« AnteriorContinuar »
on the same horizon. The Alaskan beds (p. 43) are 'probably not older than the Bathonian and certainly not younger than the Oxfordian."
E. Wilbur Berry (ibid., 84, 1914) issues a report on the Upper Cretaceous and Eocene floras of South Carolina and Georgia. Ficus is represented in the Cretaceous beds of South Carolina by five species, and the willow, the oak, the myrtle, and the laurel are among the numerous angiosperms present. The author (p. 71) regards this flora, the post-Raritan floras of the eastern States, and the major part, at least, of the Dakota flora, as Turonian and not Ceno
manian in age. The Cretaceous flora of Georgia (p. 127) is placed on the same horizon. A small Middle Eocene (Claiborne) flora of seventeen species allows of an interesting discussion of Cainozoic climate in the eastern States. The main features of the modern flora of tropical America extended as far north as latitude 33° in Middle Eocene times, and retreated later towards the West Indies. In the same paragraph on p. 161, this retreat seems to be dated as toward the close of the Tertiary," and also, on Dall's evidence from marine life, as at the close of the Oligocene." The bibliography is useful to all workers in early Cainozoic floras, including that of Bovey Tracey in Devonshire.
David White (ibid., 85-E, 1914) emphasises the occurrence of thread-like resinous casts in "mother of coal" and other Palæozoic "coals of high_rank," associated in places with megascopic lumps of resin. By the decay of the plant tissues (p. 82), these resinous infillings of secretory canals have become concentrated in undue proportion in the coals.
A paper by F. W. Clarke and W. C. Wheeler (ibid., 90-D, 1914) has a bearing on the occurrence of magnesium carbonate in rocks. It discusses "The Composition of Crinoid Skeletons," and twenty-one species, representing as many genera, are shown to utilise in their hard parts from 7-28 to 12.69 per cent. of magnesium carbonate, when organic matter is eliminated from the analyses. A specimen of Hathrometra dentata includes, moreover, 5:73 per cent. of silica, a substance usually present in quantities of about o1 per cent. When arranged by localities, it is seen that the proportion of magnesium carbonate in crinoids is in some way dependent on temperature (p. 36). Shallow water in the tropics gives the highest percentage. The calcium carbonate is always in the calcite state. The investigation of fossil crinoids, from the Lower Ordovician to the Eocene, shows nothing higher than 2.56 per cent. of magnesium carbonate, except in a Triassic form, Encrinus liliiformis, which yields 20.23 per cent. We may conclude that the matrix was in this case dolomitic. It is suggested that infiltration of calcium carbonate has reduced the proportion of magnesium carbonate present in fossil specimens. The organic matter in recent forms, often amounting to 15 per cent., would certainly allow of the substitution of some other material during fossilisation.
Ivor Thomas's first section of his revision of "The British Carboniferous Producti" (Mem. Geol. Survey of Great Britain, Palæontology, vol. i., part 4, 1914) covers the genera Pustula and Overtonia, which are here established (p. 259) on Productus pustulosus and fimbriatus respectively. But the special importance of the memoir lies in the review of the Producti generally, based on the work of several years. Doubt is thrown (p. 229) on the clasping nature of the spines of Productus, and it is suggested that a spine during growth may occasionally be diverted by an adjacent object, so as to appear to fold around it. Both external and internal features of the shells are discussed in relation to the animal as it lived, and no apology
is needed for the consideration of the general principles of mutation and the meaning of species, questions that have naturally forced themselves before the philosophic author.
It is interesting to note, in L. W. Stephenson's study of the Species of Exogyra from the Eastern Gulf Region and the Carolines (U.S. Geol. Survey, Prof. Paper 81, 1914) that the genus was first described by T. Say in 1820, the type being Exogyra costata from the Cretaceous of New Jersey. The Sby." after the name in Woodward's Manual of the Mollusca," 1851, is probably a mere misprint. E. costata is shown regularly to succeed E. ponderosa, the species having thus a zonal value. The strata are described in the same memoir.
In vol. xlii. of the Records of the Geological Survey of India (1912, p. 1), R. Bullen Newton and E. A. Smith have directed attention to the survival of a well-known Miocene oyster, Ostrea gryphoides (=crassissima), in recent marine deposits under Calcutta and in the Bay of Bengal.
