amount, or even a thousandth part of it, could have existed at any one time in our terrestrial atmosphere since the beginning of life on our planet is inconceivable, and that it could be supplied from the earth's interior is an hypothesis equally untenable.

I was therefore led to admit for it an extra-terrestrial source, and to maintain that the carbonic acid has thence gradually come into our atmosphere to supply the deficiencies created by chemical processes at the earth's surface. Since similar processes are even now removing from our atmosphere this indispensable element, and fixing it in solid forms, it follows that except volcanic agency, which can only restore a portion of what was primarily derived from the atmosphere, there are on earth, besides organic decay, only the artificial processes of human industry which can furnish carbonic acid; so that but for a supply of this gas from the interstellary spaces now, as in the past, vegetation, and consequently animal life itself, would fail and perish from the earth for want of this "food of planets."

Such were the conclusions, based on an induction from the facts of modern chemistry and geology, which I enunciated in my paper in 1878 and 1880, already quoted in the first part of this essay. I was at that time unacquainted with the Hypothesis of Newton, and with his remarkable reasoning contained in the 41st proposition of the third book of the Principia, in which he, so far as was possible with the chemical knowledge of his time, anticipated my own argument, and showed how and in what manner the interstellary ether may really afford the "food of planets," and, in a sense, "the material principle of life."

I have thus endeavoured to bring before the Philosophical Society of Cambridge, a brief history of the development of this conception of an interstellary medium, and to show that the thought of two centuries has done little more than confirm the almost forgotten views of Newton. It is with feelings of peculiar gratification that I have been able to indite these pages within the very walls of the College in which our great philosopher lived and laboured, and where, combining all the science of his time with a foresight which seems well-nigh divine, he was enabled, in the words of the poet, "to think again the great thought of the Creation."

(2) On the Upper Bagshot Sands of Hordwell Cliffs, Hampshire. By E. B. TAWNEY, M.A.

These sands appear in Hordwell Cliffs-at that part locally designated Long Mead End-and extend to a little W. of Beacon Bunny, where they run out at the top of the cliff. By old observers they were sometimes termed the "Estuarine-bed of Long Mead End" to distinguish them from the Marine Barton Clays below, and the more freshwater Lower Headon beds above. They have been recognised by most observers-from their position as identical with similar sands in the Isle of Wight which come immediately below the Lower Headon beds. We will first refer to some of the older authorities who have treated of the district.

1814. T. WEBSTER. On the Freshwater Formations of the Isle of Wight, with some observations on the Strata over the Chalk in the South-east part of England. Trans. Geol. Soc., 1st Ser. 11. p. 161, pl. 9-11. In the map of the district all the "strata above the chalk" are coloured with one tint. In his very excellent coloured section of the coast between Headon Hill and Alum Bay in the Isle of Wight, these Upper Bagshot sands are indicated by a capital letter D. No details are given concerning them [p. 184] in the letter-press.

1822. A. SEDGWICK. On the Geology of the Isle of Wight. Annals of Phil., New Series III. p. 329. Prof. Sedgwick having apparently lost his specimens and notes writes from memory, aided by memoranda made by his companion, Dr Whewell. On pp. 344, 348 he mentions that these "sandy strata contain a few marine shells, among which we remarked some very large Cerithia." He recognises the general parallelism of the Barton and Hordwell Cliff section to that of White Cliff Bay, though owing to the Barton clay at this time being universally confused with the London clay, the account of some of the beds is a little puzzling. The Lower Headon (freshwater) beds above the sands are recognised.

1824. T. WEBSTER. On a Freshwater Formation in Hordwell Cliff, Hampshire, and on the subjacent beds from Hordwell to Muddiford. Trans. Geol. Soc., 2nd Ser. 1. p. 90. After describing the freshwater Lower Headon [p. 93] we read "immediately below this formation at Hordwell is found a bed of sand from 60 to 100 feet thick, which appears first about Long Mead End, and may be well observed in the section at Beacon Bunny." The thickness is exaggerated, and the passage following as well as his section [pl. XII. fig. 1] shows that he has taken into it some of the Barton sands which are really separated by the Beacon Bunny clay, &c. The equivalence of the Hordwell "lower freshwater formation" [L. Headon] to that of the Isle of Wight is deduced from their both lying on this bed of sand, in which in some places intervenes a bed with a mixture of marine and freshwater shells,

1829. C. LYELL. On the Freshwater Strata of Hordwell Cliff, Beacon Cliff, and Barton Cliff, Hampshire. Trans. Geol. Soc., 2nd Ser. II. p. 287, pl. xxx. A detail description is given of the freshwater Hordwell beds [L. Headon], and a coloured section [pl. xxx. f. 2] of the coast shows their disposition in the cliffs. The sands in question are described [p. 290] as No. 9; "white siliceous sand, without shells, rises near Long Mead End and extends through Beacon and Barton Cliffs to the middle of High Cliff." This is a repetition of the error in Webster's section. He misses their fossil contents apparently. Classifying them in the lower freshwater, he makes them extend too far to the west. They do not extend to Barton, much less to "about the middle of High Cliff." He further is in error in stating that "no portion of the upper marine formation [M. Headon] exists anywhere in this part of the Hampshire coast."

