes being, of course, taken at 0° C. and 760 mm. atmospheric pressure. The heat generated was found to be : — For the earlier cordite, 1214 gram-units water fluid ; for the service 0'255-in. cordite, 1284 gram-units water fluid or 1189 units water gaseous; for the service 0'048-in. cordite, 1272 units water fluid or 1178 units water gaseous. From my very numerous experiments on erosion I have arrived at the conclusion that the principal factors determining its amount are : (1) the actual temperature of the products of combustion, (2) the motion of these products. But little erosive effect is produced, even by the most erosive powders, in close vessels, or in those portions of the chambers of guns where the motion of the gas is feeble or nil ; but the case is widely different where there is rapid motion of the gases at high densities. It is not difficult absolutely to retain without leakage the products of explosions at very high pressures, but if there be any appreciable escape before the gases are cooled they instantly cut a way for themselves with astonishing rapidity, totally destroying- the surfaces over or through which they pass. Among all the ex- plosives with which I have experimented I have found that where the heat developed is low the erosive effect is also low. With ordinary powders, the most erosive with which I am ac- quainted is that which, on account of other properties, is used for the battering charges of heavy guns : I refer to brown prismatic powder. The erosive effect of cordite, if considered in relation to the energy generated by the two explosives, is very slightly greater than that of brown prismatic, but very much higher effects can, if it be so desired, be obtained with cordite, and, if the highest energy be demanded, the erosion will be proportionally greater. There is, however, one curious and satisfactory peculiarity connected with erosion by cordite. Erosion produced by ordinary gunpowder has the most singular effect on the metal of the gun, eating out large holes and forming long rough grooves, resembling a ploughed field in miniature, and these grooves have, moreover, the unpleasant habit of being very apt to develop into cracks ; but with cordite, so far as my experience goes, the erosion is of a very different character. The eddy holes and long grooves are absent, and the erosion appears to consist in a simple washing away of the surface of the steel barrel. Cordite does not detonate ; at least, although I have made far more experiments on detonation with this explosive than with any other, 214 Capt. Sir A Noble. f June 21, I have never succeeded in detonating it. With an explosive like cordite, capable of developing enormous pressures, it is, of course, easy, if the cordite be finely comminuted, to develop very high tensions, but, as I have already explained, a high pressure does not necessarily imply detonation. The rapidity with which cordite gases lose their temperature, and consequently their pressure, by communication of their heat to their surrounding envelope is very striking. Exploding a charge of about If Ibs. of cordite in a close vessel at a tension of a little over 6 tons on the square inch, or say 1000 atmospheres, I have found that the pressure of 6 tons per square inch was again reached in 0'07 sec. after explosion, of 5 tons in 0"171 sec., of 4 tons in 0'731 sec., of 3 tons in 1'764 sees., of 2 tons in 3'523 sees., and of 1 ton in 7*08 sees. The loss of pressure after 1 ton per square inch was reached was, of course, slow, but the figures I have given were closely approximated to in two subsequent experiments. With ordinary gunpowder the reduction of pressure was very much slower, as was to be expected, on account of the charge being much larger ; on account, also, of the temperature of explosion being much lower. These experiments are now being continued with larger charges and higher pressures. It only remains to give particulars as to ballistics, that is as to the velocities and energies realisable by cordite in the bore of a gun, but these will be most conveniently given with similar details regarding other explosives with which I have experimented. The ballistite I have used has, like the cordite, been changed in composition since the commencement of my experiments. The sample I used for my earlier experiments was nearly exactly com- posed of 50 per cent, of dinitro- cellulose (collodion cotton) and 50 per cent, of nitro-glycerine. The cubes were coated with graphite, and the nitro-cellulose was wholly soluble in ether alcohol. The second sample was nominally composed of 60 per cent, of nitro-cellulose and 40 per cent, of nitro-glycerine. The proximate analysis gave Nitro-glycerine 41*62 Nitro-cellulose 59'05 as before the whole of the nitro-cellulose was soluble in ether alcohol. The earlier sample gave the following permanent gases under pres- sures of six and twelve tons per square inch respectively. C02 37-3 38-49 CO 27*8 28-35 H 19-1 19-83 N 15-8 13-32 CH, traces. 1894.] Researches on Explosives, 215 One gram of this ballistite gives rise to 610 c.c. of permanent gases, and to 0*1588 gram of aqueous vapour corresponding to 197 c.c. at 0° C and 760 mm. Hence the total volume of gas is 807 c.c., and the heat generated by the explosion is 1,365 gram-units (water fluid), 1,269 gram-units (water gaseous). Although I have not made nearly so many experiments on detona- tion with ballistite as with cordite, those I have made with the earlier samples (50 per cent, gun-cotton and 50 per cent, nitro- glycerine) neither detonated, nor did they show any tendency to detonate, but the case is different with respect to a sample of ballis- tite consisting of 60 per cent, gun-cotton and 40 per cent, nitro- glycerine. This sample, 0'2-in. cubes, detonated with great violence on two occasions, but I am unable, without further experience, to say whether this result was due to the change in the composition of the ballistite or to defective manufacture. The erosive action of ballistite is, as might perhaps be anticipated from the higher heat developed, greater than with cordite, but the remarks made with respect to the action of cordite apply also to ballistite. The French B.N. powder consists of nitro-cellulose partially gelatinised and mixed with tannin, with barium and potassium nitrates. When exploded under a pressure of six tons per square inch the permanent gases were found to consist of C02 28-1 vols. CO 32-4 „ H 21-9 „ N 16-8 „ CH4 0-8 vol. These permanent gases occupied at the usual temperature and pres- sure a volume of 616 c.c. ; the aqueous vapour formed occupied in addition 206 c.c., so that the total gaseous volume was 822 c.c. The heat generated was 1,003 gram-units (water fluid) or 902 gram-units (water gaseous) ; the ballistics obtained with this powder are given along with those furnished by other explosives. For purposes of comparison I have introduced among the ballistic results those obtained with amide prismatic powder, and with B.L.Gr. Particulars as to both these powders have already been given* and need not here be repeated. In a preliminary note like the present, the most convenient mode * ' Eoy. Soc. Proc.,' vol. 52, p. 125 ; ' Phil. Trans.,' Part I, 1880, p. 278. 216 Capt. Sir A. Noble. [June 21, of comparing the velocities and energies developed by the new explo- sives is by the aid of diagrams. Accordingly, in Fig. 1, I show the velocities of seven different explosives from the commencement of motion to the muzzle of the gun ; the position of the points at which the velocity is determined are shown, and on the lowest and highest curves the observed velocities are marked where it is possible to do so without confusing the diagram. Lines are drawn to indicate the velocities that are obtained with the lengths of 40, 50, 75, and 100 calibres. Fig. 2 shows the pressures by which the velocities of Fig. 1 were obtained. The areas of these curves represent the energies realised, and the lines intersecting the curves indicate the pressures at which the gases are discharged from the muzzle for lengths of 40, 50, 75, and 100 calibres respectively. The chamber pressures indicated by crusher gauges are also shown in Fig. 2, and it will be observed that the two modes of determining the maximum pres- sure are in general in close accordance. It will further be observed «that with the slow-burning powders the chronoscopic maximum pressures are somewhat, though not greatly higher, than are those indicated by the crusher gauges. This observation is not new.* It was noted in the long series of experiments with black powders carried on by the Committee of Explosives. The result is widely different where an explosive powder or a quickly-burning powder, such as R.L.Gr., giving rise to wave-pressure is employed ; the crusher gauge in such casesf gives considerably and frequently very greatly higher pressures, and this peculiarity is illustrated in the curve from .B.L.G. in Fig. 2. It is, perhaps, hardly necessary to point out that the results given in Fig. 1 have to be considered in relation to the facts disclosed in Fig. 2. Thus it will be noted that the velocities and energies realised by 22 Ib. of 0'35-in. cordite and 20 Ib. of 0'3-in. cordite are practically the same, but reference to Fig. 2 shows that with the 0'3-in. cordite this velocity and energy has been obtained at the cost of nearly 30 per cent, higher maximum pressure. A similar remark may be made in regard to the French B.N. powder if compared with the ballistite. Its velocity and energy are obtained at a high cost of maximum pressure, and it is interesting to note how the velocity curve of B.N\, which for the first four feet of motion shows a velocity higher than that of any other ex- plosive, successively crosses other curves, and gives at the muzzle a velocity of 500 f.s. under that of cordite. The velocities and energies at the principal points indicated in * Noble and Abel, ' Phil. Trans.,' TO!. 165, p. 110. t Compare Noble and Abel, loc. tit., p. 109. 1894.] Researches on Explosives. 217 218 Capt. Sir A. Noble. [June 21, ' irt r 0 el — > wTUt ajo-nbr 3 1894.] Researches on Explosives. 219 2'20 Researches on Explosives. [June 21,. ni § If € p 00 05 $ O CD 00 CO CD CD rH •^p 00 ••# CO IO O ^1 £ 1O rH t> & o Oi 00 10 OS £ . IJ S 1 OS CO 1 10 10 o rH 10 1O o IO oo rH || Ja 'o o CO CO rH oo os rH oo CO S CO oo 00 S o CO (M (M CM CM IM rH g E o S S S eo oo S3 § U5 ,0 0 «*H ,S I S kO ^ „ ^ eo rH ^1 k> . Ifs 'o i 8 CO CO o CO rH CO (M os r-1 OS (M 10 (M CO IM IO ! ~. (M d . (M r • ifl 1 0 Nr c3 S3 4 JB N CO \ 0 'E. ? •73 >b •3 0 CO 1O (M .2 "§ | O> "? oo *? o fa CO 0 E 0 0? o *~-enki '-Ctl. \ \c> \\CI irve\ a. j irve\ b. j- 'or e or e ocper Toper inve irrie n,t o nk o^ f fO'l: f zz-a 94. 94* \ \ Q£ S3 \ r* "^^ — -, C D Distance. made consistent with the evident approach of the curve to its asymptote at D. In this case, therefore, the effect of the leakage extended to a distance of about 14 ft. from the pipe, and the fall of potential on the surface appears to have been most rapid at a distance from the line immediately above the pipe equal to twice the depth at which it was buried. A third experiment was carried out on the 22nd May, 1894, with the same arrangements as in the second experiment. The objects for which this was undertaken were chiefly to obtain a record of the variation of P.D. on the surface with the charging current, which had not been previously recorded, to obtain results from charging with continuous, as well as alternating, current, and to extend the investigation as regards the latter. Unfortunately, the continuous current generator could not be used, owing to a mechanical defect, which could not be removed in the limited time available. Still, the results obtained with alternating current are, it is considered, of some value, especially as the condition of the ground was different from that which existed at the time of VOL. LTI. S 258 On Establishment of Difference of Potential, §c. [June 21, either of the two former experiments, it being wet on the surface with, rain falling during a portion of the time, and dry underneath. Curve 6 in fig. 3 appears to show by its form the effect of this state of the soil, the fall of potential from B to D being approximately indicated by a straight line. That is to say that the current flow was practically entirely confined to the surface layer. The curve is deduced from the results of experiments shown in Table III for a charging pressure of 550 volts. Table III. — Experiment of 22nd May, 1804. At machine. Readings across weights. No. Volts. Current. Exciting current. On Cardew voltmeter. On s< atical Toltmeter. Eemarks. ' Weights A and B. 1 300 13-3 65 Weather wet after 2 345 15-17 — — 75 dry. Wet surface, 4 370 16-4 — — SOi dry underneath. 6 387 17 — — 84i Weights, each, 1/56 9 430 18-8 — — 95 Ibs., and arranged 10 500 21-7 — — 107 as in second experi- 11 550 24-35 — 122k ment. 12 610 27-75 — — 135 Pipe buried to same 13 625 28-8 " 142 depth. Weights A and C. 14 470 21-79 — 40 — 15 505 22-1 — 43£ — 16 550 25-53 — 47 — 17 565 26-7 — 50 . — . 18 610 28-6 - — 56 — 19 625 29 — 59 — Weights A and D. 20 495 23-11 12 55 21 530 24-3 13-5 59i — 22 555 25-53 14-25 63 1 — 23 580 26-4 15-75 67* — 24 — 21 -79 12 — 225 25 — 24-35 13-5 — 245 26 — 25-3 14 25 — 255 27 * — 26-4 15-75 — 262 In order to obtain a direct indication of the statical P.D. between weights A and B, the voltmeter which had been used for recording 1894.] Viscosity of Water as shown by Microrheometer. 259 the charging pressure was connected to these points and observations taken, when the exciting current for the field magnets of the dynamo and the charging current to the pipe were practically the same as in a previous set of experiments, for which the charging volts were known. Both sets of experiments are shown in the table. The dis- turbance of the worms was again very distinct, and they appeared, on reaching the surface, to set their bodies parallel to the pipe, or along an equipotential line. The experiments, as a whole, show conclusively that, under very different conditions of weather and amount of moisture in the soil, it is possible to produce a fall of potential amounting to 25 per cent, of the whole pressure in the supply between points on the surface of the ground 4 ft. apart, a distance which most horses stand over, and, further, that, although the connexion of such points by a conducting body naturally reduces this potential difference, yet, when this con- ductor has about the resistance of a horse (say 400 ohms), sufficient current will pass to give a severe shock. It is hoped that, when opportunity serves, these experiments may "be extended to the case of pipes buried to greater depths, and to observe whether any difference is produced by the use of a steady, in place of an alternating pressure. VI. ''On the Viscosity of Water as determined by Mr. J. B. Hannay by means of his Microrheometer." By ROBERT E. BARNETT, A.R.C.S. Communicated by Professor T. E. THORPE, F.R.S. Received May 6, 1894. In a paper entitled " On the Microrheometer," published in the * Philosophical Transactions ' for 1879, Vol. 170, p. 275, Mr. J. B. Hannay describes an apparatus with which he made measurements of the rate of flow of water, and of some aqueous saline solutions through a capillary tube, and he deduces from the observations cer- tain relations between the chemical nature of the salt dissolved, and its effect on the rate of flow of water. Inasmuch as Mr. Hannay furnishes us with details of the dimen- sions of his apparatus, I have, at Professor Thorpe's suggestion, transformed his relative numbers into absolute measure of viscosity in order to compare his results with those of other workers. The data given are as follows : — Diameter of capillary tube = 0'0938 mm. Length „ = 21 mm. Capacity of glass bulb = 4*0530 c.c. The time of flow was measured by a stop-watch, and each number recorded was the mean of ten observations. s 2 260 Mr. R. E. Barnett. On the Viscosity of [June 21 f In calculating1 the value of the viscosity of water from. Mr. Hannay's figures, the well-known formula V/» was employed, in which — rj = Viscosity in dynes per sq. cm. in C.Gr.S measure, r = Radius of the capillary in cm. I = Its length in cm. V = Volume of liquid transpired in c.c. p = Its density at the temperature of observation. p = Pressure in dynes per sq. cm. (g = 981). t = Time of flow in seconds. On applying this formula to the experimental results given for water, the figures embodied in the subjoined table (I) were obtained. In Columns 1 and 2 are the temperatures of observation and the corresponding times of flow as recorded by Mr. Hannay. Columns 3 Table I. 1. 2. 3. **. 5. r~> • J *»•*/» . f V .P . t. 8VI 8*1 t n- 0 235-0 0 -000514 0 -000327 0 -000187 1 220-8 0-000483 0 -000348 0 -000135 2 211-7 0-000463 0 -000363 o -oooioo 3 205-0 0 -000449 0 -000375 0 -000074 4 198-5 0 -000434 0 -000387 0 -000047 5 192-5 0 -000421 0 -000399 0 -000022 6 187-1 0 -000409 0-000410 -0-000001 7 181-0 0 -000396 0 -000424 - 0 -000028 10 167-2 0 -000366 0 -000459 -0-000093 15 146-0 0 -000320 | 0 -000523 -0-000206 20 131-3 0-000287 0 -000584 -0-000297 25 115-5 0 -000253 0 -000663 -0-000410 30 106-0 0 -000232 0 -000721 -0-000489 35 96-8 0 -000212 0 -000789 -0-000577 40 88-7 0 -000194 0 -000859 -0 -000665 45 81-8 0 -000179 | 0 -000930 -0-000751 50 75 -5 0-000165 0-001005 -0-000840 60 65-0 0 -000142 0 -001162 -0-001020 70 57-5 0 -000126 0 -001306 -0-001180 80 49-8 0 -000109 0 -001499 -0-001390 85 47-1 0-000103 0 -001579 -0-001476 90 45-5 0 -000100 0 -001630 -0-001530 95 44-3 0 -000097 0 -001668 -0-001571 100 43-8 0 -000096 0 -001681 -0-001585 1894.J Water as determined by the Microrheometer. 261 and 4 contain the calculated values for the first and second parts of the formula respectively, and in Column 5 are the values for the viscosity obtained by subtracting the figures in Column 4 from those corresponding to them in Column 3. On comparing these results with the values given by Poiseuille, Slotte, Sprung, and Thorpe and Rodger as tabulated below (IT), it will be seen that Mr. Hannay's observations yield discordant, and, indeed, utterly absurd, values for the viscosity of water. At 0°, for example, the viscosity would appear to be below that of any known liquid, and at 6° it becomes nil. Table II.— Water. Viscosity Coefficients in Dynes per sq. cm. between 0° and 100°. Temperature. Poiseuille. Sprung. Slotce. Thorpe and Rodger. 0 0 -01776 0 -01778 0-01808 0 -01778 10 0 -01309 0 -01301 0-01314 0 -013025 20 0 '01008 0 -01003 0 -01008 0 -010015 30 0-00803 0 -00802 0 -00803 0 -007975 40 0 '00653 0 -00657 0 -00657 0 -006535 50 0-00553 0-00553 0 -005475 60 0 -00472 0 -00468 70 . 0 -00408 0 -00406 80 0 -00358 0 -00356 90 0 -00318 0-003155 100 0 -00285 0 -00283 As a matter of fact, it is physically impossible to pass a volume of water such as Mr. Hannay employs under a pressure of 1 m. of water through a capillary of the dimensions given in the time re- corded. At 20°, for instance, the time of flow required under these conditions would be about 4600 seconds instead of 131 '3 seconds as stated. I have carefully examined Mr. Hannay's data with the object of discovering any slip or misprint which might satisfactorily account for the discrepancy. Thus, I have tried the effect of substituting "centimetre" for "millimetre," and "radius" for "diameter" in the dimensions given, but no alteration of the kind could be made to yield values for 17 agreeing with those of other observers. In the light of these results, it would seem to be premature to discuss Mr. Hannay's observations on saline solutions, or to criticise the generalisations he deduces from them. 262 The Rotation of the Electric Arc. [June 21r VII. " The Rotation of the Electric Arc." By ALEXANDER PELHAM TROTTER, B.A. Communicated by SILVANUS P. THOMPSON, F.R.S. Received June 12, In the course of experiments made with the view of realising as a practical standard of light, the method of using one square millimetre- or other definite area of the crater of the positive carbon of an electric arc,* the author has found that the effective luminosity is not as theory would predict,")" either constant or uniform. By the use of a double Rumford photometer, giving alternating fields, as in a Vernon Harcourt photometer, his attention was called to a bright spot at or near the middle of the crater. The use of rotating sectors accidentally revealed that a periodic phenomenon accompanied the appearance of this bright spot, and although it is more marked with a short hamming arc, the author believes that it is always present. An image of the crater was thrown on a screen by a photographic- lens ; and a disc having 60 arms and 60 openings of 3°, and rotating at from 100 to 400 revolutions per minute, was placed near the screen. Curious stroboscopic images were observed, indicating a- continually varying periodicity seldom higher than 450 per second, most frequently about 100, difficult to distinguish below 50 per second, and becoming with a long arc a mere flicker. The period seemed to correspond with the musical hum of the arc, which gener- ally breaks into a hiss at a note a little beyond 450 per second. The hum is audible in a telephone in the circuit, or in shunt to it. The- current was taken from the mains of the Kensington and Knights- bridge Electric Light Company, often late at night, after all the dynamos had been shut down. The carbons were, of course, not cored ; six kinds were used. A rotating disc was arranged near the lens, to allow the beam to- pass for about 1 /1000th of a second, and to be cut off for about 1/lOOth of a second. It was then found that a bright patch, occupy- ing about one quarter of the crater, appeared to be rapidly revolving, Examination of the shape of this patch showed that it consisted of the bright spot already mentioned, and of a curved appendage which swept round, sometimes changing the direction of its rotation. This appendage seemed to be approximately equivalent to a quadrant sheared concentrically through 90°. Distinct variations in the lumi- nosity of the crater are probably due to the fact that this is only an approximation. * J. Swinburne and S. P. Thompson, discussion on paper by the author, ' Inst. Electrical Eng.,' vol. 21, pp. 384 and 403. t Abney and Testing, ' Phil. Trans.,' 1881, p. 890 ; S. P. Thompson, ' Soc. Journ./ vol. 37, p. 332 1894.] Electric Strength of Mixtures of Nitrogen, fyc. 263 The a priori theory of the constant temperature of the crater is so attractive, that the author is inclined to attribute this phenomenon, not to any actual change of the luminosity of the crater, or to any wandering of the luminous area, as is seen with a long, unsteady arc, but to the refraction of the light by heated vapour. All experiments, such as enclosing the arc in a small chamber of transparent mica, or the use of magnets, or an air blast, have failed to produce any effect. A distortion of the image of the crater while the patch re- volves, has been looked for, but nothing distinguishable from changes of luminosity has been seen. An unexpected difficulty is thus introduced in the use of the arc as a standard of light, and one which may interfere with its use under some circumstances as a steady and continuous source of light. The author is further examining this phenomenon, with the view of ascertaining its nature, and of finding practical conditions under which it is absent or negligible. VIII. " The Electric Strength of Mixtures of Nitrogen and Hydrogen." By Miss P. G. FAWCETT. Communicated by Professor J. J. THOMSON, F.R.S. Received June 21, 1894. The experiments described in this paper were undertaken at Professor Thomson's suggestion, and have beeu carried out with the advantage of his advice and help. The immediate object of the experiments was to determine the electromotive force required to produce a spark between two flat parallel metal plates in a mixture of hydrogen and nitrogen in differ- ent proportions and at different pressures. The hydrogen used was obtained by electrolysis of water, as it was found that that obtained in the ordinary way from zinc and hydro- chloric acid was liable to contain impurities which seriously affected its electric strength. The two gases were collected over water in a graduated cylindrical gas-holder, and were allowed to stand for some hours to give them time to mix before being put into the apparatus. The mixture was passed through sulphuric acid, and also through cotton wool to remove dust. The electromotive force was supplied by a battery of storage cells, each of about 2 volts, and was measured simply by counting the number of cells. The strength of the cells was measured by a quadrant electrometer. At very low pressures it was found that, unless special precautions were taken to prevent the discharge passing anywhere except between 264 Miss P. G. Fawcett. The Electric Strength of [June 21, the opposed faces of the plates, it would take a longer path, and pass between the connecting wires on the backs of the plates. When the distance was 0'047 in., the discharge began to pass between the backs of the plates when the pressure was reduced to about 2 mm. In the later experiments this was prevented by using plates em- bedded in ebonite discs, only leaving exposed the faces between which the spark was intended to pass, and by making the connexions with indiarnbber covered wires. Even then, after the plates had been used for some time, the indiarubber showed signs of giving way, and a discharge occasionally passed partly between the wires and partly between the plates. The plates were kept at the right distance apart by placing between them small flat pieces of ebonite of the same thickness (O047 in.). In the earlier experiments, the plates were in an ordinary bell- jar standing on a flat surface, the rim being greased with a mixture of bee's-wax and vaseline. Thinking it possible that there might be some vapour given off by the grease, I arranged the apparatus so that it could be made air-tight without grease. For this purpose I used a rather narrow bell-jar, closed at the bottom by an indiarubber stop- per, through which passed three glass tubes for conveying the con- necting wires, and for communicating with the air-pump and the gas- holder. The jar, with its stopper, was placed in a vessel containing mercury, so that the junction of the glass and indiarubber was im- mersed, the tubes being bent so that their ends came above the mercury. The arrangement is shown in the accompanying figure ; a, a are the ebonite discs in which the plates are embedded ; the wires, Hattery b, 6, pass through the ebonite, and are covered with indiarubber throughout their length until they come out into the open air at the ends, A, A, of the tubes, c, c, which are sealed with sealing-wax. 1894.] Mixtures of Nitrogen and Hydrogen. 265 I-H 1 oo co c^-i o o o o o o 0 A (N l-H rM r-l § CO CO O 00 Q O Q •§ pM §» , 49 g 3 • "sS ^ 13 CO 5 ^ ,-j« r4M — !« >-4e* -(« 00 Tfi rH OS CO •"* N rM r-l rM o> r n =4H 0) PH Q rH -l« ^.H« -to tw O £ rM rM (N rM rM rM _O c3 • rfl a OO CO O CO TP O O CO O G5 t>» O O t» N H So JT^ co O -<-> £ -|o «|t -It t^ C^l CO *O (N (N rM rM a rg « p a CO 00 00 O CO O O CO 00 CO 1> 00 t^ O ^ co co co co "^i oo s &c «3 rH H lO CO IN rM PM rM rM ^ coocococooocoocoo XOOQCOCOOCDiMt^cCO O W rH £ N rM rM rM Miss P. G. Fawcett. 2he Electric Strength of [June 21, The results (p. 265) were obtained when no grease* was present, and the spark was not able to pass except between the opposed sur- faces of the plates. Distance between the plates = 0'047 in. These results are represented by the continuous curves figs. 1 — 7, in which the abscissae represent the pressure in millimetres of mercury, and the ordinates the E.M.F. in volts. The dotted curves in the same figures represent the means of the results of several series of observations with the earlier arrangement of the apparatus, in which grease was present, and no precautions were taken to prevent the discharge passing otherwise than between the plates. FIG. 1. Fio. 2. Voli 1100 9. N 900 eoo 700 600 •500 400 Wl x^ / r / / I J 400 300 •20 ZQm.m. * There was always a small amount of grease on the stop-cock of the air-pump, but it does not seem probable that such a small quantity would have an appreciable effect. 1894.] Mixtures of Nitrogen and Hydrogen. 267 FlGK 3. . 4. Volts 1100 1000 600 700 600 50C 40 300 N: H= 3:2 Void N= H 600 700 COO , 500 * . > / // '/ 400 \ . //'' / 3 ^^ / 3CO- 10 20 268 Miss P. G. Fawcett. The Electric Strength of [June 21, Volt 1100 1000 900 600 700 600 500 400 300 y N: Ht= 2:3 X, / x. / 1 ,_ _/ /' 500 400 300 20 SOm.m. 1894.] Mixtures of Nitrogen and Hydrogen. 269 FIG. 7. 400 30/717^. These curves are not traced for pressures lower than about 2 mm. The two curves are nearly identical in the case of pure nitrogen ; in the other cases, with one exception, that of N:H = 2:1, the discharge seems to pass more easily when grease is present. It will be noticed that at fairly high pressures the E.M.F. required to produce a spark diminishes nearly uniformly as the pressure diminishes, but that there exists for each mixture a critical pressure at which the E.M.F. is a minimum, and that as the pressure dimin- ishes from the critical value the E.M.F. continually increases. When the pressure is slightly less than the critical pressure the E.M.F. increases with remarkable rapidity as the pressure diminishes. In fact, as the pressure falls by about ^ mm. the E.M.F. may increase by several hundred volts. The critical pressure diminishes as the proportion of nitrogen to 270 Electric Strength of Mixtures of Nitrogen, fyc. [June 21. hydrogen is increased. With spark length 0'047 in. it varies from, about 11 mm. in pure hydrogen to about 5 mm. in pure nitrogen. At high pressures the E.M.F. required to produce a discharge at a given pressure diminishes continually as the proportion of hydrogen to nitrogen is increased. At low pressures the relation between the composition of the mixture and the E.M.F. required to produce a spark at a given pressure is less simple. It is represented by the curves in fig. 8, where the ordinates give the E.M.F. in volts, and the ratio in which the ordinate divides the line NH is equal to the ratio of the volumes of nitrogen and hydrogen in the mixture. FIG, 8. 600 500 400 Ni:0 0:iH The curves are given for pressures of 20, 15, 10, 5, and 4 mm. They could not be considered accurate for lower pressures, owing to the extreme steepness of the curves (1 — 7), in consequence of which the E.M.F., which will produce a spark at a given low pressure, cannot be determined with any precision. It will be seen that at the pressures 15 and 20 mm. the general slope of the curves is downwards from the nitrogen end to the hydrogen end, but that at pressures 10, 5, 4 the slope is the other way, showing that at low pressures the effect of introducing more hydrogen is in general to increase the electric strength of the mixture. But as the proportion of nitrogen to hydrogen increases from 0 : 1 to 1 : 0, the electric strength does not diminish uniformly, but it may pass through one or more maxima and minima. It is, perhaps, hardly necessary to say that the curves in fig. 8 c-annot be regarded as accurate for strengths of mixture intermediate between those at which the observations were actually made. The t 1894.] The Asymmetrical Probability Curve. 271 curves, especially at low pressures, must be considered rather as a convenient way of showing which dots in the figure correspond to any given pressure than as an attempt at interpolation. IX. " The Asymmetrical Probability Curve." By F. Y. EDGE- WORTH, M.A., D.C.L. Communicated by Sir G. G. STOKES, F.R.S. Received June 14, 1894. (Abstract.) The asymmetrical probability carve is the second approximation — the symmetrical probability curve being the first approximation — to the law of frequency which governs the set of values assumed by a function of numerous independently fluctuating small quantities. The curve may be written 1 =*T 2 x 2x3\-\ = ——ec* * — 5- ;• •v/^c L c3 \c 3 c*/J where i/Aa; is the number of errors occurring between x and a;-t-Aar, ^12 is the mean square of errors, and j is the mean cube of errors — errors measured from the centre of gravity. This form is obtained by completing the analysis which Todhunter, after Poisson, has indicated (' History of Probabilities,' Art. 1002) ; and independently by obtaining a general form for the asymmetric probability curve, and deducing therefrom the Poissonian formula in the case when the asymmetry is slight — the only case to which that formula is applic- able. Among the peculiarities of the asymmetric probability curve are the want of coincidence between the arithmetic mean and the position of the greatest ordinate, and the descent of the curve at one ex- tremity below the abscissa — the ordinate appearing to denote negative probability. An important case of the general carve is afforded by the Binomial, for which each of the independent elements admits of only two values. The approximate form of the Binomial, obtained directly by Laplace (Todhunter, 'History,' Art. 993), is deducible from the general theory. The general, or multinomial, probability carve can always be represented by a binomial. The principle of the asymmetric probability curve affords an exten- sion of the theory of correlation investigated by Messrs. Galton and Hamilton Dickson ('Roy. Soc. Proc.,' 1886, p. 63). The symmetrical probability surface 272 Mr. R. F. Gwyther. The Differential [June 21, becomes now slightly distorted, so that the locus of the most probable y deviation, corresponding to an assigned x deviation is no longer a straight line, but a parabola. X. " The Differential Covariants of Twisted Curves, with some Illustrations of the Application to Quartic Curves." By R. F. GWYTHER, M.A., Fielden Lecturer in Mathematics, Owens College, Manchester. Communicated by Professor HORACE LAMB, F.R.S. Received May 31, 1894. (Abstract.) The object of the earlier parts of this paper is to obtain relations connecting what Halphen* calls the canonical invariants of the curve, without the intervention of what are called by him the fundamental invariants. In any geometrical investigation it is the canonical in- variants which present themselves, and the relations between the con- secutive canonical invariants, and the values of their differential coefficients must, if the method of investigation is to be used, be expressed, in terms of canonical invariants only, without the inter- vention of the other series of invariants which Halphen treats as fundamental. In the paper, the notation of Halphen's paper cited above is gener- ally followed, but the mode of initial investigation, as in a previous paper on covariants of plane curves,f is made to depend upon a homographic transformation with infinitesimal arguments. Writing £, ij, £ for the coordinates of a current point, #, y, z for the coordinates of a point on a standard curve, and yn and zn for ctPyjdaf . n ! and d"z/dxn . n !, the arguments of a covariant function are shown to be g = zz^—y—yi ?—« — with an = y^n—ynZ2jy2z3—y3z2, while yzz3— 2/3^2 is an invariant which will not appear independently. The other conditions that a function 0 (/, gr, h, an, 6». . . . ) may be a covariant function are (1) that it shall be isobaric, counting/, g, h, an, * " Sur les Invariants Difflirentiels des Courbes Gauches," ' Journal de 1'^cole 1'olytechnique,' cahier 47, vol. 28, 1880. t ' Phil. Trans.,' vol. 184 (A), 1893, p. 1171. 1894.] Covariants of Twisted Curves, &c. 273 and l>n of the weights — 1, — 3, — 2, n— 3, and n — 2 respectively, and (2) that, if d be the algebraic degree of the function, then ^ O/ {A — df — 00;, n-m+i— 2a«) -^— -(0 -2) &„_!-<) ^- -- .... 0 = 0 - o/ o/ r) 1 -S (w — 1) &«&»_,, TT -- ---- >0 = 06 J — S (m— 1) bman-m- -- -- Cbn These equations are then used to find the forms of the earlier in- variants, and to show that there are only four independent covariants, which may be written ia = 1 — VOL. LVI. 274 Mr. R. F. Gwyther. The Differential [June 21r where «4 = «4, us = fls— 3&4— a42, and The planes, whose equations are £ = 0, y = 0, % = 0, ia = 0, form a tetrahedron, which will be called the canonical tetrahedron of reference at any point on the curve, and if we put the coefficients of the expansions of ?/ and z in terms of « are the canonical invariants. We write these expansions y = x* +ft1x'+&c., Z = X? + eti,X* + aiiX1 +...., and since g = /2+ .... +bifi+&c., we see that with respect to the canonical axes, a4 = 64 = a5 = 66 = J6 = 0. To obtain the result of differentiating a canonical invariant, we must- begin with a general change of axes, and after differentiation suppose that these are the canonical axes, and put a4 = J4 = &c., = 0, a& above. For this purpose, I show that if Y — ^ v — ^ -7 — £ -A. — — , I — — , Zi — — » w w u denote any homographic transformation, we shall have A, = &c. + &c. 1894.] Covariants of Twisted Curves, &c. 275 where R contains squares and products of «i/o>, u>^ — u-'z^/f^i — w^ &c., where w« and £„ have a significance similar to that of yn and zn. In the earlier part of the paper I have shown that, if f = x + q^y and ID = l+px -f qy, where the coefficients are small, A,, = an— qi {(n— 3) yian +p {2 O — 3) xan+(n—2) «„_!} + q{x [(n— 3) 2/,an+ SmaMyB_m+1] +(w— 3) 2/a» + (w— 2) T/JCTM-I + S (m— l)aOT2 and therefore we can determine the functions 0 and ^ and get A. = t-± U}n_3 I 2 — --3- and similarly T-, f tai n>&— wo?! I ~^~2 u*-wl\ °n i—2) — — (n— l)- If we now choose the new axes to be the canonical axes, we shall have, at the origin Q6,Q83\ r 2 276 Differential Covariants of Twisted Curves, &c. [June 21, <"&— tt'2? _ / Qfi\ _ O - ~ 2- - _ _i_ -T — 5 T> 2 / — ' We can now write down the leading terms in the expansions of aa and /3,t, and we can differentiate the expressions, remembering that after differentiation we put at = 64 = . . . . = 0 a6 = a6, 07 = a7, &c. All the expressions I, J, K, &c., vanish if not differentiated, and all the part contained in R vanishes after differentiation as well. We have then only to consider that part of the differential coefficient of I, J, K, &c., which does not vanish. Writing this [I'], &c., we get [J']=12«,, [T']=.6«6, [K']=0, and if [«•„'] stands for the invariantive part of the differential co- efficient of «„, being the only part when referred to the canonical axes, we have -3)*n_2+ ---- + (n—m — 1) ftm*H-m+ ---- 2)*n_2+ .... +(n— m) »m^.iy = 155 miles per second. H£ = 154 X 4025 =158 5. One of the dark line components bears a strong resemblance to Rigel and the other to Bellatrix. The spectra of the two components can readily be separated, for the reason that only lines common to both will be doubled. Among these are. the lines of hydrogen. Lines special to either component are always single, and they retain the same relative positions with respect to one group of hydrogen lines throughout the period. In Plate 2 photographs are given to facilitate an analysis of the compound dark line spectrum. At the bottom of the diagram is a reproduction of a photograph taken near the time of second maximum (August 24, 1893), and the spectra of Rigel and Bellatrix are in- cluded in the same plate. The compound character of the dark line spectrum of j3 Lyrse at this time is shown by the fact that one group of lines corresponds very closely with those which appear in the spectrum of Rigel, and when these are subtracted from the whole spectrum, a spectrum closely resembling that of Bellatrix remains ; the latter spectrum being displaced in this photograph to the more refrangible side, as shown by the short lines drawn beneath the spectrum. The resemblance of the two components to Rigel and Bellatrix respectively, the spectra of which I have described in a previous paper,* is further shown by the following tabular com- parison, the two dark line components of ft Lyrse being called R and B respectively (p. 282). It is not intended to suggest that the spectra of the two dark line components are quite identical with those of Rigel and Bellatrix. These are simply the best known stars which they most closely re- semble, and the similarity is pointed out as an indication that we have not to deal with bodies of an unfamiliar type. Throughout the paper I shall refer to the two components as R and B respectively. The conditions at first maximum, as shown in Plate 1, are not so simple as those at second maximum, though there is evidence to show that at this point of the light curve the component B is reced- ing with respect to R. As will be seen on reference to the photo- graph of March 13, 1894, the hydrogen lines are broadened, and the * ' Phil. Trans.,' 1893 (A), vol. 184, p. 693. 282 Prof. J. N. Lookyer. [June 21, Component R. Eigel. Component B. Bellatrix. Wave-length. Wave-length. Intensity. Wave-length. Wave-length. [ntensity. 3919 2 3926 3 3933 3933 6 3933 3933 3 3963 2 . . 3963 3 3968 3968 6 3968 3968 6 3994 3994 1 . , 3994 3 4008 4008 2 4008 5 4025 4025 3 4025 4025 6 4040 2 4069 2 4071 2 4075 2 4101 4101 6 4101 4101 6 4104 2 4119 2 4120 -5 4120 -5 2 4120-5 4120 -5 4 4127 4127 3 4130 4130 3 4143 4143 2 4143 4143 5 4168 3 4172 1 . . 4172 1 4177 1 4177 1 4233 4233 2 4241-5 2 4253 2 4267 4267 2 4267 4 4340 4340 6 4340 4340 6 4345 2 4351 4351 1 4351 4351 2 4388 4388 3 4388 4388 5 4394 -3 2 4414-5 2 4417 2 4437 3 4471 4471 4 4471 4471 6 4481 4481 5 4481 4481 3 4553 3 two lines near 4471 and 4481 have approached each other, as they should do if one belongs more especially to R and the other to B. 6. When the two bodies lie along the line of sight, partial eclipses occur. This happens near the minima of the light curve. The differences in the intensities of the dark lines special to R and B, near the two minima, indicate that near the principal minimum K. is partially eclipsed by B, while near secondary minimum B is par- tially eclipsed by B. These changes will be seen on Plate 1, and again in Plate 2. In the latter we have comparisons of ft Lyrae at the two minima with Bellatrix and Rigel. If we leave the bright lines out of consideration, it will be seen that near principal minimum, 1894] The Spectrum changes in JB Lyrce. 283 X ° ? "ft /* * .2 Jj2 a** r3 ^ ' ~ ^ ft " •sill •— Tf" fl v_ -2 H _, « r^ C J ISM o * 5 9 -H 4 8 «. ^ ^.S^ », ^i C ^— - — :: ,§ J5» oo^S - 2^ a ».2° 284 Prof. J. Norman Lockyer. [June 21, the spectrum of y3 Lyrae greatly resembles that of Bellatrix, the com- ponent B in this case lying between us and component B. As the eclipse is not total, however, the lines special to B appear with reduced intensities ; the lines joining the spectrum of ft Lyrse to that of Bellatrix indicate the principal lines of 'component B. At the secondary minimum, on the other hand, component B lies in front of component B, and the spectrum consequently bears a greater resem- blance to that of Bigel. This is shown by the lines joining those of /3 LyrsD to the spectrum of Bigel in Plate 2. The difference is especially noticeable in the case of the lines near \4471, 4481, 4388, and in the group of four lines a little less refran- gible than HS. It will be seen that near principal minimum 4471 is stronger than 4481, as in Bellatrix, while about secondary minimum 4481 is stronger than 4471. If the eclipses were total, the variations of the spectrum might be expected to be still more striking. 7. In addition to dark lines there are several bright ones, which change their positions with respect to the dark ones. The photographs show conspicuous bright lines about wave-lengths 4862(H/3), 4715, 4471, 4388, 4340(HV), 4101 (H«), 4025, and 388 7 (H^). Other fainter ones also appear in some of the best photo- graphs. The line at 4471 (Lorenzoni's /) is the well-known line which appears in the spectrum of the solar chromosphere, and those at 4025 and 4715 are amongst the brightest lines photographed with the prismatic camera during the total eclipse of the sun on April 16, 1893. The displacements of the bright lines described by Pickering are confirmed in the main by the Kensington photographs. In the first seven photographs in Plate 1 taken between principal and secondary minimum, the bright lines lie on the less refrangible sides of the dark ones, at secondary minimum the broad bright lines are almost bisected by dark ones ; while from secondary minimum to prin- cipal minimum the bright lines are more refrangible than the dark ones. The investigation of the movements of the bright lines must, however, be now carried on in the light of the knowledge gained with regard to the existence of two sets of dark lines. If we consider the displacements of the bright lines with reference to the dark lines of component B, we find that they are always in the same direction as those of component B with respect to B. Thus in the first half of the period, the bright lines, as well as the dark lines of component B, are less refrangible than those of component B, while during the second half they are more refrangible. The bright lines, however, do not keep a constant position with respect to those of component B, although displaced in the same direction. 8. The bright lines are brightest soon after secondary minimum. 1894.] Photographic Spectrum of the Great Nebula in Orion. 285 If the brightness of the lines in reality remains constant, they will appear relatively brightest at the two minima, owing to the reduction of continuous spectrum which is associated with the increased bright- ness of the star at maximum, and for the same reason they should appear brighter at principal than at secondary minimum. Estimates of the brightness of the lines in relation to the continuous spectrum have been made independently by four of my assistants, and, although estimates of this kind are liable to error, the general agreement is sufficient to indicate that when all allowance is made for the varying continuous spectrum, there is a maximum of brightness of the bright lines about half a day after secondary minimum. The apparent increase of brightness near principal minimum seems to be due solely to the reduced intensity of the continuous spectrum. I have to express my obligations to Messrs. Fowler, Baxandall, Shackleton, Butler, Wardale, Crabtree, and North, who, at different times, have assisted in taking the photographs. XIII. " On the Photographic Spectrum of the Great Nebula in Orion." By J. NORMAN LOCKYER, C.B., F.R.S. Received June 13, 1894. (Abstract.) The paper consists of a description and discussion of photographs of the spectrum of the Orion Nebula, taken with the 30-inch reflector at Westgate-on-Sea in February, 1890, of which a preliminary account was communicated to the Royal Society at the time. Fifty - four lines are tabulated as belonging to the spectrum of the nebula, nine of them being due to hydrogen. Tables are given showing : — 1. The wave-lengths, intensities, and probable origins of the lines photographed in the spectrum of the nebula. 2. A comparison of the lines in the spectrum of the nebula with lines in the spectra of (a) P. Cygni, (&) bright line stars and planetary nebulae, and (c) stars in Groups II, III, and IV, of the classification according to the meteoritic hypothesis. The complete discussion has led to the following general con- clusions : — 1. The spectrum of the nebula of Orion is a compound one con- sisting of hydrogen lines, low temperature metallic lines and flutings, and high temperature lines. The mean temperature, however, is relatively low.* * ' Roy. Soc. Proc.,' vol. 43, p. 152, 1887. 286 Dr. T. Ewan. [June 21, 2. The spectrum is different in different parts of the nebula. 3. The spectrum bears a striking resemblance to that of the planetary nebulae and bright line stars. 4. The suggestion, therefore, that these are bodies which must be closely associated in any valid scheme of classification, is con- firmed. 5. Many of the lines which appear bright in the spectrum of the nebula appear dark in the spectra of stars of Groups II and III ; and in the earlier stars of Group IV, and a gradual change from bright to dark lines has been found. 6. The view, therefore, that bright line stars occupy an inter- mediate position between nebulae and stars of Groups II and III is greatly strengthened by these researches. XIV. " On the Absorption Spectra of Dilute Solutions." By THOS. EWAN, B.Sc., Ph.D., 1851 Exhibition Scholar in Chemistry in the Owens College. Communicated by Pro- fessor H. DIXON, F.R.S. Received April 7, 1894. (Abstract.) The measurements recorded in the paper were made in the hope of obtaining some information as to the molecular condition of salts in dilute solution. In order to obtain exact quantitative results it is necessary to measure the extinction coefficients of the solutions. For this purpose a new spectrophotometer was devised, by means of which it was possible to work with very dilate solutions. In this instrument, by the advice of Professor A. Schuster, F.R.S. , a Lummer and Brodhun photometric prism was used, and the photometric measurements were made by means of Abney's rotating sector. The absorption spectra of solutions of cupric sulphate, chloride bromide and nitrate, containing generally 0'003 to O004 gr. mol. per litre, were measured and found to be, within the limits of experi- mental error, identical. The solutions of cupric acetate absorb, for the same amount of copper, much more light than those of the other salts used. The difference tends to disappear as the solutions become more dilute, and it is increased by the addition of acetic acid. These facts point to the conclusion that the difference is due to the incom- plete electrolytic dissociation of the salt, and to the undissociated part having an absorption spectrum differing from that of the dis- sociated part. Dilute solutions of the potassium and ammonium salts of «-dinitro- phenol (1 ' 2 ' 4) were found to possess very nearly the same absorp- 1894.] On the Absorption Spectra of Dilute Solutions. 287 tion spectrum. The mean of the numbers obtained for these two salts was regarded as the absorption spectrum of the ion CsH^NC^aO. The solution of dinitrophenol in hydrochloric acid (containing the undissociated molecule C6H3(N02)2OH) absorbs very little light ; it is almost colourless. The extinction coefficients of dinitrophenol, and of its coloured ion, being thus known, it was possible to calculate from measurement of the extinction coefficients of a series of solutions of dinitrophenol in pure water, its degree of dissociation in these solutions. The numbers thus obtained were in very satisfactory agreement with the numbers calculated from the electrical conducti- vity of the solutions. As an example of the hydrolytic decomposition of a salt in aqueous solution, to the study of which the spectro-photometric method can be advantageously applied, ferric chloride was taken. By filtering dilute solutions of ferric chloride through a porous cell all the colloid ferric hydroxide formed can be removed ; and analyses of the solutions before and after filtration showed that the hydroxide formed in solutions containing less than O'OOS gram molecule of FeCl3 per litre contains no chlorine. The decomposition which occurs in these solutions may thus be most simply expressed by the equation FeCls+3H,0^;Fe(OH)s+3HCl. The photometric determinations of the quantity of ferric hydroxide formed in these solutions agreed fairly well with the results of the filtration experiments, though, owing to the difficulty in obtaining the solutions of ferric chloride perfectly clear, they were not so satisfactory as could be desired. The quantities of ferric hydroxide formed were not in agreement with the law of Guldberg and Waage, but agreed much better with the modified form of the law due to Arrhenius, in which the electro- lytic dissociation of the different substances is taken into account. It was observed that solutions of ferric hydroxide obtained by dissolving ferric chloride in a very large quantity of water, had a different absorption spectrum from that of solutions of ferric hydroxide obtained by dialysis. It is suggested that an explanation of this fact may be found in the differences in the complexity of the molecular aggregates existing in the different solutions. Finally, solutions of ferric chloride, to which small quantities of hydrochloric acid had been added, possess such comparatively small power of absorbing light that they cannot be regarded as containing any colloid hydroxide of iron. 288 Prof. H. G. Seeley. Researches on the Structure, [June 21, XV. " Researches on the Structure, Organisation, and Classifi- cation of the Fossil Reptilia. Part IX. Section 4. On the Gomphodontia." By H. G. SEELEY, F.R.S. Received June 21, 1894. (Abstract.) The Gomphodontia is a group of Anomodont reptiles characterised by theriodont dentition, in which the molar teeth are expanded trans- versely, more or less tuberculate, and have the crowns worn down with use, as in ungulate and other mammals. The orbit of the eye is distinct from the zygomatic vacuity, which is conditioned as in the Cynodontia, there being a long narrow parietal crest dividing the temporal vacuities. There are two well-defined occipital condyles united at the base, in a way that is closely paralleled in some mammals. The occipital plate is triangular, as in mammals, with no perforation except the foramen magnum. A deep superior notch defines the occipital plate from the lateral external squamosal bar. The malar bone, which forms the larger part of the zygoma, behind the orbit, has a slight descending process which varies in develop- ment. The hard palate terminates transversely in the middle length of the molar teeth. There is a descending transverse palatine arch situate behind the orbits. The incisor teeth are small and pointed ; the canine teeth may be inconspicuous, but are usually large, com- pressed, and serrated ; the premolars are small, circular, and usually tuberculate ; the molars are usually single-rooted, in close-set series which diverge as they extend backward, with crowns which vary in form, but are commonly wider than long, and usually have the external and internal cusps more prominent than the other tubercles on the crown. The group is based chiefly upon the genera Gomphognathus, known from skulls, a vertebra, and fragments of limb bones ; Trirachodon, known from skulls only ; and Microgomphodon, in which the canine teeth are no larger than the incisors. The last genus appears to make known the more important parts of the skeleton. These specimens, collected by the author at Lady Frere, by Dr. Kannemeyer, near Burghersdorp, and by Mr. Alfred Brown, near Aliwal North, are all from the Upper Karroo rocks, on or about the horizon of the Coal Beds. Of Gomphognathus there is a complete skull, with the lower jaw attached, about 9 ins. long, a second skull which displays the palate, and a separate lower jaw in connexion with part of the back of the skull. These specimens show four incisor teeth in each premaxillary bone, with sharp lateral serrated borders. The mandibular canine is covered when the jaws are closed. The maxillary canine is a 1894.] Organisation, and Classification of Fossil Reptilia. 