Passing to arthropods, C. D. Walcott gives us a new genus of trilobites, Saukia, with numerous species, separated from Dikellocephalus by the possession of a longer glabella and pygidium ("Dikelocephalus and other Genera of the Dikelocephalinæ," Smithsonian Miscell. Coll., vol. lvii., 1914, P. 345). On p. 363 the author explains the common retention of D. D. Owen's spelling of Dikellocephalus with a single "1," under a rule that was surely established by persons with limited glabellas. Ösceolia and Calvinella are here founded on D. osceola and D. spiniger (p. 388) respectively, though on p. 365, probably by a slip, the latter species is referred to Saukia. It seems from p. 364 that Walcott is unwilling to recognise Dikellocephalus from any locality outside the United States, and this should lead to a new examination of British and other European forms. The Devonian faunas of South Africa and South America receive a new link in the discovery by S. J. Shand of a species of the Brazilian trilobite Pennaia in the Bokkeveld Beds of the Hex River in the Cape Province (Trans. Geol. Soc., S. Africa, vol. xvii., 1914, p. 26).
Alexander Petrunkevitch has produced "A Monograph of the Terrestrial Palæozoic Arachnida of North America" (Trans. Connecticut Acad. Arts and Sciences, Yale University Press, 1913). This is a systematic review, not quite so extensive as its title would imply, of the American Palæozoic types of scorpions and spiders, involving the establishment of new genera and species. The author directs attention (p. 20) to the recent work of Clarke and Ruedemann, which indicates a relationship of the eurypterids with the succeeding limuloids, rather than with the scorpions, although the three groups may have had separate ancestors. The Carboniferous arachnid faunas of Europe and North America are stated to be distinct (p. 26); but both have a more tropical character than is found locally in their modern representatives. The excellent photographic plates are from specimens developed with much care by the delicate chiselling away of flakes of rock in order to reveal appendages.
R. Broom records (Bull. Amer. Mus. Nat. Hist., vol. xxxii., 1913, p. 563) his studies of a number of Permian labyrinthodont skulls in the American Museum, in which he has succeeded in tracing sutures. hitherto obscure, and thus in providing new descriptions of the cranial elements.
R. S. Lull describes (Amer. Journ. Sci., vol. xxxvii., 1914, p. 209) a "Fossil Dolphin from California,' which is presumed to be of Miocene age. The specimen is placed under the living genus Delphinus,
which is thus added to the few mammalian genera that have come down to us unmodified from older Cainozoic times.
A wide interest is attached to the investigation of the "Paleocene Deposits of the San Juan Basin, New Mexico," by W. J. Sinclair and W. Granger (Bull. Amer. Mus. Nat. Hist., vol. xxiii., 1914, p. 297), since these beds contain the oldest known Cainozoic mammals. The famous Puerco clays rest unconformably on a conglomerate with silicified tree-stems, below which are shales containing deinosaurs. No deinosaurs have been found in the Puerco Beds, and the faunal change is even here remarkably abrupt. Fossil plants, of no stratigraphical import, have been found for the first time (p. 306) in the Puerco Beds.
Among faunistic papers, we may note that the indefatigable C. D. Walcott reviews, with a bibliography, the 'Cambrian Faunas of Eastern Asia" (Smithson. Miscell. Coll., vol. Ixiv., 1914, p. 1).
G. A. J. C.
THE PLACE OF LAVOISIER IN THE HISTORY OF CHEMISTRY.
A SOMEWHAT novel view of "The Place of Lavoisier
in the History of Chemistry" is put forward in a paper contributed by A. Mieli to the April number of Scientia. This question has formed the subject of prolonged controversy, and has called forth the most diverse and contrary opinions. Some, with Wurtz, have boldly acclaimed the fact that "Chemistry is a French science. Its founder is Lavoisier, of immortal memory." Others have written him down as a mere plagiarist, who purloined from Priestley the discovery of oxygen, and from Cavendish the discovery of the composition of water, and thus built up a great reputation on the unacknowledged work of his English colleagues. The Italian writer asserts that these claims and counterclaims are based upon a misconception. Lavoisier's true place is not at the beginning of the period to which the atomic and molecular theories belong, but at the close of an earlier period in which the chief problems were the nature of combustion, and the composition of air and water. This period opens with Jean Rey and Boyle; John Mayow had practically reached a true solution of the main problems in 1674; but Becher and Stahl intervened, and it was only by the work of Black, Priestley, Cavendish, and Lavoisier that all difficulties and doubts were finally cleared away. Lavoisier's position in the historical sequence enabled him to use all the information and experience that had been gathered during the preceding 150 years, and it was right that he should do so, though his acknowledgments to Priestley and to Cavendish might well have been more generous.