.1838. D'ARCHIAC. Note sur les sables et grès tertiaires. Bull. Soc. Geol. Fr., S. 1. t. IX. p. 54. After a description of the French beds of Barton age notices shortly these sands at Hordwell, and attributes a marine orígin to them rather than freshwater.

1839. D'ARCHIAC. Essais sur la coordination des terrains tertiaires du nord de la France, de la Belgique et de l'Angleterre. Bull. Soc. Geol. Fr., Ser. 1. t. x. p. 169. Mentions the lower freshwater at Hordwell reposing on these sands, which are correlated with the "sables moyens" [pp. 172, 200, 208].

1846. S. WOOD. On the discovery of an Alligator and of several new Mammalia in the Hordwell Cliff, with observations upon the Geological Phenomena of that Locality. Charlesworth's Geol. Journ. 1. p. 1. It is here shown that Lyell must have missed the upper marine [M. Headon], for it was recognised as a bed 10 inches think, and its fossils were collected by himself and Mr Edwards. He next corrects Lyell's view of the extent of the freshwater formation towards Barton. Our Upper Bagshot sand is described [p. 4] as coming below the freshwater deposit, and characterised as a "greyish-white or light-coloured "sand, containing the following molluscous genera, some species being "extremely abundant: Oliva, Potamides, Ancillaria, Natica, Melania, "Melanopsis, Pleurotoma, Bulla, Mactra, Cyrena, Corbula, Sanguinolaria "[Psammobia?], Venericardia, Cytherea, Lucina, Potamomya. The bed appears to be intermediate between the London [Barton] clay and "the lower freshwater, and must be referred to an estuary formation, "for the largest proportion of its Testacea are referable to marine genera.' He then proposes to term it the "Lower Marine" in opposition to the "Upper Marine," but the inadmissibility of such a term is apparent.

66

"At

1846. J. PRESTWICH. On the Tertiary or Supracretaceous Formations of the Isle of Wight as exhibited in the Sections at Alum Bay and White Cliff Bay. Q. J. G. S. 11. p. 223. We read, p. 227, Hordwell Cliff, on the top of the London [Barton] clay, there is a bed of sand fifteen feet thick overlaid by a few inches of grey clay." In the section [pl. IX. f. 1] from Alum Bay to Heatherwood Point, Isle of

Wight, these sands are termed Headon Hill sands, and estimated [p. 228] at 120 feet. At White Cliff Bay these sands are said to be 195 feet thick [No. 20 of pl. ix. f. 2, pp. 228, 243, 253]. They are supposed to be without organic remains, though he remarks that Prof. Sedgwick had found marine shells in the equivalent bed at Hord well. The succeeding passage mentioning marine and estuary fossils at Hempstead refers of course to beds far higher in the series, and now classed as Miocene.

1846. S. WOOD. On the Fossils and Geological Phenomena of the Hordwell Cliff. Charlesworth's Geol. Journ. 1. p. 117. Merely mentioned the sands again as containing 27 genera of marine Mollusca.

1852. A. DUMONT. Observations sur la Constitution Géologique des terrains tertiaires de l'Angleterre comparés à ceux de la Belgique. Bull. Ac. R. Belg. XIX. In the table at end of the essay the Headon Hill sands are grouped with the Barton clay as equivalent respectively to the upper and lower divisions of the Belgian Laekenien, while the Headon Hill limestones and marls are placed in the Tongrien.

1852. E. HÉBERT. Comparison des couches tertiaires inférieures de la France et de l'Angleterre. Bull. Soc. Geol. Fr., Ser. 2, t. ix. p. 350. In a table in which a parallel is drawn between the deposits of the Hampshire, Thames and Paris basins, these sands are correlated with the unfossiliferous sands of Monneville and of the Mortefontaine heath [that is to say the upper part of the sables de Beauchamp].