28 powerful tooth : its extremity is worn obliquely. There appear to be six premolar teeth, all contained in a length of half an inch. The maxillary teeth are packed in close succession, as in Rodents. There are nine molar teeth. In the middle, where they are largest, four occupy the length of 1 in. The contour of the crowns of these molars is convex from front to back, as in many mammals ; and in this genus they are all behind the hard palate. The external cusp is prominent, and a ridge descends inward and backward from it upon the large flattened ledge of the crown, which is worn almost level, as though there were a rodent-like horizontal movemeut of the lower jaw. A lumbar vertebra, found in developing the back of a skull, may possibly belong to this genus. With the skulls a right humerus was found, which is of ins. long. It shows the reptilian transverse elongation of the proximal articula- tion, combined with characters which are paralleled in the marsupial mammals and Garni vora. The genus Microgomphodon is known in the first place from a skull 2^ ins. long, shaped much as in Galesaurus, but distinguished by the comparatively large size of the front pair of mandibular incisors, and the strong, conical, pointed character of the incisor teeth. The canine teeth are not differentiated from the incisors. The molars show in lateral aspect small blunt cones ; but on their palatal aspect have flattened crowns with many small cusps. All the teeth have short roots. There are three incisors on each side in both the man- dible and skull, one canine, and five molars. There is ground for associating with this genus an imperfect skeleton, which, in addition to indicating ten early dorsal ribs, and fourteen lower dorsal vertebrae and ribs in advance of the acet- abulum of the femur, shows the left bumerus, portions of right and left scapulae, portions of the coracoid, clavicle, interclavicle, the pelvic bones, all the bones of the hind limb, distal ends of ulna and radius, carpus, metacarpus, a,nd five digits. With these a fragment of a skull is associated, which has the maxillary and mandibular teeth in contact, showing the animal to be Gomphodont ; while so much as is preserved closely resembles the skull of Microgompttodon, and apparently the canine was not larger than the premolar. This skeleton demonstrates a close general resemblance of plan between the Gomphodontia and Cynodontia. The lower dorsal ribs have a transverse lozenge-shaped enlargement, which, however, is less de- veloped than in Cynognathus. The pelvis is exposed on the ventral side. As in most, if not all, South African Therosuchia, it shows no indication of median division between the pubic bones, while the ischia retain their individuality. The pubis articulated to a short tubercle on the ilium. The blade of VOL. LVI. U 290 Frof. H. G. Seeley. Researches on the Structure, [June 21. the ilium is thin, but imperfectly exposed ; and the ischia are shaped as in Pliosaurus, but the pubis does not closely resemble that of any reptile. The femur has the inferior internal trochanteric ridge only slightly developed. There is no neck defining the head of the bone from the shaft. The fibula is slender; no indication of a patella is preserved. Below the stout tibia, the proximal row of the tarsus appears to consist of two bones, an inner astragalus with hemis- pherical proximal surface, and a narrow elongated bone which appears to be the calcaneum. There were three or four bones in the distal row of the tarsus, but only one is preserved. The digits are nearly parallel with each other, and the foot has a compact character like that of Dicyncdon. The scapulae have the pre-scapula developed on the same plan as in Cynognathus, and the anterior margin of the bone reflected upward, so as to form the spine of the scapula, terminating in the acromion. The two ends of the humerns are twisted at an angle of 45 degrees, and the bone is expanded as in many Saurischian reptiles. The carpus shows three bones in the proximal row, a large reniform carpal below the ulna, regarded as the pisiform bone ; a compara- tively small middle carpal is identified as the cuneiform bone. The third bone corresponds with the scapho-lunar of Theriodesmus ; it is beneath the radius. There is no indication of any pre-pollex. There are four bones in the distal row of the carpus. There are five digits. In the pelvis and the limb bones this Anomodont type approxi- mates to the Saurischia and Mammalia, just as the Ornithischia approximate to birds in the same parts of the skeleton. Trirachodon is founded on four individuals which have the skull about 4 ins. long. Like the other Gromphodont genera, this type has the dentary bone developed so as to occupy the length of the mandible, but the lower jaw is composite, the internal bones filling the space which in mammals is occupied by the meckelian cartilage. The post-frontal and pre-frontal bones are well developed. The species differ in the character of the teeth, especially in number and form of the pre- molars. In one species from Aliwal North, the molar teeth are transversely wide, ornamented with three transverse ridges, which terminate in a slight cusp, both on the external and internal margins. There are not more than nine molars. The crown of the first pre-molar in one specimen is elongated from front to back, and shows a small coronet of rounded marginal cusps. In a species from Lady Frere the molar teeth are narrower, and the pre-molar teeth more numerous, small, and circular in the broken sections. Although these skulls are mammalian in aspect, and in some respects make new transitions towards mammals, in technical characters they retain a sufficient number of reptilian structures to 1894.] Organisation, and Classification of Fo.ssil Reptilia. 291 permit no doubt that they are true reptiles. The mammalian resem- blances in the skull being paralleled in the other parts of the skeleton, it may be affirmed that these fossils demonstrate a closer affinity between reptiles and mammals than had previously been evident. XVI. " Researches on the Structure, Organisation, and Classifi- cation of the Fossil Reptilia. Part IX. Section 5. On new Cynodontia." By H. G. SEELEY, F.R.S. Received February 13, 1894. (Abstract.) The Cynodontia is a division of the Theriodontia in which there are long and large temporal vacuities in the skull, formed chiefly by the squamosal and malar bones ; in which there is no descending pedicle to the squamosal bone ; in which the occipital condyle is crescentic and imperfectly divided into two lateral parts; and in which the hinder molar teeth, larger than the incisor teeth, develop anterior and posterior cusps, are compressed from side to side, and overlap, with shearlike action, the teeth of the mandible. The prin- cipal new genera included in this group are Cynognathus, which is known from several skulls, and one fairly complete skeleton ; and the genus Tribolodon, which does not differ in a striking way from the small Cynodonts previously known, referred to the genera Galesaurus, Nythosaurus, and Thrinaxodon. The skeleton of Cynognathus crateronotus was found at Lady Frere, near Queenstown. A single tooth of this genus had already been ob- tained by Mr. Alfred Brown at Aliwal North. The skull is between 15 and 16 in. long, 8 in. high at the orbits, and higher at the occiput, where it was about 9 in. wide. The lateral aspect is remarkably Mammalian, owing to the great development of the dentary bone, which forms a new type of lower jaw, and has a greatly developed coronoid process, and the form of the zygoma. On the palate, the palatine and transverse bones form a descending arch between the rami of the mandible, as in Crocodiles, Sphenodon and Lizards. The composite structure of the lower jaw is seen on its inner side. The pre-frontal and post-frontal bones remain distinct. There is a small quadrate bone embedded in the large squamosal bone. The latter resembles that of Mammals, both in its extension along the zygoma, and its expansion as a squamous plate on the side of the brain case. There are four incisors in each pre-maxillary ; their margins are serrated. There appear to be but three mandibular incisors on each side, so that the type resembles Cynochampsa, but there is no evidence u 2 292 Prof. H. G. Seeley. Researches on the Structure, [June 21, of close affinity with that genus. The canine teeth are large, worn on the anterior border, and serrated on the hinder margin. Remnants of canine teeth are indicated which have been replaced by those which persist. There are nine molar teeth, of which the first five are smaller than the posterior teeth. Those teeth are more than half as wide again, from front to back, as the anterior teeth. The hinder teeth have the principal cusp directed backward, with one subordinate pointed cusp on the front margin, and two subordinate cusps on the hinder margin. The crowns of the teeth stand high above the alveolar margin in this species. They are intermediate in form of crown between Canis and Zeuglodon. The nares are terminal, divided, lateral, and arch forward in front of the alveolar margin. The orbit of the eye is 8 in. behind the extremity of the snout, nearly circular, and separated from the tem- poral vacuity by the post-frontal bone. The post-frontal bones con- verge backward along the parietal crest. The malar bone develops a slight descending process on its inferior margin. There is no inter- orbital septum ossified. The type species of Cynognathus shows, on the one side preserved, a small post-orbital foramen, comparable to that of Procoloplion, and the author considers that the enlargement of this foramen makes an essential difference in plan between the skulls of Teleosaurs and Theriodonts, and regards the Mammalian zygoma as resulting from the obliteration of the post-orbital vacuity which defines the superior and inferior temporal arcades in Saurischia and other Beptilia. In general structure of palate Cynognathus resembles Lycosaurus. There is no transverse boundary to the hard palate, but the palato- nares are lanceolate. The author finds that the downward develop- ment of the bones of the palate at the posterior borders of the nares, while thoroughly reptilian, approximates to the condition in Mam- mals. The form of the lower jaw approximates to that of the older Mammals and lower Mammalian types, leading to the conclusion that the Mammalian lower jaw consists essentially of the dentary bone. The dentary bone is compared to that of Microconodon in form, and development of the angle of the jaw. The shoulder girdle consists of a large scapula, small coracoid, and compressed pre-coracoid. The scapula demonstrates the origin of a spine like that of the scapula in Mammals, by outward development of the anterior border of the scapula in Reptiles. This spine is de- fined by a pre-scapula development anteriorly. The spine may have been originally a separate ossification, such as in Pareiasaurus has been named epi-clavicle. It terminates in an acrornion which is re- flected forward. The humerus is imperfectly preserved, but has the distal con- 1894.] Organisation, and Classification of Fossil Ee^tilia. 293 dyles well developed ; and the proximal crest has a form which is seen in Marsupials, but the articular head is transverse. The vertebral column measures 37 ins. from the body of the atlas to the last lumbar vertebra ; and its total length is 45 ins., but the extremity of the tail is lost. There appear to be only six cervicals denned by the form and direction of the transverse processes for the tubercles of the ribs. The head of the rib is attached to the inter- central suture, and in the first vertebra reaches the intercentrum. There are 29 presacral vetebrse, of which 18 may be counted as dorsal and 5 as lumbar. The most distinctive feature of the vertebral column is the interlocking of the ribs in the lower dorsal and lumbar region, where the ribs become transversely expanded, and anchylosed to the side of the centrum. The neural arch in the lumbar region also interlocks, by an arrangement resembling the zygosphene and zygantrum of Serpents. No dorsal rib is completely preserved. The sacrum is small ; and the sacral ribs are smaller than the lumbar ribs, They are four in number. The middle two vertebrae are anchylosed. The caudal vertebrae are short, only four are preserved. They indicate a considerable movement. There is 110 evidence of dermal armour. The characters of the vertebral column described by Professor Cope, in Dimetrodon and allied genera, closely resemble Cynognathus. The pelvis consists of three bones ; the ilium forms an expanded plate more resembling Megalosaurus than Dicynodon. There is a large longitudinal obturator foramen, between the pubis and the ischium. The anterior transverse border of the pubis is carljilagi- nous, and there is no evidence of pre-pubic bones. The ischium is larger than the pubis. The author compares the anomodont pelvis with that of Plesiosauria, although Pliosatirus, in the form of the ilium, more closely approaches Dicynodon than Cynognathus. The femur is imperfectly preserved. It was characterised, as in all Theriodonts known to the author, by the development of an immense inferior plate or ridge at the proximal end, which distinguishes it from allied animals. In this specimen the ridge is broken away. The head of the bone is greatly expanded transversely ; and the distal end is not preserved. Under the name Cynognathus Berryi the author desciibes imperfect evidence of a smaller skull of Cynognathus, which is distinguished from G. craternotus with some doubt ; but if distinct it is defined by the relatively large size of the middle mandibular incisor, the apparent presence of ten molars, in all of which the crowns overlap each other, and the roots are barely shown at the alveolar border. In the small species the cutting margin and the cusps of the posterior teeth are better defined. If the species are identical the teeth have probably yet to be re- 294 Prof. H. G. Seeley. Researches on the Structure, [June 21, placed by a successicmal series, but no known specimen of any genus shows such replacement. The skull of Cynognathus platyceps was obtained by Dr. Kannemeyr at Wonclerboom. It is a small species distinct from Cynognathus crateronotns. The skull has lost the extremity of the snout. It is remarkable for its depression. The teeth, however, are similar to those of the larger species ; they have five denticles. The composite structure of the lower jaw is well shown, and the dentary bone be- hind the angle of the jaw retreats so as to expose the elements which form the articulation. The occipital plate of a large Theriodont skull from Lady Frere is described, which shows a circular foramen magnum, and the perfectly preserved occipital condyles which are not quite so completely separated as in Mammals, having only a median groove between them on the ventral surface. Another fragment of a skull preserved in the Albany Museum has only the pre-orbital portion preserved, and is remarkable for the small size of its incisor teeth, widely separated from each other, and for having two canine teeth parallel to each other. On both sides the crowns are imperfectly preserved. The molar teeth are on the type of Cynognathus, with a principal cusp flanked back and front by a small cusp, with a smaller accessory posterior cusp in the four hindermost teeth. As in all species of the genus the manclibular symphysis is long, oblique, and completely obliterated. There is a large pit with sharp margin, in the median line in front of the orbits, which may be a generic difference from Cynognathus, since it occurs in the area in which other specimens show indications of a thin supra- nasal ossification flanked by a pair of small hemispherical concavities. It is indicated as G. leptorliinus. Tribolodon Frerensis is the name given to a dentary bone with few three-pronged teeth widely separated from each other standing high above the jaw. With this jaw is associated a femur which shows the transverse development of the great trochanter as strongly developed at the proximal end of the bone as in Ichthyosaurus, so that the trochanter minor of Mammals only represents that of Theriodonts in miniature, the trochanter being more developed than in Saurischia or any other reptiles. With it is associated a right tibia which is some- what curved and nearly as long as the femur. These Cynodont remains have given no certain evidence of the ex- tremities of the limbs ; but with this exception they make known the entire skeleton for the first time in an African Theriodont, furnishing data for comparison with Mammals and Reptiles in every part of the skeleton preserved. 1894.] Organisation, and Classification of Fossil Reptilia. 295 XVII. "Researches on the Structure, Organisation, and Classi- fication of the Fossil Reptiiia. Part IX. Section 6. Asso- ciated Remains of two small Specimens from Klipfontein, Fraserburg." By H. G. SEELEY, F.R.S. Received June 21, 1894. (Abstract.) The author obtained parts of two skeletons from the summit of the Karroo rocks, which form the Nieuwveldt range. They resemble Theriodonts in their general marsupial characters. The fragments of skulls are not in the same slabs with the other bones. TJieromus leptonotus shows the fore-limb and some vertebrae. The humerus is determined to be Theriodont by the transverse extension of the proximal articulation. The bone is 1-^ inches long, resembling in form that of the Phalangers. The ent-epicondylar foramen is more vertical than in the marsupials ; and, as among marsupials, the radial crest if prolonged distally would be continuous with the bridge over that foramen. The vertebrae are each -^ inch long ; they show a transverse suture between the neural arch and the centrum. The anterior part of the skull, very imperfectly preserved,, indi- cates three incisor teeth with the root of a relatively large maxil- lary canine, but the region of the molar teeth is lost. There is also a posterior fragment of a skull, which makes known the bones of the palate and the base of the brain case seen from above. Enough is shown to indicate Theriodont characters, but the animal appears to diverge from the Theriodonts towards the Dicynodont type. If the base of the skull belongs to the same individual as the snout, it indi- cates a head nearly 4^ inches long. The second specimen shows 14 dorsal vertebrae, which occupy a length of 5^ inches ; each slightly exceeds ^ inch in length, so that this animal named Herpetocheirus brachycnemns, is similar in size to the fossil previously described. The centrum is deeply biconcave. There is no indication of a capitular articulation for the ribs. The ribs are slender, and the longest are 2^ inches in length. There is no trace of the transverse expansion seen in Cynognathus, although the ribs preserved indicate 20 dorsal vertebrae. The humerus is 1-j^ inches long, and is exposed on the superior aspect. It is distinguished from the type already de- scribed by wanting the tuberosity on its inner distal border, which has a convexly rounded contour. The radius is stronger than the ulna, but there is no indication of an olecranon process exposed. The ulna is no stouter than a rib. These bones are an inch long. The carpus shows one large bone below the radius; there is a smaller 296 Dr. H. Gaclow and Miss Abbott. On the [June 21, bone on its outer side, which corresponds to the distal end of the ulna, but there is no trace of a third bone preserved, and there is only one central bone preserved. There are three phalanges in a digit. The femur is l^ inches long ; its articular head appears to be small and hemispherical. There is a large internal trochanter extending down the shaft, which corresponds with the similarly placed ridge in the femur of Megalosaurs and other Saurischia. The slender character of the ribs, which are different from those in known Theriodonts, suggests the possibility that these remains belong to a group distinct from both the Cynodontia and Gomphodontia. A small badly preserved fragment of a skull found near to this fossil is described, but there appears to be no sufficient evidence for associating it with the other remains. XVIII. " On the Evolution of the Vertebral Column of Fishes." By H. GADOW, Ph.D.. F.R.S., and Miss E. C. ABBOTT. Received June 20, 1894. (Abstract.) Concerning the segmental mesodermal products the following sub- division is adhered to : — The term myotome is to be restricted to the whole rest of the protovertebra after the skeletogenous cells have been given off for the production of the sklerotomes. The sum total of the sklerotomes makes up the skeletogenous layer. The ending tome to indicate the primary, or earlier, less differen- tiated ; the ending mere to signify the final condition or product. Consequently, the protovertebrae divide into — I, Myotomes, each of which produces (1) one myomere or segment of the general mass of trunk-muscles, (2) cutis ; II, Sklerotomes which produce sklero- meres or skeletal trunk segments. Each protovertebra produces a dorsal and a ventral sklerotome ; strictly speaking, one sklerotome which consists of a separate dorsal and ventral half The protovertebral segments are not transverse " plates," but are curved into S-shape, the top end curving tail- and inwards, the middle and ventral thirds bulging headwards, the amount of curvature being (in 7 mm. embryos of Acanthias) so great that a transverse plane will cut through the dorsal and ventral third of one, and through the middle portion of the next following segment. This S-shaped curving and consequent overlapping of the proto- vertebral " plates " SS is of fundamental importance for our under- 1894.] Evolution of the Vertebral Column of Fishes. 297 standing of the formation of the vertebral column, because it explains (1) the so-called new segmentation of the axial column, (2) the almost universal occurrence of more than one dorsal and one ventral pair of arcualia (namely, arches and intercalary pieces) in each of the later vertebral segments or skleromeres. The explanation is as follows : — 1. The dorsal half of sklerotome 2 grows downwards and comes to lie behind the ventral half of sklerotome 1. 2. The ventral half of sklerotome 2 grows upwards and comes to lie in front of and below the dorsal half of sklerotome 3. 3. The formation of a physiological unit is effected by the combin- ation or fusion of the unequally numbered sklerotomic halves, in such way that the dorsal half lies behind and above the ventral half. The new skleromere I (= dorsal sklerotome 2 + ventral sklerotome 1) stands now in the following relation to the myoineres; the dorsal end of the skleromere I coincides with myomere I ; the septum between this myomere and the next previous one passes between dorsal sklerotome 2 and ventral sklerotome 3 ; this means to say right across the new skleromere I. This skleromere lies within the influence or range of action of two successive myoraeres. Taken as a whole, the skleromere is " interprotovertebral," more correctly bi- proto vertebral, because it is composed of two successive sklero tomes, namely, the ventral half of one and the dorsal half of a second. Consequently, the " resegmentation " or " neugliederung " is brought about in a manner fundamentally different to that hitherto supposed to have taken place. If A and B mean two successive sklerotomes, a and 6 their dorsal, a and ft their respective ventral A-!-B halves, then the new skleromere is composed of b + & and not of —k— > "Z , , . . , , B dorsal . A ventral because o-f « is the same as + 5 • 2 2 The formation of a skleromere by the combination of alternating dorsal and ventral halves of sklerotomes explains also the presence of eight (four pairs) cartilaginous pieces, namely, basalia (so-called dorsal and ventral arches) and interbasalia (so-called intercalary pieces) for each complete segment. The dorsal and ventral halves of the sklerotomes are pyramidal in shape, with their apices pointing respectively downwards and up- wards. Each ventral pyramid extends with its apex above the chorda, and founds there (separated from the ventral mass by the subsequent rapid growth of the chorda and its sheath) a cluster of cells which remains henceforth behind (tailwards from) the basal mass of the dorsal pyramid. The latter founds, with its down- growing apex, a colony of cells below the chorda, and in front of the basal ventral mass. Thus are produced the basalia and interbasalia, 298 Evolution of the Vertebral Column of Fishes. [June 21, each colony or cluster of cells developing into a separate piece of car- tilage. The basidorsal does not fuse with its interdorsal, because both are the offspring of two different sklerotomes, nor can the basi- dorsal fuse with its own offspring, namely, with the interventral, because both became, and remain, separated by the chorda and its sheath ; they are connected only by the indifferent connective tissue of the membrana reunions, but not by cartilage-forming cells. Concerning the formation of centra or bodies of the vertebras, we distinguish : — I. Chorda-centra, i.e., centra cut out of the full of the chordal sheath, which itself has been strengthened by invasion of cartilagin- ous cells from the skeletogenous layer. This migration of cartilage into the chordal sheath had already been hinted at by Kcelliker more than thirty years ago ; it has recently been proved by Klaatsch, and has been corroborated by us. Chorda-centra are possessed by all Elasmobranchs, potentially by Dipnoi and Holocephali. II. Arch-centra, i.e., centra formed by the skeletogenous mass which remains entirely on the outside of the chordal sheath, which latter takes no share in their formation : osseous Ganoids and Teleostei. Chorda-centra and arch-centra represent two different modes of development, each starting from an acentrous condition. This can be expressed as follows : — Chordal sheath remaining Chordal sheath strengthened by invasion of entirely chordagenous. skeletogenous cells, therefore with possibility of chorda-centra. Cyclostomata, Cartilaginous Ganoids. Dipnoi and Holocephali. Formation of Centra. Osseous Ganoids, Teleostei. Elasmobranchs. ARCH-CENTRA. CHORDA-CENTRA. The formation of chorda-centra being independent of the arcualia explains how and why the number of " centra " does not necessarily agree either with that of the arcualia or with that of the trunk- segments, e.g., Hexaiichus and tail of most other Elasmobranchs. These leading differences and their modifications have been traced in Petromyzon, Acipenser, Amia, Lepidosteus, Protopterus, Chimasra, and in numerous Elasmobranchs. In Amia calva, of which the adult and a young specimen of 57 mm. were examined, the postcentrum, i.e., the posterior, archless disk of a complete tail-vertebra, was found to be formed by the interdorsalia and interventralia of the same sklerotome, while the precentrum, i.e. the arch-bearing disk or anterior half is formed by the basidorsals of 1894.] Structure and Affinities of Heliopora cserulea, fyc. 299 the same sklerotome and the basiventrals of the next previous sklero- tome. Thus skleromere 50 is composed of a postcentrum = inter- dorsal 50 + interventral 50, and of a precentrum = basidorsal 50 + basiventral 49. The intermuscular septum runs obliquely across the precentrum, or, in other words, the precentra are bi-protovertebral or bi-myomeric, but not the postcentra. The precentra of the tail of Amia are homologous with the " pleurocentra " in the tail of the Jurassic Eurycormus, while Amia's postcentra are the same as the " hypocentra " of Eurycormus. In Lepidosteus osseus, of which adult specimens and larvae of various stages were examined, the combination of parts into one vertebral complex is superior to that of Amia, because each vertebra belongs, •with its entire anterior half (interdorsal 50 + basiventral 50), to myomere 50, and with its posterior half (basidorsal 51 + interventral 51), to myomere 51. In other words, the vertebral mass is equally divided between two successive myomeres, or the myomeres have an equal share of the skleromeres. The vertebrae are now truly bi- protovertebral or bi-myomeric, each vertebra being composed of a + b. XIX. " On the Structure and Affinities of Heliopora ccsrulea, Pall., with some Observations on the Structure of Xenia and Heteroxenia." By GILBERT C. BOURNE, M.A.. F.L.S., Fellow of New College, Oxford. Communicated by- Professor LANKESTER, F.R.S. Received May 30, 1894. (Abstract.) I have had the opportunity of making a renewed examination of the structure of Heliopora, partly through the kindness of Professor Ray Lankester, who gave me a very well preserved fragment of a colony brought by Dr. S. J. Hickson from Talisse, Celebes. I have also used some specimens which I collected and preserved in spirit in Diego Garcia, and, in studying the dried corallum, I have had the advantage of a large collection, originally the property of the late Mr. G-eorge Brook, which Mrs. Brook has very kindly handed over to me. All the specimens in my possession are referable to the only recent species known, Heliopora ccerulea, but one of them belongs to the variety tuberosa, Dana. The Soft Tissues. — These form an even sheet, investing the surface of the colony, interrupted here and there by the mouths of the polyps, which are the only apertures opening on the surface. The structures described below are entirely superficial, and there is no direct com- 300 Mr. G. 0. Bourne. On the Structure and [June 21, mnnication between the polyps and their connecting canals on one side of the colony and those on the other side of the colony. The polyps have been fully described by Moseley. They are scattered irregularly over the surface, and the only important feature presented by them, in which they differ from other Alcyonaria, is the complete introversion of the tentacles during retraction. Surrounding- each polyp, and occupying all the surface of the colony, are very numerous tubes, ending blindly below, and closed above by the sheet of superficial ectoderm which covers the exterior surface; these are the ccenenchymal caeca. They occupy cavities in the corallum known as ccenenchymal tubes, and are set at right angles to the surface of the colony. The ccenenchymal cceca communicate with one another, and with the polyps, by means of a network of canals, which lies close beneath the surface ; these are the superficial endodermic canals. At the growing edges of the colony the superficial network is not well developed, the ccenenchymal coeca are closely contiguous, and open into one another at their outer ends, either directly or by means of short, irregular, transverse passages which cross over the partitions separating adjacent ccenenchymal tubes. The cceneuchymal coeca, superficial canals, and polyps are lined internally with endoderm. Outside this is a thin layer of mesoglcea, and outside of this an irregular layer of large, dark- staining, granular, fusiform cells, which are calcigenous, and may be called calicdblasts. The calicoblasts were described by Moseley as mesodermic, but they occupy the position of ectoderm, and they are, in fact, derived directly from the superficial sheet of ectoderm. Their origin is seen in sections made perpendicularly to the surface of the colony at its growing point. Here the ectoderm cells are generally elongate and pyriform, their broader outer ends resting on a distinct external limiting membrane, their inner ends tapering and produced into long processes, which may often be traced into connexion with deeper seated fusiform cells, of more granular character. The deeper seated cells are imbedded in a thick, homogeneous, gelatinous substance which lies immediately below the ectoderm, and is the newly formed mesogloea, thicker here than elsewhere in the colony. Study of numerous sections shows that the deeper seated fusiform cells are derivatives of the elongate ectoderm cells, and that some of them are used up in the formation of the mesogloea — they appear to dissolve and to be wholly converted into a structureless gelatinous mass — whilst others increase in size, develop many refracting grannies in their interior, and become calcigenous calicoblasts. In many places the calicoblasts may be traced into direct connexion with the ectoderm. The coenenchymal cceca of Heliopora do not appear to be degenerate siphonozooids, as was suggested by Moseley, but rather to 1894.] Affinities of Heliopora cserulea, $c. 301 be specialised parts of a system of inosculating endodermic canals, such as are characteristic of colonial Alcyonaria. The corallum of Heliopora exhibits two sets of apertures, besides those due to the inroads of boring parasites, these are the calicles and the coenenchymal fenestrce. The calicle cavities are occupied by the polyps, the coenenchymal tubes, whose mouths are the fenestrae, are occupied by the cceca, which do not in the fresh condition open to the surface, the fenestroe being closed above by the ectoderm. The corallum consists of an imaginary vertical plane, occupied by vertically disposed ccenenehymal tubes, and right and left faces on which the tubes open after bending sharply from the vertical to take a short horizontal course. The vertical tubes are in section polygonal, and some of them attain, the surface at the growing edge. Those which are deflected hoi'izon bally become thickened by the formation of secondary calcareous deposits inside the primitively polygonal tubes. The calicles are formed by the arrest in growth of groups of ccenenchymal tubes as they approach the surface. The cavity of a calicle never extends into the central vertical tubes. The walls of each ccenenchymal tube are primarily formed of twelve delicate calcareous laminae, secreted by the calicoblasts covering the ccenenchymal cceca, and have this peculiarity, that each of the laminae takes a sbai'e in the formation of the walls of adjacent tubes. As seen in section, three lamina? are united at each angle of the generally hexagonal tube to form a Y's^aPe(l figure. Each arm of the Y meets, and is united by sutures with the arms of adjacent Y's> an(^ so a sor*; °f honeycomb structure is produced, which, if the symmetry of growth were perfectly regular, would consist of a series of regular hexagons. The symmetry is disturbed by the multiplication of the tubes, which do not branch dichoto- mously, as in the allied Heliolites, but increase by the addition and intercalation of new tubes amongst those previously existing. The hexagonal primary constituents of the corallum of Heliopora show very slight traces of blue colour, but as they become thickened by secondary ring-shaped deposits, the latter develop blue pigment, and give the characteristic colour to the colony. The growth of the colony is not effected, as Moseley described, by the upgrowth of an axial polyp from which lateral buds are given off, but by the rapid growth and multiplication of coenenchymal tubes. The fact that the calicles and coenenchymal tubes of Heliopora have not each their distinct and proper wall, but that their walls are common to them and to adjacent tubes, is a characteristic feature of Heliopora and its allies. I therefore propose for them the name Ccenothecalia, in contradistinction to those forms in which, as in Tubipora, each corallite is separate and distinct ; the latter group may be called the Autothecalia. 302 Structure and Affinities of Heliopora cserulea, #c. [June 21. Under the Autothecalia I class Tubipora, Syringopora, Syringolites, the Favositidce, and, provisionally, the Colnmnariadce. Under the Cosnothecalia I class Heliopora, Heliolites, Thecia, Plasmopora, Propora, Lyellia, the Chcetetidce, and, provisionally, Tetradium, Halysites, and the Monticuliporidce. The genns Heliopora is not the only Alcyonarian with a distinct ectodermic skeleton. I brought back with me from Diego Garcia two small Alcyonarians of the genus Xenia. One of the species is referable to Xenia umbellata, Savigny, var. ccerulea. The other I am describing elsewhere as a new species, under the name Xenia garcice. These forms both possess a discontinuous skeleton, formed of the minute corpuscle-like spicules characteristic of the Xeniidse. In X. umbellata, the spicules in the exsert moieties of the polyps are wholly ectodermic, and none are found in the mesogloea. In the stem the external ectoderm is filled with spicules, and the so-called ccenenchyme proves to be nothing more than the fused ectoderm of the basal moieties of the polyps, which is traversed by strands of mesoglcea binding the polyps together, and by endodermic canals which place the polyp cavities in communication with one another. The mesoglcea of the basal moieties of the polyps, as well as the connecting strands of mesoglcea, are quite free from spicules, which are, however, abundant in the mass of fused ectoderm occupying the spaces between the polyps. In Xenia garcice the spicules are, as in X. umbellata, ectodermic in the exsert moieties of the polyps, and in the ectoderm covering the stem. The basal moieties of the polyps are provided with a much thicker mesoglcea, which is, however, free from spicules, except where the mesogkeal laminae of adjacent polyps become fused, in which case intrusive ectoderm cells and spicules are found in the fused thickened areas. Elsewhere the basal portions of the polyps are separated, as in X. umbellata, by ectoderm containing spicules, the mass of which is much less abundant than in X. umbellata. There is also in X . garcice a special system of superficial endodermic canals, which lies immediately below the surface in the upper part of the stem. I have further been able to examine some specimens of Heteroxenia elizabethce, collected by the late Dr. Gulliver, at Zanzibar, and given by him to the Linacre Department at Oxford. I am able to confirm Kolliker's account of this genus, which exhibits a well-marked dimorphism, the colony consisting of fertile autozooids surrounded by more numerous sterile siphonozooids. The spicules of Heteroxenia elizabethce resemble those of Xenia umbellata and garcice, in being minute and entirely ectodermic in the exsert moieties of the polyps. The stem, however, differs considerably from that of X. umbellata, and is more specialised than that of X. garcice. Instead of the 1894.] Experimental Lesions of the Cerebellum. 303 mesoglceal laminae of the basal moieties of the polyps being distinct or only partially fused together, they are absolutely and indis- tinguishably fused, and the mesoglcea is enormously thickened, forming a coenenchymal mass resembling that of Alcyonium. The mesogloea immediately surrounding the polyp cavities is devoid of cells, but elsewhere it contains numerous intrusive cells, among which spicules are developed. The intrusive cells are derivatives of the ectoderm, and in suitable preparations numerous strands of cells are seen to pass inwards from the ectoderm, between the ramifications of the superficial set of endodermic canals, which is rather more marked in this species than in X. garcice. It seems probable that the greater part of the coenenchymal mesoglcea is formed at the expense of the intrusive ectoderm cells, very few of which develop spicules. These three species are interesting, firstly, as indicating the steps by which forms with a wholly mesoglceal spicular skeleton, such as Alcyonium, may have been derived from forms with an ectodermic skeleton ; and, secondly, as suggesting the mode in which the ecto- dermic skeleton of Heliopora may have been developed. In the Xeniidae, as in the Helioporidse, the bulk of the coenenchymal mesoglcea and the whole of the calcigenous elements are derived from the ectoderm. In the one case the mesoglceal elements pre- ponderate greatly over the calcigenous, in the other the preponde- rance of the calcigenous elements has led to the formation of a dense calcareous skeleton, the mesoglceal elements being reduced to a very subordinate position. XX. "Degenerations consequent on Experimental Lesions of the Cerebellum." By J. S. RISIEN RUSSELL, M.D., M.R.C.P., Assistant Physician to the Metropolitan Hospital. Commu- nicated by Professor V. HORSLEY, F.R.S. Received June 4, 1894. (From the Pathological Laboratory of University College, London.) (Abstract.) The paths which degenerate after ablation of one lateral lobe of the cerebellum, and after extirpation of its middle lobe, are discussed in this paper. The former operation, viz., removal of one lateral lobe of the cere- bellum, results in degeneration of all the peduncles on the side of the lesion, and in the superior peduncle of the opposite side ; but no fibres degenerate in the middle or inferior peduncle of the opposite side. The degenerated fibres in the superior peduncle on the side 304 Experimental Lesions of the Cerebellum. [June 21, of the lesion decussate in the posterior quadrigeminal region, and pass to the opposite red nucleus and optic thalamus. None could be traced beyond this point. Those in the opposite superior peduncle represent fibres which degenerate in the cerebellum, passing from the seat of lesion across to the intact half of the organ, and leaving it by this peduncle. These degenerated fibres occupy a special position in the peduncle, a part of it which is comparatively free from degenerated fibres on the side of the lesion, and a part occupied by degenerated fibres on both sides, when the cerebellum is divided into two lateral halves by a mesial incision. These facts are held to con- trovert Marchi's statement, that none of the peduncles contain commissural fibres. The degenerated fibres in the middle peduncle, on the side of the lesion, pass chiefly to the grey matter of the opposite side of the pons. Some degenerated fibres from this source pass between the pyramidal bundles, but there is no evidence to support Marchi's observation, that degenerated fibres also pass from this peduncle in the fillet and posterior longitudinal bundle to the corpora quadrigemina and peri- phery of the antero-lateral region of the spinal cord, and that some pass to the corpus striatum by way of the pyramidal tract. Of the fibres which degenerate in the inferior peduncle, the majority occupy the lateral region of the medulla, becoming more and more scattered as they pass down. These can no longer be said to form a tract below the level of the superior pyramidal decussation ; but a few scattered degenerated fibres occupy the antero-lateral region of the cervical cord, beyond which none can be traced. Degenerated fibres pass to both inferior olives from this peduncle ; but no well- marked tract to the opposite inferior olive, as described by Ferrier and Turner, was found. In accordance with these observers, how- ever, no corroboration of Marchi's results was found, in so far as he states that degenerated fibres pass from this peduncle to the ascending root of the fifth, the roots of the cranial nerves through the posterior longitudinal bundles, and the spinal nerves by the descending antero- lateral tract. In confirmation of Marchi, and contrary to the observations of Ferrier and Turner, degenerated fibres were found in all the peduncles of the cerebellum, after extirpation of its middle lobe. Those in the superior peduncle occupy all parts of it, as seen on transverse section, they decussate in the region of the posterior corpora quadrigemina, and terminate in the opposite red nucleus, beyond which point no degenerated fibres could be traced. The degenerated fibres in the middle peduncle behave much as do those which result from ablation of one lateral lobe of the cerebellum, and the same may be said with regard to the degenerated fibres in the inferior peduncle. No evidence was found to support Marchi's 1894.] On some of the Decussating Tracts of the Brain, fyc. 305 statement that degenerated fibres from this source pass to the cranial nerve roots through the posterior longitudinal bundles, and to the antero-lateral columns of the cord by way of the fillet. With regard to the well-marked antero-lateral tract, which Marchi describes as degenerating throughout the whole length of the spinal cord, it is held, in conjunction with Ferrier and Turner, that no such tract degenerates after lesions limited to the cerebellum. And in sup- port of this negative view being probably the correct one, is adduced the fact that Ferrier and Turner found a similar tract after injury to Deiter's nucleus, as did Mott also, after injury to the posterior column nuclei. XXI. " A Contribution to the Study of (i) some of the Decussa- ting Tracts of the Mid- and Inter-brain, and (ii) of the Pyramidal System in the Mesencephalon and Bulb." By HUBERT BOYCE, M.B., Assistant Professor of Pathology in University College, London. Communicated by Professor VICTOR HORSLEY, F.R.S. Received June 9, 1894. (Prom the Pathological Laboratory of University College, London.) (Abstract.) The present paper is supplementary to a paper communicated to the Royal Society, February, 1894, entitled a " Contribution to the Study of the Descending Degenerations in the Brain and Spinal Cord." It is based upon a study of the changes found in the brains and spinal cords of the animals (cats) used for that research. 1. It is found that hemisections of the mesencephalon through the superior quadrigeminal region is followed by degeneration of Meynert's commissure and Forel' s decussation, situated in front of the third ventricle and behind the optic chiasma. The degenerate fibres which go to form the decussation of Forel are large medullated fibres which ascend from the seat of injury in the tegmental region, proceed forwards and anteriorly, and then curve round in front of the third ventricle, between the latter and Mey- nert's commissure. They then pass backwards, between the optic tract and the internal capsule (pes pedunculi), and appear to end in the lateral thalamic region. This description agrees with that given by Darkschewitch and Pribytkow, who, however, state that the fibres terminate in the lenticular nucleus; by the Marchi method, on the other hand, the Author has traced the fibres past this nucleus, and across the internal capsule into the thalamus. The fibres appear to be part of the fibres constituting the " foun- tain (ventral) decussation of Forel." VOL. LVf. X 306 Some of the Decussating Tracts of the Brain, fyc. [June 21, Meynerfs Commissure. — This commissure is also invariably found degenerate, but the author has been unable to determine its exact mode of origin and termination. It would appear that the commis- sure had a wide field of origin, numerous fibres either passing through or arching round the pes pedunculi on its dorsal aspect to form it. The fibres pass to the opposite side behind the chiasma, and then descend slightly, and appear to diminish in number ; they do not appear to enter the corpus Luysii ; a few of the fibres may penetrate with the optic tract into the thalamic region, and inter- mingle with the superficial fibres of the superior fillet (compare Darkschewitch and Pribytkow, and more recently Bechterew in " Die Leitungsbahnen "). 2. Posterior Commissure. — The degenerate fibres which cross in the commissure or in the roof of the Aqueduct of Sylvius, and which result from a complete unilateral lesion of the quadrigeminal area, have not a long course, but terminate, for the most part, in the opposite corpora quadrigemina, dorsal and lateral aspects of the Sylvian grey matter, or posterior portion of the tegmentum. Degenerate fibres have never been traced into the posterior longi- tudinal bundles, as has been asserted by some authors. A special group of large superficial degenerate fibres in the anterior portion of the roof of the aqueduct have been traced from the internal capsule across the thalamus into the stalk of the superior corpus quadri- geminum and then across the commissure. These fibres alone are found degenerate in the commissure when the anterior one-third of the cat's hemisphere is removed. 3. In cases where the motor region is completely removed in the cat, degenerate fibres are found which leave the pyramidal system in the pes pedunculi, crusta, pons, and medulla. The fibres which leave the pes pedunculi and crusta pass backwards to the quadrigeminal region of the same side, those which leave the pyramid in the medulla decussate across the raphe to the opposite side, and lose themselves in the tegmentum ; they have not been traced directly ending in the motor nuclei of the cranial nerves. Muratoff has described a group of these fibres in the medulla, and supposes that they are the cortical motor fibres of the Vllth ; the author, on the other hand, has not found the fibres limited alone to this region. The decussation of the pyramid is thus not confined to the upper cervical region, but is gradually taking place during the descent of the pyramid through the bulbar segments. 1894] Presents. 307 XXII. " A Magnetic Survey of the British Isles for the Epoch January 1, 1891." By A. W. RuCKER, F.R.S., and T. E. THORPE, F.R.S. Received June 21, 1894. [Publication deferred.] XXIII. « On the Different Forms of Breathing." By WILLIAM MARCET, M.D., F.R.S. Received June 12, 1894. [Publication deferred.] The Society adjourned over the Long Vacation to Thursday, November 15. Presents, June 21, 1894. Transactions. Baltimore : — Johns Hopkins University. Circulars. Vol. XIII. No. 112. 4to. Baltimore 1894. The University. Berlin : — Deutsche Chemische Gresellschaft. Berichte. 1893. No. 19. 1894. Nos. 1—9. 8vo. 'Berlin. The Society. Berne : — Naturforschende Gesellschaft. Mittheilungen. 1893. 8vo. Bern 1894. The Society. Bombay : — Bombay Branch of the Royal Asiatic Society. Journal. Vol. XVIII. Nos. 49a— 50. 8vo. Bombay 1894. The Branch. Brussels : — Academie Royale de Medecine. Bulletin. Tome VII. Nos. 10—11. Tome VIII. Nos. 1—4. 8vo. 'Bruxelles 1893-94. The Academy. Academie Royale des Sciences. Bulletin. Tome XXVII. Nos. 1 — 4. 8vo. Bruxelles 1894. The Academy. Calcutta: — Asiatic Society of Bengal. Journal. Vol. LXIL Part 1. No. 4. Part 2. No. 4. 8vo. Calcutta 1893; Proceedings. 1893. No. 10. 1894. No. 1. 8vo. Calcutta 1894 ; Annual Address, delivered by the Hon. Sir C. A. Elliott. 1894. 8vo. Calcutta 1894. The Society. Cambridge, Mass. : — Harvard University. Bulletin. Vol. VII. No. 6. 8vo. Cambridge, Mass. 1894. The University. Dublin : — Royal Irish Academy. Transactions. Vol. XXX. Parts 11—12. 4to. Dublin 1894; Proceedings. Vol. III. No. 2. 8vo. Dublin 1894. The Academy. Florence: — Biblioteca Nazionale Centrale. Bollettino. 1891. Indici. 1894. Nos. 193—202. 8vo. Firenze. The Library. 308 Presents. [June 21, Transactions (continued). Kazan : — Imperial University. Scientific Notes. 1894. No. 3. [JRussianJ] 8vo. Kazan. The University. Leipsic : — Astronomische Gesellschaft. Vierteljahrsschrift. Jahrg. XXIX. Heft 1. 8vo. Leipzig 1894. The Society. London : — British Association for the Advancement of Science. Report of the Sixty-third Meeting, 1893. 8vo. London 1894. The Association. British Museum (Natural History). Catalogue of the Birds. Vol. XXIII. 8vo. London 1894 ; Catalogue of the Mesozoic Plants. Part 1. 8vo: London 1894 ; A Monograph of the Lichens found in Britain : being a Descriptive Catalogue of the Species in the Herbarium. Part 1. 8vo. London 1894. The Museum. Camera Club. Journal. Vol. VIII. No. 97. 8vo. London 1894. The Club. Chemical Society. Journal. January — June, 1894. 8vo. London; Proceedings. Nos. 131 — 139. 8vo. London 1894. The Society. Geological Society. Abstract of the Proceedings. Nos. 616—627. 8vo. [London 1894.] The Society. Institution of Civil Engineers. Abstracts of the Proceedings. Session 1893-94. Nos. 6—16. 8vo. [London.'] The Institution. Institution of Electrical Engineers. Journal. Vol. XXII. Nos. 108—111. 8vo. London 1894. The Institution. Middlesex Hospital. Reports of the Medical and Surgical Registrars, and Pathologist for the year 1892. 8vo. London 1894. The Hospital. Odontological Society. Transactions. Vol. XXVI. No. 7. 8vo. London 1894. The Society. Pharmaceutical Society of Great Britain. Pharmaceutical Journal and Transactions. January — June, 1894. 8vo. London. The Society. Royal Astronomical Society. Monthly Notices. Vol. LIV. Nos. 2—6. 8vo. London 1894. The Society. Royal Geographical Society. The Geographical Journal. January — June, 1894. 8vo. London. The Society. Royal Institution of Great Britain. Reports of the Weekly Evening Meetings. May, 1893. March — April, 1894. 8vo. London. The Institution. Society of Arts. Journal. January — June, 1894. 8vo. London. The Society. Society of Chemical Industry. Journal. Vol. XII. No. 12. Vol. XIII. Nos. 1-4. 8vo. London I89±. The Society. 1894] Presents. 309 Transactions (continued). Zoological Society. Proceedings. 1894. Part 1. 8vo. London. The Society. Lyons : — Societe d'Anthropologie. Bulletin. Tome XII. 8vo. Lyon 1894. The Society. Societe Linneenne. Annales. Tome XXXVIII — XL. 8vo. Lyon 1891-93. The Society. New York : — American Museum of Natural History. Bulletin. Vol. VI. Pages 161—192. 8vo. New York 1894. The Museum. Paris: — Academic des Sciences. Comptes Rendus. Janvier — Juin, 1894. 4to. Paris. The Academy. Ecole des Hantes Etudes. Bibliotheque. Fasc. 100. Livr. 1. 8vo. Paris 1893. The School. Ecole Normale Superieure. Annales Scientifiques. Tome XI. No. 6. 4to. Paris 1894. The School. Museum d'Histoire Naturelle. Tome V. 4to. Paris 1893 ; Centenaire de la Fondation du Museum d'Histoire Naturelle (10 Juin, 1793—10 Juin, 1893) : Volume Commemoratif. 4to. Paris 1893. The Museum. Societe de Biologic. Comptes Rendus. Janvier — Juin, 1894. 8vo. Paris. The Society. Societe de Geographic. Comptes Rendus des Seances. 1894. Nos. 1 — 12. 8vo. Paris. The Society. Societe d'Encouragement pour 1'Industrie Nationale. Bulletin. Novembre — Decembre, 1893. Janvier — Avril, 1894. 4to. Paris; Compte-Rendu des Seances. Janvier — Juin, 1894. 8vo. Paris. The Society. Societe Fra^aise de Physique. Bulletin Bimensuel. Janvier — Juin, 1894. 8vo. Paris. The Society. Societe Geologique. Bulletin. Tome XXI. No. 5. Tome XXII. No. 1. 8vo. Paris 1894 ; Compte-Rendu des Seances. 1894. Nos. 1—11. 8vo. [Paris.'] The Society. Societe Philomathique. Compte-Rendu des Seances. 1894. Nos. 6 — 14. 8vo. Paris. The Society. Philadelphia : — Academy of Natural Sciences. Proceedings. 1893. Pp. 377—520. 1894. Pp. 1—88. 8vo. [Philadelphia.'] The Academy. Franklin Institute. Journal. January — June, 1894. 8vo. Philadelphia. The Institute. Prague : — Gesellschaft zur Forderung deutscher Wissenschaft, Kunst und Literatur in Bohmen. Mittheilung. Nr. 2. 8vo. Prag 1894. The Society. Rome: — Accademia Pontificia de' Nuovi Lincei. Atti. Anno XLVI. Sessioni 1—3. 4to. Roma 1893. The Academy. 310 Presents. [June 21, Transactions (continued) . Reale Accademia del Lincei. Rendiconti. Serie 5. Vol. III. Semestre 1. Fasc. 1 — 8. 8vo. Eoma 1894. The Academy. Siena : — R. Accademia dei Fisiocritici. Atti. Vol. VI. Fasc. 6 — 7. 8vo. Siena 1894 ; Processi Verbali. No. 4. 8vo. Siena 1894. The Academy. Stockholm : — Kongl. Vetenskaps Akademie. Ofversigt. Arg. LI. No. 4. 8vo. Stockholm 1894. The Academy. Switzerland : — Societe Helvetique des Sciences Naturelles. Actes. 1893. 8vo. Lausanne; Compte Rendn des Travaux. 1893. 8vo. Geneve. The Society. Sydney : — Linnean Society of New South Wales. Abstract of Proceedings. March — April, 1894. 8vo. Sydney. The Society. Toulouse : — Academie des Sciences, Inscriptions et Belles-Lettres. Memoires. Serie 9. Tome I. 8vo. Toulouse 1889. The Academy. Venice : — R. Istituto Vcneto di Scienze, Lettere ed Arti. Atti. Tomo L. Disp. 4—10. Tomo LI. Tomo LII. Disp. 1—3. 8vo. Venezia 1892-94. The Institute. Vienna : — Kais. Akademie der Wissenschaften. Sitzungsberichte (Math.-naturw. Classe). Bd. GUI. Abth. 2a. Heft 1—2. Abth. 2b. Heft 1—3. 8vo. Wien 1894; Anzeiger. Jahrg. 1894. Nos. 1—13. 8vo. Wien. The Academy. Observations and Reports. Dublin : — General Register Office. Weekly Return of Births and Deaths. January — June, 1894. 8vo. Dublin. The Office. Greenwich : — Royal Observatory. Rates of Chronometers on trial, from July, 1893, to January, 1894. 4to. [London 1894] ; Rates of Deck Watches on trial, from October, 1893, to February, 1894. 4to. [London 1894] ; Report of the Astronomer Royal to the Board of Visitors, June 2, 1894. 4to. [London.'] The Observatory. India : — Great Trigonometrical Survey. Catalogue of Stars for the Epoch January 1, 1892. 4to. Dehra Dun 1893. The Survey. London : — Meteorological Office. Daily Weather Reports. Janu- ary 1 — February 17. 4to. London; Weekly Weather Report. Vol. XI. Nos. 1—21. 4to. London 1894. The Office. Melbourne : — Observatory. Records of Results of Observations in Meteorology and Terrestrial Magnetism. January — Septem- ber, 1893. 8vo. Melbourne 1893-94. The Observatory. 1894.] Presents. 311 Observations and Reports (continued). Paris: — Observatoire. Rapport Annuel, 1893. 4to. Paris 1894. The Observatory. Washington : — Department of Agriculture. Monthly Weather Review. March, 1894. 4to. Washington. The Department. U.S. Patent Office. Official Gazette. Vol. LVI. Vol. LVII. Nos. 1 — 8. 8vo. Washington 1894 ; with Alphabetical Lists of Patentees and Inventions. The Office. Journals. American Chemical Journal. Vol. XVI. Nos. 1 — 5. 8vo. Baltimore. The Editor. American Journal of Mathematics. Vol. XVI. Nos. 1 — 2. 4to. Baltimore 1894. The Editors. American Journal of Science. January — June, 1894. 8vo. New Haven. The Editors. Analyst. January — June, 1894. 8vo. London. Society of Public Analysts. Annalen der -Physik und Chemie. 1894. Nos. 1 — 6. 8vo. Leipzig ; Beiblatter. 1893. No. 12. 1894. Nos. 1—5. 8vo. Leipzig. The Editors. Annales des Mines. 1893. Livr. 12. 1894, Livr. 1—4. 8vo. Paris. Ecole des Mines, Paris. Annales des Ponts et Chaussees. 1893. Cahier 10 — 12. 1894. Cahier 1 — 4. 8vo. Paris. Ministers des Travaux Publics, Paris. Archives des Sciences Biologiques. Tome II. No. 5. 4to. St. Petersbourg 1893. Institut Imperial de Medecine Experimentale. Astronomical Journal. Vol. XIV. Nos. 1 — 6. 4to. Boston (Mass.) 1894. Smithsonian Institution. Astronomic (L') Janvier — Juin, 1894. 8vo. Paris. The Editor. Astronomy and Astro-Physics. June, 1894. 8vo. Northfield, Minn. The Editors. Ateneo Veneto. Serie XVI— XVII. 8vo. Venezia 1892-93. The Ateneo. Athenseum. January — June, 1894. 4to. London. The Editor. Builder. January — June, 1894. Folio. London. The Editor. Chemical News. January — June. 1894. 8vo. London. Mr. W. Crookes, F.R.S. Cosmos. Janvier — Juin, 1894. 8vo. Paris. The Editor. 312 Presents. [June 21, Journals (continued). Educational Times. January — June, 1894. 4to. London. College of Preceptors. Electrical Engineer. January — June, 1894. Folio. London. The Editor. Electrical Review. January — June, 1894. Folio. London. The Editor. Electrician. January — June, 1894. Folio. London. The Editor. Electricien (L'). Janvier — Juin, 1894. Folio. Pans. The Editor. Industries and Iron. January — June, 1894. 4to. London. The Editor. Meteorologische Zeitschrift. 1893. Heft 12. 