But whilst Lavoisier contributed a brilliant finale to the earlier period, his work cannot be regarded as forming in any sense an overture to the period which followed. The chief topics to be studied in the later period were those which were concerned with atoms, molecules, and equivalents. This period began with Dalton's atomic theory and the controversy between Proust and Berthoilet on the subject of fixed or variable proportions; Avogadro (like Mayow) almost solved the problem; but once again a long interval of doubt and confusion ensued, until at last the work of Dumas, Laurent, and Gerhardt, and, above all, St. Claire Deville's discovery of dissociation, enabled Cannizzaro to put forward the masterly exposition which finally dispelled the uncertainty and perplexity which had afflicted chemistry for nearly forty years.
Cannizzaro, like Lavoisier, owed much to others. His experimental work was on a much smaller scale than Lavoisier's; but he is universally honoured as the
man who cleared away the obstacles that had hindered the progress of knowledge during many weary years. Lavoisier's chief claim to immortality is of a similar character. It rests upon the fact that he was able to break through the entrenched lines of the "phlogiston " theory, and to make a broad gap through which others could enter the open plain beyond. His tragic death prevented him from reaping the full fruits of his victory over error, and it was left to others to undertake the conquest of the fertile country into which he had opened the way.
The periods suggested by the author are described in a paper communicated to the Italian Chemical Society (Rendiconti Soc. Chim. Ital., 1914, vol. viii.) as follows:
(1) The Philosophic Period, 600 to 300 B.C., including the writings of Empedocles, Aristotle, and others. (2) The Ancient Alchemistic Period, extending to about 1000 A.D., and dominated by the writings of Geber. (3) The Alchemistic Period of the Middle Ages, extending to about 1400 A.D., and including such names as Avicenna, Roger Bacon, Raymond Lully, Albertus Magnus, and the pseudo-Geber. (4) The Period of the Renaissance, including the work of Agricola (1494–1555), Bernard Palissy, and Paracelsus (1493-1541). (5) The Iatro-Chemical Period, originating with Paracelsus, and culminating in the work of van Helmont (1577-1644). (6) The Pneumatic Period, beginning with Boyle, including Stahl, Black, Cavendish, Priestley, and Scheele, and brought to its conclusion by Lavoisier. (7) The Period of the Modern Atomic Theory, beginning with Dalton, carried forward by Gay Lussac, Avogadro, Ampère, Davy, and Berzelius, and brought to completion by the exposition of Cannizzaro. (8) The Period of Organic Chemistry and of the Periodic Law, including Liebig, Wöhler, and Dumas, on the one side, Mendeléeff and Lothar Meyer on the other. (9) The Period of Physical Chemistry, originating with van't Hoff and Arrhenius. (10) The period of Radio-activity.
T. M. L.
FISHERY RESEARCH IN INDIA.1
MR. SOUTHWELL deserves the thanks of those interested in the better organisation of Imperial resources for summarising the history of fisheries research in India. That dates back only to 1906, for the work of Dr. Francis Day and Colonel Alcock was purely systematic. In 1906 economic research was initiated. Sir K. Gupta, then about to retire from the higher ranks of the Indian Civil Service, was ordered to inquire into the fisheries of Bengal. This officer tells us himself that he knew nothing of fish," and that he “had not even done anything with the rod and line." Nevertheless, he made a lengthy tour in Europe and America to see those who did know, and on his return to India a Bengal Fisheries Department was established, with Mr. A. Ahmed as Commissioner.