1852. MARCHIONESS OF HASTINGS. Description géologique des falaises d'Hordle, sur la côte du Hampshire, en Angleterre. Bull. Soc. Geol. Fr., Ser. 2, t. ix. p. 191. A measured section of Hordwell freshwater beds with their fossils. Bed 17, about 20 feet thick, is the sand in question. It is described as sand of a variable colour, greyish-green where it rises from below the sea-level, but becoming lighter westward, so that beyond Long Mead End it is pale yellow with ferruginous bands and grey spots, p. 202. The first five feet are said to be full of shells but not disposed in lines; but the thickness over which fossils extend diminishes, so that after a time fossils disappear. Remains of Lamna, Myliobates, Etobates are mentioned; but the Crocodile, Trionyx, Potamomya, come from the bed in contact with it above, and which is the bottom bed of the freshwater formation.

1852. T. WRIGHT. A Stratigraphical Account of the Section from Round Tower Point to Alum Bay on the N. W. coast of the Isle of Wight. Proc. Cotsw. Club, 1. p. 87, and Ann. Mag. Nat. Hist:, Ser. 2, VII. p. 14. These sands in the Isle of Wight, at Headon Hill, are mentioned as containing "fragments of shells too minute and waterworn to ascertain to what genus they belonged" [p. 98]. They are said [p. 99] to be probably of estuary origin. The equivalent bed at Beacon Cliff containing estuary and marine genera, together with the bones of turtles and teeth of sharks.

1852. T. WRIGHT. A Stratigraphical Account of the Section of Hordwell, Beacon and Barton Cliffs, on the coast of Hampshire. Proc. Cotsw. Club, 1. p. 120. The Hordwell freshwater and " upper marine”

[M. Headon] is described bed by bed. The sands in question are said [p. 128] to rise about 300 yards W. of Long Mead End and run out at Beacon Bunny, inclined at an angle of 2°. They are divided into No. 16, greyish-white sand scarcely coherent, the fossils not well preserved; the 14 genera cited have been previously mentioned by S. Wood: one specific name Lucina divaricata is mentioned. Thickness of the grey sand 5 feet, and it passes gradually into No. 17, of which it may be considered to form the upper fossiliferous portion. No. 17, fine white sand, non-fossiliferous; estimated at 15 to 20 feet. Much importance is attached to the fossil contents from the light which they throw on the conditions under which the [equivalent] Headon Hill sands were deposited, being probably of estuary origin. [We may add that the upper freshwater [U. Headon], as described by Dr Wright, above the ". upper marine" does not exist at Hordle; probably some of the L. Headon, the position of which has been mistaken, are referred to.]

1853. E. FORBES. On the Fluvio-marine Tertiaries of the Isle of Wight. Q. J. G. S. ix. p. 259. In this important paper, which first announced the distinctness of the Hempstead [Hamstead] series from the Headon Hill beds, and confirmed the identity of the Colwell Bay and Headon Hill marine beds, there is a mention, p. 268, of the sands below the "lower beds of the Headon series. No fossils have hitherto been re"corded from these beds or the Isle of Wight. At White Cliff Bay how"ever, to all appearance barren, they are highly fossiliferous, containing "abundant impressions of marine shells, apparently of Barton species. "The shelly matter has entirely disappeared...the specimens are quite "untransportable."

1856. E. FORBES. On the Tertiary Fluvio-marine Formation of the Isle of Wight. Memoirs of the Geological Survey of Great Britain. This Survey Memoir, giving the details on which the conclusions of the preceding paper were founded, contains the following passage, p. 86: “The "lower beds of the Headon series rest upon sands known as Headon Hill "sands, which are the equivalents of the Upper Bagshot beds...these "sands rest upon the Barton clays or highest portion of the Middle "Bagshot group."

1856. J. PRESTWICH.

On the Correlation of the Eocene Tertiaries of England, France and Belgium, Part II. Q. J. G. S. xiii. p. 89. At p. 108 we read, "In the Barton series I would now include the "siliceous sands at the base of Headon Hill and at the top of Barton "Cliff. For though they contain no organic remains at those spots, casts "of marine shells, apparently of the same species as those in the under"lying clay occur in considerable numbers in parts of the sands occupying "the same position at White Cliff Bay; whilst at Barton, as the deposit "ranges eastward towards Hordwell Cliff, these sands alternate with fos"siliferous grey clays, and form with them passage beds between the "purely marine and compact Barton clay and the freshwater beds nearer "Milford." p. 109, "At the Barton section a thin bed of sand, full of "estuarine shells, is interposed between the Barton sands and the fresh"water clays." [L. Headon.]