1894. Hefte 1—5. Sin. Folio. Wien. Oesterreichische Gesellschaft fur Meteorologie. Morskoi Sbornik. 1893. Nos. 10—12. 1894. Nos. 1—2. \JRussian.~] 8vo. St. Petersburg. Compass Observatory, Cronstadt. Nature. January — June, 1894. Boy. 8vo. London. The Editor. New York Medical Journal. January — June, 1894. 4to. New York. The Editor. Notes and Queries. January — June, 1894. 4to. London. The Editor. Observatory. January — June, 1894. 8vo. London. The Editors. Revue Generate des Sciences. Janvier — Juin, 1894. 8vo. Paris. The Editor. Revue Scientifique. Janvier — Juin, 1894. 4to. Paris. The Editor. Symons's Monthly Meteorological Magazine. January — June, 1894. 8vo. London. Mr. G. J. Symons, F.R.S. Arnoux (G.) Arithmetique Graphique: les Espaces Arithmetiques Hypermagiques. 8vo. Paris 1894. The Author. Dollen (W.) Stern-Ephemeriden, 1892-93. 8vo. Berlin, Dorpat 1891—93. The Author. Ebermayer (E.) Die gesammte Lehre der Waldstreu mit Riicksicht auf die Chemische Statik des Waldbaues. 8vo. Berlin 1876 ; Die physikalischen Einwirkungen des Waldes auf Luft und • Boden und seine klimatologische und hygienische Bedeutung. 8vo. Berlin 1873 [and Atlas. Folio]. Sir John Evans, Treas. R.S. Norman (J. H.) The Science of Money. 8vo. London 1894. The Author. 1894.] Presents. 313 Pavy (F. W.), F.B.S. The Physiology of the Carbohydrates ; their Application as Food and Relation to Diabetes. 8vo. London 1894. The Author. Rambaut (A. A.) Occultation of a Virginis observed at Dunsink, 1894, March 22. 8vo. London 1894 ; On the Great Meteor of February 8, 1894. 8vo. Dublin 1894 ; On the Proper Motion of Stars in the Dumb-bell Nebula. 8vo. London 1894. The Author. Saint-Lager (Dr.) Onothera ou CBnothera: les Anes et le Vin. 8vo. Paris 1893. The Author. Five Photographs of Nebulee. Dr. Isaac Roberts, F.R.S. vor,. LVI. 315 Third Report to the Royal Society Water Research Committee. By PERCY F. FRANKLAND. Ph.D., B.Sc., F.R.S., Professor of Chemistry in Mason College, Birmingham, and H. MARSHALL WARD, D.Sc., F.B.S., F.L.S., F.R.H.S., Professor of Botany, Royal Indian Engineering College, Cooper's Hill. Presented to the Com- mittee October 4, 1894. PART I. "Further Experiments on the Action of Light on Bacillus anthracis, and on the Bacteria of the Thames." By H. MARSHALL WARD, D.Sc.,F.R.S.,F.L.S., F.R.H.S., Professor of Botany, Cooper's Hill. In a previous Report to the Committee, I have shown that the action of light on bacteria is not only very definite, and much more pronounced than had hitherto been supposed, but that it has an im- portance in its bearing on the question of the destruction of these organisms in the water of rivers, ponds, &c., vastly greater than had ever been suspected. In this Report I offer some of the results of long continuous in- vestigations into the details of this bactericidal action of light — both solar and electric — and into the bacterial flora of the River Thames, as studied at a point where it flows below Cooper's Hill. The reader may be referred to the previous Reports* for details as to the methods of investigation employed, and as to the chief results obtained by exposing the spores of Bacillus anthracis to the direct action of undecomposed solar light ; he will also find further details in two papers presented to the Royal Society in 1892 and 1893f regarding this matter, and regarding preliminary investigations into the action of solar light which has been passed through various absorbent media. Special attention may be directed to pp. 23 — 34 of the ' Pro- ceedings,' vol. 53, and the following experimental results may in the first place be taken as supplementing those there published. * ' Proc. Roy. Soc.,' vols. 51 and 53. t ' Proc. Roy. Soc.,' vol. 52, pp. 393—100 ; and vol. 53, pp. 23—44. VOL. LVI. Z 316 Prof's. Percy Frankland and Marshall Ward. Re ,: iji -§ . a a co .2 4> O f-H te S « o g k rO 05 rrt * o -° a ' •g ° S e ? e . S>n"-S c £ o1^ cO as, 3 3 tt M « « — i 'u rn JlXof Mao W5 kd ITS ID \a •* J? S oq oj § ^ c4 eo I ' ll O O CO CO CO CO CO oq c4 5>i 01 .5 *=«— jocn=!r!3a3:::l!3&-~o 3 •** i 'C B^^i -^rC 03 ~ c-!s'!;'irr3 = •^jQQrO^O^c5-^JHC3s3c3?HHccCT1S Illwll^lllllll « ^ w v ' r 1 00 8 *"* S ^ S _e^ 53 3 s3 ^ o rS -g _ 3 0 VS S ss cr1^ o > ^ • fli tj to- '^ • SSC ^Sg®a •n ^.2 ^'s-d ^.2 c°-£ ^-.o§°-'3 • -^30 H ••.SfS S r > a" £ «_® N » ^ J -Eel ^ a c a ^ s? CD O N •* CO H ® s^sid p"tr .tr O r-t M "N ftl-sl 4 S S "* S y * « • o 4* Sil§ • -^ § O _ ° is a o - ^3 « r « reds « a _. f -* jS *© 0 "J^ 'O 3i 00 S " D 2 t ro co o O O O O O CO O CO *s | O O i-l r-i N w. O CO C > CO CO^COgq •^NTji ^ " ~ «* Q ~— cS cS S S S S B «D *> •* ~§ 'g ~ - ~ - s3 3 S3 " H w D - = - s O = -..., fT) __^- *• v_; - Cj r-l (M o fl « O JS o S CO M s -. s S5jj o S 3 o 'S O P3 ^ o rl (M CO M Report on the Bacteriology of Water. 327 or elastic bands, and the whole then so fixed that the sunlight had to traverse the coloured or other fluid before reaching the agar film in which the spores were embedded. It will, of course, be noticed that the light here traverses three plates of glass as well as the solution of the screen before impinging on the spores in the film, a fact of importance. I have summed up the characters and chief properties of the solu- tions employed in the following Table E, and need not, therefore, describe them in detail here. In all cases, excepting N"os. 3 and 5, the medium employed for solution was water ; in these exceptional cases, where carbon bisulphide and alcohol were used, it was neces- sary to have screens devoid of cement. These were met with in the form of certain small glass flasks, shaped like brandy flasks or scent flasks, with a long neck and flat sides ; they are used on the Continent for sealing up cultures of bacteria. FIQ. 2. nzr The chief objection to their use is that the flat sides are apt to be slightly uneven in thickness on the internal face ; by carefully select- ing from a large number, however, I was able to secure several very good screens of this description.* The liquid is, of course, bottled in the usual way, and the neck secured with a good cork. §VI. The following Table F summarises the results of a number of expo- sures behind these screens of coloured and other absorbent media, with particulars as to the dates of exposure, number of hours, insola- tion, and incubation, and other factors worth extracting from the notes. It will be observed that I here confined my experiments entirely to the spores of Bacillus anthracis. I did this because it became more and more evident that until I had obtained all the information possi- ble about some one species — the factor most constant in this long series of slight variables — it would be difficult to value the import- ance of specific differences later. Experience has thoroughly con- firmed the justice of this conclusion. * In later experiments I have had the side ground out flat, and glass or quartz plates cemented on. • 328 Profs. Percy Frankland and Marshall Ward. ^ 03 ~« K • H » 3 ^^ P bD*'H O , Ci Q [ *fl .£ a d Q^ ^ o 3 bCT3 ^ CB'S '? 0 o a 2 -« >> a 'S ,. d «• « "— ' ^o g r;:; a o ^ •*i ^ "o P -» 3 p 3 W » • :Ilf IE bo °.s ^ «^ d •t § . S3 ~ id j= ™ — i t^> 1*8 to bC C «T — S to ^ bo 5 5<« i S . rrr" O _3 •« C S ^ »° g d T3 2 0 bDTS o J S ^-s ^ *J3 ^ ^Ss * ? c! T3 - ^3 p d IM IS »i «j O K II a g £ o> be 3 3^ ^3 3 48 "Si 3 -O O r-l .2 d C S • E hj S3 3" 1 bo| i ^^ g-S O O -S ^2 ^ s Mt ^ |l 11 £ .2 § 3^ .2 I /~, .i Composition, Ammoniacal c pric oxide Prussian blue oxalic acid Iodine in carb disnlphide * s gd Chlorophyll alcohol 'o _o _o K. chromate (c( centrated) J 1 i-25, 11 i4 j c O [o o 3 E S bo n o 1 g 2 s 2 0 fe k o 1 1 I W 3 1 „•.„.;.„ • .-' • c Report on the Bacteriology of Water. 329 - By superposing such a screen on one of anirn. £ O 03 O 1 d S 5 3 S 2 •3 W Q — i 53 93~ 9 i o a ft-li 3 3 0 O a o 4) 93 [o o f> 03 t 1" g S T! fl 03 9> 1* 93 O £ « 93 93 All blue-violet b §Q 1,1 al* Ultra-violet an half violet Infra-red only O 1 1 O Cuts off all vio 93 2 93 S O oo 3 Infra-red to extent G T3 . o 03 «• -a g . g £ 73 _§ a O . f^ c^ F-J g Q Tl _ P£J C is "93 ^ ^ r 93 i 03 2 03 o3 9f bJO a ?, 42 3 ^ Jt 93 jj 33 bD 'S 3 ^ 2 03^ ! ^ S | ^- '*' "C J^ « s E o viole tl 93 55 93 X) 'p <1 ^ ^ «l M M w £ •^ . . , c: a . . 3 ® 9) a1 93 ,2 a I— C CT1 . 00 £ > o 03 I 1 "*""&, g 1 M • 1-4 C 'i H 03 "3 'S OD a Strong ft 1 "S I* Dilute fu h i 03 P 6 _g ^Esculin nine su^ &. i 93 93 K "S o o s o 93 "a o 93 O m P b a> 3 S rW 03 a 73 O E o 'o t. 1-1 O O "o 0 £ S B 0 35 O 1— 1 H i— ( co i-l 10 r-l CO I-l i— I 330 Profs. Percy Frankland and Marshall Ward. Table F. — Exposures behinc Number of plate. Date made. Screen employed. Date exposed. Kind of exposure. Number of hours sunshine. Date put into incubator. C (1) Feb. 12 Quinine sulph- Feb. 12 Direct 5 Feb. 13 ate and 13 I ,,20 Quinine sulph- Feb. 25 3 h. reflected : 5 „ 25 ate 2 h. direct J Iodine in CS2 . . 5 K Prussian blue M n 5 and oxalic acid L Am. cu. oxide.. • Reflected 0 M Feb. 25 Weak K. chrom- Feb. 25 M 3 Feb" 27 ate and 26 N Stroii" potass- 3 chromate 0 Picric acid .... 2 h. reflected : 3 1 h. direct P Methylene blue „ 3 M and picric acid Q M Am. cu. oxide.. Feb. 26 3 h. reflected : 5 Feb. 28 and 28 2 h. direct R " *Chlorophyll. . . " "' 5 " S Iodine in CS2 . . 5 TJ Feb" 28 Weak K. chrom- Feb! 28 Reflected 4 w ate V Strong potass- 4 " chromate , j " V w *Chlorophyll... 4 x Direct . 3 Y " Q.uinine sulph- 5, 3 " ate 1 Mar. 4 Strong potass- Mar. 4 Eeflected 2 Mar. 4 chromate 2 Weak potass- 2 chromate " 3 ,j lod. + CSj.... „ „ 2 „ 4 2 5 " Strong f uchsin Mar. 5 3 M Mar. 5 6 Dilute fuchsin 3 Report on the Bacteriology of Water. Bottle Screens, Coloured, &c. 331 Results. Bemarks. Good letter W , Nothing appeared on the plate till the third day, and then only about 200 colonies at the extreme margin. No letter in six days No letter in six days No trace of germination any- where on the plate, except extreme margin Excellent sharp letter N Letter T visible after eighteen hours incubation, but not sharp No letter C. Germination took place equally all over the plate Letter Y feebly visible after twenty-four hours No letter X. Germination equal all over plate Sharp letter Z. No trace of letter. Extremely faint B. No letter. No trace of letter. Faint letter C. No trace of letter. Very faint letter X , YOL. LVI. The first three days showed powerful inhibition- effects, and no letter was visible till fourth day : then sharp and clear. From 20th to 25th the weather was dull and cold. Temperature of plates averaged 6° C., and none had germinated on morning of 25th, when we had brilliant hot sunshine. The letter after three days was not very sharp, since the colonies around were large and not very numerous, and about six or eight were seen on the insulated area. Even on the fourth day the contrast was not sharp, the surrounding colonies being so few and so large. Closer inspection showed a faint " ghost " of letter X on third day. *The chlorophyll at first (Feb. 26) blocked out all the blue-violet, and had bands in red and green ; but during the last two hours (Feb. 28) it only cut off violet, and had feeble bands in red and green. Colour olive, in place of deep blue-green. *The chlorophyll = deepest solution ; total ab- sorption from b onwards, and deep broad bands in red and green. Temperature rather high. Sun very bright and hot, but interrupted by clouds occasionally. The exposures began before 2 P.M., and ended just after 4 P.M., and two hours expresses the maximum of sunlight. All were over plane mirrors, carefully adjusted. The experiment was of little value. I had probably not used a sufficiently large charge of spores. These dilute fuchsin screens need careful watching. The colouring matter gathers it to flocks in time, and lets much more light through. 2 A 332 Profs. Percy Frankland and Marshall Table Ward. F. — Exposures behind Number of plate. Date made. Screen employed. Date exposed. Kind of exposure. Number of hours sunshine. Date put into incubator. 11 Mar. 4 Strong cbloro- Mar. 5 3 Mar. 5 12 phyll Strong- potass- 3 15 Mar 7 chromate Dilute* fuchsin Mar. 7 • 2 Mar. 7 16 Strong* 2 17 . .ZEsculin + Reflected 2 18 quinine 2 19 Potass-chrom- 2 " 20 ate* (strong) Quinine Reflected .... 2 Ilia Mar. 10 lod. + CS, Mar. 10 1 Mar. 10 Illi 2 IIIc Strong potass- 1 Jllrf chromate 2 IV Reflected 2 V -ZEsculin + 2 VI quinine 2 Xa Mar. 11 Chlorophyll. ... Mar. 11 2 Mar. 11 XZ> CuSO4 2 7 Mar 13 lod + CS2 .... Mar 13 2k Mar. 13 9a 14 „ 15, 10 — 12 17 M 16, and 17 B 1 Mar 21 Mar. 21 2 Mar. 21 B 2 CS2 + lod 2 c Chlorophyll. , . . 3 \VI Mar 27 KKS Mar. 27 3 Mar. 27, XVII Water 3 3 P.M. Report on the Bacteriology of Water. Bottle Screens, Coloured, &c. — continued. 333 Number of clays incubated Results. No trace of letter. Q-ood X, but by no means all killed No trace. Plate evenly covered all over Extremely faint " ghost " of Z after twenty-four hours, and invisible after forty- eight hours. No trace. Germination all over No germination till second day. No letter No germination till third day. Very faint and transient letter later Active germination and traces of figures but transient only Letter T visible in twenty hours, but not all killed. Letter X visible in seventeen hours, and sharpening up in twenty hours Z visible, &c., pan passu with latter Germination equal all over plate. No letter Good letter H. Germinated evenly all over . . . No letter appeared C out sharp, but diffused Faint N, cleared by 24th, but bad outline Extremely ill-defined, and never so good on H Germinated evenly all over. Much clearance over a shield- shaped area in eighteen hours, but no clear U Very sharp and clear T. Kemarks. The chlorophyll much oxidised and olive coloured at end, but still cut out all the blue. f*Plane screen. The dilute fuchsin lets red — yellow and trace of green through, and then blocks up to one-third between F and G. Then lets a considerable proportion of blue-violet [ throuh. *Plane screen. This chromate is reciprocal to dilute fuchsin. Same plate, screens, &c. Successive windows opened. Bottle screen. Powerful inhibition on the CS2 side, and no germination there at all at first. Faint " ghost " after twenty-five hours on chromate side, but obliterated later. I Over same mirror. Both X and Z gradually oh- i- literated next day — i.e., spores not killed, only retarded. J Extremely good sun and blue sky. Exposed 1.45 to 4.15. Incubated at 25° C. One screened ; the other not. Same plate, &c. | Same plate. H screened. N not. Very hot brilliant sun and blue sky. Kept till 25th. N clearest, but bad. Very hot sun, hazy first hour, then brilliant. 2 A 2 334 Profs. Percy Frankland and Marshall Ward. If we now look at the results tabulated above, it is seen tbat tbe solar action is evident, thougb feeble, through dilute fncbsin, sesculin and quinine, and picric acid ; wbile no trace of action occurred througb potassium chroraate, chlorophyll, eosin, and strong fuchsin. On the other hand, the action was sharply defined where am- moniacal cupric oxide or water alone was employed, and also where alum dissolved in water was used. In other words, the action is most pronounced when the rays transmitted are those of the blue- violet end of the spectrum, bearing out the results already obtained more generally. During the progress of the experiments above tabulated, a number of other points of interest were observed. With carbon bisulphide and iodine it frequently happened that no letter was obtained on the plates (Expts. J, 3, Ilia, 7, 9a), but occasionally the light action was recorded by the appearance of the letter (Expts. S, Illfe, BZ). The fact is, the solution did transmit a scarcely perceptible amount of violet rays, and since I could not discover any definite relation between the times of exposure and the results, one of two possibili- ties suggested itself — either differences in the thickness of tbe glass of the plates, or differences in the degree of clearness of the atmo- sphere may account for the discrepancies. Probably both causes were effective, for, of course, they both affect these violet rays con- siderably. Another phenomenon repeatedly noticed, both in these experiments and in others, was that if the exposure to a very bright sun is con- tinued too long, and especially if the plate is not very accurately at right angles to the direction of the rays, the light may clear the plates entirely, or nearly so. This seems to be due to the reflections of the light from the glass surfaces inside the Petri's dishes ; if the light is very intense, or the exposure long, these reflected rays are sufficiently powerful to produce effects similar to those of the direct light. This seems to me to explain another phenomenon very commonly met with. In many cases of long exposure to clear hot sunshine, the first evidence of the successful light action is not a sharp well-defined letter, bat a blurred clear patch, which slowly sharpens up as incuba- tion goes on. It is evident that in such cases the action of the light has extended beyond the boundaries of the stencil letter, into parts of the film really not exposed to the direct incident rays. I explain this as due to the reflection of some of the rays from the glass surfaces in the interior of the plate. These reflected rays are not sufficiently intense to complete the bactericidal action, they only inhibit the organism more or less, or at least leave many spores still alive ; consequently, while these out- Report on the Bacteriology of Water. 335 lying spores germinate more slowly than those further away from the illuminated area — the stencil letter — they do at last germinate out, and so the previously blurred letter becomes sharp and clear in outline. The phenomenon very ranch resembles the development of the indistinct " ghosts " of letters in cases where the exposure is too short, or the light not sufficiently intense, or wanting in active rays. Such faint letters gradually become obliterated as incubation pro- ceeds, because the spores, still alive but only retarded in develop- ment, gradually germinate out to an extent so little differing from, the rest that the eye fails to detect any difference. The retarded development of a few colonies, at a late period of in- cubation, on the hitherto clear area of the exposed letter, is due to similar causes, but produced in a slightly different way. It is ex- tremely difficult (probably impossible) to thoroughly distribute the spores in the film so that some do not shelter others from the light ; consequently, when a clump of spores exists on the exposed area some of the inner spores may so far escape the bactericidal action as to be able to germinate out later, and I have had many experiences of these cases. In fact, the chief point about a good film — i.e., one which develops a sharp letter after ordinary exposure — is that the spores shall be neither too few nor too many, and thoroughly and evenly separated and distributed ; and, lastly, that the agar or other medium shall not be too thick, and thus render possible the ordering of long rows of spores one behind another (i.e., in rows parallel to the ray of incident light) which thus shelter one another from the light's action. These points, and some others, come out still more clearly in the next series of experiments. §vn. The following series of experiments were made behind superposed screens, and it must be borne in mind that the light had to traverse not only a double thickness of solution, but also five thicknesses of glass, before reaching the spores. On the whole these results may be regarded as simply confirming the previous ones, but I was (and still am) considerably puzzled by the behaviour of the iodine and carbon bisulphide screens. In several cases the plates seem to be destroyed by the light passing through this medium, and for some time I was doubtful whether there might not be a cumulative effect due to the action of the infra-red rays. It seemed extremely probable that these rays — the " dark " heat rays — do help to promote the bactericidal action, and I thought perhaps because they accelerate the chemical changes on which the action depends. 336 Profs. Percy Frankland and Marshall Ward. iiilll°|:|ji£cji*f §* Ssor.-^S^ i^-^ ^ ^'Q "SSgtl^. C0)aS^ ^>C? I § i ii§tf*li^st fz-if e|o-|=42S>,ai *>.•:"= H •^ -^ -^ S _a J '"" — -^ ^ '~ ''"• O."^ ^ C ** r "* "^ c ^ ^ :r 'hr *" S ^ ' _o" *" _S " _o a ,J° _ f , > . 3* || ar^ 3^ «C| ^§l i m§j . g~C| "*§ Ss2 •« ° £ S1? a • *• 3 'S g t* fl ^J3u*M ^"^rt °"-— n ajS^t. •*"»,ct?xt*-' § ^=i IriKifS |gi*il - _= f -a" S 1 1 ^2 ^'f -2)2-3o2«5'Soe!Sa'wIc • (§'"'" , • 'K v v , o ? tj 0 o' §•3.3 ^"g Ss '• 8 f- .5 ^^^ ^^ S a a •8 M ja E £ : s| : Jig .§"* 01 OJ H / — * — \ / — * — > *° t. "* S o3 £|j E no •° S.S-S si = S* E E J t"» E 2*.gg If 1 X5 M IN •* ^ t- CO - 3 m r ^_ jf j *~. - ss S2 "M 'N * a o 5 "• a! >• ci ».,.» s S s s * 2 * 12 N " T a * . ^'Q - gj S . 53 £ c 1 2 *' 2 " c? "* « 5* o S 2 a = s ~ a -3 •a "3 §•£• . g 5S b = a.2§ « S ».* *|S + o p -5 ts c »££ »E<'§'3 ""^ § "^. J § ^cT-2 " S + co o o°E ° o § + 0 g a § a •§ ^ °"u o o o o *- » •0 • I •8 S O jj BJ 58 jl 3 * ill IN 2 S 5 II i9 III (M 00 •* r- S «* M || 1 cr) 5 OS ||l N W O ^ U! N S5 O N M II S* e. « S — > "• •"" 5? Report on the Bacteriology of Water. 337 However, on comparing the action of a thick crystal of rock-salt with that of water and alum, I was unable to detect any such marked difference as would seem to follow if that conclusion were correct. As will be seen more clearly later, when I come to discuss the results obtained with the spectrum, the infra-red rays are themselves utterly without perceptible effect.* § VIII. The following series of experiments, designed to estimate the degree of light action on water bacteria, was carried out, under my direction and supervision, by Miss Hayward, of University College, London, and I owe it to her to state that their successful carrying out would have been impossible — on account of the numerous plates to be counted, in short periods, and involving very large numbers — but for her untiring industry and devotion to the work. Series I. About 150 c.c. of Thames water, collected at 10 A.M. on August 12, were distributed equally in three Erlenmeyer flasks, properly steri- lised, and the flasks labelled A, B,' and C ; and at 11 A.M. a 1-drop plate was made from each flask. These plates, examined and counted on August 14 at 11 A.M., gave respectively 1560, 1700, and 1080 colonies per c.c. — i.e., an average of 1446 colonies per c.c. developing in two days. On keeping the plates another day, two of them gave 3705 and 1656 respectively, while the third was uncountable and liquefied, the mean being 2080. We, therefore, assume that the water contained at the outset about 2700 bacteria per c.c., capable of developing in three days. The flasks were then placed as follows : — A was suspended by the neck so that it could be exposed to what sunshine there was, and at the same time be illuminated from below by the light reflected from a plane silvered mirror. B and C merely stood by the side of the stand supporting A, C being covered with tin-foil and black paper, while B was exposed to the light from above and at the sides. After standing thus from 12 noon to 4.30 P.M., the weather being very cloudy with only occasional bursts of sunshine, fresh samples were taken from each flask by means of sterilised pipettes, and new plates made to see if any changes had occurred of note. Taking first the plate from A, after two days' incubation, we found 1599 colonies per c.c. ; and after a further twenty-four hours, 2760 per c.c., of which 12 per cent, were liquefying forms. The number of living bacteria capable of developing in two to three days, there- * Since writing this the experiments with the spectrum have been published separately (See ' Proc. Roy. Soc.,' 1894, vol. 54, p. 472, Abstract). 338 Profs. Percy Frankland and Marshall Ward. FIG. 3. fore, was approximately the same as at the outset, and the dull light seems to have prevented the usual rapid multiplication. The plate from B, made after exposure, gave us 680 colonies per c.c. after two days' and 2550 colonies per c.c. after three days' incubation, and showed evident signs of liquefaction. Here, therefore, it would seem that the light had exercised an inhibitory action as to numbers. The plate from C gave 2268 colonies per c.c. after three days' in- cubation, but it was liquefying still more rapidly. So far, therefore, with the dull light of a cloudy day, it did not seem as if a.n exposure of four and a half hours gave results of much significance as to the numbers of bacteria ; but it did seem as if the plates from the exposed flasks showed less liquefaction. Meanwhile, the three flasks stood at a temperature of about 16° C. overnight in the laboratory, and were exposed next day from 11.30 A.M. to 4.30 P.M. — again a dull day, and practically no sunshine at all. Before exposure, however, we took samples as before, and after two days' incubation found that A had about 13,650 per c.c., B had 15,980 per c.c., and C was so badly liquefied that we could place no reliance on the numbers (1584) counted. These numbers are not very satisfactory taken by themselves, but they showed us that the matter was worth further investigation along similar lines. The following table H summarises the foregoing facts : — Report on the Bacteriology of Water. 339 I rS .S!o -0 O 00 fc 8 0 10 bit"" SCO sio M o M " . --1 a be P< 3 «^ P PQ o c 0 O 0 o fl 0) 03 "o3 O 340 Profs, Percy Frankland and Marshall Ward. § IX. On August 14bh, three flasks were prepared and exposed as before ; A' with mirror beneath, B' with no mirror, and C' wrapped up. Fairly bright sunshine prevailed during the exposure — from 11.15 to 4.30 — an occasional cloud obscuring the sun. Two samples showed that A' started with 1755 per c.c. and 1404, the mean being 1578 per c.c. After five hours' exposure, over the mirror, plates were again taken. Three plates yielded 1326, 780, and 858 per c.c., the mean being 988 per c.c., which looks as if a perceptible reduction had occurred. Two plates from B' at the start gave 680 and 2176, the mean being 1423 per c.c. After its five hoars' exposure, without a mirror, three samples yielded 918,476, and 476 per c.c., the mean being 623 per c.c., and again suggesting effect of bactericidal rays. Plates from C' at the beginning gave 1156 as the number to start with, and after the five hours side by side with the other flasks, but protecbed from the sunshine by foil and paper, samples gave 3240, 2052, and one uncountable. The mean, = 2646 per c.c., suggesting a perceptible increase. Here, again, it was evident that liquefaction took place much more rapidly on the plates from the unexposed flask than on those from the flasks exposed to light. The chief difficulty with these mixed plates is always that caused by the liquefying forms, one of which was especially troublesome, often coming on so rapidly that a plate which looked " safe " at a given time would be mined three or four hours later. The foregoing results are summarised in the following Table I. Without attempting to lay too much stress on the actual numbers in this series, it is pretty evident that if we take the totals or the means of the numbers of bacteria obtained from the water by taking three samples from each flask at each period of examination, we get at least some information as to the rate of action of the light on the total organisms. Put thus, the facts run as follows : — Of the nine samples taken at the start, four were not counted, as they liquefied too rapidly. The average of the other five gave 1434 per c.c. Flask A', after five hours' exposure over a mirror, gave 988 per c.c. as the mean of three samples. Flask B', after five hours' exposure without a mirror, gave 623 per c.c. as the mean of three plates. Flask C', not exposed, but otherwise treated similarly, gave 2646 as the mean of two plates. It seems impossible to doubt, therefore, that the exposure to light reduced the numbers by nearly one-half. But this proportion becomes Report on the Bacteriology of Water. 341 we note that cation would normally occur. much greater if we note that during the period an enormous multipli- § X. On August 15th two Erlenmeyer flasks were charged as before with Thames water, and labelled A and C. A was exposed over mirrors, and C wrapped up. We introduced the difference here of having an additional mirror behind the flask, as well as that below. The day was bright, with plenty of sunshine all the time, and A was exposed for 5| hours — from 10.30 a.m. to 4 p.m. — and then, samples taken from both. Meanwhile, the average of seven samples taken at the commence- ment gave 1644 as the number per 1 c.c. in the Thames water at starting. Unfortunately, the temperature rose during the next twenty-four hours sufficiently to soften the gelatine of the plates taken after the first 5^ hours, so we could not count these. On August 16lh — another bright, clear day — the flask A was again exposed for six hours, and C (wrapped up) beside it, both flasks having stood all night (nearly 12 hours), at 18° C in a cupboard in the laboratory. After this second exposure the plates gave — for A about 6000 per c.c., and for C over 174,000 per c.c. ; showing that the exposure to the light had kept down the numbers in A, in spite of the interval of twelve hours in a warm, dark cupboard, when, of course, the bacteria npt killed off by the first day's exposure multiplied rapidly. Here, again, I was struck by the diminution of the liquefaction on the plates from the flask A; it did not look like merely fewer liquefying forms, but as if those that were present really liquefied less rapidly and less efficiently than those from the flask not exposed to light. § XL On August 22nd two Erlenmeyer flasks, marked F 3 and F 4, were charged to a depth of 1 in. with Thames water, properly collected, &c. (see Table J). Flask F 3 was exposed to the sun with a mirror below ; F 4 stood, covered, by its side. The exposure lasted from 11.30 A.M. to 4.30 P.M. (being five hours), but only about one and a half to two hours at most was good sunlight, the day being showery with snatches of blue sky at intervals. Four samples taken at the beginning of the experiment gave 800, 1408, 748, and 1462 per c.c. as the numbers after two days' incubation. Total of four plates = 4418 ; average = 1104 colonies per c.c. to start with. Profs. Percy Franldand and Marshall Ward. tn Q* o .So •^ £ o Ut> CO "a 8 -? 3D O 00 PJ| rH i-H "^ oo i-H co os eg O P . CC 3 O O 6*0 0 J* CO 3 0 _CT< a o CO co CO U? o --1 fl -^ r- 1 M d 2" *' 2" *• »*j CD" u3 co" H 2" a 2" § £ g PH • «d • ^ *• o r-^3 d br fcjoo fcJD,-^ tJD ^^r^ bb ^ bb C r5 | 30 3 •* •^ s<1- •S rH ^ ^ -3 rH ^ *" 8 1-1 Hi T3 "*" • Tjf •*" • -*" ^ . ^ ^" • ^r a d *>+ M)S M^ i! bJD tUD pa M^ tbS B ^s ^^ ^,3 $ + ^S ^* ^rH ^ * ( rH (M r-l N rH (M rH IM a o ^ <; -tj •^ i| in 03 * ^ * PH Report on the Bacteriology of Water. 343 CD 'c O so co "j 3 en 3 0) O 3 o 3 iH CO O CD 1 eo~ . co~ . CD" «T co" y co" a co" .«TB »>( PH be Pk bo PH be Pk bO Pk bio "^ boO 3 i-l 3 CO i— i Is IS : 13 -*" • -*" ., _ « ^,a^r *ji^r ^ g ^ r-H ^ rt g 1-1 «^ a bb ^ bb Pk bb "^ bb Pk 13 ""* ^ 10 bbO bb Pk 3 ^. 3 »ft 3 CO 3 bbO bb PH 3 CO_ 3 ^ M. « W pq i— 1 O1 0 O O O ^H oq O O ^5 *Q o >o : : • : : : 1 3 § III » I'll ^ -S SH w'£ s o s 41 : 1 ^ -i bbO 3 CO i-H •* y bbO 5 P CO rH 1-1 ^' bbO = 3 CO 3 CO I— ( 3 CO E 1 Thames water » Thames water » Thames water M Thames water M « « b b b 344 Profs. Percy Frankland and Marshall Ward. After another day's incubation one plate was liquefied; the others gave 832, ]920, and 816. Total of three plates = 3568 ; average = 1189 colonies per c.c., and this may be taken as the highest number obtainable, the counting having been done and checked very carefully and thoroughly with a good lens. After the five hours' exposure, new plates were made — two from each of the flasks — and, in order to obviate as far as possible our previous difficulties with the liquefying forms, we diluted each 1 c.c. of the sample water with 9 c.c. of sterile distilled water, carefully prepared in advance. It is unnecessary to give details as to sterilisation, &c., but the following are the essential points of the plan followed : For each sample to be taken two pipettes and two test-tubes are needed. One of the test-tubes of each pair is graduated to hold 9 c.c. of the sterile water; into the other 1 c.c. of the water to be tested is dropped with one pipette, and the 9 c.c. of sterile water are then poured on to this, thus ensuring thorough and rapid mixture. The second pipette is then used to obtain the one or more drops taken to make the gelatine plate. By this method we found that two plates from F 3 (the exposed flask) gave 2880 and 1280 per c.c. on the 4th day. Total of two plates, 4160; mean, 2080 per c.c., suggesting that the light was not strong enough to prevent the bacteria from multiplying. On stand- ing two days longer these plates gave 3840 and 1600 ; total, 5440 ; mean, 2720 colonies per c.c., showing that there were a good many slowly developing germs present, and we were struck with the paucity of liquefying forms. Two plates from F 4 (the unexposed flask), made and examined at corresponding times, and in the same way, gave 660 and 2310 per c.c. on the fourth day. Total, 2970 ; mean, 1485 colonies per c.c. ; and on the fifth day (we could not go further) 660 and 3630 ; total, 4290 ; mean, 2145 colonies per c.c. So far, therefore, the numbers in the two flasks did not appear to be appreciably affected by the little sunlight that reached the water. The two flasks were now placed in an ice-safe over night, for expo- sure next day. They remained on ice about 14 hours. Next morning, August 23, these flasks were again put out at 9 A.M., and remained till about 4 P.M. (over seven hours) ; but it rained steadily all the time except the last hour, when the sun shone. Two plates, made of diluted samples as before, from each flask, were made at noon on this day, with the following results : After two days' incubation the plates from flask F 3 (exposed) showed 39,600 and 37,620 per c.c. Total of two plates, 77,220 ; mean, 38,610. After a further day's incubation the plates gave 47,190 and 41 250 ; total, 88,770 ; mean, 44,385 per c.c. Report on the Bacteriology of Water. 345 The two plates from flask F 4 (not exposed), examined at the same time, gave 10,800 and 7,200 per c.c. in two days. Total, 18,000 ; mean, 9000 per c.c. ; and, in three days, they showed 13,200 and an uncountable number, owing to liquefaction. These results were ddcidedly mystifying at first, for they showed apparently a stimulating effect of exposure to the light ; bat on going further into the matter it seems more probable that what really happens is, that (1) the sunlight was not powerful enough in blue- violet rays to produce any appreciable inhibition in the time ; and (2) the flask F 4, covered in tin foil, &3., did not become warmed so rapidly as the other, and consequently still showed the retarding action of the icing to which it had been subjected. That this explanation is right is borne out by the behaviour of the plates taken at 4 P.M. — i.e., after four hours' further exposure of the flasks — for the covered flask, although enormously increased in bacteria, was still behind the exposed one. After two days' incubation, the two plates from F 3 (exposed) at this period gave 18-4,800 and 165,000 respectively. Total, 349,800 ; mean, 179,900 per c.c. ; and, after three days, 231,000 and 207,900. Total, 438,900 ; mean, 219,450 per c.c. Whereas the two plates from the non-exposed flask gave respectively 19,500 and 12,000 per c.c. ; total, 31,500 ; mean, 15,750 per c.c., after two days' incubation, and were not counted further. Considering the enormous expenditure of time and trouble involved in making and counting these plates, we were somewhat discouraged by these negative results, which are summarised in the accompany- ing Table J (p. 346, &c.). § XII. On August 24th two flasks of Thames water labelled F 5 and F 6 were exposed exactly as the last, but the weather was fine and we had much bright jsun and blue sky, with rapidly moving white. clouds. The first exposure was from 9.30 A.M. to 4 P.M., and the arrange- ment as before. The temperature, as indicated by thermometers in control flasks, rose occasionally over 35° C, but was usually not above 30° C, and somewhat higher in the covered flask than in the un- covered one. Four plates of the water with which the flasks were charged were made at the time of starting the experiment, and were incubated as long as possible. In four days they gave us 1980, 1650, 1320, and 2970 colonies per c.c. Total of the four samples, 7920 ; average, 1980 colonies per c.c. After six and a half hours' exposure, of which about four hours was brilliant sunshine as far as we could estimate, two plates from each 340 Profs. Percy Fraukland and Marshall Ward. Table J. £ S E £ dj O | o o a . ta 0 r^ 1 "p. 1 h o PH V *8 • <3 _ I o ^ H H (1) Aug. 2 !•, 1 P.M. 49i 16—18° 25 800 7 small liquefying colonies. (•2) Aug. 25, 3 P.M. 75i » 26 832 10 liquefying colonies. (I) Aug. 24, 2.30 P.M. 51 44 1,408 6 liquefying. (2) Aug. 25, 3 P.M. 75£ „ 60 1,920 14 liquefying. (1) Aug. 24, 3 P.M. 5H „ 22 748 6 liquefying. (2) Aug. 25, 3 P.M. 75| ,, 24 816 9 liquefying. (1) Aug. 24, 3 P.M. 51J „ 43 1,462 11 liquefying. (2) Aug. 25, 3 P.M. 75 i ,, liqu efied and t icrefore unco un table (1) Aug. 26, 9 A.M. 88 ~n 9 2,880 1 c.c. diluted to 10 Heavy with sterile distilled clouds and water, and 1 drop of showers, the mixture taken. with blue None liquefying. 1 sky and mould present ad- sunshine ditionally . between. (2) Aug. 27, 12.30 P.M. 115i -. 12 3,840 Diluted. 1 mould Ditto. , (3) Aug. 28, 12.30 P.M. 139| ,, 2 small liquefying. 2 Ditto. moulds (1) Aug. 26, 9 A.M. 88 } T 4 1,280 Diluted. None lique- Ditto. fying. 1 mould (2) Aug. 27, 12.30P.M. 115.1 Q 1,600 Ditto Ditto. (3) Aug. 28, 12.30 P.M. 139i " 1 slow liquefier ap- Ditto. peared. (1) Aug. 26, 9.30 A.M. 88£ !J 2 660 Diluted. No lique- fying Plate on ice till next morning 9 A.M. (2) Aug. 27, 1 P.M. 115J 2 660 2 moulds appeared. (1) Aug. 26, 9.30 A.M. 88 k ( 7 2,310 No liquefying. (2) Aug. 27, 1 P.M. 115| „ 11 3,630 No liquefying. (1) Aug. 26, 10 A.M. 70 J» 120 39,600 Diluted. 4 small Steady , liquefiers rain all the (2) Aug. 27, 4 P.M. 100 (J 143 47,190 9 liquefiers. 1 mould morning ; appeared an hour's (3) Aug. 28, 12.30P.M. 120i . . 11 good sized lique- sun 3 — 4 fiers, and many P.M., but small ones feeble. (1) Aug. 26, 10 A.M. 70 j 114 37,620 Diluted. 5 liquefying, Ditto. all small (2) Aug. 27, 4.30 P.M. 100.V 125 41,250 Diluted. 16 lique- Ditto. fying VOL. LVI. 2 B 348 Profs. Percy Frankland and Marshall Ward. Table J— E E E -*-> B a a "3 03 O o o a ^3 J A o o 8 •M O • IM o C*H O When a 3J 'g °0 ^ o S g -u 0) made. i o o 1 1 H 1 rO O || J 1 s 0 ^ H s ^ * 121 6 * F4 Thames Aug. 22, Not 0 0 Flask left at P29 Aug. 23, water 11.30 laboratory, 12.30 P.M. (collected A.M. temperature Aug. 22, 16—18° 10 A.M.) over-night. \ in. deep 0 0 P30 F3 • V Exposed Aug. 22, 11.30—4.30 Aug. 23, Hi 2i— 3 » P31 Aug. 23, 4.30 P.M. 9—12 12—4.30 „ i, II n „ 12* 2J— 3 » P32 ,, F4 Not 0 0 P33 Aug. 23, 5 P.M. » » " " •• 0 0 » P34 " flask were made, and incubated also as carefully and long as possible, to get the maximum numbers. The two plates from the exposed flask F 5 gave 160 and 320 after four days' incubation. Total, 480 ; mean, 240 per c.c. ; whereas those from the unexposed flask gave 3400 and 5916 respectively in the same period. Total, 9316 ; mean, 4658 colonies per c.c. These numbers show very distinctly the effect of the sunshine, and are borne out clearly by what follows. After the exposure on August 24th, and after the samples for plates had been taken, both flasks were placed on ice over-night, and re- mained on ice about fourteen hours. On tlie 25th, at 9.30 A.M., the exposure was repeated, the flasks being first examined as to the effects of their sojourn in the ice-box. The numbers were found to have remained remarkably constant, being 210 for the exposed and 3136 for the covered flask. Report on the Bacteriology of Water. 349 continued. c3 ,0 o S 'S o * w "S O D 2 c 1 ^3 When examined. ew 3 o |-. «^H =+-! o 0 'o Remarks. Weather. o * £ fej S °" £. 3 o SO q s? ^3 O a s o • - H H *? fc (1) Aug. 26, 10 A.M. 69k 16—18° 32 10,800 Diluted. 8 liquefying. (2) Aug. 27, 4.30 P.M. iooJ " 44 13,200 " (1) Aug. 26, 11 A.M. (2) Aug. 27, 4 P.M. 99| >, 24 liqu 7,200 efied „ 3 liquefying. and t icrefore unco untable 1) Aug. 26, 10.30 A.M. 66 560 184,800 Diluted. 15 lique- Steady fying rain all the morning ; an hour's sun 3 — 4 P.M., but feeble. (2) Aug. 27, 5 P.M. 96J J; 700 231,000 Diluted. About 55 Ditto. liquefying (1) Aug. 26, 11 A.M. 66i „ 500 165,000 7 large and many Ditto. small liquefying. (2) Aug. 27, 5 P.M. 96£ ,, 630 207,900 Diluted Aug. 28, 11.30 A.M. 66 £ ,, 65 19,500 Diluted. 6 liquefying. 1 mould. >' >' " 40 12,000 Diluted. 1 liquefying. After this day's exposure, which lasted six hours, about half of which was more or less cloudy at intervals and the rest bright sun- shine, the plates made again showed great differences, clearly pointing to the inhibitory action of the sunlight, for the sample of exposed water gave 1584 as against 29,760 per c.c. in the unexposed flask. In other words, although flask F 5 had stood fourteen hours in the dark at a temperature not too low for increase, its exposure of about 12^ hours to the light had resulted in the reduction of its bacteria to a number below that it started with ; the unexposed flask meanwhile had its bacteria multiplying normally at the rapid rates usual for these waters. Had the sunshine been more intense during this second day, it is by no means improbable that the water could have been completely sterilised by exposure. The results discussed above are put into a tabular form in the following Table K :— 2 B 2 350 Profs. Percy Frankland and Marshall Ward. Table o § b _j "o 00 O 0) fl O & c r^ & rS O E 00 Is When made. .2 s T3 _. o "o I on 03 B 5 £ 1 O a> s o e o L J "S E 6 ^ H H w w O 5 F6 Thames Aug. 24, Not 0 0 P37 Aug. 24, 10 A.M. water 1 in. 9.30 A.M. deep. Collected Aug. 24, 9.10 A.M. » » » n •• 0 0 P38 ». „ H „ n 0 0 P39 » >, n ., N 0 0 P40 » F5 n M Exposed Aug. 24, 6* 4 P39a Aug. 24, 5 P.M. with 9.30 A.M.— mirror 4 P.M. 6i 4 P40a F'B if '» Not •• o2 0 •• P41 » N »> H » •• 0 0 •• P42 » F5 „ „ Exposed Aug. 24, 64- 4 After having P43 Aug. 25, 9.30 with 9.30 A.M.— been placed A.M. mirror 4 P.M. on ice over- night from 5.30 P.M. — 9.30 A.M. F6 » » Not •- 0 0 » P44 Aug. 25, 10 A.M F5 Exposed Aug. 24, I2i 7i 9 1 P49 Aug. 25, 3.30 with 9.30 A.M.— P.M. mirror 4 P.M. below Aug. 25, 9.30 A.M.— 10.30 P.M. F6 » " Not • • 0 0 •• P50 " § XIII. On August 25 two flasks labelled F 7 and F 8 were exposed as the last, from 9.30 A.M. to 4 P.M., weather mixed — clouds and bright sunny intervals. Four samples taken at the beginning gave 1666, 1292,1394, and 1768 colonies per c.c. ; total, 6120 ; average, 1530 germs per c.c. to start with. Report on the Bacteriology of Water. 351 K. j .H J a 1 d »i [ o o o .S § 8 When examined. •H O 5 d cj O «*-< 21 IM O >•< o Remarks. Weather. £ i d 3 o rS & H D o o 93 5 O ^ H W , B s F8 Thames Aug. 25, Not 0 0 P45 Aug. 25, 10.30 water 1 in. 9.30 A.M. A.M. deep. Collected Aug. 25, 9.10 A.M. » »- N ,. •• 0 0 P46 >. „ V M ,, • • 0 0 P47 n » » » » •• 0 0 P48 " F7 Exposed Aug. 25, 61- 4 P51 Aug. 25,4 P.M. mirror 9.30 A.M. — below and 4 P.M. behind „ J) ,, M n 6i 4 P 52« „ F8 " " Not " 0 0 P53a » » " » » -• 0 0 P54a " §xiv. On August 28 another pair of flasks (F 9 and P 10) were exposed as before, 10.30 A.M. till 6.80 P.M., and left out all night ; exposed next day from 10 A.M. to 4 P.M. The sky was cloudy on both days, but there was a good deal of sunshine also, particularly on August 29. We estimated the actual exposure to daylight as 18 hours, and about 8 hours' good sun altogether. In order to extend our numbers whence we drew the average of bacteria present at the commencement, twelve samples were taken for plates to begin with, and we diluted to 1 in 10 as before. After three days' incubation we found four of the plates too far liquefied to count, but the numbers would probably not be incon- sistent with the following : — The eight plates gave a total of 10,890 colonies, and an average of 1361 colonies per c.c. in the water at the commencement of the experi- ment. After the exposure to light and darkness from 10.30 A.M. August 28 to 4 P.M. August 29, we found about 300 per c.c. in the flask ex- Report on the Bacteriology of Water, 353 L. 03 cS j§ .sjj .S T «tH § *3 0 O p5 O O •*H E o 1^ When examined. «W •2 a O O o Remarks. Weather. 0 §•-§ ^ -U j- O ij N a a Is W*1 O O H £° > & Aug. 28, 3.30 P.M. 77 49 1666 6 liquefying ; also 1 Aug. 30, 12 noon 12H • • 100 3400 mould appeared. Aug. 28, 4 P.M. 77i 38 1292 6 liquefying. Aug. 30, 12 noon 1222 64 2716 Aug. 28, 4 P.M. 77* 41 1394 5 liquefying ; also 1 Aug. 30, 11.30 A.M. 12H 79 2686 mould. Aug. 28, 4 P.M. 77i 52 1768 9 liquefying. Aug. 30, 10.30 119| 136 4624 Aiig. 29, 11.30 A.M. 9H 3 93 Clouds and bright sunshine. M 9H •• 1 31 (1) Aug. 28,"4.30 P.M. 72J 58 1740 5 liquefying. (2) Aug. 29, 11 A.M. 91 73 2190 6 liquefying ; also 1 mould. Aug. 29, 11.30 A.M. 9H •• 71 2130 5 liquefying. posed to the light, and numbers too high to count in that covered with foil and paper ; moreover, the latter plates were so badly lique- fying that their marked contrast to the former could not escape observation. It was clear that exposure to sunlight affects the liquefying powers of the forms in the Thames water, and this apart either from the difference in numbers on the plates or because it eliminates these forms more rapidly. §XV. On August 29 two flasks were prepared as before. F 11 was exposed to sun with mirrors, &c., and P 12 put by its side covered with foil and black paper. The first exposure was from 9.30 A.M. to 12.30 A.M. — three hours' good sunshine, though with occasional clouds. The water to start with was estimated to contain about 1200 germs per c.c. After three hours' exposure F 11 gave, as the result of two plates, 354 Profs. Percy Frankland and Marshall Ward. 99 and 132 ; total = 231 ; mean = 115 per c.c. After the same time F 12 (unexposed) gave 1600 and 1920; total, 3520; mean, 1760. The flasks were meanwhile put back and exposed yet another 4^ hours to the afternoon sunshine — i.e., from 12.30 to 5 P.M. — of which about 3-ijr hours counted as bright sunshine. Two plates from F 11 made at 5.30 gave 248 and 279 ; total, 527 ; mean, 268 colonies as the number per c.c. ; while two plates made at the same time from the unexposed flask (F 12) were so badly lique- fied that, although we estimated 3300 per c.c. from one of them, we regard the numbers as really higher. These flasks stood in the laboratory over-night at a temperature of 18° C., and were then exposed next day from 11.30 A.M. to 4.30 P.M.,. about one and a half or two hours of the five being sunny. Then, at 5 P.M., fresh plates were prepared. Two plates from the exposed flask gave 640 and 3200 per c.c. re- spectively ; total, 3840 ; mean, 1920 per c.c. The numbers are not very good, as there is such a great difference between the two plates. Two plates from the unexposed flask gave 12,800 and 16,000 per c.c. ; total, 28,800 ; mean, 14,400 per c.c., again bearing out the con- clusion that the sun has powerful action on the exposed water. § XVI. To my mind one of the most important discoveries elicited from these plate cultures of Thames water was the obvious reduction of liquefac- tion on the plates made from water exposed to light, and so struck was I with the differences between these plates and those made with the water not exposed that I made an independent investigation into the matter by selecting a set of the most pronounced liquefying forms from the water and examining their behaviour when exposed to light side by side with that of non-exposed samples. I started with the commonest and most pronounced liquefying form in the Thames water at the time. I refer to it throughout as Colony ft in my notes, and write it shortly /?. I had noticed the following facts concerning it during our experiments on the action of light on the Thames water as collected — it should be borne in mind that I had already studied its characters and knew the form pretty well. In the first place the plates as a whole liquefied very much more slowly and less completely than those made from unexposed water. Secondly, although it seemed at times as if this was because the form /3 had been eliminated from the water, I suspected that a certain other form, which liquefied less rapidly and developed much more slowly alto- gether, was really the above-mentioned form /3 with feebler charac- teristics. Report on the Bacteriology of Water. 355 On September 3 a small loopful of a gelatine culture of the bacillus ft was carefully shaken in about 50 c.c. of sterile water, and distributed equally in two Erlenmeyer flasks labelled ft (1) and ft (2). Flask ft (1) was wrapped in foil and black paper; ft (2) was ex- posed over mirror from 9.30 A.M. to 5 P.M., and about five hours of these seven and a half were good strong sunshine. Thermometers in control flasks went up to 34 — 35° C. as the highest temperature regis- tered in the afternoon. At 9.30 a plate was prepared from each flask, and gave something like 3,000,000 per c.c. as the average numbers to start with. After exposure, two plates were prepared from each. Of the two plates from the exposed flask, one plate — a 1-drop plate of a l/10th dilution — yielded no colonies at all ; the other gave 21,941 per c.c., pointing to a profound light-action. Of the two plates from the darkened flask, one gave 5,400,000, and the other 5,000,000 as the nearest estimate per c.c. The two flasks meanwhile stood over-night in the laboratory at 18° C. for about fourteen hours, and at 7 A.M. next day (September 4) 1 made two plates from the exposed flask and then again put them out as before. These two plates gave 42,000 and 59,220 respectively; total, 101,220 ; mean, 50,610 bacteria per c.c., a prefectly natural rise in the numbers having occurred during the night. No plates were taken from the other flask, as I had no particular need for the numbers — "known to be very high — and wished to reserve the counting for other plates. After exposure to the bright sunshine of September 4, from 8 to 4.30, say eight hours' sunshine on the exposed flask, two new plates were made from each flask. The two plates from exposed flask gave 3050 and 4200 ; total, 7250 ; mean, 3625 bacteria per c.c., again showing a marked reduc- tion in the sunlight, and the plates were singularly free of liquefying centres, but showed many colonies of the kind I had previously sus- pected as being the representatives of ft. Of the two plates from the unexposed flask, although made from one drop each of a l/20th dilution, the numbers were again so large and the liquefaction so rapid that no reliance can be placed on them, except that they prove that no essential diminution was to be traced to the action of the water or temperature in the absence of light, but only the natural fall in numbers always found when water stands for some time. The numbers actually calculated were 660,000 and 3,300,000 ; total, 3,960,000 ; mean, 1,980,000 per c.c. The flasks stood in the laboratory at 18° C. through the night, and I repeated the exposure on September 5, putting the uncovered flask into the bright sunlight of that day without a mirror. It received a 356 Prof's. Percy Frankland and Marshall Ward. Table M. — Experiments on Insolatu Flask. When filled. Exposed or not. Time of exposure. Hours of exposure. Hours of sun. Other treatment. Plate. 0(1) Sept. 3, at 9 A.M. Not •• •• •• •• P. ft (I) 0(2) » •• •• •• •• •• P. 0(2) 0(1) >' •• •• •• •• •• Pa. ft (1)1 0(1) „ .. .. .. .. Pb. /3 (1) 0(2) " Exposed 9.30 to 5 on Sep. 3. 71 5 Stood above and in front of mirror P1 0 (2) 0(2) 0(2) " 1 " •• " Ditto. " Then = 14 hours in P" $ (2) PU1 ft (3) dark at 18°C. 0(2) 0(2) » » 9.30 to 5 on Sep. 3 and 8 to 4.30 16 18 H H P* ft (2) PV ft (2) 0(2) 0(1) » Not Sep. 4 H •'•' 1 In dark whole PV ft (2) fPc ft(. \- time = 40 J 0(1) 1! ;; •• " -. J hours In dark 64 hours lPd j8 (1 Pe 0 (l) 18(2) " Exposed 9.30 to 5 on Sep. 3, 8 to 4.30 Sep. 4, 23 19 In dark at 183 each night pvii ft (2 10 to 5 Sep. 5. good six hours' sun direct, and the temperature rose as high as 37° C. at one time, but was for the most part at 30 — 32° C. All the other conditions were as before. After exposure, a plate from the insolated flask gave 665 per c.c. One from the covered flask gave 2,600,000 as the nearest estimate I could form. It is obvious, therefore, that the bacterium referred to as Colony ft is very sensitive to the solar action, and the results obtained with the above pure cultures are summarised in Table M. § XVII. The results with a second badly liquefying form, which I call Report on the Bacteriology of Water. of Bacillus /3 in Sterile Water. 357 When made. When examined. Time of incuba- tion. Tempera- ture of incubation. Dilution or not. Quantity used. No. of colonies counted on plate. Number of bacteria per o.c. 9.30 A.M., Sep. 8 days. 5 0°. 16—18 a o c.c. "So About 5,000 3,000,000 Sep. 3 .. 1^T> •So 5,844 actually 3,506,400 counted ' (P.M., To" sV 30 squares aver- 5,400,000 p. 3 aged 500 per sq. = 15,000 ,. 1 20 TS Very similar 5,000,000 numbers " •• •• •• To *v 0 . 0 •5T 593 21,941 L.M., Sep. 15 11 i 2(J 2T 100 42,000 p. 4 . 1 "2l> JL 141 59,220 . ' P.M., 0 _j_ 122 3,050 — p. 4 TO ~5Ti 12 4,200 1 a o -33 About 1,000 660,000 i 'aTT 1 3^J About 5,000 3,300,000 5 P.M., 10 . . i 'aTJ 2ff About 5,000 2,600,000 Sep. 5 TV •$v 1 665 Bacillus vj, are not contradictory of the foregoing — indeed they sup- port them so far as they go — but are less conclusive in detail. On September 5 two Erlenmeyer flasks charged with sterile dis- tilled water, to which a loopful of the bacillus in question was added, were placed out in the usual way. Flask labelled ij (1) was covered ; flask TJ (2) was exposed over and in front of plane mirrors. A plate from each flask at the beginning gave respectively 1,470,144 and 1,699,360 ; total, 3,169,504 ; mean, 1,584,752, as the number of bacteria per c.c. to start with. The flasks were out from 10 A.M. to 4.30 P.M., the day being beauti- fully bright with a hot sun and blue sky. The temperature in the covered flask rose to 33° C. in the afternoon, that in the exposed 358 Profs. Percy Franklaiid and Marshall Ward. one to about 34° C., as shown by controls with thermometers in the water. Two samples of the water of the un exposed flask, taken at 4.30 P.M. on September 5, gave 1,575,860 and 1,285,388 as the num- bers per c.c. ; total, 2,861,248; mean, 1,430,624 per c.c., suggesting that the sojourn in distilled water at that temperature, even in the dark, causes the death of large numbers of this bacillus. Of two samples taken at 4.30 from the flask t) (2), which had been exposed for six and a half hours, neither plate gave any sign of life after ten days' incubation, whence we may assume that neither sample contained a living germ. These two flasks were put in a cool cupboard over-night — temper- ature = 15° C. — and again put out on the 6th September from 9.30 to 4.30, so that the exposed flask — ij (2) — received another good six hours of bright sun, for the day was brilliantly fine again. After the exposure, a plate was made from each flask. That from T) (2), the one exposed to the sun, gave no signs of life though incu- bated for ten days ; the other showed 560,000 per c.c. It seems probable, therefore, that in the case of Bacillus // the im- mersion in sterile water at 33 — 34° C., even in the dark, is more or less fatal, for we see the bacteria are reduced from over a million and a half per c.c. to nearly half a million per c.c. At the same time it seems pretty clear that when exposed to light at the same time the mortality of the bacilli is much greater. The inference appears fair, bat there is naturally some dissatisfaction to be felt with these negative results. The following table N" summarises these facts : — Table N. — Experiments on Insolatic Flask. When filled. Exposed or not. Time of exposure. Hours of exposure. Hours of sun. Other treatment. Plate. id) Sept. 5, Not Id) 10 A.M. f (2) lt . . . . 1(2) 9(1) ti ,, . . Covered in the open all day ') (1) " . . . . . . ,, £ i] (1) 1(2) ,, Exposed 10 to 4.30 on Sept. 5 6i 6 Over and in front I 1(2) of plane mirror 1(2) » » ?, . . „ „ II 1 (2) 1(1) ,, Kot .. Stood over-night eu(l) at 15° 1(2) M Exposed 10 to 4.30 Sept. 5, 7 6 „ III rt (2 and 9.30 to 4.30 Sept. 6 * There were 32 squares, averaging about 1 Report on the Bacteriology of Water. 