The Department then obtained the services, for a year or so, of Dr. J. T. Jenkins, and an English steam trawler, the Golden Crown, was sent out to make a survey of the fishing grounds in the Bay of Bengal. While this was going on Mr. Ahmed established a Board which met five times, after which he "ceased to be Commissioner." The result of a very imperfect survey was the formation of a Fishery Department consisting of two directors of agriculture ("whose knowledge of the fisheries is necessarily of an entirely administrative nature"), of Mr. Southwell himself (a trained zoologist), as deputy-director, and
1 Report on Fishery Investigations in Bengal, etc., with Recommendations for Future Work. By T. Southwell, Deputy Director of Fisheries for Bengal, etc. Bulletin No. 5. Department of Fisheries, Bengal. (Calcutta: The Bengal Secretariat Book Depot, 1915.) Price 6d.
of two superintendents of fisheries. One of these! latter officials graduated at Calcutta University as M.A. in light and acoustics, and afterwards carried on physical research, and the other was a student at the same University, but "failed to take his B.A. degree." The Department is allowed the use of a laboratory at the Indian Museum, and has also a laboratory on board a steam launch. Thus staffed and equipped it is proceeding with the investigation of Bengal fisheries!
The latter are fresh-water, estuarine, and marine. The edible fresh-water fishes are mainly various species of carp, and a Clupeoid fish called the hilsa. The carps breed in the rivers during the rains, and since extensive areas of Bengal are then flooded enormous numbers of fry are lost in the paddy-fields. How to make good this loss by artificial culture, and also how to deal similarly with the hilsa are obvious problems, neither of which was solved by Dr. Jenkins or Mr. Southwell. The estuarine fisheries are located in the Sunderbans-that is, the rivers, swamps, and islands formed by the deltas of the Ganges and Brahmaputra. Here there are abundant edible fishes and crustacea, but no edible mollusca. The fishing grounds are far away from the traffic routes, and the fishing boats and gear are crude and inefficient. The fauna of the Sunderbans is not even adequately investigated. Dr. Jenkins spent thirty days there with an imperfectly equipped steam launch. Mr. Southwell tells us that Dr. Jenkins showed that "the fish fauna varied greatly according to degree of salinity, depth of water, etc., and observed that nets suitable for fishing in one part of the estuaries might be unsuitable in other parts"results that might have been predicted! He also conIcluded that a 66 properly organised scheme of development of these fisheries would yield a profitable return on capital invested"-an equally indisputable con
Much more is known about the prospects of a marine fishery. Colonel Alcock regarded the fishery of the Bay of Bengal as of very great potential value. There are numerous species of edible fishes, mostly Siluroids, Scienoids, Serranoids, Pleuronectids, and Clupeoids. Dr. Jenkins and Dr. Annandale, of the Indian Museum, have made good reports on the collections made by the former. The Golden Crown was an inefficient and poorly equipped steam trawler, and was, moreover, hampered in its work by the Commissioner of Fisheries, but Dr. Jenkins showed that it was possible to trawl in the Bay of Bengal throughout the whole year. The vessel caught, on the average, 26-6 cwt. of fish a day. (The average catch per day's absence from port of an English steam trawler varies from about 60 cwt. to about 9 cwt., according to the fishing ground. The average catch per day for the North Sea, before the war, was about 17 cwt.) If it were possible to eat, with pleasure, all the species of fish caught, the Golden Crown therefore had good results. Dr. Jenkins showed also that it was possible to send fish in good condition to the Calcutta markets, and that the difficulty of navigation of the Hooghly could be evaded. There was reason for believing that European fishery methods would succeed in the Bay of Bengal.
But so far they have not been attempted, nor has the development of the fresh-water and estuarine fisheries been seriously attempted. One's impression in reading the report is that of a Department which, having the conviction that something ought to be done, yet contents itself by doing it badly. Commercial exploitation is, of course, a matter for private enterprise, but it is "up to" the Department which has modelled itself on European lines to see that scientific research is adequately promoted. J. J.
VENTILATION AND HEALTH.
NEW YORK State Commission on Ventilation has been established by the New York Association for Improving the Condition of the Poor, with the help of a grant from Mrs. E. M. Anderson. The Commission consists of the university professors respectively of physiology, chemistry, psychology, and clinical medicine, together with a ventilating engineer and an officer of the New York State Department of Health, all of whom give their time voluntarily. A Commission so constituted ought to produce results of great value. An experimental chamber has been put up and equipped with all necessary apparatus, and researches have been made into the conditions of schools, hospitals, business houses, etc. The Commission has now issued its first report. The report confirms the view that the physical, rather than the chemical, conditions of the air are of the greater importance. That temperature, humidity, and movement of the air and its freedom from dust, bacteria, and odour are the first essentials of ventilation.