350 § XVIII. On August 20, two Erlenmeyer flasks, labelled F! and F2, were charged to a depth of about an inch with sterile-distilled water with which a loopf ul of spores of B. anthracis had been thoroughly shaken up. Flask FI was exposed with a mirror below and one behind ; F2 was wrapped in foil and black paper. The exposure was from 11.30 A.M. to 5 P.M., but it was a windy and cloudy day, with a good deal of rain. Just before exposure, two plates were made to determine the number of spores introduced per c.c. in the flask Ft. One plate gave 1,950,000, and the other 2,445,000 ; total, 4,395,000 ; mean, 2,197,500 per c.c. Two plates from F2 gave respectively 2,052,000 and 2,280,000; total, 4,332,000; mean, 2,166,000. Or, if we take the average of the four plates, we get total = 8,727,000; average of the four = 2,181,750; and it will be noticed how well the four plates agreed. At 5.30 P.M. two plates were made from the covered flask, and gave per c.c. 1,700,000 and 340,000 respectively. It was noted, however, that the second plate had been badly levelled, and the colonies were heaped up to one side, and could not be properly estimated. Taking the numbers as they stand, we get total = 2,040,000 ; mean = 1,020,000 per c.c., indicating some redaction, but still enormously high numbers present. Two plates from the exposed flask F2, after the five and a half hours' insolation, gave 595,000 and 1,295,000 ; total, 1,890,000 ; mean, 945,000. of Bacill as r in Sterile Water. When made. When examined. Time of incubation. Tempera- ture of incubation. Dilution or not. Quantity used. No. of colonies on plate. No of bacteria per c.c. Days. c.c. Sept. 5, Sept. 8 3 15—18° To" Tff 2976 1,470,144 10 A.M. » !> ,, » ,, 3440 1,699,360 Sept. 5, ,, ,, ,, >> 3190 1,575,860 4.30 » „ ,, „ ,, 2602 1,285,388 '« Sept. 15 10 » A j o 0 » j» 51 0 0 Sept. 6. Sept. 9 a „ o A 22,400* 560,000 4.30 D Sept. 16 9 » 0 i an 0 0 per square as near as could be counted. 360 Profs. Percy Frankland and Marshall Ward. This seemed to show that the exposure to what was, after all, only diffused light, had very little effect in five and a half hours. Both flasks were taken in at 5.30, and put on ice for the night at 6 P.M., and remained on ice till 11 A.M. on the 21st, i.e., seventeen hours in dark and on ice. They were then put out again from 11.15 A.M. to 4 P.M., the weather being much brighter, though plenty of white cumulus clouds kept sweeping over the sun. Before putting out, two plates were made from each flask at 11.30 A.M. on the 21st. Those from Vl (unexposed) gave 1,149,000 and 646,000 ; total, 1,795,000 ; mean, 897,500 ; numbers very similar to those of the previous day, and indicating that no essen- tial changes had occurred on the ice — possibly a few had succumbed to the rapid cooling. The two plates from the exposed flask F2 gave 42,000 and 1,200,000— the last number being too high, as there were numerous invading forms on the plate rendering it difficult to count. Taking the figures as they stand we have, total, 1,242,000 ; mean, 621,000, which is a reduction on last night's figures. After exposure on the 21st, two plates were again made from each flask, with the following results. Of the two plates from the exposed flask Fl5 one gave 122,500, and the other 50,750 as the maximum numbers per c.c. Total, 173,250; mean, 86,625 per c.c. While the unexposed flask gave 1,920,000 and 640. The last low number was obviously due to some blunder ; but, even if we take it to reduce the average, we get total, 1,920,640 ; mean, 960,320 per c.c. After exposure, the flasks were again put on ice at 9 P.M., and remained there till 12 noon next day, i.e., August 22 ; they had been at 16° C. in the interval from 4.30 P.M. to 9 P.M. On the 22nd a plate was taken from each flask at 2.30 to 3 P.M., and the F! gave no anthrax colonies at all, though nursed for nearly a week. The other flask gave 990 colonies, which comes to 297,000 per c.c. On the 23rd, after another seven and a half hours' exposure, plates were again made, one from each, and gave the following num- bers— the exposed flask 2 colonies, which = 720 per c.c.. and shows that all spores were not yet killed, and the unexposed one 20 colo- nies, which = 6800 per c.c. Unfortunately we were compelled to abandon these flasks now; the mere labour of counting within the necessary periods the numer- ous plates we were making made it imperative that this series should be discontinued. I am now particularly sorry this was so, because it would have been interesting to find if, and when, the light abso- lutely cleared the water of spores. However, we could not foresee what the tabular resume brings out so clearly. (Table 0.) Report on the Bacteriology of Water. 3(51 A point of great importance arises here — not for the first time, but very vividly. That is the gradual, and much slower, but, nevertheless, determined reduction of the spores, even in the dark flask. I am convinced that the principal factor in this is the changes in temperature undergone by the water, which was warmed up to 30° C., or thereabouts, during the day, and cooled to 4° or 5° C., or even lower, during its stay on ice. §XIX. The following experiment (Table P) gives an excellent example of how much can be done by the clear sun of a hot summer day in clearing the water of living spores of anthrax. A quantity of anthrax spores were carefully rubbed up in about 50 c.c. of sterile-distilled water, on August 16, and the infected water distributed into two Erlenmeyer flasks, marked Aa; and Ba:. Aa; was exposed to the sun, with a mirror belo• 0 P4 .. » F 1 Exposed Aug. 20, 5i P5 Aug. 20, Aug. 25. 11.30 A.M. 5.30 P.M. 11.30 A.M. — 5 P.M. ,, ,, ,, „ „ „ P6 ,, Aug. 25, 12 noon F2 „ !J Not 0 P7 „ „ ,, „ 0 P8 „ Aug. 25, 12.30 P.M. Fl „ J} Exposed Aug. 20, 51 P9 Aug. 21, Aug. 24, 11.30 A.M. 11.30 A.M. 12 noon — 5 P.M. P 10 Aug. 27, 11 A.M. F2 „ • Not 0 Pll )( Aug. 26, 8A.M. „ ,, ,, „ 0 P12 „ Aug. 25, 1 P.M. F 1 f> M Exposed Aug. 20, 10i P 13 Aug. 21, Aug. 27, 11.30 A.M.— 5.15 P.M. 11.30 A.M. 5 P.M. ; Aug. 21, 11.15 . A.M. — 4 P.M. „ ,, ,. „ M „ P14 „ Aug. 27, 12 noon F2 ,, ,, Not 0 P15 Aug. 21, Aug. 26, 5.30 P.M. 7 A.M. „ „ ,, „ 0 P 16 „ Aug. 28, 11 AM. Report on the Bacteriology of Water. Snores in Distilled Water. 303 33 lH O D Hours of int-ul tion. Temperature incubation. Number of colonies on plate. Number of bacteria per c Eemarks. Weather. 94 18—20° 5000 1,950,000 1 c.c. of water from the flask diluted to 10 c.c. with sterile distilled water, and the plate made from a drop of the mixture. 71 » 6000 2,340,000 » 71* „ 6000 2,160,000 >f n M 5700 2,052,000 » 114 " 5000 1,700,000 >' Cloud all the time, and a good deal of rain. 114* J) 1000 340,000 Badly levelled plate ; anthrax colonies all H 115 . 1700 3700 595,000 1,295,000 up at one side. 72* 3350 1,149,000 Two liquefying, twenty- five large anthrax, and possibly very many smaller, very numer- ous small colonies Flask had been placed on ice over- night, 6 P.M. to 11 A.M. round the edge. Di- luted 143* " 1900 646,000 Diluted. Nine moulds and four other foreign forms. " 116* » 140 42,000 Diluted. » 97* )) 4000 1,200,000 Diluted Numerous intruders. 138i » 353 122,500 Diluted. One mould, forty-two (?) anthrax Clear sunshine, with clouds occasionally. About three hours sun. 138| » 145 50,750 Total Diluted. 109* » 6000 1,920,000 Diluted. 161* » 2 640 Diluted This plate is so ab- normal that there was obviously some blunder. VOL. LVI. 2 c 364 Profs. Percy Frankland and Marshall Ward. Table 0- 4! £ £ • . o f^ B o ^ o q • ^2 o "QJ *T3 a g & a 53 a o & o C4H 11 IH o h 1 d o " o 0 £ 1 O ,0 o C o P X 1 1 1 ? g R H R N _ Fl Anthrax Aug. 20, Exposed Aug. 20, U* P21 Aug. 22, Aug. 29, spores in 11 A.M. 11.30 A.M. — 2.30 P.M. 12 noon distilled 5 P.M. ; Aug. water 21, 11.15 £ in. deep. AM. — 4 P.M. F 2 „ „ Not .. 0 P22 Aug. 22, Aug. 27, 3 P.M. 12.30 P.M. F 2 0 P36 Fl M ,} Exposed Aug. 20, m P35 Aug.' 23, Aug?29, 11.30 A.M.— 6.30 P.M. 12 noon 5 P.M. ; Aug. 21, 11.15 A.M. — 4 P.M.; Aug. 23, 9 A.M. — 4.30 P.M. « energetic than has been commonly supposed, and the results also suggest that some spores die off very quickly, even in the dark, when put into sterile water — a fact long known. On August 18, we started a similar experiment to the last, using Thames water, freshly collected, instead of sterilised distilled water ; but this had to be abandoned owing to the difficulties with the rapidly-developing liquefying forms at the temperatures necessary for growing the anthrax. There was nothing in the results to con- tradict previous experience, but the details are of little value. §XX. On October 6 I exposed a tube of broth, infected with a loopful of colonies ft — from gelatine stab-culture — from 9 A.M. to 4 P.M. (seven hours), over a mirror, to the sun ; the sky was clear, and sun bright till about 1 P.M., and then duller. An exactly similar tube was wrapped in foil and black paper, and placed by the side of the above. At 4 P.M. a plate was made from each tube, and incubated at 15° C. ; a stab-culture from each was also made and kept at 15° C., and the original tubes were placed at 20 — 22° C. Taking the plates first. In forty hours the plate from unexposed tube showed numerous colonies, from 0'5 mm. to 1 mm. diameter, Report on the Bacteriology of Water. 365 jntinued. i IM O d 3 * •— , Q} O » ° 1^ t- sl IS. £ g a 3 S.2 o .5 -3 P 1-1 Pi 5MDO pq pq X X o" -2 . -i * a ' CO g 3 O d fl Pi Report on the Bacteriology of Water. 367 clearly that the individual colonies are weakened by the light-action, and the results with the original tubes (given below) may also be consulted. Turning now to the original broth cultures, placed at 20 — 22° C. In eighteen hours the darkened tube was distinctly turbid, whereas that exposed to the sun was perfectly clear ; and the same was the case at the end of forty hours. In sixty-four hours both tubes were turbid, but the exposed one far less so than the other. Later on it was impossible to distinguish between them. It seems to me impossible to avoid the conclusion that the light- action so weakens the metabolism of the insolated cells that they grow and divide more slowly, and dissolve the gelatine more feebly, and possibly this weakening effect is transmitted to the cells to which they give rise by division. As time passes, however, the cells gradually recover their vigour in the dark, and where plenty of food material is accessible. That this latter statement is true the following experiment proves : — On October 18 I took the two broth-tubes of Colony /3, referred to above, both of which had been in the dark, side by side, since October 6. Broth-tubes were exactly similar at the beginning, as we have seen, but one of them, had been exposed, on October 6, for seven hours to the sunlight. On making stab-cultures from these tubes, no difference could be detected between their behaviour, both began to liquefy the gelatine in forty-eight hours in the typical, thistle-head funnel form. This seems to show clearly, also, that the broth has not been in- jured as a medium for culture by its exposure to light. In order to meet the objection that the above results were due to the using of broth-cultures, I repeated them as follows on October 10 :— Two tubes of sterile water, infected from the turbid broth-culture kept in the dark since October 6, were suspended, as already de- scribed— one exposed over a mirror to the bright direct sunlight from 10 A.M. to 2 P.M., the other wrapped up, and placed beside it. At 2 P.M. plates and stab-cultures were made from each tube. The results were the same as before. The plates from the unex- posed tubes showed numerous liquefying colonies in forty-eight hours, and were completely liquefied on the third day ; the plates from the insolated tube showed only very few colonies, and these not till the third day, and they were denser, more granular, and without any signs of liquefaction for several days. To take a concrete case : a plate made with 1 drop (1/30 c.c.) from the dark tube showed about 5000 colonies in forty-eight hours, and these were already beginning to run ; the whole of the gelatine was liquefied in another twenty- four hours; the corresponding plate made with 1 drop (1/24 c.c.) 3158 Profs. Percy Frankland aud Marshall Ward. from the insolated tube showed three colonies only on the third day, and these were very small, more densely granular, and irregular in outline than the normal colonies, and showed no traces of lique- faction even on the fourth day. On the fifth day the liquefaction was beginning, but even after fourteen days one of the three colo- nies was only just breaking up. The stab-cultures gave similar results. Those from the darkened tube showed a distinct thistle-head funnel of liquefaction in three days, whereas the feeblest signs that infection had really occurred were all I could get in the same time from the insolated cultures. Here, again, however, the two sets of tubes gradually become alike, evidently because the at first enfeebled cells gradually re- gain their vigour, and once more rapidly peptonise the medium. The experiments with water were repeated on October 12, all the arrangements being as before. The exposed tube was out from 9 A.M. to 3 P.M. in brilliant sun- shine the whole time practically. After three hours' exposure, a plate was made. That from the dark tube showed colonies, visible only under the microscope, in nineteen hours, and in two days the whole of the gelatine was liquid like water. The plate from the lighted tube showed no signs until the third day ; on the fourth day six colonies had made their appearance, but these only began to soften the gelatine around some days later, and even on the twelfth day two of the six colonies were still circular depressions, though the other four had liquefied the gelatine some distance round. After six hours' exposure, further plates were made from the above tubes on October 12. As before, colonies were visible with the lens on the plate made from the dark tube in sixteen hours, and before the end of the second day the whole of the gelatine was liquefied like water. On the plate from the light tube five slowly-developing colonies had appeared by the fourth day, one of which showed feeble signs of liquefaction next day. But even after twelve days only three of these colonies were vigorously liquefyiug the gelatine, the other two being still compact and circular, though one of them lay in a slight depression. Stab-cultures were also made at the end of the six hours' exposure on October 12, with the results as before. In the case of the culture from the dark tube, the funnel of liquefaction had reached the walls of the tube, and liquefied one-eighth of an inch of gelatine in four days, whereas that from the lighted tube, in the same period and side by side, had not even begun to liquefy the gelatine, though the infection had taken. Report on the Bacteriology of Water. 369 I also, at the end of the experiment on October 12, cautiously emptied each of the water tubes, and drained it until only a drop remained, and then filled up with sterile broth. After sixteen hours the dark tube was distinctly turbid, whereas no trace of turbidity appeared in the insolated one till after forty-eight hours. Both were equally turbid on the fourth day. Of course I recognise that this last result, and that obtained with the stab-cultures, is attributable to the numbers of still living germs added to the gelatine and broth respectively, and the experiments only go to prove once more, but in a very decisive manner, what mortality the sun had occasioned in the exposed tube — for we must remember, each tube contained practically the same enormous num- bers at the start. §XXI. On October 6 a loopful of a markedly liquefying form, marked in my notes as colony 0, was placed in each of two tubes of broth, and placed in the sun from 9 A.M. to 4 P.M. One tube was exposed over a mirror, the other side by side, but wrapped in foil and black paper. The sun shone brightly from 9 to 1, and then was obscured by clouds. At 4 P.M. a plate and a stab-culture from each tube were made, and the tubes put at 20° to 22° C. in the dark incubator. The plate and stab-cultures were put at 15° C. in the dark. Taking the broth tubes first. Both were already turbid in eighteen hours, more densely so after forty hours, so that no difference between them could be detected. Of the stab-cultures, that from the dark tube had taken in forty hours, while the one from the lighted tube showed no signs. In sixty-four hours the non-illuminated culture had developed a thistle-head liquefaction funnel, whereas no trace was visible in the other. In eighty-eight hours the liquefaction had proceeded rapidly in the former tube ; only one feeble colony was visible in the tube from the insolated broth. In the course of a week or so r o further difference could be made out. With regard to the plates. That from the insolated tube showed no trace until sixty-four hours had passed, when two or three colonies •g- to f mm. diameter were seen. In eighty-six hours these were somewhat like young anthrax colonies, each with a slight depression. On the sixth day liquefaction of the gelatine was slowly evincing itself. The plate from the dark tube showed evident colonies, f to ^, and even 1 mm. in diameter, in forty hours ; and in sixty-four hours they averaged 2 to 5 mm. in diameter, and were liquefying rapidly. These circular colonies were less rapid than those of colony /3 at tho 370 Profs. Percy Frankland and Marshall Ward. same time, however, and whiter in colour. In eighty-eight hours the whole of the gelatine was completely liquefied. § XXII. On October 30 about 200 c.c. of sterilised Thames water were strongly infected with the bacillus called B. arborescens by Frank- land, taken from a vigorous culture. The infected liquid was divided equally into two Erlenmeyer flasks, one of which was at once wrapped in tin-foil and black paper. The other was supported above a plane mirror, and placed so as to obtain the maximum amount of direct sunshine available from 10 A.M. to 5 P.M. that day ; from 10 to 1 the sunshine was hot and bright, and a control flask showed that the water rose to nearly 20° C., but the afternoon was dull and cloudy, and the temperature fell to 12° C. The temperature in the covered flask, placed side by side with the exposed one, rose and fell so nearly exactly with that of the latter, that no stress can be laid on the difference. Before commencing the experiment at 10 A.M., sample plates were made of the infected water ; and at 5 P.M. two plates were made from each flask — exposed and unexposed. The plates from the freshly infected material showed the usual rapidly growing, loose, thread-like colonies in forty-eight hours, and in three days the gelatine was entirely liquefied to a watery fluid. The plates from the flask, wrapped in foil and paper, and sheltered from the direct rays of the sun, behaved similarly, the colonies being a trifle more compact in shape, but equally rapid in liquefying the gelatine completely. But the plates from the exposed flask differed from the first onwards from those not insolated. Thus no colonies were visible in forty-eight hours, a period during which the plates from the un- exposed flasks exhibited numerous typical colonies ; the colonies appeared here twenty-four hours later, clearly showing the effects of inhibition due to the light. Secondly, when the colonies did make their appearance they were more compact, and instead of shooting out in all directions and covering the plate with a meshwork of fine branches, rapidly lique- fying the gelatine, as in the case of the nnexposed specimens, the mode of growth was so affected that on the fourth day they had developed into beautifully circular yellowish colonies, zoned, and radially striated, and only just softening the gelatine. It was not, indeed, till the sixth day that liquefaction set in generally. I explain the differences as follows. The light retards the growth of the living bacilli, owing to some action on their protoplasm which induces interference with the metabolic processes on which growth Iieporl on tJte Bacteriology of Water. 371 depends ; this causes the cell-chains to be so modified in length, direction, and rapidity of development, that the colony formed from the iusolated germ is weaker and more condensed, or compact, than normally. Thus result the differences in the naked eye characters of the colonies, which may go so far that the total aspect on the gelatine plate is altered. Some clue to the action may perhaps be got eventually by following up the fact that one consequence of the light action is to weaken the enzyme action of the bacterium — for the enfeebled liquefying power is an expression of enfeebled enzyme power — either by so altering the protoplasmic machinery that less enzyme is secreted, or by so acting on the enzyme that its power of converting the medium is altered. The lessened enzyme power of course implies less power to obtain its food from the medium, and so the progeny developed from the germ started with are also feebler than the normal one. In the cases quoted, however, the colony gradually becomes more normal as regards its enzyme power, especially in the dark, because the successively developed new cells become stronger and stronger as they are fed by the nutrient gelatine, and at last the differences are equalised. The only source of error in the above conclusion that I could think of, was the possibility that the rapid running of the colonies into thin filaments in the first case is facilitated by the quicker liquefac- tion of the gelatine, and that this liquefaction is, in turn, more rapid, because there are so many more colonies per drop in the unexposed flasks, because the majority of the bacilli in the exposed flasks are killed. I accordingly made plates in which I diluted the samples from the unexposed flasks five, ten, and even twenty times as much as the samples from the exposed flasks, and so brought the numbers of colonies on each plate approximately equal. Of course it may be replied here that the differences of dilution may bring about diffe- rences in development; but experience shows that increased dilution tends to inhibition of colonies, and so I think the fact that I still get the differences in the colonies already described, strengthens rather than weakens the evidence that the alteration in the character of the colonies from exposed flasks is really due to the action of the light. On November 7, which turned out a beautifully fine day, with clear blue sky and bright sun, tubes of sterile distilled water were infected with cultures of a yellow bacillus marked *, a large -white one marked <•/, and a violet bacillus, common in the Thames. In each case, two similar tubes were prepared, exactly alike, one of which was exposed over a mirror, from 10.30 A.M. to 3.30 P.M. ; while the other was wrapped in foil and black paper, and placed by the side of the ex- posed one. The temperature recorded in control tubes was 10 — 12° C. 372 Profs. Percy Fraukland and Marshall Ward. At 4. P.M., after five hours' exposure, plates were made, and incubated at 15° C. Taking the violet bacillus first. Nothing appeared on either of the two plates made from the exposed tubes, although they were kept till November 24, i.e., seventeen days. On the plate made from the dark tube, two white colonies were visible in forty-eight hours, and on the 13th — i.e., after six days — three colonies were seen. These remained white until December 1, when one of them began to show the violet line. On December 6 this was more pronounced. The experiment, therefore, showed negative results only, and I regard it as probable that the mere immersion in water injures the bacillus. On the other hand, it is possible I did not use sufficient material in making the plates. Thus the plates of exposed tube = l/28th c.c. of 7/28 dilution, and that of dark one = l/27th c.c. of 1/27 dilution, thus giving only 2187 per c.c. even in the unexposed tube, making it probable that the last suggestion is the right one. Now, as regards the yellow bacillus, Colony *. Of the two plates made from the insolated tube, one was contaminated, and yielded no results, except that plenty of colonies appeared ; the other failed utterly. The plate from the dark tube showed innumerable typical colonies, and was completely liquefied on the fourth day. The experiments consequently must be regarded as negative. The plates of Colony <•) behaved as follows. That from the dark tube gave a typical series of colonies — about 500 = 500x28x28 = 392,000 per c.c., softening the gelatine in two days. That from the lighted tube gave far fewer— about 100 X 25 X 5 = 12,500 per c.c. — colonies, and these smaller, showing evident retard- ation ; otherwise no results. On November 12 the above experiment was repeated with Colony *, Colony 7/, and the rosy-red bacillus £, exactly as before. The day was cold and bright, though some haze appeared after 1 P.M. Ex- posure 10.30 to 3.30 as before, and thermometers = 10 — 12° C. The plates of Colony * behaved as follows. That from the dark tube was liquefied by numerous colonies, which appeared on the second day. On the fourth day all the gelatine was liquefied. The plates (two) made from the insolated tube only showed slight retardation, and liquefied also on the fourth day, though more slowly. No difference between illuminated and dark tubes could be made out in the case of Colony %, except slight retardation on the plates from the former. The results with Colony S were still more indecisive, and I could not draw any conclusions as to distinct light action. It appears probable that considerable differences will be found between the various forms in this respect. Report on the Bacteriolugy of Water. 373 § XXIII. In the following tables I give the determinations of the numbers of all bacteria found in three series of analyses of samples of Thames water in August, October, and December, 1893. As will be seen, they confirm the results of other observers, in so far that they show that the number per c.c. is considerably greater in winter than in summer. There is nothing special to note as regards the methods employed, which were those ordinarily in use, beyond the fact that we counted every colony on each plate, instead of estimating the averages from a few squares. The greatest care was taken to have the gelatine made up exactly alike in every case ; further information as to details is supplied by the tables. 374 Profs. Percy Frankland and Marshall Ward. CO O! 00 & fcc p •4 m* £ g eg »E EH « «*-< o ^H O rC O* -S E- 0, 6 g c fi .s rfs H 6 j ;-': pe of ub O Report on the Bacteriology of Water. 375 S o o o o o o CD 00 C g -t 6 i— i i— 1 i— 1 yj *o « PH ill ° o lo *s p s~i cr' Bj •e 1 -2 g -H X . r-i |1 i-i c ID 1 . X 7 . cc O O bO o •§ 1-1 o "S S OOCSOi-JNW-^llO t>t-xxxxxx 1 1 p o CO Report on the Bacteriology of Water. 4000 3000 2000 1000 /at L PI. I a-k a 15-18 C. -/5V, cr a- ft /s-/o°c a,t e-/5 .A' C1 -15 C. 48 7?. 96 DIAGRAM I.— Curves of Table Q, showing total bacteria at 12—15° C. and 15 — 18° C. in August. Abscissae = hours of incubation of plates; ordinates = numbers of colonies per 1 c.c. Trots. Percy Frauklancl and Marshall Ward. 6000 5000 4000 3000 2000 1000 _z 56 '48 58 46 84 81 63 85 76 60 79 O 24 48 72 96 120 144 DIAGRAM II. — Curves of Table Q, showing total bacteria at 16 — 18° C. in August. Abscissse = hours of incubation of plates; ordinates = numbers of colonies of bacteria per 1 c.c. Report on the Bacteriology of Water. 379 Quantitative Analysis of Thames Water in August, 1893 (Table Q). If we analyse these August numbers, we get as the average of 14 two-days' plates, 1157; of 24 three-days' plates, we have 1530; of 15 four-days' plates, we have 2208 ; of 8 five-days' plates, we get 3575 ; and of 8 six-days' plates, we have 3037, as the approximate averages. Of course these numbers can only be regarded as approxi- mations, but I think they are of use as indicating what might be looked for if the necessary large series of observations could be made over several years ; and they certainly serve to put us on our guard against placing too much confidence in the gelatine method unless a great number of observations are made, extending over a long period. Of course the great difficulty to be contended against is that of keeping the plates long enough ; if only one badly liquefying form is present, it ruins the plates before the slowly developing ones germi- nate out. Quantitative Analysis of Thames Water in October, 1893 (Table U). The following table gives the details of the analysis for October, and again I have recorded all the points. The chief feature 01 interest is the much longer time during which the plates could be cultivated, in spite of the prevalence of liquefying forms, where care was taken to dilute sufficiently. VOL. LVI. 2 D 380 Profs. Percy Franldand and Marshall Ward. > CO (M _ r of c>f cxf rHINIMi-|T}i-^iOOiMCOTi(C «4 Report on the Bacteriology of Water. 381 IN 00 O O w 10 co" oo" 06" od ^^OlO5>OTfcON-*-^ Z~T 5 •-"ha ^ £ r«ln " S Z- 5 t> 144 .. 168 DIAGRAM III. — Curves of Table E, showing total bacteria at 16 — 17° and 20 — 22° C. in October. Abscissae = hours of incubation ; ordinates = numbers of colonies per 1 c.c. 384 Profs. Percy Frankland and Marshall Ward. Quantitative Analysis of Thames Water in December. Oil December 27, 1893, a cloudy and misty cold morning following on a series of bright sunny cold days, a sample of Thames water was taken at 11 A.M., and immediately conveyed to the laboratory, and plates made as follows. The water on collection was at nearly '8° C. Twenty-two plates in all were made, of which twelve were in- cubated at 8 — 12° C., the temperature of the laboratory, and ten at 18 — 20° C. in the incubator. There were a good many changes of temperature from day to day in the case of the former, though but slight and slow alterations in the latter case. The following tables give the results of the plate cultures, every colony visible with the hand lens being counted each day as before ; and this time the low temperature cultures were continued for 240 hours (ten days), in order to see how far the numbers would eventually approximate to those of the higher temperature plates. An innovation was made in the case of the plates marked R^I^, S^, TiTa. Here I used 1 drop of the Thames water between each pair of plates. The drop was allowed to fall as usual into the tube of gelatine, then a second tube of sterile gelatine emptied into the infected tube ; then the contents were distributed rapidly into two plates, the countings of which must be taken together as is done in the table. It is noteworthy how much higher the final numbers are in these plates as compared with others, no doubt chiefly owing to the more complete separation of the bacteria from which the colonies arise, and to their having more space to develope in. Report on the Bacteriology of Water. 385 a £ 0 S ^ 10 X r-1 is 3 "o H * • • 1 1 § CO Ci <1 05 CO O 10 O ••? C7JOOO O Ot^'H-H T« CO (N CO U3 OJ CO •* -^ r-T ecember, 18 Oi O ^ £) llr ,-. OOOCOOOOOOOOONi-lOO ^OOOOOOO O O QOcsoo- •o a-S ^ ° rS 6 ft i—iOOOOOOO O O OGMCOOO t* CO CO CO ^f< 1> O5 (N N IN (N O4 ^ tO e3 b . OB 1 1 J^JV* ' ^rti^ 0$ ° *J O o ^ o 2 P p q a rH §1 ?U 08 *i HJ .£ J " o _g * «w O h 'I Temperature of incubation. Q 7 00 t h 1 o O rH 3 oo . „ ^o H "o • M as O -40 S e f« « v^-^ .«^ S g P3 M OQ'QQ' s «s <» "^-^ -^ § 8 PH W 02" a* e ^o - do - " eq C5 CD oo ••*< ONOOOOOO ^H of of of of of eo"af IH" -^r ' § g-s OOOOCOi-llNcOIN , s : II Bl o cq E cl Age 8- ) S S Report on the Bacteriology of Water. 387 r c i c e — •* ( 0 O VI OS o" CD" 3 O i oo os oo eo eo in t> co rH O O O I-H iH r-i r-l oo ,-^-, oo 1— 1 »n in • r-l r^ OOOO OS N OOOC O CO O 30 •*? t— 1 (NCOCI :> t- o o o o r-l CO *>COOOOOOO oo oo csojooootN O O M CO O OS in co m r-i CO •* •* CO "^ "^ "^ CO £*• CD Tji CD u 5 m oo co CD" CD t O r-l 00 M O5~rH i— 1 X r-l r-l r-l r-l r-l OS 00 1 i>" o" rH " O J^* CO ' OX>iDCO(MCOiOlO t^ COCD^Oii iH rH Ci| (N iH C^ Oi CO^OOOiOGOCDGO O ^f(N^r-lr- i rH^-i^O rp K>i3i 388 Profs. Percy Frankland and Marshall Ward. i O Tf< rH 1O cc^ o^ rH rH 3 i *> rH m O rH 00 O rH O O5 W ^> r-T 00*" Tj? 0>f rH rH rH -COOt>. rH CO ^? W3 GO O5 OO *O Od OO ^* o" i-T r-T cT 06" co" co" t-T -^ cf ^^-,^^-,^-^^^^ , — » — , ^^-,^-^-^^^-,r-^^ , — » — , g» t< f-t O •3 ® .2 OOO1OOOOOOO O Osl kOXrHNOOOO i> •* TfiQOlMOOOCIMIMQO IO rH (MIN^COOOr- 1 ^1 00 00 ^ M| Of ^ O" -^ Oi «J CO" rH N 0 fl * r^^V-^-l ® if] dOiOOOC^COCO^ O CO *QO1 t>* (N UD 1C CO (M rH M rH O5 (^ "^ O CD CO t^ lO iO W CO GO W O CD Oi JO CO ^ O H ^^^^^^-^^^^ 1 ' « •£ -•= H3 fl ft fl C A ol £ lO «D 00 ••# »OCDOOi-l O (N »O OOr(• O O~ eo" . rjT % r~ — ^ rH CO CO CD" rH w° o (MCOOIOCOCO^OOCOCOOOO CO X.^* C^ CO ^* O O CO 00 ^ O iO -t*» O5 CO O Q IO ^ CO ^T* IN T? in CO rH E B Oicoo5-*coiNOeococooqocDO5 CD O 00 O -5 ^^^-.^-«r^-, r-^^-,^^^^ ^-.^^ ^^2 0> O g. O lj|« OOJ>OOO>OOOO CO CD -<*IOOOO CO OOOOO OOOOO CO O CC ift O • 1C Tf< (N t^ u? Ills os o"co t>~o oT-*-* co IN ococoiooo"co cTos •* 10 •* O IN T?< 0 •* o i ^ 3 c -2 £ g ^ O &* 0 (NCOOOOSOOt-l>-OrH O C^i— ' CT!N Oi O5 (N CC O CT^CO CO O IN Tf COCO (NC^CO'2'*:3'-I fi H oo rH h Is is o o rH 1 o §D 1 o « ^5 ?35i o si, E-TEn''*1^ N* o a, CD § 00 rH ft 6 q Report on the Bacteriology of Water. 391 In the following plates I have plotted the curves obtained from the means of the foregoing illustrative series of analyses. They bring out very clearly the differences in numbers referred to, and I am strongly of opinion that much valuable information would result from a systematic series of monthly analyses of the Thames water conducted along these lines. I by no means pretend that the numbers themselves are of abso- lute value, any more than are those obtained by the ordinary methods of counting averages ; but I do think that the selected cases suggest possible lines of departure for the systematic bacteriological analysis of such a river as the Thames, if a sufficient number of other data were taken in at the same time. These data should include at least the following : (1) the temperature of the river, (2) the amount of sunshine, (3) the organic analysis of the water, (4) the rainfall. I am perfectly alive to the incompleteness of the above analyses in these respects, and they are only intended to show what I think should be done by a competent staff of assistants, if any attempt is made at a thorough investigation of the bacteriology of the Thames' and the same applies to any other water. The particular object of the above analyses was to test the view that the actual number of bacteria present in winter is less than that in summer, and they strongly confirm that ; and if we remember that this was so in 1893 in spite of (1) the river being lower in August, and therefore more concentrated as a food liquid, (2) the temperature being higher, and therefore more favourable to bacterial growth, it seems at least highly probable that the diminution in the bacteria is largely due to the increased insolation. Nor is this all (though I defer the fuller consideration of this point) that my analyses suggest. I find very distinct evidence that the bacteria in the summer water are many of them enfeebled forms, suggesting a distinct inhibition or weakening of their powers of growth. In some cases it is certain that forms obtained in August, and which afterwards turned out to be identical with forms found in the winter, at first grew so feebly that their characters on the plates led one to put them down as distinct species or varieties. I have given some experimental evidence bearing on this, and going to prove that it is due to the action of light on these forms. The matter is a very complex one, and I must refrain from further discussion of it until all the forms isolated during the year are worked out ; but it is worth while, I think, to draw the attention of investi- gators to the matter. In one or two cases, at least, there is no question that exposure to light does so affect the germination and growth of the bacteria, that the resulting colonies depart widely from the normal in many of their characters. I have in hand a large number of experimental results obtained 392 Profs. Percy Frankland and Marshall Ward. from the detailed investigation of a single species and the measure- ment of the growth of a single filament (watched continuously under the microscope) under different conditions of exposure ; they have been incidentally referred to in my lecture at the Royal Institution, in May, 1894, and I shall hope to hring them before the Royal Society shortly, They prove still more conclusively at least the main point, that exposure to sunlight does materially affect the rate and manner of growth, &c., of a bacterium rodlet or filament, as well as the germinal power of the spores. As regards further details on the action of light on the spores and bacilli, I may refer the Committee to my memoir, now in the hands of the Royal Society, an abstract of which appeared in the ' Pro- ceedings,' vol. 54, p. 472. 24 48 72 96 120 144 168 192 216 29 30 3\Jan.\ 2 3 4- 50 DIAGRAM IV.— Curves of Table S, showing total number of bacteria at 10—12° and 18 — 20° C. in December. Abscissae — hours of incubation ; ordinates = numbers of bacteria per 1 c.c. 18000 17000 16000 15000 14000 13000 12000 1 1 000 10000 9000 6000 7000 6000 5000 4000 5000 2000 1000 1053 34; '595 8127 7142 11480 6250 '4873 3315 14132 11542 •&/ Si 5262 7340 930 17515 14730 12285 6453 5917 DIAGRAM V. — Means of the foregoing curves, showing average numbers of bacteria for August, October, and December. Report on the Bacteriology of Water. 395 PART II. " The Behaviour of the Typhoid Bacillus and of the Bacillus Coli Communis in Potable Water." By PERCY FRANKLAND, Ph.D., B.Sc. (Lond.), F.R.S., Professor of Chemistry in Mason College, Birmingham, assisted by J. R. APPLEYARD, F.C.S. It has already been pointed out in the previous Reports that the only two zymotic diseases which have heen conclusively proved to be communicable, and frequently communicated by drinking water, are Asiatic cholera and typhoid fever. The behaviour in water of the particular micro-organisms which are almost universally credited with the power of exciting these specific maladies is obviously, there- fore, one of the most interesting and important questions in the whole domain of the hygiene of water supply. Inasmuch as Asiatic cholera is, fortunately, only an occasional visitor of these islands, or, indeed, of the continent of Europe, the investigation of this question with regard to this disease is certainly of less immediate consequence than is its investigation with regard to typhoid fever, which we have always with us, and which claims such a large number of victims annually from amongst our popula- tion. In the present Report I have, therefore, endeavoured on the one hand to summarise briefly what has already been done by others towards the elucidation of this subject of the behaviour of the typhoid bacillus in water, whilst on the other hand I have recorded those experiments which I have myself conducted with a view to extending the knowledge of this matter in general, and in .respect to the conditions of water-supply pertaining to this country in par- ticular. Our information concerning the behaviour of the typhoid bacillus in water is of essentially two different kinds ; firstly, this bacillus has on a number of occasions been discovered with more or less certainty in waters which were actually being used for domestic pur- poses, and to which it had therefore gained access unintentionally and in the natural course of events ; secondly, the bacillus has been purposely introduced into various waters in which its subsequent fate has then been traced by experimental observation. It will be con- venient to consider these two different kinds of information separately. VOL. LVI. 2 E 396 Profs. Percy Frankland and Marshall Ward. 1. Discovery of the Typhoid Bacillus in Natural Waters. Inasmuch as the communicability of typhoid fever by drinking water has been long recognised as a cardinal principle of modern hygiene, it was only natural that the discovery of the specific micro-organism of this disease by Eberth should have been soon followed by strenuous efforts to discover the Eberth-bacillus in potable water ; it was not until about six years afterwards, however, that successful attempts in this direction were announced. The first investigator who claimed to have discovered Eberth- Gaffky's bacillus in water was Moers (" Die Brunnen der Stadt Miihl- lieim a. Khein vom bakteriologischem Standpunkte aus betrachtet," ' Erganzungsh. zum Centralblatt f. allgem. Gesundheitspflege,' vol. ii, 1886, p. 144), who isolated the bacillus from a contaminated well supplying drinking water to a number of people amongst whom many cases of typhoid fever had occurred. This discovery was soon followed by a similar announcement from Michael ("Typhusbacillen im Trinkwasser," ' Fortschritte d. Medicin,' vol. iv, 1886, p. 353), in Dresden, who claimed to have isolated the bacillus from a well-water which was suspected of being the source of an outbreak of typhoid which declared itself at the end of the year 1885. Dreyfus-Brisac and Widal (" Epidemic de Famille de Fievre typhoide," 'Gaz. hebdom.,' 1886, No. 45) again detected the bacilli in the polluted water of a well at Menilmontant, where typhoid fever had been pievalent for some months. The typhoid bacillus has repeatedly been found in the water of the river Seine, thus Chantemesse and Widal ('Gaz. hebdom. de Med. et de Chirurg.,' 1887, pp. 146—150 ; * Centralbl. f. Bakteriol.,' vol. i, 1887, p. 682) discovered the bacillus no less than three times in this water during an outbreak of typhoid in Paris. Thoinot (' La Semaine Medical e,' 1887, No. 14, p. 135 ; 'Centralbl. f. Bakteriol.,' vol. ii, 1887, p. 39) also isolated typhoid bacilli from the Seine at Ivry, at a dis- tance of but little more than twenty yards from the point where the water is abstracted for the occasional supply of Paris with drinking water. Again, Loir ("Recherche du Bacille typhique dans les Eaux d' Alimentation de la Ville de Paris," 'Annales de 1'lnstitut Pasteur,' vol. i, p. i88) detected the typhoid bacillus in Seine water which was actually being distributed to a portion of Paris during the summer of 1887, owing to the scarcity of the Vanne water, which yields the supply under ordinary circumstances. Vincent ("Presence du Bacille typhique dans 1'Eau de Seine pendant le Mois de Juillet, 1890," ' Annales de 1'lnstitut Pasteur,' vol. iv, p. 772) again found the typhoid bacillus in the Seine water which was being similarly sup- plied to Paris during th,e summer of 1890. Report on the Bacteriology of Water. 3') 7 Beumer (" Zur Aetiologie d. Typhus abdominalis," ' Deutsche medicinische Wochenschrift,' 1887, No. 28) was able to detect the typhoid bacillus in a well-water used for drinking purposes in a country place near Greifswald, where an outbreak of typhoid had occurred. A similar discovery of the bacillus was made by Brouardel and Chauternesse ("Enquete sur les Causes de 1'Epidemie de Fievre typhoide qui a regne a Clermont-Ferrand," ' Annales d'Hygiene publique et de Medecine legale,' vol. xvii, 1887, pp. 385 — 403 ; also ' Revue d'Hygiene,' vol. ix, p. 368) in the course of an investiga- tion of an epidemic of typhoid which prevailed at Clermont-Ferrand, and in which neighbouring places using the same water supply were also involved. Finkelnburg (" Ueber einen Befund von Typhusbacillen im Brun- nenwasser," ' Centraibl. f. Bakteriol.,' vol. ix, p. 301) states that he isolated the typhoid bacillus from a well which had in all probability been contaminated with typhoid dejecta. Henrijean (" Contribution a 1'Etude du Bole etiologique de 1'Eau potable dans les Epidemics de Typhus," 'Annales de Micro- graphic,' vol. ii, p. 401) found typhoid bacilli in the drinking water of a Belgian village during an epidemic of typhoid fever. Kamen (" Zum Nachweise der Typhusbacillen im Trinkwasser," ' Centralbl. f . Bakteriol.,' vol. xi, p. 32) detected typhoid bacilli in water supplying a Russian military garrison, amongst whom typhoid fever had broken out. Pere ("Contribution a 1'Etude des Earix d'Alger," 'Annales de I'lnstitut Pasteur,' vol. v, p. 79) states that he was able to isolate the typhoid bacillus from drinking water in Algiers, where typhoid is endemic, occurring every year during the months of August, Septem- ber, and October. Martin states (" Presence du Bacille typhique dans les Eaux d' Alimentation de la Ville de Bordeaux," ' Revue sanit. de la Pro- vince,' 1891, No. 181, p. 93; 'Centralbl. f. Bakteriol, vol. xi, 1892, p. 413) that the typhoid bacillus was found by Ponchet in the public water supply of Bordeaux during an outbreak of typhoid in that city. Fodor (" Die Beziehungen des Typhus zum Trinkwasser," ' Cen- tralbl. f. Bakteriol.,' vol. xi, 1892, p. 121), in a paper read at the International Hygienic Congress held in London in 1891, describes an outbreak of typhoid fever at Budapest, during which he succeeded in detecting the typhoid bacillus no less than five times in the public water-supply. It was afterwards ascertained that the waste water from a laundry attached to the hospital gained direct access to the principal water main in the town. Kowalski (" Ueber baktsriologische Wasseruntersuchungen," 2 E 2 3? 8 Profs. Percy Frankland and Marshall Ward. ' Wiener klinische Wochenschriffc,' 1888 ; « Centralbl. f. Bakteriol.' vol. iv, 1888, p. 467) states that out of 2000 samples of water which he had examined bacteriologically, there were only five in which he was able to detect typhoid bacilli. From the above it will be seen that many investigators, since the year 1886, claim to have discovered the typhoid bacillus in potable waters, but, in the majority of cases, these discoveries, especially the earlier ones, must be accepted with considerable reserve, as, until recently, it was customary to rely for the identi6cation of the typhoid bacillus on altogether insufficient data, as it only gradually became understood that there are several other forms presenting the closest points of resemblance in their morphological characters, both micro and macroscopic, to the typhoid bacillus, with which they are not unfrequently associated, and, moreover, some of these simulatory forms are of very frequent occurrence in natural waters. Indeed, even at the present day, the identification of particular forms or " species " of bacteria is in a transitional and highly unsatisfactory state, as it is daily becoming clearer that the characters, both morphological and physiological, of one and the same micro-organism are often liable to the profoundest modifications through changes of environment and other causes, whilst there are almost daily being discovered in nature new forms which differ only from already well-known forms or " species " in what appear to be the most minute, trifling, and insig- nificant particulars. Tinder these circumstances, it is practically certain that some of the bacilli discovered in water, and believed to have been typhoid bacilli, must, in reality, have been only forms closely simulating the more striking characters of the typhoid bacillus. On the other hand, it is equally certain that a great many waters which have been sub- mitted to examination for typhoid bacilli may have contained them without their being discovered, for, as will be pointed out later, the ordinary method pursued in the bacteriological examination of water, in which a few drops, or at most a cubic centimetre or two, of the water is submitted to plate cultivation, can only, under the most exceptional circumstances, lead to the detection of typhoid bacilli. Thus, whilst there is considerable evidence that the typhoid bacilli have been found on a number of occasions in waters which had been convicted of distributing typhoid amongst their consumers, the failure to discover these bacilli in other waters equally guilty need excite no surprise when the very imperfect methods of examination which are commonly employed for their discovery are taken into consideration. In connection with the above discoveries of typhoid bacilli in pot- able water, I will only at this stage further remark that, in by far the majority of cases, the waters accused of containing these bacilli Report on the Bacteriology of Water. 399 were well waters, and, as is well known, it is just this kind of water which has most frequently been conclusively convicted of distribut- ing typhoid fever. The above discoveries of the typhoid bacillus in natural waters have, in almost all cases, been made not by means of the ordinary method of plate cultivation, which affords little or no chance of a few- typhoid bacilli being discovered amongst a host of common water bacteria, the colonies of which generally grow with great rapidity and not unfrequently cause rapid liquefaction of the gelatine, but by special methods of treatment which have been devised to oppose the proliferation of the water bacteria whilst not materially interfering with the growth and multiplication of the typhoid bacilli, the latter thus acquiring a large numerical preponderance, if not entirely excluding the other forms present in the water. Unfortunately, these conditions which foster the growth of the typhoid bacillus to the exclusion of the ordinary water bacteria are equally propitious to other microbes which are invariably associated with the typhoid bacillus, and which, in fact, resemble it so closely, especially in morphological characters, that they may easily be mis- taken for the typhoid bacillus, and it is this circumstance which causes so much doubt to attach more especially to the earlier of thoso alleged discoveries of the typhoid bacillus in potable water which are recorded above. The particular micro-organism, which is especially liable to cause confusion in this respect, is the so-called Bacillus coli com.rn.unis. This organism was described by Escherich, and is found regularly in the human intestinal tract and faeces, as well as in the excreta of other animals. It is regarded as identical with the Bacillus neapolitanus (Emmerich), and the " Faeces bacillus " described by Weisser, whilst by recent experiments I have shown that it is closely allied to, if not identical with, the Bacillus ethacetosuccinicus previously described by me. In all cases, therefore, in which water is supposed to have been infected with the dejecta of typhoid patients, the B. coli communis may be expected also to be present. In order, therefore, to ascertain definitely whether the typhoid bacillus is present in any given water, care must be taken that the B. coli communis is not mistaken for the former, and to guard against this it would be a desideratum to have some method which would, whilst revealing the presence of the typhoid bacillus, effectually eliminate or separate out its almost constant attendant, the B. coli communis. Unfortunately, this is a consummation which has not yet been realised in fact, for not only is the vitality of the B. coli com- munis in water, as will be shown below, superior to that exhibited by the typhoid bacillus, but in all attempts which have so far been made to suppress the vitality of other organisms, and yet permit of the 400 Profs. Percy Frankland and Marshall Ward. development of the latter, the B. coli communis has shown itself to be possessed of greater powers of resistance than the typhoid bacillus itself. Hence, although the addition of various chemical substances in suitable proportions may effectually destroy or retard the growth of other organisms, yet the B. coli communis survives and remains present along with the typhoid bacillus ; indeed, in many cases it has been shown that such additions have actually destroyed the typhoid bacillus, and left the B. coli communis alone master of the field. It is true that, growing in artificial cultures side by side, there are certain differences observable between these two organisms, for the B. coli communis grows more luxuriantly in the various culture-media em- ployed than does the typhoid bacillus, or to use the expressive language of a French observer, the B. coli communis grows as though it were well, and the typhoid bacillus as though it were ill; but yet on the gelatine plates of each there are nearly always colonies which are indistinguishable from those of the other, whilst even in the potato-cultures, which used to be regarded as the crucial test for the typhoid bacillus, the B. coli communis may and does exhibit, under certain conditions, growths which resemble in every respect those produced by the typhoid bacillus. In two media, however, as has been pointed out by D unbar (" Ueber den Typhusbacillus und den Bacillus Coli Communis," ' Zeitsch f. Hygiene,' 1892, 491), a marked difference is found in the behaviour of these two organisms. Thus, when inoculated into sterile milk, the typhoid bacillus renders the liquid slightly acid, but never causes its coagulation ; the B. coli communis, on the other hand, at the temperature of the body coagu- lates the milk in from 24 — 48 hours, rendering it at the same time strongly acid. Again, when grown in sterile fluid meat extract or broth, the B. coli communis at 37° C. produces, in the course of from 3 — 12 hours, a quantity of gas (consisting of hydrogen and carbonic anhydride), whilst no formation of gas has, under the same conditions, ever been observed in the case of the typhoid bacillus. This latter mode of distinguishing between the B. coli communis and the typhoid bacillus I have reduced to the following extremely simple and handy form suitable for their rapid differentiation : — The organism under investigation is inoculated into a test-tube containing ordinary gelatine peptone in a melted state, the latter is shaken to distribute the bacilli throughout the liquid, which is then allowed to solidify and maintained at the ordinary temperature (18 — 20° C.). The tube, if it contains the B. coli communis, will invariably, after 12 — 48 hours, exhibit numerous conspicuous gas bubbles distributed through the solid medium, whilst no such bubbles make their appearance in similar tubes containing the typhoid bacillus. The test possibly depends upon the meat extract containing sufficient dextrose (derived from the post-mortem transformation of the glycogeu Report on the Bacteriology of Water. 401 in the blood) for a visible fermentation by the B. coli communis to take place. The bubbles of gas are certainly independent of any ingredients present in either the gelatine or in the peptone, for I have found them to form also in agar-agar-peptone, and also in meat- extract gelatine to which no peptone had been added.* The great convenience of the test depends upon its involving only the use of a medium which must invariably be at hand at all times in every bacteriological laboratory, and also on its dispensing with the use of an incubating temperature, whilst it has the farther advantage over Dunbar's original broth-bubble test that the bubbles of gas being fixed in the solid gelatine, the tubes can be examined at leisure even days or weeks after inoculation, whilst with the broth-bubble test, if the tubes are not examined at the right time, the fermentation may have ceased ; besides, in the broth, of course, the bubbles are not nearly so conspicuous. Extensive use has been made of this method during the present investigation, and for rapidly and certainly dis- tinguishing between the typhoid bacillus and the B. coli communis I have found it unequalled ; on the other hand, it must be borne in mind that it does not serve to distinguish between the B. coli communis and many other fermenting organisms. A farther but less certain distinction which should also be em- ployed for differentiating between the typhoid bacillus and the B. coli communis is the so-called indol-reaction. This test is best ap- plied in the following manner, as recommended by Kitasato : — To 10 c.c. of the culture in ordinary alkaline peptone-broth of the organism under examination, and which has been growing for twenty-four hours in the incubator, add 1 c.c. of a solution of potas- sium or sodium nitrite (containing 0'02 gram in 100 c.c.) and then a few drops of concentrated sulphuric acid. If indol is present, a rose to deep-red coloration is produced, depending on the interaction of nitrous acid with indol to form nitrosoindol nitrate which is of red colour. On applying this test to the B. coli communis an indol-reaction should be obtained, whilst the typhoid bacillus gives invariably a negative result. In practice I have found it advisable not to apply the indol- test until the broth- culture has been forty-eight hours in the in- cubator. Although the B. coli communis is generally supposed to give the indol-reaction, this character would appear not to be so constant as is commonly imagined. In my own experiments I have known one and the same culture- series of the B. coli communis not to give the indol- reaction at one time, and yet subsequently to become possessed of this power, although I have not been able to determine the cause which leads to the loss of the indol-producing capacity. Thinking * Dunbar found also that no bubbles were formed in a solution of peptone -with- out meat extract. 402 Profs. Percy Frankland and Marshall Ward. that it might possibly be due to the growth of the bacillus having become enfeebled, I tried growing it under unfavourable conditions, viz., in phenol-broth, in which I left the bacillua for months with- out transplanting, but even by this severe treatment I found no diminution in the indol-producing power. The absence of indol- production by the B. coli communis has also been noticed by Dunbar, who in his exhaustive memoir (Joe. cit.) compares the behaviour of a culture of the typhoid bacillus with a culture of the B. coli communis, and found that neither bacilli gave the indol-reaction. It is obvious, therefore, that indol production is not a necessary attribute of the B. coli communis, and that too much reliance must not be placed on it as a means of distinguishing between the typhoid bacillus and the B. coli communis. As convenient, for purposes of reference, I have collected in the following two tables the principal characters of these two bacilli : — Typhoid Bacillus. (Bacillus typhi abdominalis.) Authority. — Eberth, ' Virchow's Archiv,' vol. 81, 1880; also ibid., vol., 83, 1881. Gaffky, " Zur Aetiologie des Abdominal typhus," ' Mittheilungen a. d. Kaiserlichen Gesundheitsamte,' vol. 2, 1884, p. 372. Where Found. — In the blood, urine, faeces, as well as in the organs of typhoid patients. Found by numerous investigators in water. Microscopic Appearance. — A short, plump bacillus about three times as long as broad, with rounded ends. It occurs in the tissues usually singly, but in artificial cultures it grows frequently into long threads. It is very motile and is provided with numerous cilia, which are attached to both the sides and ends of the bacillus. To stain the cilia add 22 drops of caustic soda to 16 c.c. of the mordant (Loeffler). It is not stained by Gram's method, and stains less readily with aqueous aniline solutions than most bacteria. Gunther recommends heating the cover-glass, after the dye has been poured on it, for a few seconds until it begins to steam, and then washing off the stain as usual. It does not form spores. Cultures : Gelatine Plates. — The colonies on the surface form large spreading greyish-white iridescent expansions with jagged and ir- regular edge. Under a low power they exhibit a brownish shimmer and a characteristic woven structure. The depth colonies are darker, with regular edge, and are covered with delicate irregular lines. No liquefaction takes place. Gelatine Tubes. — Grows chiefly on the surface, producing a delicate greyish-white iridescent expansion with irregular edge. Agar-agar. — Forms a greyish-white moist expansion. Report on the Bacteriology of Water. 403 Potatoes. — Produces an almost invisible greyish-white growth after forty-eight hours, but on touching the moist-looking surface with the needle a tough resistant pellicle is found. On some potatoes, how- ever, its growth is more apparent, so that the above is not the only appearance to which it may give rise. Blood Serum. — Produces a milk-white expansion restricted to the path of the needle. Broth. — Benders it turbid. Milk. — Grows abundantly, rendering it slightly acid. No coagula- tion takes place. Remarks. — Tt grows best at about 37° C. Kitasato states that it produces no indol reaction. It produces sulphuretted hydrogen in iron-gelatine, the needle-track after from five to six days becoming intensely black in colour. In iron-agar, at from 33® to 34° C., this black colour appears at the end of 24 hours (Fromrne). It produces sulphuretted hydrogen in broth with or without peptone ; compara- tive tests made with the B. coli communis revealed no difference either in the degree of the reaction (as shown by the lead-paper test) or in the rapidity with which it took place in the case of the two organisms. The typhoid bacillus never produces gas in any artificial media. It is destroyed when heated for ten minutes at 60° C. Injection into the aural vein of rabbits causes death in 24 — 28 hours (Fraenkel and Simmonds), guinea-pigs into which the cultures are introduced by the mouth, as described for cholera, are also killed (Seitz). Opinion is, however, still divided as to whether death is due to mere intoxication by the bacterial products present in the cultures, or to actual multiplication of the bacillus within the animal. In this connection, see Petruschky (' Zeitsch. f. Hygiene,' vol. 12, 1892, p. 269). Bacillus Coli Communis. Authority. — Escherich, ' Fortschritte der Medicin,' vol. 3, 1885, Nos. 16 and 17. Also Dunbar, " Ueber den Typhus-bacillus und den Bacillus Coli-communis," ' Zeitschrift f. Hygiene,' vol. 12, 1892, p. 485. Also Luksch, " Zur Differen/ialdiagnose des Bacillus typhi ab- dominalis (Eberth) und des Bacterium Coli-commune (Escherich)," * Centralblatt f. Bakteriologie,' vol. 12, 1892, p. 427. Where Foimd—In the intestinal tract of man and animals. Found often in water by numerous investigators, and frequently mistaken for the typhoid bacillus. Microscopic Appearance. — The typical form is a short bacillus 0'4 ;t broad and 2 to 3 ft long ; it is, however, very variable, oval indi- viduals and forms resembling cocci being also found. It exists chiefly as a double bacillus arranged in groups. It is slightly motile, and is provided with 1 to 3 cilia, whilst the typhoid bacillus has 8 to 404 Profs. Percy Frankland and Marshall Ward. 12 cilia (Luksch). Nicolle and Morax mention that the coli bacillus has invariably fewer cilia than the typhoid, that whereas the former rarely possesses more than six, the latter usually exhibits ten to twelve, whilst the cilia of the former are also far more fragile (' Annales de 1'Institat Pasteur,' vol. 12, 1893, p. 561). It does not form spores. Cultures : Gelatine Plates. — Forms round, and very often oval, smooth-rimmed granular colonies in the depth, which later become yellowish-brown in colour. On the surface it forms flat, irregular, pale white expansions, which under a low power exhibit a farrowed appearance due to the unequal thickness of the colony in its different parts. The colony also presents a distinctly wavy lineal structure parallel to the periphery. No liquefaction ensues. Gelatine Tubes. — Grows somwhat abundantly in the depth, pro- ducing small white pin-head colonies, whilst on the surface it forms an expansion resembling the growth on gelatine plates. Agar-agar. — Grows abundantly on the surface, producing a dirty- white, faintly shining expansion. Blood Serum. — Forms a milk-white expansion. Potatoes. — Produces a slimy yellow expansion on some potatoes, on others grey-white, whilst in some cases it resembles the typhoid bacillus in being hardly visible. Broth. — Renders it turbid. Milk. — Renders it acid, and at 37° C. coagulates it in from twenty- four to forty-eight hours. Remarks. — Both cultures of twenty-four hours' age generally exhibit considerable evolution of gas ; ordinary gelatine or agar stab- cultures also generally exhibit bubbles of gas in the solid medium. Such bubbles can invariably be obtained by inoculating into ordinary melted gelatine, which is afterwards allowed to solidify (Percy Frankland). The addition of dextrose to the gelatine is quite un- necessary for this purpose. Exhibits indol reaction after twenty- four to forty-eight hours' culture in peptone broth. Is capable of exhibiting very different degrees of pathogeneity according to its origin, cultures made from diseased tissues in which it is present on being intraperitoneally inoculated into rabbits cause peritonitis, and the bacilli are found in pure culture in the heart's blood. (Alex. Fraenkel, ' Wiener klin. Wochenschr.,' 1891, N"os. 13—15.) 2. Behaviour of Typhoid Bacilli experimentally introduced into Potable Water. The second kind of information concerning the behaviour of the typhoid bacillus in water has, as already mentioned, been gained by Report on the Bacteriology of Water. 405 direct experiment, i.e., . by purposely introducing the bacillus into water, and in this manner the conditions which are favourable and unfavourable to the vitality in water of the typhoid bacillus can, of course, be much more readily ascertained than by the study of such chance cases as those already enumerated above, in which the bacilli had, in the natural course of events, gained access to water. I have below expressed in a tabular form, the principal results of the numerous investigators who have already availed themselves of this method of experimenting on the behaviour of the typhoid bacillus iii water. From this table it will be seen that different observers ascribe very different degrees of vitality to ' the typhoid bacillus in water, nor is this to be wondered at when it is remembered that the typhoid bacilli introduced into the waters may have been possessed of very different degrees of initial vitality according to their age and previous history ; whilst, secondly, the waters experimented with were, of course, not the same ; thirdly, the amount of food-material introduced into the water along with the bacilli must have been subject to the very greatest variations, and again the temperatures and other condi- tions under which the infected waters were preserved were equally variable. Thus, taking the experiments made with distilled water, in which, therefore, there is the most chance of the water having been of uniform quality, Braem found the introduced typhoid bacilli still alive after 188 days, whilst the longest duration, of vitality in this medium observed by Hochstetter was five days, Meade Bolton, Slater, Straus and Dubarry, and Wolffhiigel and Riedel giving periods intermediate between these two wide extremes. These dis- crepancies in the case of distilled water are doubtless to be accounted for partly by the difference in initial vitality possessed by the different typhoid bacilli employed and partly by the difference in the amount of culture-material imported into the distilled water along with the bacilli themselves, whilst the actual numbers in which the bacilli were introduced may also greatly influence the degree of longevity observed. This wide divergence in the results obtained by previous observers would alone call for a reinvestigation of this subject with a view to ascertaining the longevity of the typhoid bacillus in definite types of British potable water, and taking more into consideration the exact chemical composition of the waters experimented with. There is, however, another point arising out of the results arrived at by previous investigators which still more urgently demands re- investigation with a view to its confirmation, qualification, or direct contradiction, and this is the relatively far greater longevity of the typhoid bacillus in sterilised than in unsterilised potable water, 406 Profs. Percy Frankland and Marshall Ward. which has been affirmed more especially by Kraus and subsequently by Karlinski. This point is obviously of the very highest import- ance from a practical hygienic point of view, as it is with unsterile potable water that we are in practice alone concerned, and the duration of vitality ascribed to the typhoid bacillus in such water by both these observers is of very limited extent — not more than seven days. The experiments of Kraus are so striking, and have attracted so much attention, that I will give them in more detail in the following table : — Typhoid Bacillus. Description of Water. Number of Days after Inoculation when Examined. 1. 2. 3. 5. 7. 9. 20. 150. (I) Munich water supply (Mangfall) 57,960 57,000 56,000 0 0 0 Numbt 50,400 50,840 35,910 Numb 0 0 0 rof Ty 15,680 32,643 10,010 er of Wt 0, 490 280 >hoid I 9,000 h,900 7,060 Lter Bai 80 lost 500 acilli fou 0 0 0 :teria fou 288,000 300,000 256,000 nd in 1 c 0 0 0 nd in 1 c. 400,000 427,000 lost c. of Water. 0 0 0 c. of Water. 970,000 innumerable 456,000 0 0 0 1,080 1,980 1,050 (2) Well-water, Munich (3) „ (I) Munich water supply (Mangfall) (2) Well-water, Munich (3) These results indicate, therefore, that, on introducing the typhoid bacilli into the potable waters in question, which were almost naturally sterile, the typhoid bacilli promptly disappeared as soon as the water bacteria had undergone extensive multiplication, which had taken place in each of the three experiments by the seventh day after the importation of the typhoid bacilli. These interesting experiments cannot, however, by the light of our present knowledge, be accepted without criticism, for there cannot be the slightest doubt that, when only the ordinary method of plate- cultivation is employed in such investigations on unsterilised water, the typhoid bacilli will be generally overlooked unless they are present in large numbers. Again, the attempts which have been made, both by Kraus and other experimenters, to count the typhoid colonies on plates containing such mixtures of colonies, and the numerical estimates given of the typhoid bacilli in such unsterilised waters, must be wholly illusory, for the number of typhoid colonies which develope what may be called a typical appearance (i.e., one which enables them to be readily recognised with reasonable cer- tainty) depends on a variety of different circumstances, amongst which may be mentioned the age of the plate, the extent to which Report on the Bacteriology of Water. 407 the colonies are crowded together on the plate, very probably, also, the nature of the other colonies on the plate, and certainly the degree of vitality possessed by the typhoid bacilli themselves. Thus, if the numerical estimate of the typhoid colonies on a plate is made by counting as such the characteristic surface expansion colonies only, the result must be entirely fallacious, as nothing is, in my experi- ence, commoner than to find only a vanishing proportion of the total number of typhoid colonies, even on a pure plate, giving rise to these surface-expansions at all. In. looking for typhoid bacilli in such artificially-infected unsterile waters it is, in fact, necessary to employ special methods for their detec- tion similar to those which, as already pointed out, had to be devised for the examination of natural waters for the typhoid bacillus, and it is only when such special methods have been employed with a negative result that the conclusion can be legitimately drawn that the typhoid bacillus is not present in the living state in the particular volume of water operated on. In the present investigation, the uniform practice has been made of examining all unsterile waters by means of Parietti's method of phenol broth-culture (see description below) in order to ascertain the presence or absence of the typhoid bacillus or of the B. coli com- munis. Method of Detecting the Typhoid Bacillus and Bacillus Coli Communis in Unsterile Waters. It will not be necessary to describe the various methods which have been devised for discovering the presence of typhoid bacilli in water, but it will be sufficient to point out that these are nearly all based upon the fact that the typhoid bacillus is, in comparison with most bacteria, but little affected by small doses of either phenol or dilute acids, so that, by adding suitable quantities either of phenol alone or of phenol in conjunction with acid to the culture-media, the growth and multiplication of the typhoid bacillus is not materially interfered with, whilst the proliferation of most, if not of all, the water-bacteria is suppressed. Of these various methods, the one which I selected for the purposes of this investigation was that devised by Parietti (" Metodo di ricerca del Bacillo del tifo nelle aque potabili," ' Kivista d' Igiene e Sanita pubblica,' 1890). This method, which consists in adding phenol along with hydrochloric acid in certain proportions to neutral broth is carried out as follows : — A series of test-tubes containing 10 c.c. of neutral broth, each receive from 3 to 9 drops of a solution having the following composi- tion : — 408 Profs. Percy Frankland and Marshall Ward. Phenol 5 grams. Hydrochloric acid (pure) 4 „ Distilled water 100 „ (In practice, I generally employ some tubes to which 3 drops (= 0'25 c.c.) and others to which 5 drops (= 0'4 c.c.) of this solu- tion have been added.) To the tubes thus treated, from 1 drop to several cubic centimetres of the water under investigation are added, and, after thoroughly mixing the contents, the tubes are placed in the incubator at 37° C. As soon as the tubes become turbid (which in the initial presence of many typhoid bacilli will occur already in twenty-four hours, but if only few are present, may be postponed for forty-eight, seventy-two, or even more hours) they are submitted to ordinary plate cultivation in three dilutions, the second and third dilutions only being actually poured on to plates or into Petri dishes, whilst the first dilution gelatine-tube should be preserved to see if gas-bubbles develope in it. The gelatine-plates thus prepared are frequently found to yield nothing but typhoid colonies, whilst in some cases the latter are mixed with the colonies of water-bacteria, and in some cases, again, there are only colonies of water-bacteria on the plates, In no case must it be concluded, from the mere appearance of the colonies that typhoid is present, but the colonies must always be submitted to the further tests of 1. Microscopic examination. 2. Inoculation on to potatoes, and comparison of growth with that of simultaneous cultures of the typhoid bacillus on the same potatoes. 3. Inoculation into broth, and examination of the broth-culture after forty-eight hours' growth in the incubator at 37° C. for indol, which, if it is the typhoid bacillus, should be absent. 4. Inoculation into milk, which should not subsequently become coagulated on keeping for one week in the incubator. 5. Inoculation into a tube containing melted gelatine-peptone ; on distributing the bacilli in this and then congealing the gelatine, no gas-bubbles should be formed on keeping at 18 — 20° C., •whilst in the case of the B. coli communis bubbles will make their appearance in from twelve to forty-eight hours. The B. coli communis, as already pointed out, is even less sensitive to phenol and acids than the typhoid bacillus. In the case of those unsterile waters infected with the B. coli communis, the above method was similarly employed for its detection. The above outline will show that a systematic investigation of the behaviour of the typhoid bacillus in un sterilised waters is attended with considerable difficulties, and involves an enormous amount of Report on the Bacteriology of Water. 409 labour, which is, however, well worth bestowing in consideration of the very great importance attaching to the question at issue. FIRST SERIES OF EXPERIMENTS. The Vitality of the Typhoid Bacillus and of the Bacillus Coli Communis in Thames Water. As already indicated above, I have endeavoured to make these ex- periments as far as possible comparable with those previously con- ducted by me on the B. anthracis and its spores recorded in the Second Report. The Thames water was collected by me personally from the river, close to the intake of the Grand Junction Waterworks, near Hamp- ton, on May 4, 1893 ; this spot was selected as being in that region of the river from which the supply for London is abstracted. This water was submitted to chemical analysis with the following results : — Results of Analysis expressed in Parts per 100,000. Total solid matters 26'SO ^ Organic carbon "1 , , , . f 0'247 > by combustion < nitrogen J I 0'038 Organic nitrogen (by Kjeldahl method) 0'041 Ammonia (free) 0'013 ,, (albuminoid) 0'016 The sample was turbid Oxygen consumed by organic )> and free from poison- matter 0-102 ous metals. Nitrogen as nitrates and nitrites. . 0'124 Total combined nitrogen 0'173 Chlorine T65 Temporary hardness 13'8 Permanent ,, 4'4 Total , 18-2 _ The analysis shows that the water is chemically a typical sample of Thames water as found in this part of the river at this season of the year. In this series of experiments I proposed introducing the typhoid and coli bacilli respectively into (a) Thames water in its natural and unsterile condition; (6) Thames water sterilised by filtration through porous porcelain ;* (c) Thames water sterilised by steam,\ and to com- * The Chamberland filter (see 2nd Eeport, ' Boy. Soc. Proc.,' vol. liii, p. 183) was sterilised in the steamer on two successive days, 1£ litres of the Thames water being then passed through it immediately before infection. t 1500 c.c. of the Thames water were placed in the steamer for two hours on each of three successive days. 410 Profs. Percy Frankland and Marshall Ward. pare the respective behaviour of the two organisms in these three different states of the Thames water. The infection of the several waters was made on 11.5.1893, as follows : — I. Typhoid. — Forty needle-loops were removed from the surface of an agar-culture of the typhoid bacillus which had been grown at 18 — 20° C. for fourteen days, great care being taken to carry as little as possible of the culture-material along with the growth. This growth was introduced into 50 c.c. of steam-sterilised Thames water and violently shaken for fifteen minutes in a sterile bottle to thoroughly disintegrate the masses. This water-dilution was then employed for the infection of the larger quantities of water as indi- cated below. II. B. coli. — The water-dilution of the B. coli communis was made in exactly the same way as that of the typhoid bacillus described above, excepting that only twenty-five needle-loops of the growth were taken, as owing to its greater thickness it was detachable in larger masses from the surface of the agar. The culture employed was of just the same age (fourteen days), and had been grown at the same temperature (18 — 20° C.) as the typhoid bacillus. With the water-dilutions thus prepared the following infections were made : — Typhoid bacillus. B. coli communis. Unfiltered Thames Water. 2000 c.c. received 8 c.c. of 1000 c.c. received 3 c.c. of the water dilution. the water dilution. Porcelain-filtered Thames Water. 750 c.c. received 3 c.c. of 750 c.c. received 2 c.c. of the water dilution. the water dilution. Steamed Thames Water. 750 c.c. received 3 c.c. of 750 c.c. received 2 c.c. of the water dilution. the water dilution. The waters thus infected were each subdivided amongst a number of smaller sterilised flasks plugged with sterile cotton-wool, and these were respectively placed in the incubator or refrigerator, according as they were to be exposed to a winter or a summer temperature. Thus : INFECTED. Typhoid. B. Coli. Unfiltered Thames Water. 3 flasks incubator. 3 flasks incubator. 3 „ refrigerator. 3 „ refrigerator. Report on the Bacteriology of Water. 411 Porcelain-filtered Thames Water. 3 flasks incubator. 3 flasks incubator. 3 „ refrigerator. 3 „ refrigerator. Steamed Thames Water. 3 flasks incubator. 3 flasks incubator. 3 „ refrigerator. 3 „ refrigerator. UNINFECTED CONTROL WATERS. 2 flasks incubator. 2 „ refrigerator. N.B. — The convention is adopted throughout the text of the Report of designating the flasks 1 1, 21, 31, and IE, 2E, 3 B, ac- cording as they have been kept in the incubator or refrigerator respectively, and in this manner the individual flasks are readily identified. In order to ascertain whether the infection of the water had com- municated any large amount of organic matter to the water, some of the infected waters were submitted to a partial chemical analysis. This is a matter which has, unfortunately, been almost entirely neglected in previous observations on the vitality of pathogenic bacteria in potable waters, and it is obvious that if such investiga- tions are to have anything but a negative value, the waters, after infection, must not differ materially in their chemical composition from that which they possess in their natural state. These chemical analyses yielded the following results : — Results of Analysis expressed in Parts per 100,000. Unfiltered TJnfiltered Unfiltered Thames water Thames water Thames water* infected infected uninfected. with typhoid. with _B. coli. Ammonia (free) 0-013 0'013 O'OIS (albuminoid) 0'016 0'020 0"028 Oxygen consumed .... 0'102 O'lOl Chlorine T65 170 170 Thus the infection, especially in the case of the typhoid bacillus, had caused but very little increase in these ingredients. The examination of the several flasks was conducted on the fol- lowing principles : — 1. The unfiltered infected Thames water was examined by plate- cultivation from time to time in order to ascertain the changes in the * The full analysis of this water is given on p. 4O9. TOL. LVI. 2 F 412 Profs. Percy Frankland and Marshall Ward. total number of micro-organisms present, but without any hope of counting or even identifying the typhoid or coli colonies, as this is, for the reasons already given (see pp. 406, 407) in general, quite out of the question. On the other hand, the presence or absence of living typhoid and coli bacilli was periodically determined by cultivation with phenol- broth (see p. 407), a method which, of course, does not permit of an estimation of their number, but which, as will be seen, is often able to throw light on their relative abundance or on their relative degree of vitality. 2. The sterile (porcelain-filtered and steamed) infected Thames waters were periodically examined both by plate cultivation and by the phenol-broth method : so that in the case of these waters, in which the typhoid or coli bacilli were not mixed with other bacterial forms, not only could their presence or absence in the living state be determined, but their actual numbers ascertained. 3. The unfiltered uninfected Thames water was periodically ex- amined by plate cultivation in order to follow the increase or decrease in the numbers of the water bacteria, whilst examinations by the phenol-broth method were also made in order to ascertain whether there were any forms amongst the water bacteria which might be confounded with the typhoid or coli bacilli, and thus to check the diagnoses made in the case of the unsterilised infected waters. 1. Bacteriological Examination of the Uninfected Unsterilised Thames Water. (First Series.) It will be most convenient to consider first the behaviour of the control waters which were placed under the same conditions, in the incubator and refrigerator, as the infected ones. The results of the gelatine plate cultivations of these control waters are summarised in the following table : — Report on the Bacteriology of Water. 413 OS 00 S3 3 p. £ eg W I CO o O <4H o I Q 02 •s o 0 CO rH 3 g h o _o O 2 O, O «*-. E CO CD JN^" cT r£j 95 0 CD CO 95 Qfl a rH i 'S '3 ti 95 2 2 tf 0 'o3 S ao fe (M '5 'o 1 11 1 8 O O O 1 8 o S lO o" oo" •^H ,£> 3 ' CO •IN •* O 3 rt 1 11 ivation. 10 12 |o Ip -b-fci P H 10 |0 -4=1 HO H H o 1| |0 10 M|H~|2 a a 11 0 95 r* ft H15 o o Jp S C 03 03 S c 03 03 |« ^ 11 8 3 H fa H-lo "H "H ! "Ei •* •* -S-ft £ 95 0 ^ •73 b °° 95 95 _£ "o3 "c * 43 •<* CO CO CO « (N (M (M N *J ' O T3 95 O E D !50 a o FM PS rH PH S* 'S 95 95 .£; iM W 2 S s h d £ .2 "o3 95 rH M HH M E 3 FR rH i-H rH rH O " p hH ,S S O O pC -^ ® CO CO CO CO CO £ > ' ^J 03 o oo oo K 'JO OS CO C5 X 0 JS j-i rH E rH E rH E rH S 0 "^ 95 in 1O 10 «5 CO m o 95 ^ (4 95 r-! CD oi O 2| P= 2 r 2 414 Profs. Percy Fraukland and Marshall Ward. The sample of Thames water which between the date of collection (4.5.1893) and the date of first examination (11.5.1893) had remained in bottles almost completely filled up to the stopper, and at a tempera- ture of about 10 — 12° C. exhibited in the first instance an unusually small number of bacteria (only 290 in 1 c.c.). There can be little doubt that the original number present must have been greater than this, and have become diminished during this period of residence in the stoppered bottles, for on introduction into the flasks plugged with cotton-wool they underwent enormous multiplication. In the flask kept at 19° C. the multiplication was doubtless most rapid, but had already fallen again to 45,000 per c.c. on the second examination, whilst in the flask kept at 6° C. multiplication and subsequent decline were probably both less rapid, so that on the occasion of the second examination the number present was still 563,000 per c.c., which underwent continuous diminution during the remainder of the time that this flask was kept under observation. These phenomena of initial multiplication followed by decline have been already frequently called attention to, both in the former Reports and by other observers, so that there is no necessity to dwell further upon it here beyond pointing out that it shows that the water- bacteria in this sample of water employed were in an active and flourishing state under the conditions maintained during the experiment. Examination by Plate-Cultivation of the Unsterilised Thames Water infected with Typhoid and B. coli communis respectively. Having in the previous pages traced the numerical changes which took place in the bacterial contents of the control uninfected un- sterilised Thames water, I will now proceed to describe what occurred in the case of the same unsterilised Thames water which was infected with typhoid and coli respectively (in the manner indicated on p. 410). The flasks containing these infected unsterilised Thames waters were kept at a winter (6 — 8° C.) and a summer (19° C.) temperature respectively, and were examined from time to time by gelatine plate cultivation, and the results of these periodical examinations are recorded in the following table : — Report on the Bacteriology of Water. 415 M 1. 4) a> m (0 tXi'>.-W)»H«'T3e8(>i ." O ,£3 > 3 <» cS^Sc^^^ ' — ' • — t 'S ^"^ §Q.2 "S '?» s S 'S '?»"«* '• .5 * 'S 1 -2 2 § J3 "CM °S s K,2 °S ^ § -2 "S i> O "'i^^ ° « 1* fi B * ^43 "S ^ e^ 2 q 43 P- 3 ® 'S "-S^''5-^ ^ -'-'otw'w §"S S °° S3 S 9 cs » r o S ^ 43 ^ bfi1 — ' 3 *^ ID s3 'S a ^ ^2 O) O Is FH s so 5.1 3* I^^^l P-l PH ^ O t^-fi !ltll!i jfidi';!fj|f O -*i °S ^1 «4-l .2 0 (3 bC ^•all-Hi ^^1 S.I §^ i llllliiilll 43 § 43 P,PR 0 43 S £ 43 S H >i O H b .2 °- "S • 8 fli o JH .9>J «jT in" on 1 || | 0 eo $ *$£ | co Qi m *W b t~ -a S ° 5 • 11 ,0 *« fil g« i-T CO cT co q i— i p 1 l-g •K § § .£ M 11 13 S3 S3 51 "o ** o -^ !« HJO> ]O P P ^ H'4" « •1 "P"|5 P -,s 111 || « CO CO CO CO jC °Q PH O " fl •a 1 0 O M M «3 W) la pgj ^ •g _>j S3 >> •O <§ 43 *3 i ^H a O CM • •43 a § « O 1 M M PH 43 0) B Hn . a i— i 5 S « co co CO o S ''El/^ r* oo co co g J7 C3 S r-J r-l = i— i « V(5 U5 g *1 * i-H tH -H 416 Profs. Percy Frankland and Marshall Ward. r * SV. K-%'1 If; ^ o £| o ^3 ^d o . H i _cr oo o |5o o-2 3 « * r> , , c»- «> O =*H t. .-rtf J O®"-! ^ § ° 5 '3 P4 H H T3T3 •rt P? PI - Report on the Bacteriology of Water. 417 From the above table it will be seen that the unsterilised Thames water, which contained remarkably few bacteria (only 290 per c.c.) at the time, was infected with a very large number of typhoid bacilli (about 78,000 per c.c.). The total number of bacteria in the water kept at the winter temperature of 6 — 8° C. underwent enormous multiplication followed by decline, as in the case of the uninfected unsterilised water. In the water kept at the summer temperature of 19° C., on the other hand, the numbers found exhibited almost con- tinuous decline, and also closely resembled those found in the unin- fected unsterilised water preserved under similar conditions. In both cases, however, there must have been a great multiplication of the water-bacteria, for whilst the gelatine plates prepared from these waters during the first week after infection admitted of the ready recognition of typhoid colonies, in the subsequent examinations this was altogether impossible, so that the large number of colonies present on these later plate cultivations must have been derived from the extensive multiplication of the comparatively few water- bacteria present in this unsterilised water at the outset of the experiments. I must, however, again emphasize what I have stated before, that whilst the recognition of typhoid colonies on such plates containing the colonies of numerous water-bacteria is often difficult and attended with much uncertainty, any estimation of the number of typhoid colonies on such plates, as has been attempted by some observers, is altogether illusory and calculated to lead to the most erroneous con- clusions. For whilst the surface colonies of the typhoid bacillus are even liable to be confounded with the surface colonies of some other bacteria, in the appearance of the depth colonies (and, of course, in ordinary gelatine plates the majority of the colonies are beneath the surface) there is nothing to distinguish them from an immense number of other forms common in water. Thus, whilst in the above series of examinations I have no hesitation in saying that on the plates prepared on the llth, 16th, and 17th May, typhoid colonies were present, I rely for the determination of their presence or absence after those dates entirely on the results of the examinations by phenol broth-culture which will be given below. Again, even in the case of those plates which obviously contained typhoid colonies, I do not con- sider that any estimate of their number could be justifiably made, as such an estimate could only include the surface colonies which had developed the characteristic expansions. Thus the examination by plate-culture of these unsterilised waters does not enable us to ascertain whether the typhoid bacilli under- went any numerical increase in these waters, but from the fact that no such increase was observed in the case of the typhoid bacilli similarly introduced into steam-sterilised Thames water (see p. 451), and in which, therefore, the conditions for their multiplication were 418 Profs. Percy Frankland and Marshall Ward. Eemarks. Nearly all the colonies on the plates had the typical appearance of those of S. coli communis. All the plates exhibited a large number of liquefying colonies, showing that the few water bacteria originally present in net have undergone extensive multiplication. There were also a great many small colonies, doubtless to a large extent those of the B. coli communis. bo D IM O 0) "a be . '— -^ '—' aj rf ^ «'§ o3 "n tD o There were no surface colonies on these plates resembling those of the S. coli communis. Flask 1 I was not used, because a little of the cotton-wool stopper got into the water on the occasion of the last examination. All the plates exhibited a large number of colonies, causing liquefaction of the gela- tine, and hence necessitated early counting. All the plates again exhibited a large num- ber of colonies, causing liquefaction of the gelatine, and hence necessitated earl v counting. (H .8 *• O 1 o o o t« ^ o 4 o a O i— 1 6C 83 ^ r-T t-T j>T 0 (M "3 g *H' M o oo co £ o-S O o • "S^j | A O CO ||*8 si h O ^ ^ 1 | 8 o O 8 8 o P -w t-T to o" 0» •^ fe o fl^ QO iH , .,-< ^ § OS - 03 03 03 03 3 os "° ® «2 ""3 1-, P P P P ^P JO gje P P ^ -*-» o H i- TH^I 1- H h 1- i- S3 ?> 'S "2 _2 -2 2 "S •* co co co co eq co o " ira ' 0 ' CO ' f5ii«| « i— 1 ^H CO 1— 1 CO si T}«' 1— < £ 5 it Report on the Bacteriology of Water. 419 far more favourable, there can be no reasonable doubt that they did not undergo any increase but only decline ; this supposition is, more- over, corroborated by the results of the examinations by phenol broth-culture, to which I shall presently refer. I will now turn to the similar examinations made by gelatine plate- culture of the same unsterilised Thames water infected with the B. coli communis, the results of which are recorded in the table on p. 418. The results recorded in the above table for the B. coli communis are almost precisely parallel to those recorded in the previous table for the typhoid bacillus. There is again, in the case of the water kept at the winter temperature of 6 — 8° C., the enormous multiplication in the total number of bacteria present, followed by rapid and almost continuous subsequent decline. In the case of the water kept at the summer temperature of 19° C., a slight increase was observed on the occasion of the second examination (but, as pointed out in the case of the typhoid table, a great increase followed by rapid decline may have taken place in the interval between the first and second examina- tions), after which there was a great decline followed by some re- crudescence at the end. In the case of the waters kept both at the winter and the summer temperatures respectively, however, it is obvious that extensive multiplication of the water-bacteria must have taken place, owing to the very large increase in the number of colonies causing liquefaction of the gelatine which was observed. For the same reasons as stated in the case of the typhoid bacillus (see p. 417), it is impossible to form any estimate of the numbers in which the coli bacilli were present after the day (11.5.1893) of their introduction, nor as to the length of time over which they persisted in the living state in these waters. From the corresponding experi- ments, however, made with the steam-sterilised Thames water, it is quite possible that the B. coli communis, unlike the typhoid bacillus, may have undergone some multiplication in the water. It is to the examinations by the method of phenol broth-culture that we must again have recourse in order to ascertain how long the coli bacilli remained alive in these unsterilised waters. There is a point which is brought out very strikingly in these tables, and to which I would draw attention at this stage, and that is that the total number of bacteria present in these unsterilised waters at the end of the period over which these experiments extended, was, both in the case of the uninfected waters (see table, p. 413) as well as in that of the typhoid (see table, p. 415), and in that of the coli (see table, p. 418), greater in the water maintained at the summer than at the winter temperature respectively. The probable explanation of this phenomenon would appear to be that at the lower temperature (6 — 8° C.) many of the bacteria present may be unable to form spores, 420 Profs. Percy Frankland and Marshall Ward. r3 i y s 1 1 | jj g i s • eg t, . 1 s S I .s^ oi 1 3 1 S si H 1 1 § S el 11 d |g o 'C 1 1 52 C oj S? * * O t- *+ _ feg^a 1 ll^laf-s** fill 1 i P.^ S PH" J3 03 Bj^'Ba • BS^* Sr> *• S CO M 173 c rs .« £2 ^3 O *J M ° «3 C 11 1 1 1 1 Is •0 1 1 t 5 00 h '•gS'g 1 1 1 1 2 .£ '« '"5 1 1 1 1 oB S CM •o oc*.-o§.S .• ^w*3 3 tc .i c - o ^ !I 1 1 1 1 i||l||f|i i f. " 1 1 1 1 o u ?— M I"fcj CO •-! i l?rg-ll d ° A CO ^ "S ^> o3 *j oj y J4 9 be ra ••I — . *^ bo rt C«g^ ^* e) B 2 ?" O 1 | 2 o "3 A C§ C3 "" o bC J, ec "2 a- 1 ^"S P B . °; > M 1 I !cs 1 i c (S £ S s Report on the Bacteriology of Water. 421 *Hochstetter state* that he could detect no diff.-rence in the be- haviour of the cul- tures grown at differ- ent temperatures. tSterillwd Berlin tap- water. §Sterilised distilled water . fSeltzer- water. 0 -C eo *3 _^ fcT a: « 13 01 « -*S * -*- -*— *Onsterillsed Innsbruck drinking water. *Unsterilised well-water. tUnsterilised Munich Mangfall water. Con- sidered a very pure water. Leitmeritz town wa-(v sterilised. 1 1 1 1 1 If!!! 1 1 1 1 «"= °. Ill If 1 1 1 1 •J 5 > ,0 •a o • >> . * E 3 j fi >> w T* •> oo •a •a B o I t o 1 X? O « T3 CO 00 II*' 2 2 o C C-a i B * £5^, 'S— >, 1 1 5- 5 J&sf * 1 1 rs l>- o; 3 "O ® OE -J5 CO Isi sS" 1 'llll^llfjii 1 i i 1 1 d o J ll 1 O O d 00 d w 0 0 d B 00 d i 00 ppll1 1 St~ ^ x c ^ || i«|feMs * 1 g | I 1 1 1 1 IllijS-el^S 1 2 tet.co'J*'S'5'n •^ 3 0. H w i 1 be 00 i ^-^ 8 wST 0 o 8 ^ c c. Sx. % 9 ^^ Sj ''o — r^~ 0, a *j J 'S ^5 §2 1 3 9] S'5 — 9Q ij ^ S S 422 Profs. Percy Frankland and Marshall Ward. Beraarks. •a j, c g i.i s | §;§ °S 5 =^SM T3 *g ^j .Q H .^ ^ tl |f| ^c *~ S be iflSillli Sterilised well-water. Sterilised Ourcq water. Sterilised Vanne water. The latter has less or- ganic matter than the Ourcq water. Sterilised distilled water. Well waterin Rostock, unsterilised. * * H- ur. * SgSa • X n 1111 i 1 \ 1 1| i 1 \ 1 °3 a « ,*§«§*» ^cSo*^ 1 •3 C5 0? 1 JP I •5 — '£ J 1 "o*. la *. *7 * *. aj s s a 1 05 J k u C1! CO «-• t— « •* ooco £ o p I-d 1*1 1 1 1 1 1 1 1 0 t~> "^ a! p O 3 a M OH ll ll §11:! 1 1 1 1 1 ** 013 III*! *„ S'S ba a o, o •- * '3 O Q 1 O OO H s f £ S C-) CO 1 § sis ? I3* *~* ^ B °s° a .s a o 1 *flM «?l«l .f= „ 5 j= * liifi «^a° 1 1 O *"So3 S *• • S 5 5 f og § 5- «£""|£go'2 3 ^^ E 5* *^» "^ e^*'" ^ u ^ rt — ' 2 -rt O 03 O Sr""- ••* £^'-a *rf o S 0) •£ ^ g S 0 •O ~ 0? § * I £" 00 £— a o o a> o 00 ^ 00 ^ 3 £.3 •^ i '-' a ||f £ x if ll « >, 11 I • s" £ D Report on the Bacteriology of Water. 423 •Sterilised Berlin tap water. 1 tsteriliscd distilled water. §Sterilised highly pol- luted Biver 1'anlie water. tStcrile distilled water. "Simple aerated non- sterile. tSterile soda-water. §Sterile soda-water, non- ae rated. g 4 = I , <* i p. N oT i if s 0 M ai ? 1 1 3 g ^x 5- '-3> d >- a< oo "°5 1 cT® • *t^ c •> **c3 « co — e» ••*-• p, OQjco*0*3-""^' 1 0 1 2 ^^x"^^ ^ 2^*0"® x*S S ^ KM" I'i^l *"-". 2l3p g ipji •-.•ltiilli"f J 'e"-S t"gco hS>'-fe'?5t:«™'et2 1 1 1 13 §S'.a* Wg-n-iCS ^ o S ^ e r ^ g »3 ; i;g-£ = ;S"£££3 •= sSSlrf S-alilllrf*! a -^ - "Ec.^ ^ .E;,i'§s'o-c'Soicy W § ^ ~ 1 « ' /•§ ^ = ' -- «" 2 "1 ? * fe a a t-ei'" ai't'(:"™^L|WM 1 'J! 1 0° 0 $ S 1 00 tO .5 P* 2 | ||||? |'l|l||tll| One needle-point from a gelatine culture intro- duced into 50 c.c. of the water. One needle-loop from a broth-culture. One needle-point from a gelatine-culture intro- duced into 10 c.c. of the water. CO ** •"•' ° 111 £ !N 0 1 lliif IllflillSS § iail'S d|^-§|^-?sss $ IrlSg riililjilll ? i-P-gi lIMpil! ! silli nllilfii] S g c a p 3 c c-g g esa s-g fe§ •S 3o>-> 11 » '&:| > j2 g g -g J D-gDDDD(5 ^DD(^w&C«iS£ |i I -| VNo turbidity even on 6.6.1893. (4) » • • 1-0 » J Unsterilised Typoid- infected Thames. (5) Flask 11 .. 0-5 » Did not become turbid. (6) » '« 1-0 » Turbid in 48 hours. fT) Flask 1R.. 0-5 » Turbid in 24 hours. (8) »i • • 1-0 » >» >j The results recorded in the above table indicated that on 29.5. 1893, whilst there were still no bacteria in the uninfected unsterilised Thames water to interfere with the phenol broth test, this test pointed to the presence of living typhoid bacilli in the typhoid-infected Thames waters, which had been kept both at 6° C. and at 19° C. (flasks 1 B and 1 I). The results of the test, moreover, indicate that these typhoid bacilli were now less numerous or in a less active condition in the flask 1 I (19° C.) than in the flask 1 B (6° C.), because both phenol broth tubes prepared from 1 B became turbid already in twenty-four hours, whilst of the two similar tubes prepared from flask 1 I, only the one in which 1 c.c. of water was employed for cultivation 426 Profs. Percy Fraukland and Marshall Ward. became tuvbid, and then only after forty-eight hours, whilst the tube in which only 0'5 c.c. of water was employed did not become turbid at all. It must not, however, be supposed that the diagnosis of typhoid bacilli in these waters was allowed to rest on such slender evidence as the mere clouding of these phenol broth-cultures, but the latter were submitted to gelatine plate cultivation to see if the characteristic typhoid colonies made their appearance, and these colonies were further confirmed by inoculation (a) on to potatoes for exhibition of the characteristic growth, (&) into gelatine tubes to see if bubbles of gas would make their appearance, (c) into broth for the indol test, and generally also (cT) into milk to see whether coagulation of the casein would take place. Thus in the case of the above phenol broth- cultures commenced on 29.5.1893, the final confirmation of typhoid was not obtained until 12.6.1893, or a fortnight later. The phenol broth test was again applied to the waters on 5,6.1893, with the following results (p. 427). The plate cultivations made from the phenol broth tubes, referred to in the table (p. 427), yielded the following results : — Broth tube. (21.) Typhoid-infected Unsterilised Thames, Flask 1 I. (Typhoid Present.) The presence of typhoid was confirmed by the typical appear- ance of colonies, growth on potatoes, negative indol test, and negative gelatine bubble test. (23.) Typhoid-infected Unsterilised Thames, Flask 1 R. (Typhoid Present.) The presence of typhoid was confirmed by the typical colonies, growth on potatoes, negative indol, and negative gelatine bubble tests. (37.) and (45.) Coli-infected Unsterilised Thames, Flask 2 I. (B. coli Present.} The presence of the B. coli was confirmed by typical colonies, growth on potatoes, positive indol, and positive gelatine bubble tests. (39.) and (47.) Coli-infected Unsterilised Thames, Flask 1 R. (B. coli Present.) The presence of the B. coli was confirmed by typical colonies, growth on potatoes, positive indol, and positive gelatine bubble tests. Report on the Bacteriology of Water. 427 Examination by Phenol Broth-culture on 5.6.1893. ft | £} O -0 || Water used for cultivation with phenol broth 'Quantity of water taken in c.c. Q uantitj of pheno solution added to 10 c.c broth. ' Eernarks. Unsterilised Typhoid-infected Thames. (21) Flask 11.. 0-5 3 drops Turbid in 48 hours. Plates poured 9.6.1893. (22) (23) Flask IE.'.' 1-0 0-5 " Turbid in 48 hours. Turbid in 24 hours. Plates poured 6.6.1893. (24) .. i-o » Turbid in 24 hours. Unsterilised Coli-infected Thames. (37) Flask 21.. 0-5 3 drops Turbid in 24 hours. Plates poured 6.6.1893. (38) (39) Flask IE.! i-o 0-5 » Turbid in 24 hours. Turbid in 24 hours. Plates poured 6.6.1893. (40) „ 1-0 » Turbid in 24 hours. Unsterilised Uninfected Thames. \ (41) Flask 11.. 0-5 3 drops Turbid in 48 hours. Plates poured 9.6.1893. (42) (43) Flask IE!! 1-0 0-5 " Turbid in 48 hours. Turbid in 48 hours. Plates poured 9.6.1893. (44) „ 1 -0 » Turbid in 48 hours. Unsterilised Coli-infected Thames. (45) Flask 21.. 0 '5 5 drops Turbid in 24 hours. Plates poured 6.6.1893. (46) (47) (48) Flask IE.. 1-0 0-5 I'O „ , Turbid in 24 hours. „ Turbid in 24 hours. Plates poured 6.6.1893. „ Turbid in 24 hours. Broth tube. (41.) Uninfected Unsterilised Thames, Flask 1 I. The only colonies on the plate which bore any resemblance to typhoid, yielded pink growths on potatoes, and developed green fluorescence on being grown in gelatine tubes ; they were thus iii reality wholly unlike and different from typhoid or coli. vor,. LVI. 2 G 428 Profs. Percy Frarikland and Marshall Ward. Broth tube. (43.) Uninfected Unsterilised Thames, Flask 1 R. Colonies bearing some resemblance to typhoid yielded growths on potatoes, which were also not unlike typhoid, but on inocula- tion into gelatine tubes a green fluorescence was obtained, con- clusively proving that it was another organism and not typhoid or coli. From these examinations then it was apparent that on this day, 5.6.1893 — 1. The unfiltered Thames water infected with typhoid on 11.5.1893, or twenty-five days previously, still contained living typhoid bacilli, both in that portion of the water which had been preserved at 19° C. as well as in that kept at 6° C., the number and vital activity of the typhoid bacilli being apparently greater in the latter than in the former. 2. The unfiltered Thames water infected with the B. coli communis on the same date also still contained these bacilli in a living state, both in that portion of the water kept at 19° C. as ivell as in that maintained at 6°G. 3. The unfiltered uninfected Thames water ivhich had been maintained under exactly similar conditions, contained no bacteria which after careful examination could be mistaken either for the typhoid bacillus or the B. coli communis. The unfiltered infected waters were again examined on 14.6.1893, with the following results : — Examination by Phenol Broth-culture on 14.6.1893. Number of broth tube. Water used for cultivation with phenol broth. Quantity of water taken in c.c. Quantity of phenol solution added to 10 c.c. broth. Remarks. Unsterilised Typhoid-infected Thames. (85) (86) Flask 11.. Flask JR.. 1-0 1-0 3 drops )> Turbid in 48 hours. Plates poured. »> » Unsterilised Coli-infected Thames. (87) (88) Flask IT.. Flask IE.. 1-0 1-0 3 drops M Turbid in 24 hours. Plates poured. •) »> The plate cultivations made from the above phenol broth tubes gave the following results : — . Report on the Bacteriology of Water. 429 Uroth tube. (85.) Unsterilised Typhoid-infected Thames, FlasJc 1 I. (Typhoid Absent.) The plate exhibited many liquefying colonies as well as a large number of very small colonies ; the latter were placed on potatoes, and yielded light brown growths ; on inoculation into gelatine tubes liquefaction took place, therefore certainly not typhoid. (86.) Unsterilised Typhoid-infected Thames, Flask 1 B. (Typhoid Absent.) The small colonies on the plate which alone exhibited any resemblance to typhoid were examined as in No. 85 above, and yielded exactly similar results, therefore certainly not typhoid. (87.) Unsterilised Coli-infected Thames, FlasJc 1 Incubator. (B. coli Present.) Almost pure cultivation, with numerous typical extension colonies like B. coli ; these yielded characteristic growth on potatoes, gave the indol reaction, and the gas-bubbles in gelatine tube. B. coli, therefore, present. (88.) Unsterilised Coli-infected Thames, Flask 1 Refrigerator. (B. coli Present.) Exactly similar results with this as with No. 87 above. B. coli, therefore, present. From these examinations it appears that on 14.6.1893, — 1. The typhoid bacillus was no longer demonstrable in the Unsterilised Thames water, which had been infected with it thirty-four days pre- viously. It had disappeared both in that portion of the water which had been preserved at a winter, as well as in that kept at a summer, tempera- ture. 2. The B. coli communis, on the other hand; was easily demonstrable in similar water which had been preserved under precisely the same con- ditions. These infected Unsterilised Thames waters were again submitted to examination on 21.6.1893, with the following results : — 2 o 2 430 Profs. Percy Frankland and Marshall Ward. Examination by Phenol Broth-culture on 21.6.1893. CU fl) O.S * s ° -n 63 § 2 te^ Water used for cultivation with phenol broth. Quantity of water taken in c.c. Quantity of phenol solution added to 10 c.c. broth. Eemarks. Typhoid-infected Unsterilised Thames. (86A) Flask 11 .. I'O 3 drops Turbid in 48 hours. Plates poured 23.6.1893. (87A) Flask IE.. I'O » _ Not turbid on 4.7.1893. Coli-infected Unsterilised Thames. (94) Flask 2 I .. 1-0 3 drops Turbid in 24 hours. Plates poured 22.6.1893. (95) Flask IE.. 1-0 » » » » » The plate cultivations made from the above phenol broth tubes gave the following results : — Broth tube. (86A.) Typhoid-infected Unsterilised Thames, Flash 1 Incubator. (Typhoid Absent,) The colonies were inoculated on to potatoes, on which a light brown growth extending over the whole potato was obtained. On inoculation into gelatine tubes liquefaction followed, therefore certainly not typhoid. No bubbles of gas appeared in the gelatine. (94.) Coli-infected Unsterilised Thames, Flaslc 2 Incubator. (B. coli Present.) The plate had the appearance of being a pure cultivation of B. coli, with the characteristic colonies, yielding characteristic growth on potatoes, also indol reaction, and gas bubbles on inoculation into melted gelatine tubes. B. coli, therefore, present. (95.) Coli-infected Unsterilised Thames, Flaslc 1 Refrigerator. (B. coli Present.) Exactly similar results were obtained with this as with No. 94 above. B. coli, therefore, present. Report on the Bacteriology of Water. 431 Thus, on this day (21.6.1893) as on the occasion of the previous examination (14.6.1893), the typhoid 'bacilli were no longer demonstrable in the unsterilised Thames water into which they had been introduced on 11.5.1893. The B. coli communis, on the other hand, was again easily discoverable under the same circumstances, i.e., forty days after its intro- duction into the unsterilised Thames water. The typhoid-infected unsterilised Thames water was again simi- larly examined on 26.6.1893, with the following results : — Examination by Phenol Broth-culture on 26.6.1893. •S.S *" p s> — f r- Water used for cultivation Quantity of water taken in Quantity of phenol solution added Remarks. S -s ? 2 \4£> phenol broth. c.c. to 10 c.c. broth. Typhoid-infected Unsterilised Thames. (96) Flask 11.. 1-0 3 drops' Turbid in 48 hours, Plates poured. (97) Flask 1 R . . 1-0 " The plate cultivations made from the above phenol broth tubes yielded the following results : — Broth tube. (96.) Typhoid-infected Unsterilised Thames, Flask \ Incubator. (Typhoid Absent.) The plates contained some small colonies presenting some resemblance to typhoid ; on being transferred to potatoes they yielded light brown growths unlike typhoid, and on being inoculated into gelatine green fluorescence without liquefaction was obtained. These colonies were, therefore, not those of typhoid. (97.) Typhoid-infected Unsterilised Thames, Flask 1 Refrigerator. (Typhoid Absent.) The plates exhibited two types of colony, firstly, liquefying ones which could not be typhoid, and, secondly, small depth colonies, which on transference to potatoes gave thick greyish- brown growths wrinkled in parts, and on inoculation into gelatine tubes caused subsequent liquefaction. These were, therefore, not typhoid. 432 Profs. Percy Frankland and Marshall Ward. This examination, made on 26.6.1893, confirmed, therefore, the two previous examinations of 21.6.1893 and 14.6.1893, which showed that the typhoid bacillus, introduced on 11.5.1893, was no longer discoverable in the unsterilised Thames water. The final examination of these waters was made on 5.7.1893, thus — Examination by Phenol Broth-culture on 5.7.1893. Number of broth tube. Water used for cultivation with phenol broth. Quantity of water taken in c.c. Quantity of phenol solution added to 10 c.c. Eemarks. broth. Unsterilised Uninfected Thames. (106) Flask 1 I .. 1-0 3 drops Turbid in 48 hours, Vlates poiired 7.7.1893. (107) » IB. 1-0 „ Ditto. (119) Flask 11.. 1-0 5 drops Not turbid in 6 days. (120) „ IE.. i-o » Ditto. Typhoid infected Unsterilised Thames. (108) Flask 11.. 1-0 3 drops Turbid in 72 hours. Plates poured 8.7.1893. (109) „ IE.. i-o » Turbid in 48 hours. Plates poured 7.7.1893. (121) Flask 11.. 1-0 5 drops Not turbid in 6 days. (122) „ IE.. 1-0 » Ditto. Thus, whilst the broth tubes containing only 3 drops of phenol still became turbid when inoculated with both uninfected and in- fected waters, the broth tubes to which 5 drops of the phenol solution were added remained clear after inoculation with both waters. The plate cultivations made from the above broth-tubes, which became turbid, yielded the following results : — Broth tube. (106.) Unsterilised Uninfected Thames, Flask I Incubator. The plates exhibited a large number of highly fluorescent expansion colonies without liquefaction ; some of the least fluorescent of these colonies, and bearing therefore a faint resemblance to typhoid, were transferred to potatoes, on which they gave rise to light brown, sharply demarcated growths, quite unlike typhoid. No gas bubbles were produced on in- oculation into melted gelatine tubes. Report on the Bacteriology of Water. 433 Broth tube. (107.) Unsterilised Uninfected Thames Flask 1 Refrigerator. The plates exhibited an apparently pure cultivation of a liquefying organism, the colonies having a strong resemblance to those of B. liquidus (see 2nd Report, p. 186). Some of the colonies which had not yet caused liquefaction were trans- ferred to potatoes, on which they gave rise to thick, whitish, waxy growths, quite unlike those of typhoid. (108.) Typhoid-infected Unsterilised Thames, Flask 1 Inctibator. (Typhoid Absent.) The plates contained fluorescent expansion, liquefying and small dot colonies. All three types of colony were transferred to potatoes, on which the small dot and fluorescent expansion colonies gave rise to light brown sharply marked growths, and the liquefying colonies to shining slimy colourless growths ; none of these were, therefore, similar to those of typhoid. (109). Typhoid-infected Unsterilised Thames, Flask 1 Refrigerator, (Typhoid Absent.} The plates appeared to be pure cultivations ; the depth colonies were small dots, often oval in shape, and the surface colonies resembled very small " milk-drops ; " even on the plate in which the colonies were few and far between there were none of the typical expansion colonies resembling typhoid. On transference to potatoes, shining slimy growths, having the appearance of drops of water, were obtained. This final examination of these Unsterilised typhoid-infected Thames waters therefore again confirmed the previous results, which may be thus summarised: — (1.) The Unsterilised uninfected Thames water, collected at Hampton on 5.5.1893, contained throughout the entire course of the series of experiments no bacteria resembling either the typhoid bacillus or the B. coli commtmis. (2.) The Unsterilised Thames water infected with typhoid on 11.5.1893, was still found to contain living typhoid bacilli on 5.6.1893, or twenty-five days after infection, whilst on 14.6.1893, thirty-four days after infection, they were no longer demonstrable. (3.) These remarks apply equally to the waters preserved at a winter and a summer temperature respectively, 434 Profs. Percy Frankland and Marshall Ward. although there is some evidence that on the last day of their detection the typhoid bacilli were present, either in larger numbers or in a more active state in the water maintained at the winter than in that at the summer temperature. (4.) The same unsterilised Thames water infected with the B. coli communis on the same day, and kept under pre- cisely similar'conditions of temperature, &c., was still found to contain living coli bacilli forty days after their introduction. These bacilli doubtless' persisted in the living state, even for a much longer period of time than this, no later examinations being made; and on the occasion of their last detection they did not appear to have lost any of their original vitality, as they still promptly reacted with the phenol broth test. Experiments on the Influence of the Addition of Salt to the JJnst&riliseA Typhoid-infected Thames Water. In the recent cholera epidemic at Hamburg, it is now almost universally recognised that the most important agent in distributing the zymotic poison was the highly polluted and nnfiltered water of the River Elbe. This water, moreover, during the epidemic was found to be unusually rich in salt, in fact at times it was distinctly brackish in character. Thus a sample of Hamburg water sent to me by Mr. Ernest Hart in October, 1892, and which I submitted to analysis, had the following composition : — Sample of Hamburg Water received from Mr. Ernest Hart. • Results of Analysis expressed in Parts per 100,000. Total solid matters. Organic carbon. Organic nitrogen. Ammonia. Nitrogen as nitrates and nitrites. Chlorine. Hardness. Free. Albumin- oid. Tempo- rary. Per- manent. Total. 