Stagnant air at the same temperature as fresh air, even when it contains twenty or more parts of carbon dioxide, and all the organic and other substances in the breathed air of occupied rooms, has, so far, shown no effect on any of the physiological processes, except that the appetite for food may be slightly reduced.
Here we have confirmation of the view recently expressed in these columns by Prof. Leonard Hill, that stagnant air by reducing the metabolism of the body impoverishes the health and vigour of the body. Over-heated rooms produce a slight but distinct elevation of body temperature, increase the rate of the heart in the reclining posture, and its acceleration on rising from the reclining to the standing posture, and slightly lower the systolic blood pressure. The increased heart rate and diminished blood pressure found in the standing position. show how the heated atmosphere relaxes the tone of the body, and tends to make the blood sink down into the dependent parts, and so produces sensations of fatigue, and reduces the inclination to work. The physiologist of the Commission, Prof. Fred S. Lee, determined after exposing cats to over-heated rooms that their excised muscles were more easily fatigued than those of the controls-14-26 per cent. less work was done. The sugar in their blood was also diminished.
In a commercial establishment employing about 4000 clerks, it was found that the building was grossly overheated. The fault lay in large part with the disrepair into which the thermostats had been allowed to fall. In certain ducts designed for fan ventilation no fans had been installed, while in others the fans were running at only a fraction of their efficient speed. In other ducts the register openings were so badly adjusted that while some rooms received more air than needed, others received less.
In one hospital in New York notorious for its overheating, records of 70°, 74°, and two 75° and above were obtained. In the children's wards five out of eleven records were more than 70° F. In one hospital the children's ward was 76° in the daytime. The Commissioner regards anything below 70° F. as free from over-heating!
Abominable conditions are proven then in certain institutions in New York, conditions which sap the health and vigour of the young, and turn them, so to speak, into weak hothouse plants. General recognition is required that the chief aim of ventilation is to provide a moving current of cool air, to remove the heat produced by human metabolism, and by the
combustion of illuminants. The Commission, so far, has added a considerable amount of useful information and evidence confirmatory of the views held by physiologists in this country.
The Commission points out how in the case of a school building the chosen architect strives to outdo the others in the size and ornamentation of the building, and to satisfy the excessive requirements of the school committee. Then begins the process of trimming, and the heating and ventilating plant being the biggest single item of equipment, comes in for the most attention, with the worst results. And this, too, despite the fact that the ventilating plant is really the lungs of the building, and counts most for the comfort and efficiency of the occupants. 'But, of course, there must be so many rooms, just so many gargoyles, and just so much marble. For these things are seen and read of all men."
THE UTILISATION OF SOLAR ENERGY. IN N the first part of the paper referred to below are given particulars of the various apparatus which have been used to obtain power from solar radiation. In doing so the author was able to describe in fair
spheric) pressure, an efficient engine for the use of such steam had to be designed, and in this Shuman was uniquely successful, for when the author tested a Shuman low-pressure engine at Erith he found that its steam consumption was only 22 lb. of steam at 16.2 lb. sq. in. abs. per b.h.p. hour, the b.h.p. being 94.5. This beats all the old atmospheric engines, and indeed everything to date for steam at that pressure. Shuman's 1910 sun heat absorber had an area of only 15 sq. ft., and the tests made with it showed that at its best 300 sq. ft. of it would be required to produce enough steam for one b.h.p., allowing the aforementioned 22 lb. per b.h.p. hour.
In 1911 Shuman made another absorber like that of 1910, except that it had two plane silvered glass mirrors, one attached to the upper edge of the "hotbox" and one to the lower, and so arranged that 2 sq. ft. of solar radiation were concentrated on to I sq. ft. of boiler surface. The hot-box (originated by H. B. de Saussure, the Swiss geologist, physicist, and naturalist, who died in 1799) was 3 ft. wide, 6 in. deep, and 66 ft. long. There were twenty-six such sections. The back was formed of -in. millboard, on top of which was 2 in. of cork-dust, covered with-in. millboard. The laminar boiler (about in.
detail the construction of such apparatus, but he stated that he had been unable to get any but the most meagre information as to the results obtained with the various plants, and much of that appeared to be untrustworthy. The case was very different in regard to Shuman's work since and including the year 1910, for the author himself had conducted the experiments with Shuman's plants.