78-00 0-926 0-088 0-030 0-047 0 31-3 4-1 13-7 17-8 Oxygen consumed by organic matter, as measured by reduction of a solution of permanganate acting for three hours in the cold = 0-366. The water was very turbid, depositing a quantity of brown suspended matter. The high percentage of salt is due to the waste liquors which are discharged from the Stassfurth salt works and other factories into the Elbe and its tributaries. Now it has been more recently shown (Trenkmann, " Beitrag zur Biologie des Kommabacillus," ' Centralbl Report on the Bacteriology of Water. •135 f. Bakteriologie,' vol. 13 (1893), p. 313) that the addition of sodium chloride, and of other salts in certain proportions, to water containing the cholera bacilli causes a most remarkable multiplication of the latter, as may be seen from the following table, given by Trenkmann, which exhibits the effect of making such additions to a sterilised well water, which had been purposely infected with, cholera bacilli. Addition of various Salts to Sterilised Well Water containing Cholera Bacilli. The waters were maintained at 21 — 24° C. (Trenkmann.) Number of cholera bacilli in 1 needle loop. After 24 hours. After 8 days. (1.) (2.) (3.) (4.) (5.) (6.) (7.) (8.) (9.) (10.) (11.) (12.) (13.) (14.) (15.) (16.) (17.) 10 c.c. sterile well „ +1 drop* ] + 2 drops + 3 „ + 1 drop + 2 drops + 3 „ + 1 drop + 2 drops + 3 „ + 1 drop -t- 2 drops + 3 „ + 1 drop + 2 drops + 3 „ + 2 „ 1 drop 580 520 6,120 9,240 15,000 1,740 6,600 17,160 8,040 6,660 20,940 3,360 7,560 6,540 7,440 28,680 31,560 54,720 } • 12,480 19,560 10,440 10,920 1,460 2,260 4,040 14,760 16,080 0 per cent, sodium chloride .... » i) )> • • • • » » » .... , nitrite » » >> disodium phosphate . » >» • ) » !) • sodium carbonate . . . i » » • • • » » » • • • ,, „ chloride and 10 per cent, disodium phosphate * 25—27 drops = 1 c.c. These results show that the presence of an unusually high propor- tion of salts may not improbably have played an important part in the distribution of cholera by means of the water of the Elbe in the recent Hamburg epidemic, and possibly also by means of the Thames water in some of the former London epidemics, as at that time the metropolitan supply was in part derived from the tidal portion of the river. In view of these circumstances, it appeared to me to be of con- siderable interest to ascertain whether and in what way the behaviour of the typhoid bacillus in water is affected by additions of salt, and, with this object, the following experiments were undertaken : — 436 Profs. Percy Frankland and Marshall Ward. Preparation of the Saline Waters. — The saline waters employed were of three different sti-engths — (a.) Containing O'l per cent, sodium chloride. (6.) „ I'O (c.) „ 3-0 „ Three portions of pure sodium chloride, weighing respectively 0'5, 5'0, and 15'0 grams, were sterilised in the air-oven at 150° C. for several hours, and then placed in three sterile 500 c.c. measuring flasks. To each of these was added sufficient of the typhoid-infected unsterilised Thames water (p. 410) to dissolve the salt, and the solution was in each case further diluted to the 500 c.c. mark with the same infected unsterilised Thames water. These saline waters were then distributed in smaller flasks plugged with cotton-wool, and two of the latter of each particular strength were placed in the refrigerator at 6 — 8° C., and two in the incubator at 19° C. Thus there were Unsterilised typhoid-infected Thames water f 2 flasks refrigerator + O'l per cent, salt 1 2 „ incubator Unsterilised typhoid-infected Thames water f 2 flasks refrigerator + 1'0 per cent, salt L 2 ,, incubator Unsterilised typhoid-infected Thames water f 2 flasks refrigerator + 3'0 per cent, salt 12 „ incubator These saline waters, like the others, were prepared on 11.5.1893, and were examined by gelatine plate cultivation at frequent intervals subsequently. The results obtained must obviously be compared with those from the typhoid-infected unsterilised Thames water, to which no salt was added, and which have already been recorded in the Table on pp. 415 and 416, but which are again given, so as to fcicilitate comparison, in the first column of the following table : — Report on the Bacteriology of Water. b O | ^^ §^ O "M" O ^ O *~* Timents 93 O O SH L •s'J I § g >, co ^ oo^ 2e O o3- of II • 8 ft M O b H CO O li | II 8-^3 It II || ->i "S ? C3 0 03 oo" co IS 8e^ * l-l co" co"^' 10" co" C 0 a o O 5 a o 0 «M O 4* o .a°o3 8 co" ij? 11 8=2, O^ e3 It 0 -4- 0 T O d e o h o I" 1-1 •H 3 _o [-5 i o r-l g ^ 0^ O co ^ 0^. j ^ '•a "JS h 03 ^ — m O o o S> O !•*"* O k» CD 03 11 a 1 CD •5 •+3 CO j 3 1 J 03 H 03 C h- 1 cxf IS »O CO ^S CO (M & 1 0 Oi 0 00 rC ^ .-S o £ rH rrl rH colonies 0 h O jj .&>§ 1 II §'^D> 41 Q ^ §>» „ 03 £ - C 'f i | S ^ •> a S co" CO w o co S? N S- i c3 o h 6 V N^ ^ CO *s_^ Ol ^ — ^ ^ ^— S ^ " 0 o ftj M -*; o .•s g S ft ti ^ ^ fl £3 O g ^ Q^ O ^ _ a ^ 0 « 8 li oo" 8,1 S>-> 08 O_ a* |1 f- i T: 1 1 O 33 M 00 co co co_ «e -*0^ •s 1 ^: a> S fe 1 c6 ;- Q S •a O c3 O^ S'lo' §'cox §5 a EH S^g o^ ^ O_ 03 - * „ ^ ^ _^ ' O • ^> 83 otT VO *^ Oi *"^ (^ 'O ^5 ^^ r 0 c3 t§=a t>- ^ w. In =S 2s, S^ CD ^5 3 M •— I -w O li , 1 «H S GO '^ 1 o Id S 03 §* ^ §c? oo^ W eo, 11 i-l CO s** 11 iM N, * -*i 0 0) <+-< 6 • rH -3 C p-j 0 O 'o 3 '43 a- •5 o3 T: co CO C5 CO i CO OS .£: ° oo oo co 00 co >» c 2 £ 0 S o id XO US VO CD C_( J° *: i— i ^ CM OJ d OJ ® i— i i-H CN Turbid in 48 hours. Turbid in 24 hours. Plates poured 14.6.1893. (80) » • • 0-5 » Turbid in 48 hoursr Typhoid-infected unsterilised Thames + I'O per cent. NaCl. (75) Flask 11.. i-o 3 drops Turbid in 24 hours. Plates poured 14.6.1893. (76) (81) Flask 1 E . . 0-5 1-0 » Turbid in 48 hours. Turbid in 24 hours. Plates poured 14.6.1893. (82) „ ..! 0-5 » Turbid in 24 hours. Typhoid-infected unsterilised Thames + 3 per cent. NaCl. (77) Flask 11.. 1-0 3 drops Turbid in 48 hours. Plates poured 15.6.1893. (78) (83) (84) Flask IE.'! » • • 0-5 1-0 0-5 3) » Turbid in 48 hours. Turbid in 48 hours. Turbid in 48 hours. Plates poured 15.6.1893. The six turbid broth tubes, indicated above as selected for further examination by gelatine plate culture, yielded the following results :-- Broth tube. (73.) Typhoid-infected Unsterilised Thames + O'l per cent. NaCl, Flask 1 Incubator. (Typhoid Absent.) The colonies yielded pink growths on potatoes and liquefied gelatine ; they were therefore not those of typhoid. (79.) Typhoid-infected Unsterilised Thames + O'l per cent. Nad, Flask 1 Refrigerator. (Typhoid Absent.) The colonies presented some resemblance to typhoid, as did also the growths on potatoes obtained from them. No indol reaction. On inoculating into gelatine tubes, the latter were found to very slowly liquefy on long keeping. On inoculating VOL. LVI. 2 H 444 Profs. Percy Frankland and Marshall Ward. from such a liquefied tube on to potatoes, the same typhoid-like growth was obtained. This organism, which was again sub- sequently met with, might easily lead to a false diagnosis of typhoid, unless the gelatine tubes were preserved for some time. Broth tube. (75.) Typhoid-infected Unsterilised Thames + 1 per cent. Nad, Flask 1 Incubator. (Typhoid Absent.) The colonies on transference to potatoes gave rise to light- brown growth, not like typhoid. (82.) Typhoid-infected Unsterilised Thames + 1 per cent. NaCI, Flask 1 Refrigerator. (Typhoid Absent.) The colonies liquefied the gelatine, and gave rise to easily visible growths on potatoes. (78.) Typhoid-infected Unsterilised Thames + 3 per cent. NaCl,. Flask 1 Incubator. (Typhoid Absent.) The colonies gave rise to pink growths on potatoes ; there- fore certainly not typhoid. (84.) Typhoid-infected Unsterilised Thames + 3 per cent. NaCl,. Flask 1 Refrigerator. (Typhoid Absent.) Same results as with No. 78 above. Thus in the case of none of these saline waters could a diagnosis of typhoid bacilli be made on 13.6.1893; by reference to p. 433 it will be seen also that on 14.6.1893 the typhoid bacilli were no longer demonstrable in the Unsterilised Thames water to which no salt had been added. These saline waters were again examined by phenol broth-culture on 21.6.1893, with the following results :• — Report on the Bacteriology of Water. Examination by Phenol Broth-culture, 21.6.1893. 445 Number of broth tube. Water used for cultivation with phenol broth. Quantity of water taken in c.c. Quantity of phenol solution added to 10 c.c. broth. Remarks. Typhoid-infected Unsterilised Thames + O'l per cent. M" aCl. (88) (91) Flask 11 .. „ 1R-. 1-0 1-0 3 drops Turbid in 48 hours. Plates poured 23.6.1893. » 5> )l » Typhoid-infected Unsterilised Thamet + 1^0 per cent. NaCl. (89) (92) Flask 11.. ii IK- 1-0 1-0 3 drops » Turbid in 48 hours. Plates poured 23.6.1893. » >» >» » Typhoid-infected Unsterilised Thames + 3'0 per cent. NaCl. (90) (93) Flask 11 .. „ IK.. 1-0 1-0 3 drops » Turbid in 48 hours. Plates poured 23.6.1893. j> » )> » Thus, on this occasion, all the waters reacted in forty-eight hours with the phenol broth-solution. The following results were obtained on plate cultivating the turbid broth tubes : — Broth tube. (88.) Typhoid-infected Unsterilised Thames + O'l per cent. NaCl, Flask 1 Incubator. (Typhoid Absent.) The colonies on transference to potatoes yielded light brown growths, which were certainly not due to typhoid. (91.) Typhoid-infected Unsterilised Thames + O'l per cent. NaCl, Flask 1 Refrigerator. (Typhoid Absent.) The potato growths from colonies somewhat resembled typhoid ; there was also no indol reaction, but gelatine tubes were slowly liquefied by the organism, which was, therefore, not typhoid. (See similar experiences with No. 79, p. 443.) (89.) Typhoid-infected Unsterilised Thames + 1 per cent. NaCI, Flask 1 Incubator. (Typhoid Absent.) The potato growths from colonies were thick and brown in colour, quite unlike those of typhoid. 2 H 2 446 Profs. Percy Frankland and Marshall Ward. Broth tube. (92.) Typhoid-infected Unsterilised Thames + 1 per cent. NaCl, Flask 1 Refrigerator. (Typhoid Absent.) The potato growths from colonies were pink in colour, and, on inoculation into gelatine, liquefied it. Certainly not typhoid. (90.) Typhoid-infected Unsterilised Thames + 3 per cent. NaCl, Flask 1 Incubator. (Typhoid Absent.) The potato growths from colonies were pinkish-white in colour and much too conspicuous for typhoid. (93.) Typhoid-infected Unsterilised Thames + 3 per cent. NaCl, Flask 1 Refrigerator. (Typhoid Absent.) The potato growths from colonies were of a dirty white colour, and on inoculation into gelatine the latter was lique- fied. Therefore, certainly not typhoid. Thus, again, on this occasion, 21.6.1893, it was found impossible to demonstrate the presence of typhoid bacilli in any of these saline waters. The saline waters were again examined by phenol broth-culture on 26.6.X893, with the following results :— Examination by Phenol Broth-culture, 26.6,1893- Number of broth tube. Water used for cultivation with phenol broth. Quantity of water taken in c.c. Quantity of phenol solution added to 10 c.c. broth. Eemarks. Typhoid-infected Unsterilised Thames + O'l per cent. NaCl. (98) (101) Flask 11.. 1-0 3 drops Turbid in 48 hours. Typhoid-infected Unsterilised Thames + I'O per cent. NaCL (99) (102) Flask 11.. 1-0 3 drops Turbid in 48 hours. Typhoid-infected Unsterilised Thames '+ 3'0 per cent. NaCl. (100) (103) Flask 11.. 1-0 3 drops Turbid in 48 hours. These turbid broth tubes were submitted to plate cultivation on 28.6.1893, and the following results obtained: — Report on the Bacteriology of Water. 447 Broth tube. (98.) Typhoid-infected Unsterilised Thames + Ol per cent. NaCl, Flask 1 Incubator. (Typhoid Absent.) The plates exhibited liquefying and small colonies respec- tively. From the small colonies, potatoes were inoculated, a brown and slimy growth being obtained, on inoculation from which into gelatine tubes, a blue, fluorescent, non- liquefy ing growth resulted. Thus nothing like typhoid was present. (101.) Typhoid-infected Unsterilised Thames + O'l per cent. N"aCl, FlasJc 1 Refrigerator. (Typhoid Absent.) The potato-growths obtained from the colonies bore some resemblance to typhoid, although the surface was rather too shining. On inoculating into gelatine tubes, it was found that very slow liquefaction took place. This was, therefore, the same organism which had been several times before met with in this flask. Under the microscope the bacilli also present some re- semblance to typhoid, but they have squarer ends. (99.) Typhoid-infected Unsterilised Thames + 1 per cent. NaCl, Flask 1 Incubator. (Typhoid Absent.) On the plates there were small colonies forming surface ex- pansions with green fluorescence ; these gave rise, on potatoes, to thick, greyish-brown growths quite unlike typhoid. (102.) Typhoid-infected Unsterilised Thames + 1 per cent. NaCl, Flask 1 Refrigerator. (Typhoid Absent.) The plates contained both liquefying and small colonies, the latter, on transferring to potatoes, gave strong, conspicuous, and flesh-coloured growths quite unlike typhoid. (100.) Typhoid-infected Unsterilised Thames + 3 per cent. NaCl, Flask 1 Incubator. (Typhoid Absent.) The plates contained small colonies forming surface expan- sions, which gave light brown growths on potatoes quite unlike typhoid. (103.) Typhoid-infected Unsterilised Thames + 3 per cent. Nad, Flask 1 Refrigerator. (Typhoid Absent.) The plates contained liquefying and small colonies respec- tively, the potato-growths from the latter were strong, thick, waxy, and greyish- white, quite unlike typhoid. Thus, on this occasion (26.6.1893) again, the presence of typhoid bacilli could not be demonstrated in any of these saline waters. The final examination of these saline waters was made on 5.7.1893, with the following results : — 448 Profs. Percy Frankland and Marshall Ward. Examination by Phenol Broth-culture on 5.7.1893. AAf of ov Quantity •si jj vv drei used for cultivation Quantity of water taken. of phenol solution added Remarks. So 3 b ** with phenol broth. c.c. to 10 c.c. broth. Typhoid-infected Unsterilised Thames + O'l per cent. NaCl. (113) Flask 11.. 1-0 3 drops Turbid in 48 hours. (123) » • • » 5 „ Not turbid in 6 days. (116) Flask IE,. j) 3 „ Turbid in 48 hours. (126) » • • 9) 5 „ Turbid in 6 days. Plates poured. Typhoid-infected Unsterilised Thames + 1 per cent. NaCl. (114) Flask 11 .. 1-0 3 drops Turbid in 48 hours. (124) » » 5 „ Not turbid in 6 days. (117) Flask IE.. » 3 „ Turbid in 24 hours. Plates poured 6.7.1893. (127) » • • » 5 „ Not turbid in 6 days. Typhoid-infected Unsterilised Thames + 3 per cent. NaCl. (115) Flask 11.. 1-0 3 drops Turbid in 72 hours. (125) » • • li 5 „ Not turbid in 6 days. (118) Flask IE.. » 3 „ Turbid in 48 hours. (128) » • • 5 „ Not turbid in 6 days. Plate cultivations were made of the turbid broth tubes with the following results : — Broth tube. (113.) Typhoid-infected Unsterilised Thames + O'l per cent. NaCl, FlasJc 1 Incubator. (Typhoid Absent.) The plates had the appearance of being pure cultivations of an organism producing fluorescent expansion colonies without liquefaction. The colonies in which fluorescence was least conspicuous and which had, therefore, most chance of being typhoid, were transferred to potatoes, on which they yielded a greyish-brown growth sharply distinguishable from the potato and quite unlike typhoid. (116.) Typhoid-infected Unsterilised Thames + O'l per cent. NaCl, Flask 1 Refrigerator. (Typhoid Absent.} The plates exhibited a number of small colonies giving rise to surface expansions something like those of typhoid. These Report on the Bacteriology of Water. 449 colonies, on being transferred to potatoes, again yielded incon- spicuous colourless growths very similar to those of typhoid. Negative results were also obtained with the milk, indol, and gas-bubble reactions. It was found, however, that the gelatine-tubes inoculated from these typhoid-like colonies underwent very slow liquefaction. This was, therefore, ob- viously the same organism which had been repeatedly obtained before from this flask, and the superficia? resem- blance to typhoid of which has been already referred to (see pp. 445, 447). Broth tube. (114.) Typhoid-infected Unsterilised Thames + 1 per cent. NaCl, Flask 1 Incubator. (Typhoid Absent.) The plates contained fluorescent non-liquefying colonies. The less obviously fluorescent colonies on transference to potatoes yielded growths which were not sufficiently different from typhoid to decide, whilst the milk, indol, and gas-bubble tests were also negative. On inoculation into gelatine tubes, however, the latter became fluorescent. (117.) Typhoid-infected Unsterilised Thames + 1 per cent. Flask 1 Refrigerator. (Typhoid Absent.) The plates exhibited a pure cultivation of bacillus, giving rise to small, cup-shaped, liquid colonies, probably 13. liquidus (Percy Frankland), which is very frequently found to survive in the 3-drop phenol broth-cultures. (115.) Typhoid-infected Unsterilised Thames + 3 per cent. NaCl, Flask 1 Incubator. (Typhoid Absent.) The plates exhibited small, milk-drop colonies, with slight tendency to expand. Microscopic examination showed them to be due to bacilli thinner than the typhoid bacillus, and on potatoes they yielded light yellow, shining growths, unlike those of typhoid. (118.) Typhoid-infected Unsterilised Thames + 3 per cent. NaCl, Flask 1 Refrigerator. (Typhoid Absent.) The plates contained a number of liquefying colonies, apparently B. liquidus (Percy Frankland), also some small colonies, some of which gave rise to very small surface ex- pansions and very small milk-drops. The potatoes inoculated from the milk-drop colonies yielded flesh-coloured growths, whilst those from the expansion colonies were not decisive. 450 Profs. Percy Frankland and Marshall Ward. Plates were again ponred from these potatoes, and very small surface-expansion colonies again obtained, which, however, again were not like those of typhoid, and gelatine tubes in- oculated from these yielded, in course of time, brown surface growths quite unlike typhoid. The results of these experiments with the unsterilised Thames water, to which O'l, 1, and 3 per cent, of common salt respectively had been added, are instructive in more ways than one. Thus : — (1.) They show that the addition of the salt stimulated the growth and. multiplication of some of the water bacteria to an enormous extent, the effect being the most marked with the largest proportion of salt (3 per cent.), whilst the water to which only 0-1 per cent, of salt was added, behaved almost exactly like the untreated Thames water. (2.) This multiplication was, as usual, followed by decline, but the saline waters remained, even after six weeks, more densely, and with the larger proportion of salt much more densely, populated than the Thames water to which no salt was added. (3.) As regards the effect of the salt addition on the typhoid bacilli present in the water, the experiments show that they were most prejudicially influenced. Thus whilst on the eighteenth day after infection the typhoid bacilli were easily demonstrable in the ordi- nary unsterilised Thames water to which no salt had been added, and also in that which had received 0-1 per cent, of salt and which had been kept at 6 — 8° C., they were not discoverable in any of the waters to which 1 and 3 per cent, salt had been added, nor in that which had received only 0-1 per cent, salt, but which had been kept at the summer temperature of 19° C. (4.) In the case of the 3 per cent, salt addition, I am of opinion that the rapid disappearance of the typhoid bacilli is largely due to the direct action of the salt, whilst in the case of the smaller proportions it may also be due to the great multiplication of some of the common water bacteria. Further experiments on the behaviour of typhoid bacilli in Thames water, to which salt had been added, were subsequently made (see pp. 530, et seq.), they proved entirely confirmatory of the results just recorded above, both as to the stimulation of the multiplication of the Report on the Bacteriology of Water. 451 water bacteria, and as to the more rapid disappearance of the typhoid bacilli. The directly prejudicial action of the salt on the typhoid bacillus was farther demonstrated by the addition of salt to steam- sterilised Thames water containing typhoid bacilli. Behaviour of the Typhoid Bacillus and of the B. coli communis in Steam-sterilised Thames Water (First Series of Experiments). It will now be interesting to consider the behaviour of these bacilli in the precisely parallel series of experiments made with the same sample of Thames water which had been previously sterilised by steam. These experiments are of importance more especially because they enable us to ascertain whether the water contains the necessary food materials for these particular bacteria, as, owing to the absence of other forms, it is now possible to determine how the actual numbers of these bacteria are affected by residence in the water. The infection and distribution of these steam-sterilised waters has already been described on pp. 410 and 411, so that I can pass at once to their subsequent examination, made both by gelatine plate and phenol broth-culture, at different intervals of time. In the table (p. 452) the fate of the typhoid bacilli introduced into the steam-sterilised Thames water on 11.5.1893, is followed over a period of seventy-six days, and it will be seen that during this time their numbers underwent an almost continuous decline. Thus, whilst at the outset they were present to the number of, in round numbers, 70,000 per cub. cm., at the end of this period their presence was only just demonstrable by gelatine plate cultivation at all, and not more than from 6 — 12 were discoverable in 1 cub. cm. The most important feature in this chronicle of their deportment is the circum- stance that they exhibited no multiplication or increase in numbers during their residence in the water, clearly showing, therefore, that the latter did not afford the nutriment and other conditions necessary for the proliferation of the typhoid bacilli. In fact, the decline from the commencement is an unbroken one, with the exception of the solitary observation of an increase from 27,000 on 22.5.1893, to 42,000 on 29.5.1893, in the case of the water maintained at a winter tem- perature. Whatever may have been the cause of this increase, it is not sufficiently great to be comparable with that extensive reproduc- tion which takes place in the case of those bacteria which are the natural inhabitants of water. Moreover, that the water was not suited to the well-being of the typhoid bacilli was further testified to by the fact that the colonies on the gelatine plates became smaller, feebler, and more degenerate as time went on, until on the occasion of the last few examinations, they were, with difficulty, recognisable a.s typhoid colonies, and exhibited a marked disinclination to form the characteristic growths expanding over the surface of the gelatine. 452 Profs. Percy Frankland and Marshall Ward. o g Jfc> d i— i 03 3 J « hied from rigerator fl O 1 lO 1 O | W o 1 »o r~- 0 •a ^ E c3 O a, « ^ o ~% ^H fA o S 'o 2 II *G 3 o O 1 If 11 0 S tao • 2 =4-1 I> 2 r*3 o *b G 0 rS 'o 2 r3 'O ® 'o S3 r&i '&"&! ^H 03 o ~~~ ^» ^> Q^ 03 _, C4H O -*J i~& S-S o h 3 1 1 <£ tS ^S tD 03 92" (>« ^ 10" 10" ^W «*H «*H 3 rH '® i-Q — -* Kl 5SJ 1-1 03 C3 03 rO g C JK ^> ^> & >j S^ s to (3 h- 1 0 "a "a •" OO O ^ o IM d ^ _o o o '-2 1 H|0 JO JH "i-o "|2"|2 -!2-£ H2-|2 H»H» "P"fc * 11 o ® '43 3 o 6 s3 11 S 03 -1*0 11 rt 83 o « 83 03 « C w o3 G C 83 83 O O 1 co t* >H o> MH "I- HO-IO -|0-.|iO -««|n rH r-l "e3 'QI P E C TJ IJ o "o3 -^2 ^ .0 ** co eo -* CO •* •* Tfl -^1 CO X X X •* CO ia S ^ H 0 ^ ™ t«9 o3 •S 1 o 8} ^ o _O H M rt M PH M & N rt g '1 1 1-1 in "s S to o 1 1 IB B n i 1 O AH a i— i 1 tf 13 c 11 .- " t^ " t^ 6 r-l s c3 g o o o o g ill 1 t5 ° fl 1 1 E o pH oo" s 05 oo" pO ^( (JC1 oo O5 jm^ O o £3 ^ | g CD (M s lO in to " CD t> " *> " t; S u S I— J CD co Q to CD 00 MB "3 •£ ki 1-1 r-l < (57) Flask 1 I 0-5 5 drops Turbid in 24 hours. Plates poured 6.6.1893. (58) 5) 1-0 „ Turbid in 24 hours. (59) Flask 1 E 0-5 „ Turbid in 24 hours. Plates poured 6.6.1893. (60) .. i-o » Turbid in 24 hours. These examinations by phenol broth-culture show, therefore, that, on 5.6.1893, the typhoid-infected steam-sterilised Thames water reacted already in twenty-four hours with the test, irrespectively of whether 3 drops or 5 drops of phenol solution were added to the 10 c.c. of broth; whilst, by referring back to p. 427, it will be seen that the typhoid-infected unsterilised Thames water only reacted in twenty- four hours, even with the 3 drops of phenol solution, in the case of the water which had been kept at the winter temperature, whilst the summer temperature water only reacted after forty-eight hours. From these comparative tests it can be inferred, therefore, that, on the date in question, the typhoid bacilli were in a less vigorous con- dition in the unsterilised than in the sterilised water. Of the plate cultivations made from the turbid broth tubes Nos. 61, 57, and 59, those from No. 61 yielded typical typhoid colonies which were confirmed by growth on potatoes and by negative results with the indol and gas-bubble tests ; the plates from Nos. 57 and 59, Report on the Bacteriology of Water. 457 on the other hand, yielded the typical colonies of the B. coli com- munis, and these were further confirmed by growth in potatoes, and by positive results with the indol and gas-bubble tests. The above phenol broth-culture tests were all made with 0*5 and I/O c.c. of the water, but on the same day (5.6.1893) some further experiments were made to incidentally determine whether much smaller volumes (a single drop) of water would react, and, if so, whether with equal rapidity. Thus — Examination by Phenol Broth-culture of Small Quantities of Infected Water, 5.6.1893. •SJ5 f-l 3 0> ,-. li 11 Water used for cultivation with phenol broth. Quantity of water taken in c.c. Quantity of phenol solution added to 10 c.c. broth. Eemarks. Typhoid-infected Steam-sterilised Thames. (65) (66) Flask 11.. » • • 1 drop >» 3 drops 5 „ Turbid in 24 hours. Very slightly turbid in 24 hours ; tur- bid in 48 hours. (67) (68) Flask IE.. » • • 5) )) 3 „ 5 „ Turbid in 24 hours. Not turbid in 24 hours ; turbid in 48 hours. Coli-infected Steam-sterilised Thames. (69) (70) (71) (72) Flask 11.. Flask IE.. M "• • 1 drop j> » » 3 drops 5 „ 3 „ 5 „ Turbid in 24 hours. )> )> j» j) » » The interest attaching to the phenol broth examinations consists in the circumstance that the actual number of typhoid and coli bacilli present in the volumes of water used can be calculated from the results of the plate cultivations made on the same day (see pp. 452 and 454). Thus, it will be seen from the tables on pp. 456 and 457, that there was, in nearly all cases, practically no difference in the time which elapsed before the phenol broth tubes became turbid, irrespec- tively of whether 0'5 c.c., 1 c.c., or only 1 drop of the same water was employed ; for, even in the 1 drop of the water, it is apparent from the plate cultivations (p. 452) that there must have been upwards of K.OOO typhoid bacilli present, and a still larger number of coli bacilli n those waters infected with this bacillus. 458 Profs. Percy Franldand and Marshall Ward. Another examination by phenol hroth- culture was made of the typhoid-infected steam-sterilised Thames water about one month later, on 5.7.1893 and on 6.7.1893, and for the last time on 25.7.1893. Thus— Examination by Phenol Broth-culture. Number of brotli tube. Water used for cultivation with phenol broth. Quantity of water taken in c.c. Quantity of phenol solution added to 10 c.c. broth. Remarks. Typhoid-infected Steam-sterilised Thames. 5.7.1893 (110) (112) Flask 11.. 5J • ' 1 drop » 3 drops 5 „ Turbid in 24 hours. 48 „ 6.7.1893 (129) (130) (131) (132) Flask IE.. v • • 1 drop 0-5 c.c. » 3 drops 5 „ 3 „ 5 „ Remained clear. » j> Turbid in 72 hours. » )) 25.7.1893 (199) (200) Flask 11.. „ 1R.. 1 -Oc.c. 3 drops 3 „ Turbid in 48 hours. >! » Coli-infected Steam-sterilised Thames. (195) (196) Flask 11 .. „ 1R.. 1-0 c.c. » 3 drops : Turbid in 48 hours. From the above it will be seen that even on 25.7.1893, when the plate cultivations (see p. 452) were only yielding about twelve colonies per c.c., and these colonies of a very feeble and degenerate character, the phenol broth-cultures of 1 c.c. of the water still became turbid in forty-eight hours, and thus revealed the presence of living typhoid bacilli with the greatest facility. On the other hand, when the number of typhoid bacilli in the water is small, it may very easily happen that a phenol broth tube now and again may fail to go turbid (as in the case of Broth tubes 443 and 444, see table above), and it is very necessary, therefore, to exercise great caution, and not to draw conclusions from a single observation, but only after a number of repeated trials. The examinations by phenol broth-culture of these infected steam- sterilised Thames waters thus entirely substantiate the results arrived at Report on the Bacteriology of Water. 459 by the direct method of plate cultivation, and show that both the typhoid and coli bacilli were still present in a living state in this water, irre- spectively of whether it had been preserved at a summer or a winttr temperature, for a period of seventy-five days. Behaviour of the Typhoid Bacillus and of the B. coli communis in Thames Water sterilised by Filtration through Porous Porcelain. (First Series of Experiments^) The preparation and infection of this water has already baen described (see pp. 410 and 411), and, as already indicated, the infected water was placed under precisely the same conditions of tempera- ture, &c., as the steam-sterilised and unsterilised waters. In the periodical examination of this water the following results were obtained : — VOL. LTI. L' I 460 Profs. Percy Frankland and Marshall Ward. o ' r-i 1 a 1 1 8 0 o •S T3 1 S 'a! S •HP 2 'J * QJ o ^Q j^ PH 0 03 o_ 'j^ 00 o t2 S i '3 o *^ 13 o ^0 — -i o PH o ° Hi * ci co oo x (N 00 T—fl eo eo oo III i-H l-t US I-t 10 w" OJ M «i, 1 Report on the Bacteriology of Water. 461 The above results were most unexpected, for they show that, ilthongh as many as 75,000 typhoid bacilli per 1 c.c. were introduced into this water, they were entirely destroyed in five days at 19° C., and had undergone a very large reduction in number at 6 — 8° C., whilst in twelve days they were no longer discoverable in this water kept at the low temperature. On 2.6.1893 sterile broth was added to the flasks, which were then placed in the incubator at 38° C., but even this treatment did not lead to any revivification of the typhoid bacilli, which could neither be detected by plate cultivation nor phenol broth- culture. The same rapid destruction of the B. coli communis was observed in this water, as will be seen from the following table : — 2 i 2 462 Profs. Percy FranHand and Marshall Ward. & CO 3 fo Tj CO S °> .52 co la-S iC o ^^ 1 ,d H rt •8 OS o rM ?H a a fa 1 o 3 O O * I 1 0 -^ '3 ^ O c3 > 1 8 Om O> , QO 1 i 03 n O 0 O O Kj H B 3 H O 3 1 i M =2 d ^ ^» 33 Hf -^-p JH O a> o •* i**i °''a TJ TS c d c d l^g 'o o> ° °H« - 03 »o J-1 ^^ £-1 ® *H rH — -2 "^ r- 1 "i ^ I OJ J5 "§ X> rp (35 «O *^t^ i-l rH — rH l|1 t b O • f E o f .1 BCD 2 i-H rH • *s ^ 44 PH 1 Cj b l! ^ 5 £ O 'I 3 m H i-H ,_, C! O PH C! M •2.2 • 't; "= TJ ? >• c3 c ]« a S 00 1-H § 00 I— 1 CO C5 00 t— 1 00 ° 3 o to LO kO 10 c o 1— 1 O rH CO (M 8 03 "S ? Q ^ AH Report on the Bacteriology of Water. 463 Tims, in the case of the B. coll communis again, there was the same disappearance in five days of the bacilli in the water kept at 19° C., the great diminution in numbers in the same time in the water kept at 6 — 8° C., followed by complete disappearance of the bacilli in this water also by the twelfth day. Similar attempts made by the addi- tion of sterile broth to resuscitate the bacilli in these waters also proved unavailing. These results, showing that the typhoid and coli bacilli were more rapidly destroyed in the porcelain-filtered than in the unsterilised, and far more rapidly than in the steam-sterilised, Thames water, were so surprising that it was necessary to banish every suspicion of some accidental disturbing cause having arisen in these experi- ments. The most obvious suggestion was that the filter itself might have introduced some antiseptic substance into the water. This was, however, highly improbable, as the filter in question had only been previously used for the similar sterilisation of Thames and Loch Katrine waters. In order to abolish this objection, however, the porcelain cylinder was thoroughly scrubbed externally with a tooth- brush, and then upwards of 30 litres of distilled water passed through it. The filter was then stefm sterilised and employed for the filtration of some more of the same Thames water, which was infected with typhoid and coli as below. Thus — (•>« o Q^ »S £ 1 it 6 5 11 •3 0 § < C 8 - - 00 c?o . 2!« ^s^-i o 1 sj-2 CO 00 00 ^i co 00 00 "a 43 a |^-s *, Bq TJ 1 • i * Pi 6 •S° | P3 c ^o 1 •s 1 T '? 3 S S *§ p Q4 DO lu o h ^ b 5 £ *o H- 1 6 I— 1 •J S r-H m H 83 j3 PH a ^s S .° .2 • 'S *^ ® w CO eo CO fe ^ ^^ 05 as O5 -38 oo f— 1 00 !_4 oo oo o'S o CO CO* CD CO ||I 10 H 0 N id i— i O Report on the Bacteriology of Water. 465 These results, therefore, entirely confirm those previously obtained, the disappearance of both the typhoid and coli bacilli being even still more rapid than on the former occasion. Results of a similar character were also subsequently obtained with t)ther waters sterilised by filtration (see pp. 479, 483, 502), in seme of which, moreover, totally different filters, constructed of infusorial earth, were employed. SECOND SERIES OF EXPERIMENTS. The Behaviour of the Typhoid Bacillus and of the B. coli communis in Loch Katrine Water. Having, in the first series of experiments, determined the behaviour of these bacilli in a typical calcareous surface water like that of the Thames, which receives the drainage from cultivated land, I pro- ceeded in the next instance to carry out a somewhat similar series o experiments with Loch Katrine water, which may be taken as a type of an upland surface water derived almost exclusively from unculti- vated land, and of a somewhat peaty character. The sample of Loch Katrine was collected from a tap on the main in the Broomielaw, Glasgow, on Ju»e 30th, 1893. Submitted to plate cultivation on the spot, it was found to contain , 112 bacteria in 1 c.c., whilst on chemical analysis it yielded the following figures : — Loch Katrine water uninfected. Loch Katrine water infected with typhoid. Loch Katrine water infected with _B. coli. 2-60 Organic carbon "1 by combus- Organic nitrogen J tion .... Organic nitrogen (by Kjeldahl 0-185 0-019 0-013 Ammnnifl, (frpfi) -,,----,,,,,, 0 o „ (albuminoid) 0'006 0-013 Oxygen consumed by organic matter 0-144 0-151 0-140 Nitrogen as nitrates and ni- 0 006 Total combined nitrogen 0-025 0-65 0-65 0-65 Temporary hardness 0 Permanent „ 0-8 Total 0-8 Proportion of organic carbon to organic nitrogen in sus- pended organic matter 8-14 : 1 466 Profs. Percy Frankland and Marshall Ward. Infection of Loch Katrine Water with Typhoid and B. coli commands, 4.7.1893. The cultures of the bacilli employed were on agar, and in both cases twenty-eight days old. In the case of the typhoid bacillus 40 needle-loops, and in that of the coli 25 loops, were taken from the surface of the agar, removing as little of the culture-material as possible, and introduced in each case into 50 c.c. of steam-sterilised water, which was then violently shaken to ensure disintegration of the bacterial masses. The experimental waters were then infected from these water-attenuations as follows : — Typhoid bacillus. Bacillus coli commitnis. Unsterilised Loch. Ka- trine water 2,000 c.c. infected with 8 c.c. of water attenua- tion 1,000 c.c. infected with 3 c.c. of water attenua- tion. Steam-sterilised Loch Katrine water 750 c.c. infected with 3 c.c. of water attenua- tion 750 c.c. infected with 2 c.c. of water attenua- tion. Porcelain-filtered Loch Katrine water 750 c.c. infected with 3 c.c. of water attenua- tion 750 c.c. infected with 2 c.c. of water attenua- tion. These infected waters, after thorough agitation, were then, as in previous experiments, subdivided amongst a number of small sterile conical flasks, plugged with sterile cotton-wool ; in each case some of these flasks were placed in the incubator at 19° C., whilst others were kept in a refrigerator at 6 — 8° C. The uninfected unsterilised Loch Katrine water was also put into similar flasks, which were kept under precisely similar conditions for control. 1. Bacteriological Examination of the Unsterilised Uninfected Loch Katrine Water. The control-waters were submitted to periodical examination both by gelatine-plate and phenol-broth culture, with the following results : — Report on the Bacteriology of Water. 467 0> CO ot Oi 00 o CO .S g - - _ ;.•"{ o I a 6 ^ 00 CO CD >f o 5 S S S 7 1 •s0 a | 2 '3 0 WDrH tJD 0 S « O S S f colonies obtained of water. bp 'S CM 3 121 • few liquefying col •S '-< .S t*> >> CM «M 2 * "3 ^ (Few liquefying col 1 ying colonies more (Few liquefying col ying colonies more o h *^> « O c3 a g IO (N co 5, ^ S ^ a I *- h3 13 & O ^^^ v~*' * 1— 1 i JO JO ^ q} n Ts-fi ^jtn1^ o "S "P i fl _o J° H d ^^ H|OJS CM c e3 o3 c3 c3 1 [iH (D ^ "S IP o § jSjS jSjS ^r-J S « 2 1 J>» i>. 00 U -S ® E • * d rH r-l r-i (M fi -2 i 468 Profs. Percy Frankland and Marshall Ward. In this unsterilised uninfected Loch. Katrine water it will be seen then that distinct, but only very restricted, multiplication took place, which was, as usual, followed by subsequent decline. The examinations by phenol-broth culture of the uninfected un- sterilised Loch Katrine water will be best considered along with the similar examinations made of the infected unsterilised waters (see p. 472, et seq.). 2. Bacteriological Examination of the Infected Unsterilised Loch Katrine Water. The Loch Katrine waters, infected with typhoid and the B. coli communis respectively, were periodically examined, both by gelatine- plate and phenol-broth culture, with the following results : — Report on the Bacteriology of Water. 469 6D • bC g » be 2 •« •^ 2^3 C .r-< O •S fl fl £ cS aber of colonies obtained from 1 of water. ubator flask. Kefrigerator fl; 690 (Few liquefying colonies.) (Numerous liquefying lonies ; no colonies un- stakably like typhoid.) 8125. (Very numerous lique:! colonies ; no colonies nu'stakably like typho (Very numerous lique- ng colonies; no surface .onies with marked re- nblance to typhoid.) 5000 (Very numerous liquei colonies; no surf ace col with marked resemb] to typhoid.) (Very numerous lique- ;ng colonies.) 4105 (Vory numerous liquei colonies.) 0 3 10 o '3 N § a O f^ o o> 1O V* O QQ o=E? (M ^B rH IN ** J> fl o °o

"fo -g ^ -t 2 QJ O ^ • ^ T3 ^ rrt rrt T3 T I S '5,^H * fl fl fl fl fl fl • 1 3 P 2 O 03 03 c3 03 03 03 1 i ? y J* -is -e JO Jo -P " 1 3, » 0 (4 *H ® »J CO CO •£"03-° •* M CO IM 1 (M a 3, a IM fr i 3 rH • >H -2 O fl M PH PH ^L, So .2 iH rH r- % a '*•* .al V •a ^ •3 03 m 1 ~ o 5.2 SMC «3 C 0 9 *o O r\ \n of P»I CN| O -*^ 00 'g (D "o o -~? 0 0 ^5 « I'-'g § Number of colonies obtained f of water. Incubator flask. Refriger; 3350 (Only few liquefying colonies ; surface colonies like those < or coli.) 1900 (Few liquefying colonies; only few surface colonies like typhoid or coh'.) (Numerous liquefyi no surface col typhoid or coli.) 63 (Very few colonies at all ; one surface colony just like typhoid or coli.) (Few liquefying col surface colonies 1 or coli.) 72 (Few liquefying colonies ; no surface colonies like typhoid or coli.) (A number of lique nies, and some si nies like typhoid o • «*•< a "s ^'i HS "!2 -|2 -12 -|2 Hot "H fl ^~* '-£> 3 c .3 ^H fi o 00^ *1 I i a P. 03 03 1 "S co c3 |i "* HO f) E o O ^ 'O «- S * J, III 1 CO CO rft CO CO CO CO ts ^ >3 , 03 TS j 0 1 ri A (4 M rp| •S° ^o r- 1 f"H i—t g *s 'to § 1 , 1 • "3 ^3 CD h ij 2 "3 ^ «2 O 03 1 1 o M rH r-l pH r-l & a i— i [B O »3 * -2 CO CO CO CO £ 2 la o> s2 1 H oo oo rH IH $ ° ? •*' — J ad rH s * M 3 o O O 472 Profs. Percy Frankland and Marshall Ward. In the water, therefore, kept at the summer temperature there was a continuous decline in the total number of bacteria, nor was there, apparently, any multiplication of the water forms; whilst in the water kept at a winter temperature, not only was there a slight numerical increase, but also, obviously, a considerable multiplication of the water-bacteria as evidenced by the increase in the number of liquefying colonies. In the following tables are recorded the results of the examinations by phenol-broth culture of the several unsterilised Loch Katrine waters, both infected and uninfected : — Examination of Unsterilised Loch Katrine Waters (First Series) by Phenol Broth-culture, 8.7.1893.. Number of broth tube. Water used for cultivation with phenol broth. Quantity of water taken in c.c. Quantity of phenol solution added to 10 c.c broth. Remarks. Uninfected Unsterilised Loch Katrine. (120) (129) (121) (130) Flask 11 .. Flask IE..' 1-0 3 drops 5 „ 3 „ 5 „ Turbid in 48 hours. Did not go turbid. Turbid in 48 hours. Turbid in 48 hours. Plates poured 10.7.1893. Typhoid-infected Unsterilised Loch Katrine. (122) (131) Flask 11 .. 1-0 3 drops 5 „ Turbid in 48 hours. Turbid in 48 hours. Plates poured 10.7.1893. (123) (132) Flask IE.. n 3 „ 5 „ Turbid in 48 hours. Turbid in 48 hours. Plates poured 10.7.1893. Coli-infected Unsterilised Loch Katrine. (126) (135) Flask 11.. 1-0 3 drops 5 „ Turbid in 24 hours. Turbid in 24 hours. Plates poured 9.7.1893. (127) (136) Flask IE.. » H 3 „ 5 „ Turbid in 24 hours. Turbid in 24 hours. Plates poured 9.7.1893. From the above table it will be seen that all the waters rendered the phenol broth-tubes turbid, those infected with the B. coli com- Report on the Bacteriology of Water. 473 munis in twenty-four hours, the uninfected and typhoid -infected waters in forty-eight hoars. Of the plate cultivations prepared from these turbid broth-tubes, it need only be stated that the plates from the tubes which had gone turbid with typhoid-infected water, yielded characteristic typhoid colonies which satisfied all the several confirmatory tests ; similarly the plates prepared from those broth-tubes which had been rendered turbid by coli-infected water, yielded the characteristic colonies of the B. coli communis, and also satisfied the various confirmatory tests. On the other hand, those phenol broth-tubes which had become turbid through the uninfected water, yielded colonies on the plates which were small in the depth, and formed small pin-heads on the surface of the gelatine, but did not give rise to the characteristic expan- sions, whilst on transferring these to potatoes, strong, highly -raised, greyish growths, much more conspicuous than those of typhoid, were obtained. Thus on July 8, 1893, four days after infection, both the typhoid bacillus and the B. coli communis were proved to be still alive in the unsterilised Loch Katrine water. The second examination by phenol broth-culture was made on July 15, 1893, or eleven days after infection, with the following results : — 474 Profs. Percy Frankland and Marshall Ward. Examination of Unsterilised Loch Katrine Waters by Phenol Broth- culture, 15.7.1893. ew 0 0 ,0 a *> _, 11 I* Water used for cultivation with phenol broth. Quantity of water taken in c.c. Quantity of phenol solution added to 10 c.c. broth. Eemarks. Uninfected Unsterilised Loch Katrine. (140) Flask 11 .. i-o 3 drops Turbid in 48 hours. Plates poured 18.7.1893. (146) (141) Flask IE!! M M 5 „ 3 „ Did not go turbid. Turbid in 48 hours. Plates poured 18.7.1893. (147) » • • » 5 „ Did not go turbid. Typhoid-infected Unsterilised Loch Katrine. (142) Flask 11.. 1-0 3 drops Turbid in 72 hours. Plates poured 18.7.1893. (148) (143) Flask IE,.. » » 5 „ 3 „ Did not go turbid. Turbid in 24 hours. Plates poured 18.7.1893. (149) „ » 5 „ Did not go turbid. CoU-infected Unsterilised Loch Katrine. (144) (150) Flask 11 .. j> • • 1-0 » 3 drops 6 „ Turbid in 24 hours. Turbid in 48 hours. Plates poured 17.7.1893. (145) (151) Flask IE.. »» • • » » 3 „ 5 „ Turbid in 24 hours. Turbid in 48 hours. Of the plate cultivations made from the above turbid broth-tubes, it will be sufficient to say : — 1. That from the uninfected water only liquefying colonies were obtained, probably B. liquidus (Percy Frankland) ; as already men- tioned, this organism is very frequently obtained in phenol broth- cultivations in which only 3 drops of phenol solution has been added. 2. The plates from the phenol broth-tube which had only been rendered turbid in seventy-two hours by the incubator flask of the typhoid-infected water, yielded also only liquefying colonies, and nothing like typhoid colonies was discoverable on the plates. The corresponding broth-tube from the refrigerator flask, on the other hand, which had become turbid in twenty-four hours, yielded numer- ous small and expansion colonies which were undoubtedly due to typhoid. Report on the Bacteriology of Water. 475 3. The broth-tube, which had been rendered turbid in twenty-four hours with coli-infected water, yielded plates containing numerous depth and surface-expansion colonies, which were undoubtedly those of the B. coli communis. Thus from these examinations it was apparent that on July 15, 1893, or eleven days after infection, the Bacillus coli communis was still alive in the unsterilised Loch Katrine ivater, as was also the typhoid bacillus in similar water which had been kept at the winter temperature of 6 — 8° 0., whilst in the water Icept at a summer temperature of 19° C. the typhoid bacillus was no longer discoverable. Another examination was made of these uninfected and infected unsterilised Loch Katrine waters on July 21, 1893, with the following results : — Examination of Unsterilised Loch Katrine Waters by Phenol Broth- culture, 21.7.1893. Number of broth tube. Water used for cultivation with phenol broth. Quantity of water taken in c.c. Q.uantity of phenol solution added to 10 c.c. broth. Remarks. Uninfected Unsterilised Loch Katrine. (175) (176) Flask 11.. „ IE.. 1-0 5) 3 drops Only turbid after 8 days. Only turbid after 4 days. Ttiphoid-infected Unsterilised Loch Katrine. (177) (178) Flask 11.. „ IE.. 1-0 3 drops Turbid in 4 days. Plates poured 25.7.1893. Turbid in 48 hours. Plates poured 23.7.1893. Coli-infected Unsterilised Loch Katrine. (179) (180) Flask 11.. 1-0 3 drops Turbid in 48 hours. Plates poured 23.7.1893. Turbid in 48 hours. Plates poured 23.7.1893. The plate cultivations made from the above turbid broth-tubes yielded the same results as those on July 15, 1893 ; thus no typhoid colonies were obtained on the plates from broth-tube No. 177, whilst they were easily discoverable and confirmed on the plates from broth-tube No. 178 ; again the colonies of the B. coli communis were VOL. LVI, 2 K 476 Profs. Percy Frankland and Marshall Ward. readily detected and confirmed on the plates from both broth-tubes Ubs. 179 and 180. From these examinations it was evident, therefore, that the B. coll -communis was still alive in the unsterilised Loch Katrine waters (kept both tit winter and summer temperature) on July 21, 1893, or seventeen days after infection ; the typhoid bacillus was also still alive in the similar water kept at the winter temperature (6 — 8° (7.), whilst it was again, as on the previous occasion (July 15, 1893), proved to be extinct in the same water kept at the summer temperature of 19° C. 3. Bacteriological Examination of the Infected Sterilised Loch Katrine Waters. With the preceding results must now be compared the behaviour of the typhoid bacillus and the B. coli communis in the Loch Katrine water which had been previously sterilised by steam and by filtra- tion through porous porcelain respectively. These infected sterile waters were, as before, intended to show whether these bacilli are capable of multiplication or not in water of this character when the disturbing influence of the simultaneous presence of other micro-organisms is removed. Report on the Bacteriology of Water.. 477 c3 ^ M ^3 a ci o 02 'E. !-~> £-• 0 ^ o «§ i o G o Is o oo f! C5 f-l « o C II 0 cS £ o M o • i 1 ^ 0 1— 1 t— 1 s i-i 5 2 K 2 478 Profs. Percy Frankland and Marshall Ward. Thus in this steam-sterilised Loch Katrine water the typhoid bacilli underwent rapid degeneration, the rate of their decline being more rapid at the summer than at the winter temperature ; for at the higher temperature they ivere no longer demonstrable seventeen days after in- fection, whilst at the lower temperature they were still just discoverable even after twenty-one days. Essentially similar was the behaviour of the typhoid bacilli in the Loch Katrine water which had been sterilised by filtration through porous porcelain, thus : — Report on the Bacteriology of Water. 479 0) 1 CO rs "o o 4 3 T3 o r-l H 1 CO o ^"2 'a - bO o o O •3 t. o «*-! 4J '3 fe • 'o O 'o ^ to I %<» co I •"u -2 • o •§ g] _g ber of col< ator flast, rH s .S rH S § 0 :ontamina S d "5 • O •§ £ •i PH ^ lsd o 10 ^r^.2 'V ~H~H -I-M* "H "H 6 6 "o ^"§ q'S a a 0 Ct c d « « O O • — ' O cS 03 03 c3 03 C3 C^ S 'cL'~' H etJp io|oio|a OH »|o 0 0 r5 S § Kt-i •|rt [H HH rH rH O 0 umber of days plates were incubated. 0 ^fc 00 00 to us CD CD te ij O o 1 r2 1 g P5 rt S rt | °s is 1-1 1-1 1-1 1-1 0 03 BJ^ •M HH n3 • *s h . « __rt ^ .2 _5 S ^ to N o3 A M r~l "M M U u 1-1 r-l r"1 rH e -g § •| | fg §2 ^ S 90 rH OO rH CO rH 1 rH eo rH • e E 1*8 03 -w P5 5 O rH rH rH S A 480 Profs. Percy Fraukland and Marshall Ward. Thus again in the LocJi Katrine water, sterilised by filtration througTa porous porcelain, the typhoid bacilli underwent rapid degeneration, more especially in the water which was preserved at the summer tempera- ture, in which they were no longer found by plate cultivation twenty-one days after infection, whilst in the same water Icept at the winter tempera- ture they were still easily recognisable, althoiigh in greatly diminished numbers, on that day. It is particularly noteworthy that the behaviour of the typhoid bacilli was practically identical in this Loch Katrine water, irrespectively of whether it was employed in the unsterilised or in the sterilised condition, and irrespectively of whether the sterilisation was effected by steam or by filtration through porous porcelain. In all cases, 'moreover, the effect of temperature on the typhoid bacillus was very marlced, the longevity being much greater in the Loch Katrine water, unsterilised or sterilised, Jcept at the ivinter than in that kept at the summer temperature. The behaviour of the B. coli communis in these sterilised L. Katrine •waters is recorded in the following tables : — Report on the Bacteriology of Water. 481 •s o T3 O .2 1 OQ O O O ^ o 1 a £ I sq OR O o 0 *4H o i— i •8 .sp .s d! '.fl fn' (§ Q ^ S CD £ ri .2 «_, N~ 0 •7^ iM o CS •H r-j O J-5 o "i* "H O O oq cq O «!•= 0 0 "o £ ^ 03 o c ^ ji S 3 O 4*tH o o «o »o o a o o 1 1 1° I' f-H P9 r-H r-l P^ § O M d 'S S3 ^2 *a\ a h o E 13 h _O I Pi hH rt r-l ^ -^ 1-1 [3 § ^ -5 O CO CO CO CO CO co fe 03 ^3 O5 Oi S .^ 3 00 oo X X GO 00 §3 " . g t> s 5 i— i 06 .J >O 0 || 1 rH 1 ' 482 Profs. Percy Franklaiid and Marshall Ward. Thus in the steam-sterilised L. Katrine water, the B. coli communis disappeared in a surprisingly short time, being no longer demonstrable on the Ylth day after infection. In this case also the decline was more rapid in the water kept at a summer than in that at a winter tempera- ture. Very similar again was the behaviour of the B. coli communis in. the L. Katrine water previously sterilised Iby filtration through porous porcelain, thus : — Report on the Bacteriology of Water. 483 o •8 6 _2 r-i P _o 1 S M3 i-l co t-i i-i TJ §0 .5 S o -S ^ H 'o *" o 03 fe o» o .S t*-, C 0 o 'o ,!£ o £ «*-i 53 O O Q 0 O O <• o fi -^ s ^ Id ^5 S o 1— ( 45 l> *n m C£ O 0 'M -| ^ O 'o 5 'H ,.9 P i^ jg _ ti o 0 -u 0 o g 0 «M _« "E, 6 .S3 .2 pj rt 2 JJ SB^ i *s •« 'o jy p^ J3 O S 1 fa £ cS li jS | 0 5 PQ 1— 1 ^ 1-4 M 1— 1 1 0 M 'o c II c '-Z o 00 06 rH 00 S 00 CO a 00 • OS 00 eo O ^ m o -*- a 2 s ^ 3 06 £5 IO co 1 drop Steam-stei 3 drops •ilised Loch Katrine. Turbid in 24 hours. 17.7.1893 (162) (163) Flask 11.. 5) • • 0-5 1-0 3 drops Turbid in 24 hours. 20.7.1893 (172) (173) Flask 11 .. 0-5 1-0 3 drops Did not go turbid. )J )! 25.7.1893 (207) (208) (209) (210) Flask 11.. „ IE.. Typhoid- Flask 11.. 1-0 infected P 1-0 3 drops orcelain-jt 3 drops Turbid in 4 days. Did not go turbid. Itercd Loch Katrine. Did not go turbid. Turbid in 48 hours. 26.7.1893 (220) (221) Typho Flask 11.. „ IE.. id-infect e 1-0 I Steam-s 3 drops 'erilised Loch Katrine. Did not go turbid. Report on the Bacteriology of Water. 485 Examinations of Coli-infected Sterilised Loch Katrine Waters by Phenol Broth-culture. Date and number of broth tube. Water used for cultivation with phenol broth. Volume of water taken in c.c. Quantity of phenol solution added to 10 c.c. broth. Eemarks. 21.7.1893. Coli-infected Steam-sterilised Loch Katrine. (195) (196) Flask 11.. 2-0 1-0 3 drops » Did not become turbid. >j » Coli-infected Porcelain-filtered Loch Katrine. (197) (198) (199) (200) Flask 11.. Flask IE.. 2-0 1-0 2-0 1-0 3 drops j) » Did not become turbid. Turbid in 48 hours. Did not become turbid. 25.7.1893. Coli-infected Steam- sterilised Loch Katrine. (203) (204) Flask 11.. „ 1R.. 1-0 1-0 3 drops >» Did not become turbid. j» » Coli-infected Porcelain-filtered Loch Katrine. (205) (206) Flask 11.. „ IE.. 1-0 1-0 3 drops » Turbid in 48 hours. » )» 26.7.1893. (222) Flask 11.. 1-0 3 drops Did not become turbid. These examinations by phenol broth-culture substantiate the results obtained by gelatine plates, and show that only a very small number of the bacilli were still living in the waters on the later dates. Thus, in the case of broth-tubes Nos. 199 and 200, it is evident that in No. 199, in which 2 c.c. of water were employed, at least one living B. coli communis was introduced, for the broth-tube became turbid; whilst in No. 200, in which only 1 c.c. of the same water was employed, no living bacillus can have been introduced, as the broth-tube did not become turbid. On referring to the table of gelatine plate examina- tions (p. 483) it will be seen that on the same day (21.7.1893) in the same water there was found only one B. coli communis colony per 1 c.c., so that it might easily happen that any particular 1 c.c. of the water might not contain any bacillus, as was apparently the case in the 1 c.c. of this water added to the phenol broth-tube No. 200. 486 Profs. Percy Franklaiid and Marshall Ward. It is in. this way that particular interest attaches to a comparison of the results obtained by gelatine plate and phenol broth-cultur-e in the case of these infected sterile waters, as the two methods can be made to control each other, whilst in the case of the infected un- sterilised waters the method of phenol broth-culture has to be exclu- sively relied on for the detection of the typhoid and coli bacilli. Behaviour of tlie Typhoid Bacillus in the Loch Katrine Water. (Second Series of Experiments.) In the first series of experiments with the L. Katrine water re- corded above, the number of typhoid bacilli initially introduced was so small that it would obviously not be possible to directly compare the results with those previously obtained with Thames water in which a much larger number of typhoid bacilli were initially intro- duced, as I have found in previous investigations of the same kind that one of the factors determining the longevity of pathogenic bacteria placed in water, or for the matter of that placed in any un- favourable surroundings, is the absolute number in which they are present. In other words, amongst, for instance, 1,000 bacteria taken from a given source there may be some individuals which will resist a particular adverse influence, whilst amongst 10 bacteria taken from the same source there may be none capable of resisting the adverse influence in question. When, therefore, I found that such a small number of typhoid bacilli had been introduced into the L. Katrine water in the first series of experiments, I immediately started a second series of ex- periments with the same water, but introducing a much larger number of typhoid bacilli. In this second series of Loch Katrine experiments, which were begun on 7.7.1893, or three days after the first, only unsterilised Loch Katrine water was infected with typhoid, thus : — Infection of J/. Katrine Water in Second Series of Experiments. — 25 needle-loops were taken from the surface of an agar-culture of the typhoid bacillus, 11 days old, and introduced into 20 c.c. of steam- sterilised tap-water, which was then violently shaken for 15 minutes ; 10 c.c. of this water-attenuation were then added to 1500 c.c. of unsterilised L. Katrine water. After thorough mixture this was divided up amongst a number of sterilised flasks plugged with sterile cotton-wool, which were placed in the incubator (19° C.) and re- frigerator (6 — 8° C.) respectively. Control flasks containing the same unsterilised L. Katrine water, but uninfected, were placed under precisely similar conditions. The results of bacteriological examination of the uninfecied control L. Katrine water are given in the following table : — Report on the Bacteriology of Water. 487 o «M rt '5 O H3 tc 6 ^ =£* »i M b 5 -g 1 O "§ 3 (M M3 rs CO §? g o d to c| |:l ' J.2 P^ O S 2 c, a 3 ^^ 0} Ci.j ^1 O 'a ° S ^* o ^ 03 rt 1 =4-1 O h o Cj 1 d 0 'O 0 CO r^l >»0 11 i-Q S 1 _o 9 0 3 C M gl 8 «2 a ^ JSJO Jo jgjj •3 P^ J3 53 ta b O £ Jj _o « 5 "3 O '-g ^3 3 pq i-l 1 M i-t 1-1 £ Hi -^ ~ .2 o . Hi ss a co Ci co 1-1 CO o 00 1-1 00 i-H CO g 3 » JO 13 1= T" 03 0 ° • r3 r-j r— !•> "^ 1*3-2 - fl r* r- J~ £< C C •2 S 3 c ca c3 cS <{~ £ C3 CS 000 Jo |l» |O If 1* !•" > *.£ "l« "pi > ""l" "151 0 •- I* o 1 E "S t TJ • m O O ° "sj co co co •* co ^"Jj _S o -§ .2 3 * '" fl •-^ o "» "S " 0 '£/f WH ^ ;'\- .S S co s -i 2 00 ^ 1O t, ^1 s's Tj ^2 ., r^ CB ^^ .3 BTJ 5* S — • o 10* 53 C -° ^H 03 0 1O O c3 o CO t, rH "I- C^l ^ ~^"^ CO o 4> rH 0-5.3 si O „ gj'&g CO i. s'-i"* ^ ?^l H> o "S £ ^ « S J g'S » . /--» •w J 05 "-• ?^* -5 'V *p, o JS •S 3 1 6 "5"! ell T 2'^ s ^ o a O v. **-j iaff '? ° tc 2 c ^ * c« o 1 09 .fc ° ^ 43 T2 s P— • O 5^ lira i cm O +iv) S> .:; B'UJ o •i Il'l ' (N t- ? ^ I ^ 1 .S -2 ^ -• ° f 1 2 1 J -. a '— ^ ^ ~ 3 c IS s ^ 1 H H H * HQH tlllod S TL S, 'I SH IS S = - - S .= •£ 0 ^15 & co \a co us CO «3 CO O CO IO CO vO O* o « « aj it J||S O 9 ?-t rH jg . . . . fl — s | .2 | i— i PH M ^ M P? 2 •§ | -| -a i— 1 rH i— 1 rH ^ ^ •^ s 3 § § " S § " ? " i =§ " u H E S S S F-H ^^ t-, "• co CO fli ^i c^ ^.^^ __ .^^s x-"s x— >— ^0 o 491 O or: , !-, -S •- ' P-i eo us co o r i "3 VOL. LYf. .. » ' HOHO 2 L 492 Profs. Percy Frankland and Marshall Ward. Thus ivhen the typhoid bacilli were introduced into the L. Katrine water in large numbers, they were still easily discoverable by phenol broth- culture on the fourteenth day, although from the eliminations by plate- cultivation {see p. 488) it is obvious that their numbets had undergone enormous diminution. They doubtless persisted even longer than this, but the experiments had to be interrupted. Thus when introduced in large numbers their persistence is greater than when only small numbers are employed, for in the previous experiments they were no more demonstr- able in the unsterilised water ivhich had been Jcept at a summer tempera- ture (19° (7.) for 11 days (seep, 476). COMIARATIVE BEHAVIOUR OP THE TYPHOID BACILLUS IN THAMES, LOCH KATRINE, AND DEEP WELL WATEK. The previous experiments had clearly shown that the typhoid bacillus, although unable to multiply in either ordinary Thames or L. Katrine water, even when these waters are deprived of other competing or inimical bacteria, is yet able to remain alive for con- siderable periods of time in these waters, not onty when they are previously sterilised, but even, although for a distinctly shorter period, in their unsterilised condition and in the presence of an abundant bacterial population. Inasmuch as the access of typhoid bacilli to potable water of all kinds is one of the most ever-present dangers to the public health, it becomes a matter of pressing hygienic importance to determine whether the particular kind of water into which they may gain access affects the chance of their reaching the water-consumer in a living state. The population of the United Kingdom is chiefly supplied with one or other of three different kinds of water, of which the Thames, L. Katrine, and deep-well water of the Kent Company may be taken as types, and it is with these three types of water that I have, therefore, instituted the comparison in question. From the experiments which I have detailed above, it is obvious that the longevity of the typhoid bacillus in any particular water is subject to very considerable variations, according to the initial vitality of the typhoid bacillus employed, and according as a rela- tively large or small number of the bacilli is introduced into the water. In order, therefore, to institute a comparison between several different waters as to their relative capacity of maintaining the typhoid bacilli in a living state, it is absolutely essential that the typhoid bacilli placed in the several waters should be taken from one and the same cultivation, and that they should be introduced in each case in as far as possible the same numbers. These were the conditions which were secured in the series of com- pnrative experiments made with these three different types of water, and which are now to be described. Report on the Bacteriology of Water. 4 Do Simultaneous Infection with Typlwid of Thames, Loch Katrine, and Deep Well Water. Each, of these waters was, in this comparative series, simulta- neously experimented with in the natural unsterilisecl state, also after sterilisation by steam, as well as after sterilisation by filtration through a porous cylinder composed of baked infusorial earth. These nine different kinds of water were all infected at one time with the same quantity of typhoid bacilli taken from one and the same culti- vation. For this purpose an agar-cultivation of sixteen days' age, and grown at 18 — 20° C., was employed. Forty needle loops were care- fully taken from the surface of this cultivation and thoroughly mixed by prolonged agitation with 50 c.c. of sterilised tap-water. Of the water- attenuation thus prepared 4 c.c. were added to 1000 c.c» of each of the nine different kinds of water. ID this manner wasj therefore, secured the equal infection both qualitatively and quanti- tatively of each of the nine experimental waters. Each of these in- fected waters was subdivided amongst several sterile flasks. Th0 mouths of these flasks instead of being plugged with cotton-wool, were in this series of experiments simply covered with sterile beakers, an arrangement which is in many respects preferable for purposes of this kind. All these flasks, together with similar flasks containing each of the three unsterilised waters not infected, were placed in a dark cupboard in which there prevailed an almost uni*- form temperature of 9 — 12° C. In this series of experiments, besides determining the relative longevity of the typhoid bacilli in the several different types of potable water, I have also endeavoured to ascertain the effect on the bacteria of keeping the waters at rest and in motion respectively. To this end, in the case of each water, one flask was kept at rest and only shaken up when a sample was to be taken from it, whilst the; other flask was daily submitted to violent agitation over a period of five minutes, this agitation being repeated two or three times on the same day. The convention will be adopted in the following pages of referring to the flasks kept at rest by the letter A, whilst those which were subjected to daily agitation are distinguished by the letter B. The various waters were periodically examined both by plate- cultivation and phenol-broth culture on the same lines as described for the previous series of experiments. The uninfected waters yielded the following results on cheinical analysis : — 2 L 2 494 Profs. Percv Frankland and Marshall Ward. Results of Analysis expressed in Parts per 100,000. Thames water collected 5.10.1893. Deep well water (Kent waterworks) , collected 6.10.1893. Loch Katrine water, collected 13.10.1893. 26-20 44-40 2'40 Organic carbon "1 by combus- Organic nitrogen J tion .... Organic nitrogen (Kjeldahl process^ 0-172 0-036 0-40 0-058 0'013 0-189 0-024 0-016 0 '005 o 0 0-013 O'OOl O'OOl Oxygen consumed by organic 0-083 0-013 0-147 Nitrogen as nitrates and ni- 0 '175 0 '467 0-007 Total combined nitro°ren 0-215 0-480 0 -031 1-75 2'50 0'65 4-2 21-1 0 ' 13 '7 8-6 1-0 Total 17-9 29'7 1 -0 Turbid Clear Yery slightly turbid Report on the Bacteriology of Water. 405 d 'S d "3 s 1° 5 1 ^ •a ^2 o 1 'o3 •a "a o o" rH CD" •~ . iM M 11 8 0-S O o3 m 5 O CO Oi _r s '3 'o ' '-?1 o _ T o (H f o (-. t£ ^ a S-i o •g £» p O o * ,0 ^( (^ ^^* ^^ a 3 •a C3 * l-H * s o C -£ S- 0 III -1 1" -K 1= to "P-!O P H o ® *3 0^3 •rS T3 'H T3 o o "3 S "o ° 1 03 is 5 P1 o •Sis J|2 JO Jo T»T« -iS-'i? O 4i J3 O i* •** III 00 JO o o T3 V 'So d e pq m 8 r§* "£ 1 ft 3 h j 0 a e E =2 •o 49 6 ts OS pq ^ j 1 «, PM 6 M _. "i X c .0 0 . _s '-3 ® 'S 03 TJ ? t» 03 c « S CO C3 i—( g s « s » g 2 6 1-1 q i-i i— i £ v % OS +* * ci F— I -; e8 H TJ c 'Hn 1^: °i a . a 12 c HS J= l HS«| 1- H2HS 1= lo •-•lo-io 1- H ^-!S pq 49(5 Profs. Percy Frankland arid Marshall Ward. The above tables show that the bacteria in the uninfected Thames •water underwent very considerable multiplication, and in the typhoid-infected water, although the total number of bacteria de- clined, it is certain that the water-bacteria underwent considerable multiplication, as there was a great increase in the number of lique- fying colonies ; the diminution in the total number of bacteria was due to the disappearance of the typhoid bacilli which were initially present to the extent of about 16,000 per c.c. It will be slwivn (.p. 508) that typhoid bacilli were no longer demonstrable in this water by phenol broth culture after 28.10.1893, or nine days after their intro- duction into the Thames water. Report on the Bacteriology of Water. 497 j Q T3 r^ nd O o O -^ a a rH • rH 'a 'a S SO O o O 3 M 3 8 SH r-H CO P ^ S '3 •3 r-T 0 s | '3 t: 5 ° to r2 10 •S 43 o "3 .2^ K IM o_ "o izi ^p § ° in 01 rH o o "3 o "&. N O 10 o o o S S O^ *« "* (M ^ ^ r-T rH S 'In 3 S3 ! <3 d < a 10 JO |0 Jo Jo |o lo i> ^ 0 "V n<= ,o >, , c O ^ a^ S r» 3 ^O ultivati o 5 'C ^ 3 3 •£•£ a a 3 3 'o'B o 4. > a 0 • £ 1 • *o £ 03 CD £H a CD "^ CD ^ ^? ^ d»» ^ 3 . _. I1 O 'Pn T3 B 1 'So • _o « w S '3 '£ r* P S J 3 r| I E O 4> V 1 9 "Pn rH ^ << -«1 * M 'o _c • a CO O5 00 O5 X CO Oi oo CO a oo co Ci 00 0^ a 0 O rH rH rH rH r-i rH rH 03 «. ll rH o ^ 4 Ci i— I rH eq 55 "i « a 's 'a 3 O O 8 S M CO Gi rH rH 8 O ra ,a "S =H B HH o ^ o S o O ' § o •* B O 00 QO_ CD" C? CD" t>T rH "S 0 PI g HSJS "|S -IS Ho -IS • rH P •a 5 B a 11 03 ti Ti 9 M ^ -SHS rtHS •* Jo !« o o> rH 1 OQ C co co co •« co | co co IM T3 e • tfl rS a 'o a PH *P H 3 GO E o d! B PQ ^ ^ H ^ CO 1 r— ^ CO C35 X i-H 8 00 I— 1 co rH co oo rH o" r-i rH r-J r-l I-H' r-l r-i i-H ci r-i r-i rH 71 498 Profs. Percy Frankland and Marshall Ward. From the above tables it will be seen that the bacteria in the uniu- fected Loch Katrine water underwent no multiplication, but, on the contrary, slight decline. In the infected water the initial number of typhoid bacilli must have amounted to about 22,000 per c.c., and although the total number of bacteria in this infected water under- went a great decline, there can be no donbfc that the water-bacteria multiplied considerably, as was evidenced by the increased number of liquefying colonies, the diminution in the total number of bacteria being doubtless due to the disappearance of the typhoid bacilli. It will be shown on p. 511 thai the typhoid bacilli were discovered for the last time by the method of phenol broth culture on 7.11.1893, or nineteen days after their first introduction. Report on the Bacteriology of Water. 499 13° «' ^ t- ^ .j O CD O 1—4 "3 t "2 r- C S 'So — «4H 4-1 o CM T3 O 2 ^3 8 cc O CM o" i 's -2 3 'a 8 i— r C5 O 8 oT o 13 0 _3 5 « CD 3 1 ^ a d CM 01 si C3 1* o - C3 IO * CD O "o o 0 — 0 CD 49 O .2«r. i-T ^* ^* (N " "^ C C "co "o E o CfH o * o o 8 IO 0) 0 IO g o o o PJ o_ 00__ o o o u ^4 rH co"~ rJ" 10" co" Tj« o g" 5 "a CO 3 s c cS 03 a B — a C 3 r* -. ^ JO |o Jo Jo Jo lo JO PH lo |o |o |c o o 10 ^ o ^1*0 ^1^ |W3 |O ""I® l« ^ •2 ^»j •2 O -H P EH ,> ,M m f § 2 13 • 3 3 ^3 CO en h i o O .£ 2 E £v 3 &i ,p =2 _« H a o CD A ro ^. ,: p2 ,. ^, _) ^J E -. a ^5 ^ ^ **? ^ ^ "1 ^5 (5 D I 4 a 0 S i rS * S ? - j O5 oo S oo C3i 00 CO Oi X CO C5 00 CO § g 00 CO C5 00 CO o 00 co ob 3 '-I rH rH I— 1 rH rH rH rH rH rH 0 ^ i d r-J ,_; rH ,_; O rH rH rH rH rH i— 1 rH rH rH rH rH rH rH || < i i i O5 rH i-i SO ai rH r-i ^ ^H £ LI 500 Profs. Percy Franklaiid and Marshall Ward. From the above tables it will be seen that the water-bacteria in the deep well water underwent much more extensive multiplication than in either the Thames or Loch Katrine water. This extensive multiplication of the bacteria in such deep well water was first called attention to by me in 1886 (' Proc. Roy. Soc.'). The in- fected water must have contained, initially, about 28,000 typhoid bacilli per c.c., and in this infected water the increase in the total number of bacteria was. even more marked than in the uninfected, the increase in the number of liquefying colonies being altogether enormous. The behaviour of the typhoid bacilli in the sterile deep well water (see p. 505) clearly shows that they cannot have partici- pated in this bacterial multiplication observed in the infected ua- sterilised deep well water, but on the contrary they must have un- dergone great diminution, as it will be shown on p. 515 that the typhoid bacilli were no longer discovered by phenol broth culture in this water after 21.11.1893, or thirty -three days after their first intro- duction. Report on the Bacteriology of Water. 501 O I H o -3 i t 2 * S -fol a d rH "§ I|J-2J g^J O -^ 'So • o o o O C 'K O """ 5 •^ bC £n *o „ -— himber of colonies obtained i of water. 1 i o ^ O 00 Ci o" 1-1 t-- o 0 1— 1 00^ of 00 i— i r-H o o Ci GO oT he plates contained no colon bling typhoid, but a very lar of small yellow colonies very colonies found in the \infilte The organism giving rise colonies had probably passe the filter in small numbers then undergone enormoiis i tion in the filtered water. f* r=H ^ H 01 c j^ 'o'^ •J3 1- 1- on •^ 03 O >J 1-|U! "V "f° CV CIH HO -I-3 0 H« ,--\s r3 £ -£i « § 'c ^ 'el 1 sj a a s c ce a z O 0 I X!>H *n g*. H HM K\" "• 'ft o> ^ FH (D p a d °o ^ 'd *s o ,£ "§ "QJ *7=' * ^3 1O i>-i> CD CD cs oo p lO X 30 s ^ H § 0 £ « 2 PH a _j V 0) •a 1 c2 C! • r-* o Co * o 'Sc c w n 'o _0 n c _>> .3 0, • ^H u •^ ft |g EH 13 3 *3 ^3 09 ti o O d ! ' 1 <2 01 0 'S "8 M ? C3 o 00 OJ oo o oo 1 Ci uO o co o X rH O "jj S d d FH ^ ^ d d 09 t QJ N a> d GO d ,^ Q> d *S 4. ^- H i— i CO CO 502 Profs. Percy Fraukland aud Marshall Ward. From the above tables it will be seen that in the steam sterilised Thames water the typhoid bacilli underwent no -multiplication, but, on the contrary, steady diminution, being last discovered on 20.11.1893, or thirty-two days after their first introduction. In the Thames water, sterilised by filtration, their disappearance was far more rapid, for they were no longer discoverable eleven days after their introduction, and they had doubtless died off even before this. These results entirely confirm my previous experiences recorded on p. 464, and the confirmation is of the more importance, as the filters used in the two cases were entirely different ; thus whilst that used on the former occasion was a Chamber-land cylinder of porous porcelain, the one used in this latter instance was a small porous cylinder con- structed of infusorial earth. These infusorial earth cylinders are much more porous than the porcelain ones, and pass the water far more rapidly. Report on the Bacteriology of Water. 503 o o r-l 1 I H 'So g C3 o 0 E b-. g o t— 1 •a '3 tC co" d "3 Is O c3 S S 1 'd 'o J s -| B | "S S g 8 O 0 h 5 "Oi s >M S- «r " IN a ,U d S3 * . 5 _3 d _, _o -si ,"§ nt H|»H- Hia°'£ HO «'» »|o> 0 0 b» §1 "o « 0 gl d d d d es - 1 L 03 r* ^ V "H '' xirt V P4 a 03 "5 ^J p I i _ *o Is ^ 0 j "S •S"* §^ ^ ^ y to t-i> O CO l" ao OS Sa " .d ' TJ T3 I OJ 1 "Ei 1° d pq m a t^ 00 V •3 JJ •a p s 03 s h 01 2 o d K o M ^5? ^ , ,. *^"t 83 rM ^^ ^H ^H ^R ^S 1 "1 M _. a 1 rd 1 5 . CO CO CO CO CO CO rd •* . Q> OJ Ci OJ Ci * 1 3 " - 03 oo 00 00 CO 00 oo d* o *; 3 3 d d .n' 1— J F-J N I! c * '1 1 ^ Ci d co GO d (M Si ci 1 S 1 o cT ~ \ ~ 1 S O r-T T—l o o o o g O o §_ I CO of 1 (N r~> "S .^ '-B rf,n H>H ** «t>£ H" 6 as w •B d C3 d d 93 33 d c d c 03 03 ^ 1 ^ •>'*" Irf U Jn B1-1 J-i^H w o CJ H 1^ 1^ j^ [FH r4 M ^rt * , O 1 .2 .5 '5 U5 l>t- O CO l>t> 00 0 o E o PM a> -4-9 o •? d _o PQ M w 1 '> PH 3 >^ s •" 2 «2 V <1 ^ ^ • I'Sj 3 a 3 "3.8 « r OD » s^-§ 4=^-0 M n o - Q o 0 o o ™ 1— 1 o O CD 18 1 HO ->:: 00 oq M o. cS G 5 c3 ~ 5 S S H|M I-. 1-1 0 0 P "V i- CO CO "d o> M •£> .s "S 1— 1 O CO CO q .9 "S _o pn B 'o ]S y/TJ P & .3 EH "1 o 1 o 1 "1 co 00 CO 1 rH CO 00 d O "* rH s = ci O CO 06 500 Profs. Percy Franklaud and Marshall Ward. From the above tables it is seen, that in the steam-sterilised deep well water also the typhoid bacilli were incapable of multiplication, and, on the contrary, undenoent continuous decline in numbers; they icerc last discovered on 8.11.1893, or twenty days after their introduction, whilst on 20.11.1893, or after being in the water for thirty-two days, they were no longer discoverable by plate-cultivation. In this connection it is particularly noteworthy that the typhoid bacilli were still dis- covered on 21.11.1893 in the unsterilised deep well water, thus shoic- ing that in this deep u-ell water their longevity was ^t,naffected by the circumstance of whether the water was sterilised or unsterilised. In the case of both the Thames and Loch Katrine waters, on the other hand, the longevity of the typhoid bacilli was much greater in the sterilised than in the unsterilised water. This circumstance is particularly instructive and important, inas- much as it was just in this typhoid-infected unsterilised deep well water that the water bacteria present multiplied most extensively, and yet this large multiplication of the common water forms did not prejudicially affect the typhoid bacilli. This deep well water, on the other hand, is in the sterilised con- dition less favourable to the longevity of the typhoid bacilli than the sterilised Thames and Loch Katrine waters, for in these three steam-sterilised waters the introduced typhoid bacilli disappeared first in the deep well and last in the Loch Katrine water, their longevity in the steam-sterilised Thames water being greater than in the deep well and less than in the Loch Katrine water. (For further remarks on this behaviour see p. 517.) In the deep well water sterilised by filtration through porous por- celain (in this case again infusorial earth), the typhoid bacilli again disappeared with remarkable promptitude, being no longer dis- coverable eleven days after their introduction. In order to ascertain whether these waters sterilised by filtration through porous cylinders owed the rapid disappearance of the typhoid and coli bacilli which almost invariably occurred in them to the presence of any antiseptic substance possessing general bacteri- cidal properties, the following experiment was made : — The typhoid-infected porcelain-filtered deep well water referred to above, and in which the typhoid bacillus was proved to be extinct on 30.10.1893, and 8.11.1893 respectively (see Table, p. 505), was on 11.11.1893 treated with three drops of the unsterilised uninfected deep well water. These three drops of unsterile water must have contained about 3000 water bacteria, as calculated from the results of plate cultivation given in the table on p. 499, and, as the volume of filtered water to which these three drops were added was about 100 c.c., the latter must have acquired about thirty water bacteria per 1 c.c. by the addition. Report on the Bacteriology of Water. 507 This porcelain-filtered deep well water, to which the tliree drops of msterile deep well water were thus added on 11.11.1893, was ex- lined by plate cultivation on 14.11.1893, or three days after the lition, and was then found to contain 10,462 bacteria per 1 c.c. ; it was jain examined on 23.11.1893, or twelve days after the addition, and then contained 603,900 bacteria. It is obvious, therefore, that in this same water in ivhich the typhoid bacilli were destroyed with such remarkable rapidity, some at any rate of the common water bacteria present in the unsterile deep well water were able to multiply both to an enormous extent and with wonderful celerity. This result dismisses, in my opinion, the last lurking suspicion which might still remain of any antiseptic substance having accidentally gained access to the water in the process of filtration through these porous cylinders. VOL. LVI. 2 H 508 Profs. Percy Franklaiid and Marshall Ward. ' *C3 C3 3 .1 «$g §^ "S 2 s O •^ ^-. — c?1 ^ c .5 f° u S 2'§ *"• "5 o S* * p Q 5 S t«^ S •- S2 ga ifill «*" 0.1 M . "S S •0 g -S2 .2 £• || Q tc '5<« .2 S t2 o •sl^'-g 1 . .2 Is "g'° « C 0,q 3 1 1 «§ l%*k §•9*1 a 3 c »• 2 Si _£« ~ *o '58 2 a & tfi 03 .2 O "w Oi ^ "3 §••2 C " ftf "S -fe *O O ^ I SI "2 *c S3— rf"?< 5? 3 "o o< ag .S a> ^S a 2 g .2 £ « -a 10 co »o co xo eo \a eo xo eo « 5 ;occu5 I * _c ~'o S" '-5" »o to i-it-00 00 Oil-- X 33 t» OS 2 u ;i 510 Profs. Percy Franklaiid and Marshall Ward. _ • ^ **5 3 ^ .2 "S 3 ^ "S "a. O •3 I E g 1 1 o ^3 § 5 ^ 2 *; 2J e »: ^J 1*3 's 'c -ifi o o S 02 C) ;j | || ° "o -S /* S 3 -"3 -S c *" 5 »-< PM 0) 5< "5 1*8 ll ll ? 5g S S 'S c 3 A 9 *• •§ 5 O O) ^c 1^1 *r^ O O eg S 0. 2 2 ^3 1 p *a 0 *» S S J 'S -g 'i S S cS . ^ — ^ oj w g . • M B K ^.S ' 2 - ' £-<° • £ ™ HH 1 S B 0 § 3 JH£"^ 3.« 3O2 3 O 5 •£ 2 ° ° ° ^*S_e*2 O^^tfi O r^ ^ a j: s: ss si JSO2» 3'£3'S='ai< TO ™ Q W (N N W ^Cot^ oC^Q^N^1 S •2 M 5 £ .5 .S .Ssboa MC^MC e*: .S 12 o -a -o -o -c •e':i'o -o STSWO -a-a*" cS •£ c 3 3 3 3 3 . s< a 3 c 3 » a 3 3 S! J^- S3 S 3 ^ 3 3 £3.^2 --2 3^^— 3 --"^ ^" EH Q EH fn c-, EH EH QHQEH QHEH '» ^"* O 1 M 1 : :S2| i : : i :3 SgS 0 c -"* N/^^s O QQ (JQ ij ^srs^'i o ^o '-3 ^ lua =s i rs ^ ^ - ^ 2 •— - 8 ^ -S ** 5 ^ o -2 «! H H H £ EH O O l^-l J 00 ^_^^ CO § __ * fe -^ »M ^™^G^V'^viO "^"^ 00 r"' I-H t~» CO Ci ^ O ^O CN JS "e ^ o SSci^ci^' ci O O : CD CO CO CO 512 Profs. Percy Frankland and Marshall Ward. S c i ec « O 03 C *aJ |.s ^ sl « a 'S > 0 (8 •3 be .2? 50 || 1 1 *0 si lii •ti fl s 'S •g s I'5 8 8*1 5? Sac •o •fr-2 s M » g m 0 3 W c o- >- S r| = " " " 0"S « «rt n ta co MS CO °o « 9 o l| § <— i CD •S : :|S | c .• 5"^ 22 Z _o "^ ^^^-—^ ^S ^Ji ° rw "^ »> 'C S rQ OJ l| 1 t/j ^ 5* •r1 & 1 .2 | $ § J o "o ID ?i *•«£ *T3 O S t» Ci '7 S .§ "o T3 " rg 6 <§ "'§ 'o 1 P H a, I*8 41 § 6.3 CO O5 ^^^^^^^.^ I- o ^o ^ . ""1 oo o> oo oi • o -g g *5 (M — 1 i-H r-l rH fN (M fifj ^ n^ Report oil. the Bacteriology of Water. 513 •g a .S £2-1 1 2 3 fcs , •~ ^"-1.— 's .. '*M OJ 3 t*> Q O*^ »rt "" '5 -g 1 **§" §^ g S § ill §•! P. _ - O ^0 2 - S o. J J P. ^ 5; SI "8 .§ ^-f 2 '1^, « ,S v- •c So ft • A .2 ~ "So! ^- c cT *c3 -- S'3 c -2 ,2 "2 c 5 jj 5 "3 is J! S tr^ t- « 3 •s!i S g »> § •C 3 **" 73 ^ g S o> T w II Hil -J 1 C3 .2 o ^0 ""S ? fc £-3 1 ^= ^£ a is s~ |S || ^ 1 | a 11 11 f| jljl l| t's~ g^- if Of i« S 2 8 o. 'S"5 §8 "3. s to o n "-^ • 2 ss <""S .fcj'w C; •— ' -^ ^.' o ^ CB j ^5 rt,j ^ _- ~ D r-; CJ ^SS-sH^S^ So ,•1? / ^ _ 3 .S3 .^ _-g _ _5 g be 3 g « S 3 £-33 Ss2'o<£ £«' S C *^ 3 3 3 ^" 3 g _. 5 -S" 3 >, S 3 GJ ||| % - * | 1111 "^"^ 11 "g g eS w Jj* .S .S .S ** 3 ^ * .S 3 .S S a _c ,3 •00*3 *3 S > ^ 3 = 3 3 t£.r- 3 3 EH O EH H 3 c 3 g s 3 £3 S3 3-p. t--e fa.S'S 5p.boi3 tt^ 3-^H 3^,-« p^ — -3oo3 ™ *j H G H Q H EH HEn A -IS''© a 3 i 1 -3 •« 3 ^ O A •* «w • »s & § -3 's 3 ° "S 13 -5 " " o- eoioco wsco us coua ® f-t & fc| ^i fl O S e* o • o ^ o O •f *'i 2 rH • ^ : '"321 1 : : : :33«S ft '"^ I I 'd ~ ^H 11' I ^ a o [^ - — v — '.,H ;/. . ai ^—'S~*' ^^"^'tl "|,= Qj -4^ ** S H 1 "S 2 "o ^ ^ a S g O 0 0 I 1 i k 4 "a 'c "S _g '71 S'7 '? 1 S «V« s * ° . 2 ^j? ;"? ^- ; O.T5 & ;p | M" 1 IS! „ a * 3 * • .2 2 .S 5 g ^ en s 2S D C H 3 o, c 3 a 3 C5 EH lUlll •S.S § Report on the Bacteriology of Water. o oo-e « o~£ E " P< H P •§ "3 S S'S 03 S f g — •« *•" T3 3 -r £ P. £• Sir i'a^ M.a := he **" ® '& "3 =3 o_s 3 a &•- oT»w" • S ^ k> S. 0 •£ 0. "3 -5 3 i? pi "S'S 5 .2* g" ^ c £ £ o P "iN^I H| ii -13 — SI £ bC p, ^ ^ 3 *" OS ^- rf ° 3 .S ^>J> j; -g .|S£ C.M *a 1)3 "S JP S P '2 •° — >.= c >> 3 .5 3 J3 ^, S *g <§ ^ so '3 § := 3 .£ £? 2 2 = o tc ts o P< is c o *•* o 02 &?*o £1 >. O S.'5 ° o S S,S 'g 2 •O"4) C p-'5i a J3 * o ae "~ § C'— "Si C^ '§; s "^ 5 TS c " •3^3 * § a •3 "C i 0 „ 15 ii 0> c C V 3 S S^"3S 2 8 *•= c i ^ -2 !s 1 * e «r 1||1 fa llglS ^ o cT • J | 3^ g IP. i f | *| A o be ? Ho e 'S'a"S,J-1 S _. !T ^ .*•' -^ JlJJjjj "^ *3 i? "vS ||| cScll I B «5g . S ^ ^ ^ c ^ ~!2° "^ e 2 * a 0 -3 ^ § g.1- 2 ."3 — "= 2 '= VIH -g ^ •? 5 '•? ;= ^ " 13 3 "E « — -g S '^ S 3 ^*i H=|S§ 1 |gS-5 .— 3 p, B 3 5 an H s. O, Si Baas O - - - 8 s = s T3 'S T3 ro 10 n o CO O TO «5 TO «O 05 lO 9 9 9 r-l i-H i— i : :S 2 t !T3 : :S§ i ;3 S /"x x-v 0 5"^.2 ^ G o *.£ pj 0 "o -S Ti _c •"C C o *& <£ ""S J -i .° "Z .2 "a. "s ^ *3 S^ "5 >^ 'c S^ P H' P H P H w TO §____ 1 TO OJ 00 ^-v . » N_«'S.^'^w_^^»-*' . X_^N™^N^^ >^X s a a •S I S }f o ca o3 S1 §2 J1 8. ts . £ ^ 1 § § 11 g> s ta S 1 2 £ 1 1 o ^ "§<».• '1 1 3s- o. S "" *O oT w .2 ~ 43 S 3 § P.S> oj ° I « *S I, -rt 12 •o-s 0 0 0 *S o? •- 1 J * ^ & "i 1 £•'§ •§, o< M •2 - 3 "° o *^ * "S ?1 « 2 fl® 00 *^3 " ,_ a ^ 1 51 if ii &tj ^ *>; rt •g'a, £ M '» "*"* r*" -^* Si 0 « •tn "" fl> . Q3 it 15 i ! || || 1 & s^ S'3^- "3. jj^ £ | i K 1 '• 2« ! la ^ o> S *Z PH ^ ^ -^ •^ E ^ PH cS Bs -j 5 .- .- -— ^— , • ^ P^ &. •a «> g S •e • -1s- .•~ 3 ' * > M- S a.' ^ S 5* £ £ o! § ^ £S goo'co';a'g S^Ti" 3 "~! " i 3 c -| I | 1*1 o i III l|| * = 1 S'3>5 g JS M "S ^ ^ O cl O •S g § -13 «5 jg "S 2 'i 73 • *3 SL | fl 43 2 09 ° 0 O^ "I ^ co 10 co us CO »O CO U5 •V.V-J co'-JJ a) t< a| g o O O |i1.s r— 1 l-H 7T «H «4H .< • o> ° 0) O s 2 .2 5 iS 73 °if -2 ^ Si .2 'fj £LI a 1 ^ -s ° rj-j 7TJ 09 O J 1 •=! r£ "o ll 'B "? i a s 'ij — -4J QJ 73 09 1 I I1 f J R rt £ £ -r 9 s rrt rrj r^ -— Q 5 55 QJ 'S » 1* 39 "» "Oi S-^ o * «2 5 ^ *S 09 0) 3 09 o o o -* q> to cy ^H !s ^s ^sl^ § O3 •3 .a p SH • ^3 T3 T3 ^ "C I 73 ! 73 ^ ^ 'S 09 *fi • 'o • 'o '§ g -2 09 * 'o " 'o 'o s ^ 1 B 0 J 0 if rx,-» 1 IJIJ I'll 1 1 P P H'3 H>GH'"H SS & a H£> H C H ft 73 O •* •*' (2 .. _, . § S^ » -21: 2 "s 00 ^ r-j 0 OOO CO CC 00 C?O O? i^ OOO O CO CO CO CO 00 T}< 1-1 rH "-< $2, CO CO CO CO CO CO CO o ^" c 2 o -S o •s -*3 f>-H 3 s ^.'s S? S^ ~ (^^ 3 «3 *" S •" S *"SS o" o «s 'sl.'gi. "sl, = o ^. ^ §,-§ 'g-.-s '»:§ 'O.TJ 8 '£..-? 0 WJ aj "^ *"* 2 *"" 2 *"* ^ "o V. •4J ,^??i "^^oa, o"a. +? > ^t pn 8 V S cS* c* "^ rf tS1 •" 3 OS 1^* s •S §" to" tc" tc i -0 1 , !§> *Ill ll §5 5 i S 6 |S 01 3 r^ i| |^ j£ 1^- p4 -^ >. m v 0 £•= » pi _• K E 0,, c^> >•» ^3 fr* B PH «"^ j^ ,,^-A ^__^ 'a *c * x 0 . "s "2 . IO »C • • • • 1 s .§<"£ 3 ^ 3 S n 5 £J /o 0 O^ c o "^ o H Ot p>t 00 *-*J oco»oeo'O eoio <£> Qr Q *" » Sn || g 0 O O uu 03 -f M a iH H 99 -u Sj t>°S 0} J A : : i i-S .§ ° ! : ? •^ ^ ^ §" • ,„_! r^2 a; O a. M _o C S 0 "00 r^ ' 1 1 •?•§ «4-l O 05 11 0 T ? § . : rH — C3 S kl 1 I -s=s : ^ Q g O tti • ^ f 3 -jilt! -i_. (B O rf) o c ^ .2 °£ a 05 X_> si 1 1 |l| •| 1 3 11 1 S o g o JE; M ^ § "o 'S *C C "^ 'u fl o y "® 48 ^ g a r^ rg rji ^C ^ F— j 173 *73 TT ^ '^ 03 3 1 S '§ 'o J 1 -2 S 5 ^.i§ &.^ ?^.2 | •§ CO 0> PRHftHRH P PftHfiHPH — d •-c o vji S -'— C3 O 1o ^ Ol CC •-" l ^. '" ^ ^— < i w^ 00 'M c^i CC !>• O ^* !^ ^^ ^— . Q-\ ,0 O ""B" 5 l^l ! 55 T) g *»*7 § C *o 111 s!-!3-5! "1 Si 1 «r 1 - J | « i- s || p« s* o c § -H i Js P< " *O G o ;5 QJ c*3 ^j^^E^-2*? 1 c -"^ ° "Z 'S •a" §H j; 5-1 |g "o ° ^ "* ^s 8 J3 5>< ; 53 ^ O ^" ^ ^ *^ S'O'H ^ OUCfl^5 'S *- ^ p« * . I'S S £-3 I 'C^-p, "°'|B SH^'5 8. "a "sj O O^£y-,H<~-ol:G5 153 S o s ,- "° "° *** -S . 5 B C O O O C ev^^X0-1*0-^ .C4J **j'O ^ *S -"3 .ir- C*H Sf *3 TR o 5 Q S ^C'^o'3^'3^ "S Q ^ ® -M"+-T ® S *™ § ^J" £ ^* 3 o aj^aJbCO^pS *o "8 5 P--^ "° >^3 ^ § BiS f |^ Is^f s^l«e •§ § g * "i"? -g s 1 a 1 'S. c8 — £• n, ** e I d}1^ o^-^—* S* o> f 3 S >> S jj 3 "§ jT C •^ !|!|SHit|J HI pin if s •§ •1 if! a, e 7j ™ ~ ~ j2 § ^i" M 'S ,2 -a .2*3^ S .5 "S. ^ jj £ "S. "p, «-l •2 . "* "o. § 5^ SSSS^gino'^wg .~^° .^So-SsS^" O S S * e 3 « •G 'c H "^ te <-* ** tu *o^l£^"fo S^1^ ",•*)» ^ *3 •rt r] pi S « .2 ^ .2 »J a 'S 3 o S. O S *" ~ '2 g o <» _ s S • ^ |S r- § £ B 3 8 = § - C 'S -~ ^ B £oo'3s 5'a!° tj'"00^ £ | 3"? S •* 3 ^ ^ojsj'ckubo'g w o'S s*^ •§ » 1 0.2-S a o » w £,§ O J3 | x^x^O gj l^ 1 S"i«I ™ 5"* «^s3 xl S 00 O -». "X S *M ^ S % .5 5 c S .E 2' * 2* ° * '" 2 .5 "2 o * -S '|-S .5 g"3 « -S 5 1 a 1 -^1 C •o e !3.5<3.£ H Q H O EH E- H H 0 H S H 1 — * — 1 ^ * ' 01 • ^ £H o, £4 I = 2 = | S S S 5 1 8 1 1 s K CO co CO CO 9999 o o o o o o o o O O lH CO ^ CO rH 03 i-1 CO i-l H £ * ^H CO i ^l*8 rrt ^ ® I -a OJ a 5 A ~ 'T ->.S ^ -^ H H P 4 ^ Tfi . i ^^..^ Oi — ^^ ^--^-v ^-,^-s CO --^^-^ S ^\^*~v ^0 ?T r-< CO 30 O'rH IN CO 00 Cl i— i CD co co co >ra 10 *""} CO QO oo ^ co~" to" i-i oo oo oo 00 t- 00 i— 1 00 OO M W^ CO CO CO een furnished with the Thames water bacteria by in- oculating with a few drops of unsterilised Thames water, the typhoid bacilli were still demonstrable on 19.2.1894, or thirty- four days after their introduction, whilst seven days later, on 26.2.1894, they were no longer discoverable. (5.) Thus, not only was the longevity of the typhoid bacilli far greater, as usual, in the sterilised than in the unsterile waters, but of the two unsterile waters, the one naturally so, and the other rendered unsterile by the inoculation of a few drops of unsterile Thames water, the naturally unsterile one proved to be decidedly more antagonistic to the vitality of the typhoid bacillus, than the water rendered artificially unsterile, as we may call it, by inoculation with unsterile Thames water. (6.) This result is the more significant and important, inasmuch as it was shown (see p. 521) that in the naturally unsterile water no multiplication of the water bacteria took place, whilst, in what we may call the artificially unsterile water, it was shown (see p. 523) that a very large multiplication of the introduced water bacteria certainly did take place. (7.) In the previous series of experiments (see p. 516) it was equally clearly shown that the typhoid bacilli enjoyed a greater longevity in the unsterile deep well water than iu Report on the Bacteriology of Water. 529 either the unsterile Thames or Loch Katrine waters, although it was precisely in the deep well water that the water bacteria underwent multiplication. (8.) These experimental observations lead me to the conclusion that the antagonistic action of the unsterile waters on the typhoid bacillus is not to be attributed to the multiplication of the water bacteria leading to the suppression of the typhoid bacilli " by competition in the struggle for existence," to use the common phraseology of many writers on these subjects, but through the existence in the unsterile waters of conditions (due, doubtless, to a great extent to the presence of chemical products elaborated by water bacteria) which are inimical to the vitality of the typhoid bacillus. That conditions inimical to the vitality of the typhoid bacillus can be generated by the water bacteria alone is demonstrated by the above experiments in which a few drops of unsterile Thames water were added to the typhoid-infected steam-sterilised water, with the result that the longevity of the typhoid bacilli in this water was far less than in the steam-sterilised water itself. That the longevity of the typhoid bacillus is still less in the natur- ally unsterile water than in that rendered unsterile by inoculation, I attribute to the fact that in the naturally unsterile Thames water, countless generations of water bacteria have flourished before the water is made the subject of experiment at all, and it must, therefore, be more or less saturated with those bacterial products which are prejudicial to the vitality of the typhoid bacillus, and which, in fact, frequently hamper or even inhibit the further multiplication of the water bacteria themselves. The deep well water, on the other hand, in its natural condition is in a very different state ; as my numerous former examinations (see Second Report, pp. 178 — 180) of this water have shown, it is drawn from its subterranean source in an almost perfectly sterile condition, having never since its exhaustive filtration through the porous strata of the earth, in which it has altogether altered its chemical composition, harboured any micro-organisms at all, so that the abundant bacterial multiplication, which, as I have shown, it exhibits in the laboratory, is really the first time that it is subjected to the influence of bacterial growth, and it takes a correspondingly longer time, therefore, for the conditions inimical to the typhoid bacillus to be established. In fact, in the experiments with deep well water (see pp. 505, 515) the typhoid bacillus was actually discovered in the unsterile water a little later than in the steam-sterilised water. That the multiplication of the bacteria in the deep well water does not lead to conditions so antago- nistic to the vitality of the typhoid bacillus is also, no doubt, attri- 530 Profs. Percy Frankland and Marshall Ward. butable to the circumstance that the number of different kinds of bacteria in the deep well water is much more limited than in ordinary surface waters. For the practical hygienic application of these experimental observations, reference should be made to the final conclusions at the end of this Report, see p. £>43. Further Experiments on the Influence of the Addition of Common Salt to Water containing the Typhoid Bacillus. The several waters prepared for the last series of experiments were also made to serve the purpose of verifying the remarkable results previously obtained (see p. 434, et seq.) by the addition of common salt to typhoid-infected waters. These experiments were commenced on 12.2.1894, on which day' some of each of the following waters (for full particulars concerning them, see p. 519) received 1 per cent, of pure sterile sodium chloride respectively : — (a.) Typhoid-infected steam- sterilised Thames water. (6.) Typhoid-infected steam-sterilised Thames ivater, inoculated with a few drops of unsterile Thames water. (c.) Uninfected unsterilised Thames water. (d.) Typhoid-infected unsterilised Thames water. These waters, to each of which 1 per cent, of sodium chloride had been added, were preserved in a dark cupboard at a temperature of 9 — 11° C., and were submitted to periodical examination along with the several waters of the last series, with which they were to be compared. Uninfected Unsterilised Thames Water to which 1 per cent, of Sodium Chloride was added on 12.2.1894. Dates Number Volume of Number of on which plate cultivations were made. of dajs plates were in- cubated. water employed for plate cultivation. colonies obtained from 1 c.c. of water. Kemarks. c.c. 14.2.1894 5 TO and ^ 2,150 Only very small number of liquefy- ing colonies. 19.2.1894 3 To-acd^ 607,000 Very large number of liquefying colonies. 24.2.1894 3 ToandTib 1,200,000 Numerous liquefying colonies. Report on the Bacteriology of Water. 531 On comparing this table with the one on p. 520, which refers to the same water, only without the addition of salt, it will be seen that the effect of the salt addition was to cause an enormous and rapid multi- plication of the water bacteria. The liquefying colonies underwent great multiplication, but there was also an enormous increase in other non-liquefying colonies of various kinds. Typhoid-infected Unsterilised Thames Water to which 1 per cent, of Sodium Chloride was added on 12.2.1894. Dates on which plate cultivations were made. Number of days plates were in- cubated . Volume of water employed for plate cultivation. Number of colonies obtained from 1 c.c. of water. Remarks. 14.2.1894 5 c.c. i TOT 5,000 Only a very small number of lique- fying colonies. 19.2.1894 3 -So and a'o 404,000 Considerable number of liquefying colonies, but special increase in small non-liquefying colonies. 24.2.1894 3 1 OT,r] 1 To anu TO 0 407,000 Comparatively few liquefying colo- nies, but an enormous number of small non-liquefying colonies. On comparing the above table with that on p. 521, it will be seen what an enormous multiplication of the bacteria present was caused by the addition of the salt ; in this case the multiplication was not so much shared in by the liquefying organisms as by the others. Typhoid-infected Steam-sterilised Thames Water to which 1 per cent, of Sodium Chloride was added on 12.2.1894. Dates Number Volume of Number of on which plate cultivations were made. of days plates were in- cubated. water employed for plate cultivation. colonies obtained from 1 c.c. of water. Remarks. c.c. 14.2.1894 5 i and TV 17,385 Typical pure cultivation of typhoid. 19.2.1894 8 A and TL 360 Ditto. 24.2.1894 16 f and T\ 6 Only depth colonies on the plate. 532 Profs. Percy Frankland and Marshall Ward. On comparing this table with that on p. 522, it will be seen that the addition of the salt was highly prejudicial to the typhoid bacilli, the latter exhibiting a most rapid diminution in number, and degene- ration in their vitality. Typhoid-infected Steam-sterilised Thames Water inoculated with a few drops of Unsterile Thames Water, to this 1 per cent, of Sodium Chloride 'was added on 12.2.1894. Dates Number Volume of Number of on which plate cultivations were made. of dajs plates were in- cubated. water employed for plale cultivation. colonies obtained from 1 c.c. of water. Eemarks. c.c. 14.2.1894 3 5^ and ^ 25,725 Numerous liquefying colonies and large number of small depth colonies. 19.2.1894 2 and TTO 296,500 Ditto. 24.2.1894 3 s-V and T£T> 340,000 Ditto. On referring to the table on p. 523, it will be seen that in this water, before the addition of the salt, a considerable multiplication of the water bacteria had taken place, but at the time the salt was added, the total number of bacteria only amounted to from 2000 — 5000 per 1 c.c. ; after the addition of the salt, however, an enormous multipli- cation followed. We must now see how the vitality of the typhoid bacilli was affected by this enormous bacterial multiplication which took place in the unsterile waters after the 1 per cent, of common salt was added to them. This was, of course, ascertained by the method of phenol broth-culture, the results of which are recorded in the following table : — Report on the Bacteriology of Water. 533 *O M _^ • _fl rQ f^ >, JJ J- M ee S S" 2 c P S O tf £ £ P-'o ^* Wi *C3 *S P.O >>!*'M a o *a "^ ft 5a Sj§ 8 8 "^'S ^ o S ^ 'o. P< "p. as ^-, _*3 .?! « ^ ^ S1 o * -a ^ • £? ^»'AJ *> ** 00 M « • oj ^ *^ i— i .£* C £ ^ — ^ o ^* ft 03 |t ;|li ^ g. •| B " 1 8 s v ^^ S °° ^'^ °^ R ^ '^ m ^8 "H^^1 **" ** ^ •S WE !2 5 a 'S .^ <» 5 ^ £ -^ ll- l^sl «1 is 2 * a p i1 KOJ pu-'5§ S^ £1 "S •o *o ^ S **~l ' — ^ P* .is^0*' §? ^ * 3 2 5 * *S 5 iP s § a ^ ® f^ f g S § _g 5 2-|^r|'e-s 2 ^ ^O<^*-»'5ioC'cc OOco o 1 I ffteij ^t|^iM'Or*'WL^O'3t*'^^~' ^ 'f *^ r^ 0> O ^1.^fo^'CQj .9 oq £ 3 .£ J ^ 3 ? -S 'l -S -S 1 IS "° 2 '- &~ * •- ^ l2^'~"5'^Cw®3'^ai 'C'oss 3 O 0 -0 »',C £ -^ Q IS»c*J^Q^a''*J)Q"3 )3i3 "*^ S H 2a>3-J=c |£8|£8S§|J£JJ,S' o a a &§si.£:2 O Q E- O .fe>ro d o . 5 a 2 "^ " -a & & O »x .. K » «b- M »4 |£l|3JS COWCO »« «« U3 CO CO CO CO ||r t-l 73 «4- "3 =4-1 TJ d llllll IB! ^**-^2i^r: tHT3^ 13 T303 fio C^5 dja.'^ r^wflrQSrQ ^t..^ M 1 lisii siiii^i 1'Sl c3'--r9fr;^rd «^^fl *3 r^ ^* O ^ O ^"> ** • p ° PH"" pE^^fiH1 °P t> °H""H * H* •a^j! 1 1 III OOO •<* i-l VDQCOQO3 CO £q CO CO CO CO CO CO CO CO GO CO •!>• "^ *O i— i CO CO CO CD i*q 00 CO CO CO 03 § "g H *S 1— 1 05 ^" 534 Profs. Percy Frank! and and Marshall Ward. •a •= r,; .-B •= ps ^ ^ ^ *o o o *o ^ P. £ 0.0. g 3 ' ia"5. "'S £ 3 « „ « » 'S K 3 j & S * 0 'S o 2 jj 00 c -3 8 g ?> .2 § T3 "3 ^ H '•§§!§ -s 1 g -1 5 •r 8 g S 'S "§ 1 "S ti ^ ° s 1-1 ~ 3 9 00 1 u '3 « P. ;~ g-g, 3i? >, l^f s' H ed S ff . * n rO H >> a i O> cj JA O 1 . Si B 1 2 08 G .« Sc^r "^ % 0 "| a| | m '-° -2 ^ J ^ ^ A 2 Is Z; EL, ^ _2 6 BH ' ' a | S • » § ' « o5 >» >t «| S3~-a >>.§ o s S o » a S £^ 5 f. f. 3 ^ S * .g*_g * .S -S a ^i ^s S "^ s ^ •« 2 " H 3 •|^' |^|^ alial 2 3 3 3 ^S .0 cp ^2 ^ s s s 1 •*-* *-* a B a 2 E-I 3 s t- ^a *- o 3 OS 3 a " o -N P H H H H H £5 fz; •i 0 H •:•='§ .§ •" o A o cl p. o - o q ^f- Oj r~^ ^ r-rt x~* si *"O 4! •i ^ .RtZT'o ^ r-l rd CO CO CO CO CO CO CO ^ § O* o <° * ~" § as *i a Q "S S • -H o o o o 9 -< ^ ri SE ® o co i-l CO i-l CO 3 bE «} *** o -« s 3 .J B i "o a P, 1 ! 1! 3: i | «-| 43 PH _C g "~J 1 IS- s i-.i * §s" "S Hi o h « PH D 2-. £ a 0 p< *S *^ ' x * T^ •-1 TJI •" g N IslJ^ != « •§ ^ •^3 is ^5 (M* H .2 *P /^ci tn O ^ *j^ _.; _l t^i "^ ~^J 1^ l- T3 Oi . Q 7^ ^S Qj ^j "§ • ^ ""* OS •;" "oo ^ ' j3 "B „ ^ *" "^ "ft » 5 S "o P * 1 llll iiri S^ S "S CD c3 ^ c3 C ^ II -H a 0 f° °° rcSSo3'°°'o® "fe * "S o "S «3 1 1 | |l||:||a I'illa Q; C-J CJ W £J 11 ll — « 1 rjj * fl ~ O jrt "^ "e !§ S ^ lo H £ h 3 •§-§^5 s ^J g 12 2 'H ^ ° 03 «£'| 'S .2 'S jg ^ ° -3 2 Ji 43 *S 3 -i D ill „> a .5 O &-> oj ^ ^ O P H H ^» M >> ^ O H H >^ m *-H H P h* *- /3 _, "g 3S Profs. Percy Frankland and Marshall Ward. Steam-sterilised Peaty Water, infected 30.10.1893 with — continued. Typhoid bacillus. Plate cultivations made _B. coli communis. 6.11.1893 , • ^ 1044 colonies from 14,031 colonies from 1 c.c. (plates 1 c.c. (plates incubated 7 days). incubated 7 days). 15.11.1893 432 colonies from 17,580 colonies from 1 c.c. (plates 1 c.c. (plates incubated 5 days). incubated 5 davs). 23.11.1893 220 colonies from 14,457 colonies from 1 c.c. (plates 1 c.c. (plates incubated 6 days). incubated 4 days). Peaty Water Sterilised by Filtration, infected 30.10.1893, with— Typhoid bacillus. Plate cultivations made S. coli communis. 3010.1893 , . _^ 1169 colonies from 21,603 colonies from 1 c.c. (plates 1 c.c. (plates incubated 6 days). incubated 4 — 6 days). 6.11.1893 r- * : x No colonies on plates No colonies on plates (plates incubated (plates incubated 9 days). 9 days). Thus even in this water, heavily charged as it was with vegetable organic matter, the typhoid bacilli failed to undergo any multiplication, but, on the contrary, as usual, suffered a continuous numerical decline. The B. coli communis, on the other hand, remained practically unaltered in numbers during the period of upwards of three weeks over which these observations extended, the slight numerical increase being too insignificant to be regarded as evidence of true multiplication. In the peaty water which had been- sterilised by filtration through porous porcelain there was again the same phenomenon, so frequently referred to before in this Report, of the extraordinarily rapid disappear- ance of the introduced tijphoid and coli bacilli. On the possible Adaptation of the Typhoid Bacillus to active life in Potable Water. In none of the experiments recorded above was any multiplication of the typhoid bacillus observed, although these experiments have been made with waters varying from the deep well water of the Kent Company, which is almost wholly destitute of organic matter, to the peaty water referred to on p. 537, which contains about the maximum amount of organic matter met with in water used for drinking .purposes. Report on the Bacteriology of Water. 539 Some previous observers, on the other hand, record the multiplica- tion of the typhoid bacillus in potable waters with which they have made similar experiments ; whilst others, again, have found no multiplication. It appears to me highly probable that in most, if not in all, cases in which multiplication has been observed, it has been occasioned through the introduction of an appreciable amount of food-material along with the typhoid bacilli; for, as already pointed out, most investigators have exercised very little care in respect of this highly important factor. But although the typhoid bacilli taken directly from an ordinary cultivation and plunged into potable water may not be able to pro- liferate in the latter, it appeared to me quite possible that if the environment of the typhoid bacilli were gradually, instead of sud- denly, changed, the requirements of the bacilli might perhaps be thereby so far modified as to undergo multiplication in the aqueous medium. For by gradually changing the surroundings, it would be anticipated that those individuals most capable of nourishing under the altered conditions would be propagated, and that each successive generation of typhoid bacilli would thus become more adapted to the new medium. To ascertain whether this process of education could be actually accomplished, the following experiments were made. Education of Typhoid Bacilli for Aquatic Life. A gelatine-culture of the typhoid bacillus was, in the first instance, inoculated into sterile broth of the ordinary strength, and kept at Ifi — 20° C. ; turbidity ensued in twenty -four hours. From this broth cultivation an inoculation was made into 50 per cent, broth (broth mixed with its own volume of water); this also became turbid in twenty-four hours at 18 — 20° C. From the 50 per cent, broth culti- vation an inoculation was made into 10 per cent, broth (1 volume of broth mixed with 9 volumes of water) ; this liquid only became visibly turbid in from two and a half to three days. After four successive generations of cultivation in this 10 per cent, broth medium had been carried on, the time elapsing between the inoculation and the appear- ance of turbidity gradually diminished until, with the fifth generation, turbidity already set in in twenty-four hours. From this 10 per cent, broth, inoculation was then made into 1 per cent, broth (1 volume of broth mixed with 99 volumes of water) ; this became turbid in twenty-four hours. Continuous cultivation in this 1 per cent, broth medium was then carried on for a period of two months, after which it was employed for infecting steam-sterilised Thames water, thus : — On 31.1.1894, 2 drops of a 1 per cent, broth cultivation (three VOL. LVI. 2 o 540 Profs. Percy Frankland and Marshall Ward. days old) of the typhoid bacillus were added to 600 c.c. of steam- sterilised Thames water, which was, on the day of infection and on several subsequent occasions, submitted to plate cultivation with the following results : — Number of days Dates on which plates plates were Number of typhoid bacilli were prepared. incubated. in 1 c.c. of water. 31.1.1894 8 4,895 2.2.1894 ...... 7 15,372 5.2.1894 10 11,184 12.2.1894 7 6,558 23.2.1894 7 5,795 6.3.1894 6 6,068 10.3.1894 9 4,093 These figures show that unquestionable, although not very exten- sive, multiplication of the typhoid bacilli took place in the water thus infected ; but in order to ascertain whether this proliferation was effected at the expense of the very small quantity of culture- material necessarily introduced along with the bacilli, or at the ex- pense of the organic matter pertaining to the Thames water itself, the following further experiment was made : — On 23.2.1894, 10 c.c. of the above-infected water, which on that day contained 5795 typhoid bacilli per 1 c.c., were added to 20 c.c. steam-sterilised Thames water, and the latter was then and several times sutsequently submitted to plate cultivation with the follow- ing results : — Number of days Dates on which plates plates were Number of typhoid bacilli were prepared. incubated. in 1 c.c. of water. 23.2.1894 7 1830 26.2.1894 7 1842 2.3.1894 6 375 6.3.1894 6 268 From these figures it is equally evident that in this case no multi- plication but only numerical decline of the typhoid bacilli took place. If, however, the typhoid bacilli can proliferate at the expense of the organic matter belonging to the Thames water, they should have multiplied in the above experiment, as they were imported into a quantity of water, the organic matter of which had not since sterili- sation been exposed to bacterial life ; but from the fact that they did not multiply, but, on the contrary, only fell off in numbers, it becomes almost certain that the distinct multiplication observed in the former experiment was effected at the expense of the small quantity of food- material originally introduced into the water along with the typhoid bacilli. Report on the Bacteriology of Water. 541 Another experiment was made on similar lines to the first half of the above experiment, to see whether the multiplication there observed could be confirmed, thus : — On 27.3.1894, 4 c.c. of a 1 per cent, broth cultivation (three days old) of the typhoid bacillus were put into 50 c.c. of steam- sterilised Thames water, the mixture being violently shaken up for fifteen minutes ; 2 c.c. of this mixture (equivalent to -$T c.c. of the original 1 per cent, broth culture) were then added to 200 c.c. of steam-sterilised Thames water, which was then submitted to plate cultivation as follows : — Number of days Dates on which plates plates -were Number of typhoid bacilli were prepared. incubated. in 1 c.c. of water. 27.3.1894 6 37,515 29.3.1894 6 61,566 31.3.1894 9 50,935 4.4.1894 6 27,818 11.4.1894 8 20,130 In this case again, therefore, there was a smalt but distinct multi- plication. From these experiments it appears that typhoid bacilli ivhich have undergone a prolonged and gradual training in more and more aqueous culture-media do exhibit distinctly more vitality in potable ivater than bacilli which are at once transferred into icater from highly nutritive solid media like agar or peptone jelly. On the other hand, there is con- siderable reason for believing that the slight but distinct multiplication which these trained bacilli undergo in potable water, is effected at the expense of small quantities of food-material introduced along with them, at the time of infection, and not at the expense of the organic matter belonging to the water itself. The result of the experiments with these specially-trained typhoid bacilli greatly fortifies the opinion which I have expressed above, that the extensive multiplication of typhoid bacilli in potable waters which has been observed by some investigators was most probably due to the importation of appreciable quantities of food-material along with the bacilli themselves. SUMMARY. The investigation which has been detailed in the foregoing pages is divisible into the following sections : — 1. A series of experiments in which the vitality of one and the same culture of the typhoid bacillus was observed in one and the same sample of Thames water, using the latter under the following conditions : — 2 o 2 542 Profs. Percy Frankland and Marshall Ward. (a) In its natural unsterile state ; (&) Sterilised by steam ; (c) Sterilised by nitration through porous porcelain. In ea'ch case the effect of temperature was studied by pre- serving the infected waters at winter and summer temperatures respectively. 'Description of experiments, see pp. 409 — 433, 451 — 465 ; summary of conclusions, pp. 433, 434.) 2. A perfectly similar series of experiments, carried on side by side with the above, in which the Bacillus coli communis was em- ployed instead of the typhoid bacillus. (Description of experiments, see pp. 409 — 433 ; summary of conclu- sions, pp. 433, 434.) 3. A series. of experiments in which the effect of the addition of common salt in various proportions to unsterile Thames water was studied, the unsterile Thames water being employed for this purpose both uninfected and infected with the typhoid bacillus. (Description of experiments, see pp. 434 — 450 ; summary of conclu- sions, p. 450.) 4. A series of experiments perfectly similar to No. 1 above, in which Loch Katrine water was employed instead of Thames water. (Description of experiments, see pp. 465 et seq. ; summary of con- clusions, pp. 476 — 486.) 5. A further series of experiments made with the same sample of Loch Katrine water, only introducing a much larger number of typhoid bacilli into a given volume of water ; in this series of experiments only unsterilised Loch Katrine water was em- ployed. (Description of experiments, see p. 486 ; summary of conclusions, p. 492.) 6. In order to compare the relative longevity of the typhoid bacillus in the more important types of potable water, a long series of experiments was carried out in which typhoid bacilli from one and the same cultivation, and in as far as possible equal num- bers, were introduced into Thames water, Loch Katrine water, and the deep well water of the Kent Company respectively. Each of these waters was employed : — (a) In its unsterile natural condition ; (5) Sterilised by steam ; (c) Sterilised by filtration through a porous cylinder of infusorial earth. Report on the Bacteriology of Water. 543 In this series of experiments the influence of rest or agitation was also studied. (Description of experiments, see pp. 493 et seq. ; summary of con- clusions, pp. 516 — 518.) 7. A series of experiments made in order to ascertain whether tho bactericidal properties of unsteriltsed surface water can be artificially induced by inoculating steam-sterilised Thames water with a few drops of unsterilised Thames water, and thus giving rise to a bacterial population in the previously sterile water. (Description of experiments, see pp. 519 et seq. ; summary of con- clusions, pp. 528 — 530.) 8. Further experiments on the addition of common salt to typhoid- infected Thames water, both sterile and unsterile, with a view to confirming or contradicting the results of the experiments referred to under No. 3 above. (Description of experiments, see pp. 530 et seq. ; summary of con- clusions, pp. 535 — 536.) 9. Experiments made to ascertain whether the typhoid bacillus and the Bacillus coli communis, multiply in potable water which is very highly charged with vegetable matter (peaty, upland, surface water). (Description of experiments, see pp. 536 et seq. ; summary of con- clusions, p. 538.) 10. Experiments made to ascertain whether the typhoid bacillus can, by prolonged preliminary culture in more and more diluted media, be trained for aquatic life in potable water. (Description of experiments, see pp. 539 et seq. ; summary of con- clusions, p. 541.) For the detailed conclusions arrived at from the results of these several series of experiments, the reader is referred to the summaries which are appended to the descriptions of each series of experiments, as indicated above, whilst the general conclusions which arise out of the entire investigation may -be summarised as follows : — Summary of Conclusions. 1. Typhoid bacilli from ordinary agar-agar- and gelatine-cultures on being introduced into steam-sterilised potable water in such num- bers as not to materially alter the composition of the latter undergo TIG multiplication. This result was uniformly obtained irrespectively of whether surface water like that of the Thames, which has received the drainage of manured land, or upland surface water like that of Loch Katrine, the organic matter in which is very similar in absolute 544 Profs. Percy Fraiikland and Marshall Ward. amount to that in Thames water, but almost exclusively derived from vegetable sources (peat) ; or, again, other upland surface water much more highly impregnated with peaty matter ; or, lastly, deep well water containing the merest traces of organic matter, was employed. In all cases, of course, special precautions were taken to prevent, as far as possible, the importation of culture- material along with the bacilli. 2. By first submitting the typhoid bacilli to prolonged culture in more and more aqueous media, and then introducing them into steam-sterilised Thames water, slight but distinct multiplication of the typhoid bacilli was observed, and, although, perhaps, by this method of training the typhoid bacilli had become more adapted to aquatic life, it appears probable that the multiplication observed took place at the expense of the minute quantity of culture-material neces- sarily introduced with them ; for on transferring some of this infected water in which multiplication had taken place to a larger volume of the same steam-sterilised Thames water, no further multiplication was found to occur, showing that the organic matter belonging to the steam-sterilised Thames water itself was not capable of ministering to the growth and proliferation even of these specially educated typhoid bacilli. 