Shuman set out with the idea that the principle of former workers in this field of using high-pressure steam was wrong. He argued that high-pressure steam meant a high temperature, and therefore a large loss due to radiation. We shall see later that though this is true, a better overall efficiency is attained when the steam pressure is higher and the boiler area is reduced, so as to save loss due to radiation. It is also a thermal advantage to have a high concentration of the solar radiation, but Shuman started first of all without any concentration (which has its advantages by reason of simplification); then he used a concentration of two to one, and finally 4.6 to one.
As steam was to be generated at a low (say atmo1 Abstract of a paper read before the Royal Society of Arts on April 28 by A. S. E. Ackermann.
thick) was fixed in front of this, leaving an air space of an inch between it and the millboard. In front of the boiler was another air space of an inch, then a sheet of window-glass, another air space of 1 in., and finally the top sheet of window-glass. 10,296 sq. ft. of solar radiation were thus collected, and the best hour's run gave 816 lb. of steam at a pressure of 14.2 lb. sq. in. abs., equivalent to 26-8 b.h.p. and a thermal efficiency (of the absorber alone) of 29.5 per
The orientation of these reflectors was east and west, and they did not "follow the sun," consequently the output of steam fell off considerably in the morning and evening. The solar radiation was received on only one side of the 1911 boiler, while the other side lost some of it, but, due to Prof. C. V. Boys, F.R.S., the boiler which the author tested in Egypt in 1913 received heat on both of its sides, and its top edge as well, and the concentration was 4-6 to 1. The orientation of the reflectors was north and south, and they were made automatically to "follow the sun." Each reflector was trough-shaped, parabolic in cross section, 13 ft. 5 in. across the top and 205 ft. long, and there were five such sections. Hence 13,752 sq. ft. of solar radiation were collected. The
average power for the five hours' run on that day (August 22, 1913) was 594 b.h.p. per acre, and the minimum on the same day was 52-4 b.h.p. per acre, a decrease of only 16.8 per cent. The maximum thermal efficiency of the absorber alone was 401 per cent., or 36 per cent. better than the thermal efficiency of the 1911 absorber, while its steam production was 33 per cent. better.
The author's experiments in Egypt show that a decrease of 20 per cent. in the humidity of the atmosphere caused an increase of 30 per cent. in the steam production.
Nearly all the technical part of the paper is contained in three appendixes, while a fourth consists of the bibliography of the subject.
In appendix i. it is shown that when the 1913 absorber was tested with naked boilers, the solar heat not used, and expressed in B.T.U. per hour per sq. ft. of boiler surface per 1° F. difference in temperature between the boiler and the air, is nearly constant and equal to 8.68.
In appendix ii. is derived the equation to thermal efficiency of a solar heat absorber and the efficiencies calculated by means of it are compared with the actual thermal efficiencies.
In appendix iii. the equation to the thermal efficiency of the absorber is combined with the equation to the thermal efficiency of a Carnot engine, thus giving the overall thermal efficiency. From this it is shown that the theoretical maximum overall efficiency of the 1913 Egyptian plant was 5.9 per cent., while the actual efficiency was 4.32 per cent. Thus 73.2 per cent. of the maximum possible efficiency was attained.
The equation to the thermal efficiency of the absorber is :
Inserting the values of the known quantities for the Egyptian plant gives :
no=0'71-404T-1+9'45 x 10-10T3 - 1664 x 10-12T4 (3) where D=the width in feet of the reflector; p=the perimeter in feet of the boiler; r=the efficiency of silvered glass as a reflector of heat; s=the solar constant in B.T.U. per square foot per min. =7.12; a=the coefficient of atmospheric transmission; T=the absolute temperature in degrees F. of the boiler; A the absolute temperature in degrees F. of the
FIG. 3.-Curve showing the relation of the overall efficiency of the 1913 Shuman-Boys sun-heat absorber (with naked boiler), combined with a Carnot engine, to the absolute temperature of the boiler steam.
reflectors; 568=the absolute temperature in degrees F. of the condenser; k=the coefficient of radiation, conduction, and convection=10-10 x 0.36 B.T.U. per sq. ft. per min., this value having been determined by the author in 1899 under almost the same conditions as those to which it is now applied.
The graph of equation (3) is given in Fig. 3, from