3. Although no instance of multiplication of the introduced typhoid bacilli in these steam-sterilised potable waters was observed, on the other hand the bacilli were found to be possessed of very considerable longevity in them, thus : — Description of steam-sterilised water. Thames water (11.5.1893) kept at 6— 8°C. . „ „ 19° C. Loch Katrine water (4.7.1893) kept at 6—8° C. Duration of life of typhoid bacillus. Upwards of 76 days! Still just recognis- „ „ J able. Upwards of 21 "I Only a small number of days I typhoid bacilli was in- Loch Katrine water (4.7.1893) kept at ! Between 13 f troduced into these 19° C. Thames water (19.10.1893) kept at 9—12° C. Loch Katrine water (19.10.1893) kept at 9—12° C. Deep well water (19.10.1893) kept at 9—12° C. Thames water (16.1.1894) kept at 9—12° C. Peaty upland surface water (17.10.1893) kept at 9—12° C. and 17 days J waters. Between 32^ Typhoid bacilli from and 39 days j one and the same Upwards of 51 ! source and in the same days j numbers were intro- Between 20 j duced into each of and 32 days J these waters. Upwards of 48 days (still abundantly present) . Upwards of 24 days (still abundantly present). In no case was the duration of vitality a very limited one ; its exact length in any particular water is doubtless dependent on the initial vitality of the bacilli and the numbers in which they are introduced. In the strictly comparative experiment on steam- sterilised Thames, Loch Katrine, and deep well water, it is seen that the longevity of the typhoid bacillus is distinctly greatest in the Loch Report on the Bacteriology of Water. 545 Katrine, and least in the deep well water, and intermediate between the two in Thames water. Of these three waters, also, the Loch Katrine contains the most, the deep well the least, and the Thames an intermediate amount of organic matter. Not improbably these circumstances are connected together. 4. The experiments distinctly show that in these steam-sterilised potable waters a summer temperature of 19° C. is more prejudicial than a winter temperature of 6 — 8° C. to the duration of life of the typhoid bacillus. 5. Inasmuch as the numerical estimation of typhoid bacilli in unsterile potable waters is practically impossible, the duration of life of the typhoid bacilli in such waters has alone been made the subject of study. This enquiry has involved an enormous amount of labour, as the certain detection, by means of the special methods em- ployed, of the typhoid bacillus, even in a single specimen of water, may entail work extending over several weeks. The duration of life of the typhoid bacilli introduced into the various unsterile waters in the several series of experiments was as follows : — Description of unsterile water. Thames water (11.5.1893) kept at 6—8° C. ,, 19° C. Loch Katrine water (4.7.1893) kept at 6—8° C. Loch Katrine water (4.7.1893) keot at 19° C. Loch. Katrine water (7.7.1893) kept at 6—8° C. Loch Katrine water (7.7.1893) kept at 19° C. Thames water (19.10.1893) kept at 9—12° C. Loch Katrine water (19.10.1893) kept at 9— 12° C. Deep well water (19.10.1893) kept at 0— 12° C. Thames water (16.1.1894) kept at 9—12° C. Duration of life of typhoid bacillus. >• Between 25 and 34 days. Upwards of) Only a small number of 17 days I typhoid bacilli was in- Between 4 f troduced into these and 11 days J waters. f Upwards of 14 days, after which, no ) further examinations were made. (A. much larger number of typhoid bacilli was introduced in this than in the above experiments.) Between 9~") Typhoid bacilli from one and 13 days Between 19 -, uuu and 33 days j bers were Between 33 | into each and 39 days J Between 20 and 27 days. and the same source and in the same nuin- introduced of these waters. On comparing this table with that given under No. 3 above, it will be seen that in all cases, excepting one, the duration of life of the typhoid bacillus was greater, and often much greater, in the steam- sterilised than in the corresponding waters unsterilised. The single exception to this general rule was in the case of the deep well water in which the typhoid bacilli lived about the same length of time, irrespectively of whether the water was sterile or not. The table also shows that, as in the case of the steam-sterilised waters, the exact length of time that the typhoid bacilli endured residence in one and the same type of unsterilised water was subject 546 Profs. Percy Frankland and Marshall Ward. to great variations in the different experiments, doubtless largely in consequence of the different initial vitality of the typhoid bacilli em- ployed, and also, doubtless, in consequence of the different numbers in which they were introduced in the several series of experiments. Of principal interest is the comparative experiment made with Thames, Loch Katrine, and deep well water, in which typhoid bacilli from one and the same source, and in the same numbers, were intro- duced into these three types of water, and in which the duration of life was found to be shortest in the Thames water (9 — 13 days), longest in the deep well water (33 — 39 days), and intermediate iu the Loch Katrine water (19 — 33 days). This result is of very great practical importance as indicating the greater danger of typhoid bacilli gaining access to deep well than to surface water. This danger is, in actual practice, further enhanced by the fact that well water is almost invariably consumed without storage, whilst surface- waters are often stored for days or weeks, and in the case of upland surface water the storage frequently extends over many months. The effect of temperature on the duration of life of the typhoid bacillus was well illustrated in the series of experiments with Loch Katrine water (4.7.1893), in which it was found that at 19° C. the typhoid bacilli had already disappeared in 4 to 11 days, whilst at 6 — 8° C. they were alive for upwards of 17 days in the same water. The effect of agitation or rest on the typhoid bacilli in these waters was not very pronounced, but the evidence on the whole, both in the case of the sterilised and unsterilised waters, goes to show that the agitation or aeration of the water is unfavourable to the typhoid bacilli, and that in the unsterilised waters it occasions a more rapid multiplication of the water bacteria. 6. The greater bactericidal power of unsterilised than steam-steri- lised surface waters is not apparently due to the multiplication of the water bacteria in the unsterile waters, bringing about a competition or " struggle for existence " between these aquatic forms and the typhoid bacilli, but rather to the elaboration of products by these aquatic bacteria (and very possibly also by other vegetable life present in surface waters) which are inimical and prejudicial to the welfare of the typhoid bacilli. Thus, in the typhoid-infected unsterilised deep well water an enor- mously greater multiplication of the common water bacteria took place than in the unsterile Thames and Loch Katrine waters ; yet, notwithstanding the typhoid bacilli not only lived much longer in this unsterile deep well water than in the unsterile Thames and Loch Katrine waters, but there was practically no difference between the duration of life of the typhoid bacilli in the sterile and unsterile deep well water respectively. on the Baet&nology of Water. 547 Of course it may be urged that, the unsterile deep well water possibly does not contain those water bacteria which are particularly- fitted for entering into successful competition with the typhoid bacilli, and that perhaps such water bacteria are only to be found in the unsterile surface waters. 7. The series of experiments summarised in the following table show that unsterile surface water, like that of the Thames, possesses bactericidal powers irrespectively of any further multiplication of its contained water bacteria, thus : — Uninfected unsterilised Thames water, kept at 9 — 12° C., exhibited but little change in the number of its contained bacteria over the period of five weeks from 16.1.1894 to 24.2.1894. (The numbers only varied from 5500 — 2825 per 1 c.c.) The same unsterilised Thames water, infected with about 170,000 typhoid bacilli per 1 c.c., exhibited a continuous decline in the total number of. bacteria present in it over the same period. The same Thames water, after sterilisation by steam, was infected with upwards of 100,000 typhoid bacilli per 1 c.c., and at the end of the same period (16.1.1894 — 24.2.1894) there were still upwards of 5000 typhoid bacilli per 1 c.c. present. The same typhoid-infected steam-sterilised water was inoculated with a few drops of unsterile Thames water to communicate to it the Thames- water bacteria, and the latter underwent very extensive multiplication in this water. Notwithstanding, the typhoid bacilli lived between thirty-four and forty-one days in this water, whilst in the unsterile Thames water, in which no multiplication of the water bacteria took place, they only lived between twenty and twenty-seven days. This shorter duration of life of the typhoid bacilli in naturally unsterile Thames water than in that rendered unsterile by inoculation I attribute to the circumstance that in the naturally unsterile Thames water countless generations of water bacteria must have flourished before the water is made the subject of experiment at all, and it must, therefore, be more or less saturated with those bacterial products which are prejudicial to the vitality of the typhoid bacillus, and which, in fact, frequently hamper or even inhibit the further multiplication of the water bacteria themselves. Thus it is obvious that the unsterile water in question was already, at the outset of the experiment, in such a condition as to prevent any multiplication of its own water bacteria, whilst, after it had been steam sterilised, the same bacteria multiplied abundantly in it. But again, at the outset of the experiment the unsterile water was in such a condition as to cause a comparatively rapid disappearance of the introduced typhoid bacilli, whilst after steam sterilisation it only became again endowed with this power of destroying the typhoid 2 o 3 548 Profs. Percy Frankland and Marshall Ward. bacilli when the introduced water bacteria had undergone extensive multiplication. 8. The addition of common salt to unsterile Thames water, in the proportion of O'l, 1, and 3 per cent., causes the. enormous multiplica- tion of the water bacteria present, the most striking result in this respect being obtained with the largest addition of salt. 9. The addition of common salt to typhoid- infected unsterile Thames water diminishes the duration of life of the typhoid bacilli in this water, thus : — Duration of life of typhoid bacillus. Unsterile Thames water kept at 6-8P C.. . •••••••••} Between 25 and 34 days. Ditto with O'l per cent, salt" kept at 6—8° C ....... „ 25 „ 33 „ „ „ 19° C ....... Less than 18 days. Ditto with 1 per cent, salt, kept at 6 — 8° C ......... „ „ _ Ditto with 3 per cent, salt, kept at 6 — 8° C 18°C 10. This more rapid disappearance of the typhoid bacilli in the unsterile Thames water to which salt was added cannot be wholly, but only in part, attributed to the resulting multiplication of the water bacteria, as the addition of salt in similar proportion to typhoid- infected steam-sterilised Thames water also caused an exceedingly rapid disappearance of the typhoid bacilli, although the disappearance was not so rapid as in the same water to which a few drops of un- sterile Thames water had been added, and in which, therefore, a great multiplication of the water bacteria took place, thus : — Duration of life of typhoid bacilli. Steam-sterilised Thames water with 1 per cent, salt, Between 12 and 19 days. kept at 9—12° C. Ditto, to which also a few drops of unsterile Thames Less than 5 days. water were added, and in which extensive multi- plication of the so-introduced water bacteria took place 11. The Bacillus coli communis, taken from ordinary agar-agar cultures and introduced into steam-sterilised Thames water, under- goes considerable multiplication, when under precisely similar con- ditions the typhoid bacillus does not multiply. The behaviour of the B. coli communis in the steam-sterilised waters may be summarised as follows : — Report on the Bacteriology of Water. 549 Description of steam-sterilised water. ] Duration of life of -B. col! communis. Thames water (11.5.1893) kept at' 6-8°C. Thames water (11.5.1893) kept at 19° C. Loch Katrine water (4.7.1893) kept at &-8°C. Loch Katrine water (4.7.1893) kept at" 19° C. ( Still abundantly present, after consider- able multiplication, on the 75th day. Between 14 and 17 days. The coli bacilli were introduced in this case in much smaller numbers than in that of the Thames water above. No multi- plication was observed. Yery peaty water (30.10.1893) kept at ! Upwards of 24 days. Still present in 9 — 12° C. uncliminished numbers after very slight multiplication. 12. The Bacillus coli communis introduced into unsterile water persists in the living state for a much longer period than the typhoid bacillus. Thus : — ; Duration of life of the -B, coli communis. Unsterile Thames water (11.5.1893)1 -,-,• 1 ,. ir. -. , •, , ,, . a oo rt Upwards of 40 days, and doubtless much K6Dt at D — o O. If c.i • TT £ -i mv. i 111 - iooo\ f\ longer, but no further examinations Unsterile Thames water (H.o.1893) ( kept at 19° C. j Unsterile Loch Katrine water (4.7.1893) "| kept at 6—8° C. Unsterile Loch Katrine water (4.7.1893) kept at 19° C. made. Upwards of 17 days. No further ex- aminations made. 13. In the numerous experiments made on the behaviour of the typhoid bacillus and of the B. coli communis in water (Thames, Loch Katrine, deep well, and peaty water) which had been sterilised by filtration through cylinders of porous earthenware and baked in- fusorial earth, a most remarkably rapid disappearance of both bacilli was observed in all cases, excepting that of the Loch Katrine water. Further experiments will have to be made before any definite con- clusions can be drawn from these unexpected results. APPENDIX. The Behaviour in Potable Water of Anthrax Bacilli taken directly from the Animal Body. By PERCY FRANKLAND, Ph.D., B.Sc., F.R.S., and CHARLES TEMPLEMAX, M.D., B.Sc. In the 2nd Report to the Royal Society Water Research Committee, " On the Vitality and Virulence of the Bacillus anthracis and its Spores in Potable Water," the enquiry was mainly confined to the deportment either of anthrax spores alone, or of such mixtures of bacilli and spores which are found in the usual cultivations of anthrax 550 Profs. Percy Frankland and Marshall Ward. on artificial media. It is, however, obviously a matter of great importance to ascertain how anthrax bacilli, entirely free from spores, as they are found in the tissues of animals which have succumbed to this disease, behave when they are introduced into potable water. There is the more urgency for this investigation, as it was pointed out in the introduction to the Report that the experiments previously undertaken by others in this direction have led to highly discordant results. The experiments which we have made on this subject were incidental to another investigation, on which we propose reporting later on, but as these experiments should have been made in connection with the 2nd Report, had time and opportunity permitted, we are bringing them forward now to fill up without further delay the hiatus in that Report. First Series of Experiments. The spleen of a white mouse, dead of anthrax, was excised under the usual aseptic precautions, and transferred to a small sterile bottle containing 20 c.c. of sterilised tap water, in which it was completely broken up by bruising with a sterile glass rod. More sterile water about 50 c.c. in all, was added to the bottle, and the whole violently shaken for fifteen minutes, so as to ensure even distribution of the bacilli throughout the water. 10 c.c. of this water attenuation were then added to about 400 c.c. of steam-sterilised Dundee water and thoroughly mixed, after which this infected water was distributed amongst a number of sterile tubes plugged with cotton wool. This infected water was submitted to plate cultivation on the same clay, with the following results : — f c.c. water yielded 8,190* anthrax colonies per 1 c.c, 7,871 {1 Q Q C)OK V " 8325 TT » " 0,0/0 „ ,, Tube 18 {t *. To " " O, LUU ,, ,, Tube 19 (V " 10.248 L-jV ,, ,, 12,810 ,, ., Tube 20 { ^ 9'408 L-fV » >> 6,050 ,, ,, 90,949 Average 9,095 Thus, on the day of infection the water contained about 9000 anthrax bacilli per 1 c.c. * All these plates were incubated for 5 days at 18 — 20° C. / I ^ A *- Report on the Bacteriology of Water. 551 The tubes containing this infected water were kept in a dark cup- board at 12° C., and submitted to plate cultivation at intervals, thus : — • Number of Volume of Number of Plate cultures Number of days plates water used for anthrax prepared. tube. were incubated plate colonies per at 18— 20° C. cultiyation. 1 c.c. of water. 2 days after ~[ infection J 19 9 c.c. / T=o I T3o 6 10 .» 20 » / & I T*o 236 135 5 days after \ infection J 19 9 r -s- ~\ J 10 i ^ r V7 ? All the plates were free » 20 >» 1 1 To J from anthrax. These tubes were again examined on two subsequent occasions and again yielded sterile plates. Thus these anthrax bacilli, taken directly from the dead mouse and introduced into sterile Dundee water in such large numbers as 9000 per 1 c.c. of water, all died within five days, the water being kept at 12° C. Second Series of Experiments. In this series of experiments the anthrax bacilli from the dead animal were introduced both into steain-sterilised Thames and steam- sterilised Dundee water respectively, and these waters were preserved at different temperatures, in order to ascertain the influence of this factor on the result. The spleen of a white mouse, which had died of anthrax twenty- eight hours after inoculation, was broken up with sterilised tap water in the same way as already described above, and with this water attenuation larger volumes of sterile Thames and Dundee waters respectively were infected, and these infected waters were then dis- tributed amongst a number of sterile tubes plugged with cotton wool. On the day of infection some of these tubes were submitted to plate cultivation, in order to ascertain the number of anthrax bacilli introduced, thus : — f Tube No. 1 j J c-c- Dundee waters I » Thames water •< i. '{ft water yielded 4840 anthrax colonies per 1 c.c. 4860 6897 7722 6136 5875 6480 6552 552 Profs. Percy Frankland and Marshall Ward. The average number of anthrax hacilli in these infected Thames and Dundee waters at the outset was, therefore, about 6000 per 1 c.c. Some of these tubes were then placed in a refrigerator at 5° C., others were kept in a cupboard at 13° C., whilst others were put in an incubator at 19° C. The water in these tubes, kept afc the different temperatures specified, was submitted to plate cultivation at intervals, with the following results : — Water kept at 5° C. Plate cultures prepared. Description of water. Number of tube. Number of days plates were incubated at 18— 20° C. Volume of water used for plate cultivation. Number of anthrax colonies per 1 c.c. of water. 1st day after! infection J Thames •< 9 10 6 51 c.c. { A {ITS A 1028 2470 2550 2031 " Dundee < 3 4 H f 10 I S {5 10 A 3386 3733 3456 3335 2 days after 1 infection J Thames •< 9 10 6 i t f TJJ L ill 834 660 860 485 1624 » Dundee 4 3 4 » i ? L Tu 1995 1864 1497 ' 5 days after 1 infection J Thames \ 9 10 6 ; T\ i A / Tfe I T^T None „ Dundee < 3 4 » ( i£ H L -5-9 I A 134 235 None » n r 6 1 -0 None 14 days after ~| infection J Thames < i 9) 0-5 1-0 L 10 [ 0-5 » Dundee -j 4 [ „ 1-0 0-5 1-0 0-5 Thus, in the Thames water maintained at 5° C. the anthrax bacilli (introduced to the number of 6000 per 1 c.c.) had already undergone very Report on the Bacteriology of Water. 553 considerable diminution in numbers on the day following their introduc- tion ; after two days the numbers had still further diminished, whilst five days after introduction they were no longer discoverable at all. The fate of the anthrax bacilli in the Dundee water was quite similar, their dis- appearance being, however, a little less rapid; thus a few bacilli were still present in one of the two tubes on the fifth day after infection. The following table exhibits the results obtained with the same waters kept at 13° C. : — Water kept at 13° C. Plate cultures prepared. Description of water. Number of tube. Number of days plates were incubated at 18— 20° C. Volume of water used for plate cultivation. Number of anthrax colonies per 1 c.c. of water. 1st day after ! infection J Thames < 7 8 0 c.c. ; TV I T% { * 2640 2870 3564 3505 » Dundee •; 1 2 „ { s { I 4422 4048 3331 3153 2 days after "1 infection / Thames •< 7 8 6 I To L T2o / TV I TV 1690 2196 2035 >' Dundee •< 1 2 » J 1 I t ; fo I TV 2442 2848 2806 2785 5 days after "1 infection J Thames •< 7 8 6 { S J -5_ I 10 910 1096 1240 '• Dundee •< 1 2 " L -fr { S 1427 2283 1965 r 7 6 1-0 None 14 days after! infection J Thames ^ 8 " 0-5 1-0 0-5 r 1 1-0 „ Dundee i 2 ' 0-5 1-0 ' 0-5 Thus at the higher temperature of 13° C., the anthrax bacilli were • markedly more persistent than at 5° C-, although they had in all cases 554 Profs. Percy Frankland and Marshall Ward. become largely reduced in numbers by the fifth day after their introduc- tion into the water, and by the fourteenth day they had all disappeared. In the following table are recorded the results which were obtained with the same waters maintained throughout at a temperature of 19° C. :— Water kept at 19° C. Plate cultures prepared. Description Number of of water. tube. Number of days plates were incubated at 18—20° C. Volume of water used for plate cultivation. Xumber oi anthrax colonies per 1 c.c. of water. 1st day after "1 infection j Thames •< 11 12 c.c. « { t 1 * I t 3,400 3,200 2,786 2,830 e 5 1 5 6,110 » 6 „ { i L 12 4,908 5,327 5,718 r « J To 3,530 2 days after 1 infection J Thames •< 1 12 ) s — 2_ 4,420 4,066 4,253 r C / M 5,125 » Dundee •<, p 6 1 A J T5o » 12. L To" 10,000 5,200 5,745 5 days after 1 infection J Thames < 11 12 1 f 10 G L Tff ; * L -fs 30,428 36,020 • 32,448 28,194 r 5 J 1°0 46,200 » Dundee •< 6 » { t 45,825 32,370 25,140 14 days after"! infection J Thames <. 11 12 C) { S / M I A 207,238 163,750 Dundee < 5 " 06,581 L 6 » J T2 L T^ 54,300 42 days after") infection J r Thames < 11 12 3 i S 99,937 109,590 102,180 M Dundee •< 5 » I 5 J To * {T^ 66,240 51,227 L T% 49,610 Report on the Bacteriology of Water. 555 Thus at the temperature of 19° C., the behaviour of the anthrax bacilli in these same waters was entirely different ; far from their undergoing rapid diminution in numbers followed by early disappearance, they only exhibited slight diminution during the first few days after their introduc- tion, upon which there folloived an enormous multiplication. This multiplication was already observable, in the case of one tube, on the second day after infection, whilst on the fifth day it was very pronounced in all the tubes, and on the fourteenth day the numbers reached were very large, remaining practically unaltered even on the forty-second day. The explanation which naturally suggests itself for this entirely different behaviour of the anthrax bacilli at the higher and the lower temperature respectively, is that at the higher temperature of 19° C. the anthrax bacilli can form spores, whilst at the lower temperatures this sporulation cannot take place. With the appearance of the spores, however, the longevity of anthrax in sterile potable water becomes, as was shown in the Second Report, practically indefinite. In order to test the validity of this hypothesis that sporulation had taken place in the waters kept at 19° C., the following experiments were made : — (1.) 1 c.c. of the contents of Tube 11 (Thames water, see table above) was kept at 70° C. for ten minutes, in order to destroy anthrax bacilli ; on subsequent plate cultivation, innumerable anthrax colonies were obtained. (2.) A similar experiment made with 1 c.c. of the contents of Tube 12 (Thames water, see table above) gave exactly the same result. (3.) A similar experiment made with 1 c.c. of the contents of Tube 5 (Dundee water, see table above) gave the same result, innumerable anthrax colonies being obtained on the plate. (4.) A similar experiment made with 1 c.c. of the contents of Tube 6 (Dundee water, see table above) gave 23,352 anthrax colonies. Thus in the case of all these waters kept at 19° C., it is evident that practically the whole of the anthrax microbes present at the end of forty-two days were there in the condition of spores, showing as they did no appreciable 'diminution in numbers by the process of heating to 70° C. for ten minutes. These experiments show, then, very clearly that the fate of virulent anthrax bacilli passing from an anthrax victim into potable water will be dependent on the temperature of the latter ; if the tempera- ture of the water is below that at which sporulation of anthrax can take place, then the bacilli will perish in the course of a few days ; whilst if the temperature is high enough to admit of sporulation, 556 Report on the Bacteriology of Water. then anthrax spoi-es will be formed, and these, as is now well known, may persist in a living and virulent condition for an almost indefinite period of time — for months and probably even for years. From these experiments it fm*ther appears that the lowest tem- perature at which spore formation of anthrax in water will take place lies somewhere between 15° and 19° C. According to Koch, the lowest temperature at which anthrax spores are obtained is 16° C., and the most advantageous temperature for the production of the hardiest spores is 20 — 25° C. Report of Magnetical Observations at Falmonth Observatory. 557 Report of Magnetical Observations at Falmouth Observatory for the Year 1893. Latitude 50° 9' 0" N. and Longitude 5° 4' 35" W. ; height, 167 feet above mean sea-level.* These observations have been made by instruments purchased from the Government Grant Fund administered by the Royal Society. The Observatory having been comparatively recently established, the Vertical Force self-recording instrument is not yet in thorough working order. It is hoped in future to publish complete records of all three elements. Photographic curves of Magnetic Declination and of Horizontal Force variations have been taken regularly throughout the past year, and the magnets have worked satisfactorily. The scale values of the instruments were determined in April last. The following values of the ordinates of the photographic curves were then found : — Declination, 1 cm. = 0° ll'-7. Bifilar, April 7th, for 1 cm. 8 H., = 0'00055 C.G.S. unit. This latter value not being in accordance with the prescribed standard scale, the sensibility of the magnet was increased, and a second series of deflections made on April 10th, when the value was determined for 1 cm. £ H. = Q'00050 C.G.S. unit. No violent magnetic disturbances have been recorded during the year ; the principal movements occurred on the following dates : — February 5, March 26, July 16, August 6, 7, 18, and November 1, 2. Observations with the Absolute Instruments have been made monthly, of which the following is a summary : — Determinations of Horizontal Intensity, 34. ,, Inclination, 35 sets of four. „ absolute Declination, 33. The results in the following tables, Nos. 1, 2, 3, 4, are deduced from the magnetograph curves which have been standardised by observations of deflection and vibration. These were made with the Collimator Magnet marked 66A, and the Declinometer Magnet marked 66c in the Unifilar Magnetometer by Elliott Brothers, of London. Table No. 5 is deduced from these observations. * The records of the Falmonth Magnetic Observatory hare hitherto been published in the 'Journal of the Royal Cornwall Polytechnic Society.' The committee of management having obtained leave to communicate their annual magnetic report to the Eoyal Society, it will henceforward be printed in the ' Proceedings.' The results are worked up in the same way as those obtained at Kew, and the reports of the two observatories will in future appear simultaneously. — E. VOL. LVI. 2 P 558 (19° + West.) Report of Maynetical Observations at Table I. — Hourly Means of Declination, at the Falmouth on five selected quiet Days in Hours Mid. 1 2 3 4 5 6 7 8 9 10 11 Winter. 1893. Months. Jan. . . 7-2 7-7 8-4 9-1 9-0 9-0 10-2 9-1 8-9 8-7 9-2 11-0 Feb. .. 7-3 7-3 7-5 7-4 7-2 6-9 6-5 6-0 5-6 5-8 7-0 9-4 March . 7-3 7-2 6-8 6-7 6'2 5-7 5-5 5-1 3-6 3-1 4-5 8-0 Oct. .. 3'6 3-7 3-5 3-6 3-9 3-8 3-4 2-3 0-6 0-2 1-7 4-9 *Nov. .. 3-9 4-7 4-4 4-4 4-4 4'4 4-4 3-5 2-9 2-2 3-6 6-7 Dec. .. 4-7 5-5 5-8 6-0 5-7 5-7 5-5 5-5 52 4-6 4-4 6'4 Means 5-7 6-0 6-1 6-2 6-1 5-9 5-9 5-2 4-5 4-1 5-1 7-7 Summer. April . . 6-8 6-8 6-8 6-6 6'4 6-1 5-2 3-7 2-1 1-8 2-7 6-5 tMay .. 5-9 6-5 6-6 6-0 5-5 3'9 2-6 0-8 0-3 0-6 2-8 6-4 June . . 6-1 5-6 5-5 5-3 5-0 3-3 1-1 0-7 0-3 0-7 3-0 5-9 July .. 4-5 4-6 4-2 4-1 3-7 2-4 0-8 -0-6 -0-9 -0-2 2'4 5-9 Aug. .. 4-5 4-6 4-3 4-1 3-4 2-6 1-0 -0-5 -0-5 0-3 2-8 6 '4 Sept. . . 3-2 3-2 3-5 2-6 2-2 2-2 1-2 0-3 -0-6 0-3 2-7 5-9 Means 5-2 5-2 5-1 4-8 4-4 3-4 2-0 0-7 o-i 0-6 2-7 6'2 * Mean derived from 7th, llth, and 21st. f Mean of four days, 2nd, 14th, 21st, and 28th. Table II. — Solar Diurnal Range of the' Falmouth Hours Mid. 1 2 3 4 5 6 7 8 9 10 11 Summer mean. -0-6 -0-6 -0-7 -1-1 -1-5 -2-4 -3-9 -5-2 -5-8 -5-3 -3-1 + 0-4 Winter mean. , , i , , , , , , , , , -1-5 -1-2 -1-1 -1-0 -1-1 -1-3 -1-3 -2-0 -2-7 -3-1 -2-1 + 0-5 Annual mean. , , i /• i / , , / / , -1-0 -0-9 -0-9 -1-0 -1-3 -1-8 -2-6 -3-6 -4-2 -4-2 -2-6 |+0-4 NOTE. — When the sign is -f the magnet Fctfntovth Observatory for t/te Year 1893. Observatory determined from the Magnetograph Curves each Month during the Year 1893. Noon 1 2 3 4 5 6 7 8 9 10 11 Mid. Winter. 12-6 14-0 13-6 12-7 12-5 11-9 11-3 10-7 10-0 9-5 9-2 8-7 8-8 11-4 12-5 12-6 11-5 10-2 9-2 8-7 8-6 8-2 8-0 7-6 7-3 7'4 12-0 14-9 15-5 13-9 11-8 9-2 8-7 8-1 7-6 7-4 7-5 7-5 7-4, 8-5 10-2 10-2 11-0 7'4 6'2 5-3 4-8 4-5 3-7 3-4 3-6 3-2 9-3 10-6 9-9 8-5 7-7 6-7 6-7 6-6 5-2 4-7 4-3 4-5 5 '1 8-6 8-7 9-8 9-5 8-8 7-7 7-3 6'9 5-6 5-2 5-0 5-1 4-2 10-4 11-8 11-9 11-2 9-7 8-5 8-0 7-6 6-8 6-4 6-2 6-2 6-0 Summer. 10-8 14-0 15-1 13-7 11-5 9-7 7'9 7-5 7'4 7'3 7-4 7-0 6-9 11-1 13-1 13-7 12-2 9-9 8-3 7-1 6-6 6-6 6-6 6-6 6-6 6-8 9-8 12-2 12-4 11-2 9-8 8-2 7-2 6-1 5-8 5-5 5-9 5-7 5-9 9-3 10-9 11-9 11-6 9-7 7-4 5-9 4-7 4-6 4-6 5-3 5-1 4-8 10-1 12-7 12-7 11-1 8-9 6-6 4-9 4-7 4-5 4-3 4-2 4-3 4-5 9-8 11-8 11-8 10-2 7-7 6-0 5-0 4-6 4-2 4-2 3-9 3-4 3-2 10-2 12-5 12-9 11-7 9'6 7-7 6-3 5-7 5-5 5-4 5'5 5-3 5-3 Declination as derived from Table I. Noon 1 2 3 4 5 6 7 8 9 10 11 Mid. Summer mean. + 4-4 + 6-7 + 7-2 + 5-9 + 3-8 + 1-9 + 0-5 -o-i -0-3 -0-4 -0-3 -0-5 -0-5 Winter mean. + 3-2 + 4-6 + 4-7 + 4-0 + 2'5 + 1-3 + 0-8 + 0-4 -0-4 -0-8 -1-0 -1-0 -1-2 Annual mean. + 3-8 i + 5-6 + 5-9 + 4-9 + 3-1 + 1-6 + 0-6 + 0-2 -0-3 -0-6 -0-6 -0-7 -0-8 points to the west of its mean position. 560 Report of Magnetical Observations at 0-18000 + (C.G-.S. units.) Table III. — Hourly Means of the Horizontal Force at Falmouth (corrected for Temperature), on five Hours Mid. 1 2 3 4 5 6 7 8 9 10 11 Winter. 1893. Months. Jan. . . 430 430 433 435 437 439 440 440 439 435 429 423 Feb. .. 469 469 469 469 470 470 470 470 467 456 446 444 March . 447 446 446 444 444 444 445 441 433 421 412 411 Oct. .. 472 469 469 471 471 470 468 467 458 449 440 435 | NOT. .. 460 460 461 459 455 462 463 463 459 446 434 430 1 Dec. .. 455 454 454 457 458 460 464 464 461 456 449 443 Means 456 455 455 456 456 457 458 457 453 444 435 431 Summer. April . . 473 471 471 472 471 472 475 475 470 459 440 429 May . . 471 471 472 470 470 468 464 457 444 434 429 429 June . . 478 475 474 474 473 473 468 459 452 446 442 441 July .. 464 462 460 459 460 460 456 448 439 429 422 422 Aug. .. 464 464 464 464 465 463 459 449 439 430 422 422 Sept. . . 465 463 464 461 463 459 456 451 442 429 424 427 Means 469 468 467 467 467 466 463 456 444 438 430 428 (C.GKS. units.) Table IV. — Diurnal Range of the Falmout' Hours Mid. 1 2 3 4 5 6 7 8 9 10 11 Summer mean. + •00008 + -00007 + •00006 + -00006 + -00006 + •00005 + -00002 - -00005 -•00017 - -00023 - -00031 - -00033 Winter mean. + •00004 + •00003 + -00003 + -00004 + •00004 + -00005 + -00006 + •00005 + •00001 -•00008 - -00017 - -00021 Annual mean. + •00006 + •00005 ' + •00005 + •00005 + •00005 + -00005! + -00004 •00000 - -00008 -•00016 - -00024 - -00027 NOTB. — When the sign is + th Falmouth Observatory for the Year 1893. Observatory as determined from the Magnetograph Curves selected quiet Days in each Month during the Year 1893. Noon 1 2 3 4 5 6 7 8 9 10 11 Mid. Winter. 424 431 435 435 436 436 439 442 442 443 443 443 4*2 447 452 457 460 462 463 468 471 472 472 470 470 470 411 420 431 439 445 442 445 448 450 450 450 449 448 442 446 451 459 462 464 471 472 474 476 476 475 475 433 439 447 454 458 464 466 467 471 472 471 468 468 442 445 448 453 455 456 460 464 464 465 463 460 457 433 439 445 450 453 454 458 461 462 463 462 461 460 Summer. 427 434 446 461 469 474 474 476 479 476 477 476 475 431 448 460 471 476 480 482 482 483 483 479 479 481 446 454 463 470 478 484 489 493 490 488 488 484 483 431 436 443 455 462 46G 475 476 471 472 469 466 465 432 443 454 459 466 466 468 473 473 474 472 469 468 434 441 451 455 455 458 464 466 470 471 471 471 467 433 443 453 462 468 472 475 478 478 477 476 474 473 Horizontal Force as deduced from Table III. Noon 1 2 3 4 5 6 7 8 9 10 11 Mid. Summer inenn. -•00028 -•00018 - -00008 + -00001 + -00007 + -00011 + -00014 + '00017 + -00017 + -00016 + •00015 + -00013 + -00012 Winter mean. - -00019 - -00013 - -00007 -•00002 + -00001 + -00002 + -00006 + -00009 + •00010 + -00011 + -00010 + -00009 + -OOC08 Annual mean. -•00024 - -00016 -•00008 - -00001 + •00004 + -00007 + -00010 + -oooi a + •00014 + -00014 + •00013 + •00011 + -00010 reading is above the mean, 562 Report of Magnetlca.l Observations at The Inclination was observed with the Inclinometer No. 86 by Dover, of Charlton, Kent, and needles 1 and 2, which are 3^ inches in length, the results of which appear in Table VI. The Declination and Horizontal Force values given in Tables I to IV are prepared in accordance with the suggestions made in the fifth report of the Committee of the British Association on com- paring and reducing magnetic observations, and the time given is Greenwich mean time, which is 20 min. 18 sec. earlier than local time. The following is a list of the days during the year 1893 which were selected by the Astronomer Royal, as suitable for the deter- mination of the magnetic diurnal variations, and which have been employed in the preparation of the magnetic tables : — January 7, 8, 15, 25, 26. February 1, 11, 13, 26, 27. March 10, 13, 18, 19, 20. April 4, 9, 21, 22, 23. May 2, 14, 17, 21, 28. June 8, 13, 17, 22, 24. July 5, 6, 10, 30, 31. August 1, 9, 16, 17, 27. September 4, 7, 13, 23, 24. October 9, 11, 16, 21, 22. November 7, 11, 15. 20, 21. December 7, 13, 18, 21, 22. The following are the principal results of the magnetic elements for the year 1893 :— Mean Westerly Declination, 19° 6''4. Mean Inclination, 67° 5''3. Mean Horizontal Force, 0'18455 C.G.S. unit. The Declination and Horizontal Force are deduced from hourly readings of the photographic curves, and so are corrected for the diurnal variation. The Inclination is the mean of the absolute observations, the mean time of which is noon. In Table V, X is the mean of the absolute values observed during the month (generally three in number), uncorrected for diurnal variations and for any disturbance. Y is the mean of the products of the tangents of the Dips and the corresponding values of X. The whole of the instruments have been maintained in good order. The Magnetic Chamber has been kept in a satisfactory state of dryness, and the Magnetic Hut in the garden has been newly roofed during the year. EDWARD KITTO, Magnetic Observer. Falmouth Observatory for the Year 1893. 563 Table V.— Magnetic Intensity. Falmouth Observatory, 1893. C.G.S. r aeasure. 1893. X or Horizontal force. Yor Vertical force. 0 '18447 0 '43621 February 0 -18457 0 '43662 March 0 -18438 0 ' 43688 April 0 -18467 0 "43682 0 -18470 0 "43675 0 -18482 0 "43696 July . . 0 -18466 0 "43683 0 -18460 0 "43679 0 -18434 0 " 43653 ''October 0-18445 0 "43658 0-18451 0 -43644 December 0 '18444 0 '43628 0 '18455 0 "43664 Table VI. — Observations of Magnetic Inclination. Falmouth Observatory, 1893. Month. Mean. Month. Mean. January 28 0 / 67 3-7 July 28 0 / 67 5 "8 30 67 5-0 29 67 4 "4 31 67 5-5 31 67 5"0 February 25 67 4-7 67 3-5 August 27 67 5-1 67 5-7 27 67 6 "2 28 67 5 "2 28 67 5 '7 (V7 ^ •&, 67 5-1 67 8 "5 March 25 67 6-2 27 67 6'1 28 67 8'0 28 67 4-7 April 15 67 7"! 67 4-1 October 28 67 6-4 67 5-4 28 67 5 "6 30 6'7 6 '5 29 67 5 '3 31 67 5-6 May 25 67 5-0 67 5 -1 67 5-8 67 5'1 26 67 2-9 29 67 5-0 27 67 5'8 30 67 5"0 June 27 67 4-6 67 3 "8 December 27 67 5-0 67 4 "7 28 67 3 "9 28 67 5-9 29 67 5-4 29 67 4-3 67 4-4 67 5-0 INDEX TO VOL. LYI. ABBOTT (Miss E. C.) and H. G-adow, on the evolution of the vertebral column of fishes, 296. Abney (W. de W.) measurement of colour produced by contrast, 221. Air, on the electrification of (Kelvin and Maclean), 84. Allen (J. B.), E. Threlfall, and J. H. D. Brearley, researches on the electrical properties of pure substances. No. I. The electrical properties of pure sul- phur, 32. Alternating current system, note on the possibility of obtaining a unidirec- tional current to earth from the mains of an (Cardew), 99. on an instrument for in- dicating and measuring difference of phase between E.M.F. and current in any (Cardew), 250. Anthrax bacilli taken directly from the animal body, the behaviour in pot- able water of (Frankland and Tern pieman), 549. Appleyard (J. E.) and P. F. Frankland, the behaviour of the typhoid bacillus and of the Bacillus coli communis in potable water, 395. Aston (Emily) and W. Eamsay, the molecular surface-energy of mixtures of non-associating liquids, 182. the molecular surface-energy of the esters, showing its variation •with chemical constitution, 162. Asymmetrical probability curve, the (Edge worth), 271. Atmospheres which extinguish flame, the composition of (Clowes), 2. Atmospheric temperature, on rapid variations of, especially during Fohn (Buchanan), 108. /3-Lyrse, the spectrum changes in. Preliminary note (Lockyer), 278. Bacilli taken directly from the animal body, the behaviour in potable water of anthrax (Frankland and Temple- man), 549. Bacillus ardhracis, further experiments on the action of light on (Ward), 315. VOL. L7I. Bacillus coli communis and the typhoid bacillus in potable water, the behaviour of (Frankland and Apple- yard), 395. Bacteria of the Thames, on the (Ward), 315. Bagnold (Major) and P. Cardew, on the difference of potential that may be established at the surface of the ground immediately above and at various distances from a buried mass of metal charged from a high prestsui-e electric light supply, 252. Barnett (E. E.) on the viscosity of water as determined by J. B. Han- nay by means of his microrheometer, 259. Bateson (William) elected, 130. admitted, 162. Bessemer process, the spectroscopic phenomena and thermo-chemistry of the (Hartley), 193. Bidwell (3.) on the effect of magnetisa- tion upon the dimensions of iron rings in directions perpendicular to the magnetisation, and upon the volume of the rings, 94. on the recurrent images following visual impressions, 132. Blood, the influence of intra-venous injection of sugar on the gases of the (Harley), 148. Boulenger (George Albert) elected, 130. admitted, 162. Bourne (Or. C.) on the structure and affinities of Heliopora ccerulea, Pall., with some observations on the struc- ture of Xenia and Heteroxenia, 299. Boyce (B.) a contribution to the study of (i) some of the decussating tracts of the mid- and inter-brain, and (ii) of the pyramidal system in tbe mesencephalon and bulb, 305. Boys (C. V.) on the Newtonian constant of gravitation, 131. Bradford (John Eose) elected, 130. admitted, 205. Brain, a contribution to the study of (i) some of the decussating tracts of the mid- and inter- brain, and (ii) of C VI INDEX. the pyramidal system in the mesen- cephalon and bulb (Boyce), 305. Brearley (J. H. D.), J. B. Allen, and R. Threlfall, researches on the elec- trical properties of pure substances. No. I. The electrical properties of pure sulphur, 32. Breathing, on the different forms of (Marcet), 307. (Title only.) Buchanan (J. Y.) on rapid variations of atmospheric temperature, especially during Fohn, and the methods of observing them, 108. Callendar (Hugh Longbourne) elected, 130. admitted, .162. Candidates for, election, list of, 1. Cardew (P.) note on the possibility of obtaining a unidirectional current to earth from the mains of an alternating current system, 99. on an instrument for indicating and .measuring difference of phase between E.M.F. and current in any alternating current system, 250. and Major Bagnold, on the differ- ence of potential that may be estab- lished at the surface of the ground immediately above and at various distances from a buried mass of metal charged from a high pressure electric light supply, 252. Cerebellum, degenerations consequent on experimental lesions of the (Russell), 303. Cheyne (William Watson) elected, 130. admitted, 162. Chree (C.) the stresses and strains in isotropic elastic solid ellipsoids in equilibrium under bodily forces de- rivable from a potential of the second degree, 26. Clark (Frederick Le Gros) obituary notice of, i. Clowes (F.) the composition of atmo- spheres which extinguish flame, 2. Colour produced by contrast, measure- ment of (Abney), 221. Congruences, on the singular solutions of simultaneous ordinary differential equations and the theory of (Dixon), 277. Copper, measurements of the absolute specific resistance of pure electrolytic (Swan and Rhodin), 64. on determining the thermal con- ductivity of (Gray), 199. Curves, the differential covariants of twisted, with some illustrations of the application to quartic curves (Gwyther), 272. Cynodontia, on new (Seeley), 291. Davison (C.) on the Leicester earthquake of August 4, 1893, 19. Delepine (S.) and A. Ransome, on the influence of certain natural agents on the virulence of the tubercle-bacillus, 51. "Differential covariants of twisted curves, the, with some illustrations of the application to quartic curves (Gwy- ther), 272. • equations and the theory of con- gruencies, on the singular solutions of simultaneous ordinary (Dixon), 277. Dixon (A. C.) on the singular solutions of simultaneous ordinary differential equations and the theory of congruen- cies, 277. Earthquake of August 4, 1893, on tlis Leicester (Davison), 19. Eclipse of the sun, April ] 6, 1893, pre- liminary report on the results obtained with the prismatic camera during the total (Lcckyer), 7. • report on results obtained with the slit spectroscopes (Hills), 20. Edgeworth (F. Y.) the asymmetrical probability curve, 271. Election of Fellows, 1. Electric arc, the rotation of the (Trot- ter), 262. light supply, on the difference of potential that may be established in connexion with a high pressure (Car- dew and Bagnold), 252. strength of mixtures of nitrogen and hydrogen, the (Fawcett), 263. Electrical current to earth from the mains of an alternating current sys- tem, note on the possibility of obtain- ing a unidirectional (Cardew), 99. • properties of pure substances, re- searches on the. No. 1. Pure sulphur (Threlfall, Brearley, and Allfn), 32. Electrification of air, on the (Kelvin and Maclean), 84. Electromotive force and current in any alternating current system, on an instrument for indicating and measur- ing difference of phase between (Cardew), 250. Ellipsoids in equilibrium, the stresses and strains in isotropic elastic solid (Chree), 26. Esters, the molecular surface-energy of the, showing its variation with chem- ical constitution (Ramsay and Aston), 162. INDEX. Vll Ewan (T.) on tlie absorption spectra of dilute solutions, 286. Explosives, researches on. Preliminary note (Noble), 205. researches on modern. Prelimin- ary communication (Macnab and Eistori), 8. Falmouth observatory, report of mag- netical observations at, for 1893, 557. Fawcett (Miss P. G.) the electric strength of mixtures of nitrogen and hydrogen, 263. Fellows elected, 130. admitted, 1, 32, 162, 205. Fishes, on the evolution of the vertebral column of (Gadow and Abbott), 296. Flame, the composition of atmospheres which extinguish (Clowes), 2. spectra at high temperatures. Part II. The spectrum of metallic manganese, of alloys of manganese, and of compounds containing that element (Hartley), 192. Part III. The spectroscopic phenomena and thermo-chemistry of the Bessemer process (Hartley), 193. Fluids, on the dynamical theory of in- compressible viscous, and the deter- mination of the criterion (Reynolds), 40. Fohn, on rapid variations of atmospheric temperature, especially during (Bu- chanan), 108. Foraminifera, contributions to the life- history of the (Lister), 155. Fossil reptilia, researches on the struc- • ture, organisation, and classification of the (Seeley). Part IX. Section 4. On the Gomphodontia, 288. — Section 5. On new Cynodontia, 291. — Sec- tion 6. Associated remains of two small specimens from Klipfontein, Fraserbur?, 295. Frankland (P. F.) and J. E. Appleyard, the behaviour of the typhoid bacillus and of the Sacillus coli communis in potable water, 395. and C. Templeman, the behaviour in potable water of anthrax .bacilli taken directly from the animal body, 549. • and H. M. Ward, third report to the Water Eesearch Committee of the Eoyal Society. — Part I. Further ex- periments on the action of light on Sacillus anthracis, and on the bacteria of the Thames (Ward), 315.— Part II. The behaviour of the typhoid bacillus and of the Bacillus coli communis in potable water (Frankland and Apple- yard), 395. Froude (Eobert Edmund) elected, 130. admitted, 162. Gadow (H.) and Miss E. C. Abbott, on the evolution of the vertebral column of fishes, 296. Geological time, Niagara Falls as a chronometer of (Spencer), 145. Gomphodontia, on the (Seeley), 288. Gravitation, on the Newtonian constant of (Boys), 131. Gray (J. H.) on a method for determin- ing the thermal conductivity of metals, with applications to copper, silver, gold, and platinum, 199. (T.) on the measurement of the magnetic properties of iron, 49. Gwyther (E. F.) the differential co- variants of twisted curves, with some illustrations of the application to quartic curves, 272. Hannay (J. B.) his determination of the viscosity of water (Barnett), 259. Harley (V.) the influence of intra- venous injection of sugar on the gases of the blood, 148. Harmonics, on certain functions con- nected with tesseral, with applications (Leahy), 45. Hartley (W. N.) flame spectra at high temperatures. Part II. The spec- trum of metallic manganese, of alloys of manganese, and of compounds con- taining that element, 192. Part III. The spectroscopic phenomena and thermo-chemistry of the Bessemer process, 193. Heliopora ccerulea (Pall.) on the struc- ture and affinities of (Bourne), 299. Heteroxenia and Xenia, observations on the structure of (Bourne) , 299. Hill (M. J. M.) elected, 130. admitted, 205. Hills (E. H.) the total solar eclipse of 16th April, 1893. Eeport on results obtained with the slit spectroscopes, 20. Hopkinson (J.) and E. Wilson, propa- gation of magnetisation of iron as affected by the electric ciirrents in the iron, 108. Hydrogen and nitrogen, the electric strength of mixtures of (Fawcett), 263. Iron, on the measurement of the mag- netic properties of (Gray), 49. propagation of magnetisation of, as affected by the electric current s c 2 Vlll INDEX. in the iron (Hopkinson and Wilson), 108. Iron, the effect of mechanical stress and of magnetisation on the physical pro- perties of alloys of (Tomlinson), 103. rings, on the effect of magnetisa- tion xipon the dimensions and volume of (Bidwell), 94. Isotropic elastic solid ellipsoids in equi- librium under bodily forces derivable from a potential of the second degree, the stresses and strains in (Chree),26. Jones (John Viriamu) elected, 130. Kelvin (Lord) and M. Maclean, on the electrification of air, 84. Leahy (A. H.) on certain functions con- nected with tesseral harmonics, with applications, 45. Leicester earthquake of August 4, 1893, on the (Davison), 19. Liquids, the complexity and the disso- ciation of the molecules of (Ramsay), 171. the molecular surface-energy of mixtures of non-associating liquids (Ramsay and Aston), 182. Lister (J. J.) contributions to the life- history of the Foraminifera, 155. Lockyer (J. N.) on the photographic spectrum of the great nebula in Orion, 285. preliminary report on the results obtained with the prismatic camera during the total eclipse of the sun, April 16, 1893, 7. the spectrum changes in j8-Lyrse, preliminary note, 278. Love (Augustus Edward Hough) elected, 130. admitted, 162. Lydekker (Richard) elected, 130. Lyginodendron Oldhamium (Will.), the root of (Williamson and Scott), 128. Maclean (M.) and Lord Kelvin, on the electrification of air, 84. Macnab (W.) and E. Ristori, researches on modern explosives. Preliminary communication, 8. Magnetic properties of iron, on the measurement of the (Gray), 49. Magnetic survey of the British Isles for the epoch January 1, 1891 (Riicker and Thorpe), 307. (Title only.) Magnetical observations at Falmouth Observatory for the year 1893, report of, 557. Magnetisation, on the effect of, upon the dimensions and volume of iron rings (Bidwell) , 94. and mechanical stress, the effect of, on the physical properties of alloys of iron and nickel, and of man- ganese steel (Tomlinson), 103. of iron, propagation of, as affected by the electric currents in the iron (Hopkinson and Wilson), 108. Manganese (metallic), alloys of man- ganese, and compounds containing that element, the spectrum of (Hart- ley), 192. steel, the effect of mechanical stress and of magnetisation on the physical properties of alloys of (Tom- linson), 103. Marcet (W.) on the different forms of breathing. (Title only), 307. Mascart (Eleuthere ISlie Nicolas) ad- mitted, 32. Mendeleeff (Dmitri Ivanovitch), ad- mitted, 1. Metals, on a method for determining the thermal conductivity of (Gray), 199. Microrheometer, on the viscosity of water as determined by Hannay's (Barnett), 259. Molecular surf ace- energy of mixtures of non-associating liquids (Ramsay and Aston), 182. of the esters, (Ramsay and Aston), 162. Nebula in Orion, on the photographic spectrum of the great (Lockyer), 285. Newtonian constant of gravitation, on the (Boys), 131. Niagara Falls as a chronometer of geo- logical time (Spencer), 145. Nickel, the effect of mechanical stress and of magnetisation on the physical properties of alloys of (Tomlinson), 103. Nitrogen and hydrogen, the electric strength of mixtures of (Fawcett), 263. . Noble (Sir A.) researches on explosives. Preliminary note, 205. Obituary notice of a Fellow deceased : — Clark, Frederick Le Gros, i. Peach (Benjamin Neeve) admitted, 1. Penrose (Francis Cranmer) elected, 130. admitted, 162. Presents, lists of, 28, 81, 128, 160, 203, 307. INDEX. IX Probability curve, the asymmetrical (Edgeworth), 271. Ramsay (W.) the complexity and the dissociation of the molecules of liquids, 171. and Emily Aston, the molecular surf ace -energy of mixtures of non- associating liquids, 182. • the molecular surface-energy of the esters, showing its variation with chemical constitution, 162. Ransome (A.) and S. Delepine, on the influence of certain natural agents on the virulence of the tubercle-bacillus, 51. Recurrent images following visual im- pressions, on the (Bidwell), 132. Reptilia, researches on the structure, organisation, and classification of the fossil. Part IX. Section 4. On the Gomphodontia (Seeley), 288.— Sec- tion 5. On new Cynodontia, 291. — Section 6. Associated remains of two small specimens from Klipfontein, Fraserburg, 295. Reynolds (O.) on the dynamical theory of incompressible viscous fluids and the determination of the criterion, 40. Rhodin (J.) and J. W. Swan, measure- ments of the absolute specific resist- ance of pure electrolytic copper, 64. Ristori (E.) and W. Macnab, researches on modern explosives. Preliminary communication, 8. Riicker (A. W.) and T. E. Thorpe, a magnetic survey of the British Isles for the epoch, January 1, 1891. (Title only), 307. Russell (J. S. R.) degenerations conse- quent on experimental lesions of the cerebellum, 303. Salomons (Sir D.) on some phenomena in vacuum -tubes, 229. Scott (Dukmfield Henry) elected, 130. admitted, lf>2. and W. C. Williamson, the root of Lyginodendron Oldhamium ("Will.), 128. Seeley (H. G.) researches on the strtfc- ture, organisation, and classification of the fossil reptilia. Part IX. Sec- tion 4. On the Gomphodontia, 288. — Part IX. Section 5. On new Cynodontia, 291.— Part IX. Section 6. Associated remains of two small specimens irom Klipfoutein, Fraser- burg, 295. Smith (Frederick John") elected, ISO. Smith (Frederick John), admitted, 162. Solutions, on the absorption spectra of dilute. (Ewan), 286. Spectra of dilute solutions, on the ab- sorption (Ewan), 286. Spectroscopic phenomena and thermo- chemistry of the Bessemer process, the (Hartley), 193. Sper-trum changes in /3-Lyrse, the. Pre- liminary note (Lockyer), 278. of metallic manganese, of alloys of manganese, and of compounds containing that element, the (Hart- ley), 192. of the great nebula in Orion, on the photographic (Lockyer), 285. Spencer (J. W.) Niagara Falls as a chronometer of geological time, 145. Sugar, the influence of intra-venous in- jection of, on the gases of the blood (Harley), 148. Sulphur, the electrical properties of pure (Threlfall, Brearley, and Allen), 32. Sun, preliminary report on the results obtained with the prismatic camera during the total eclipse of the, April 16, 1893 (Lockyer), 7. the total solar eclipse of 16th April, 1893. Report on results ob- tained with the slit spectroscopes (Hills), 20. Swan (Joseph Wilson) elected, 130. admitted, 162. on some voltaic combinations with a fused electrolyte and a gaseous de- polariser, 56. and J. Rhodin, measurements of the absolute specific resistance of pure electrolytic copper, 64. Temperature, on rapid variations of atmospheric (Buchanan), 108. Templeman (C.) and P. F. Frankland, the behaviour in potable water of anthrax bacilli taken directly from' the animal body, 549. Tesseral harmonics, on certain functions connected with, with applications (Leahy), 45. Thermal conductivity of metals, on a me'hod for determining the (Gray), 199. Thermo-chemistry of the Bessemer pro- cess, the (Hartley), 193. Thorpe (T. E.) and A. W. Riicker, a magnetic survey of the British Isles for the epoch January 1, 1891. (Title only), 307. Threlfall (R.), J. H. D. Brearley, and J. B. Alien, researches en the elec- trical properties of pure substances. INDEX. No. I. The electrical properties of pure sulphur, 32. Tomlinson (tl.) the effect of mechanical stress and of magnetisation on the physical properties of alloys of iron and nickel and of mangr.m;se steel, 103. Trotter (A. P.) the rotation of the electric arc, 262. Tubercle-bacillus, on the influence of certain natural agents on the yiru- lence of the (Ransome and Delepine), 51. Typhoid bacillus, and Bacillus coli commitnis, the behaviour of the, in potable water (Frankland and Apple- yard), 395. Vacuum-tubes, on some phenomena in (Salomons), 229. Veley (Victor Herbert) elected, 130. admitted, 162. Vertebral column of fishes, on the evo- lution of the (Gadow and Abbott), 296. Viscosity of water as determined by J. B. Hannay, on the (Barnett), 259. Viscous fluids, on the dynamical theory of incompressible, and the determina- tion of the criterion (Keynolds), 40. Visual impressions, on the recurrent images following (Bidwell), 132. Voltaic combinations with a fused elec- trolyte and a gaseous depolariser, on some (Swan), 56. Ward (H. M.) and P. F. Franklancl, third report to the Water Research Committee of the Royal Sosiety. — Part I. Further experiments on the action of light on Bacillus anthracis, and on the bacteria of the Thames (Ward), 315. Part II. The be- haviour of the typhoid bacillus and of the Bacillus coli communis in potable water (Frankland and Apple- yard), 395. W ater, on the viscosity of, as determined by J. B. Hannay's microrheouieter (Barnett), 259. Water Research Committee of the Royal Society, third report to the. Part I. Further experiments on the action of light on Bacillus anthracis, and on the bacteria of the Thames (Ward), 315. Part II. The behaviour of the typhoid bacillus and of the Bacillus coli communis in potable water (Frankland and Appleyard), 395. Williamson (W. C.) and D. H. Scott, the root of Lyginodendron Oldhamium (Will.), 128. Wilson (E.) and J. Hopkinson, propa- gation of magnetisation of iron as affected by the electric currents in the iron, 108. Xenia, and ffeteroxenia, observations on the structure of (Bourne), 2a9. ERRATA. Page 12, line 8, after the word " grams " insert " of ballistite." „ 252, „ 10, for " charge" read " charged." END OF FIFTY-SIXTH VOLUME. HA1UUSON AMi> SONS, PK1NTEUS IN OEDINAKV TO HEU MAJESTY, ST. MAJtlllS'S LAKE. Q Royal Society of London 41 Proceedings L718 v.56 Phyjicai * Applied Sci Seriali PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET UNIVERSITY OF TORONTO LIBRARY