of water. Observations were commenced after the new pipe and cistern had been in use for several weeks, and were continued for about a month, the water being changed every twenty-four hours. The average quantity of lead dissolved in twenty-four hours was

The greater amount dissolved in the new pipe than in the cistern depends upon the larger extent of surface exposed, in the one case 31, in the other only 23 square feet to each cubic foot of water.

The quantities of lead dissolved, although considerably below the proportions considered actually dangerous, are, nevertheless, somewhat alarming; but it must be borne in mind that the water seldom remains in the pipes more than ten hours, and that during the night when the temperature is lowest; and that in Glasgow the water is usually supplied direct from the street mains, without passing through cisterns.

The results obtained by Dr. Wallace may be reduced to the following summary:1st, the water-supply of Glasgow is very sensibly harder, and acts with considerably less energy on lead than the water of Loch Katrine; and 2nd, the amount of lead taken up from pipes and cisterns at the present time is not such as to give rise to serious apprehensions.

On an Apparatus for the rapid Separation and Measurement of Gases. By Drs. WILLIAMSON and RUSSELL,

GEOLOGY.

Address by Sir RODERICK IMPEY MURCHISON, D.C.L., LL.D., F.R.S., Director-General of the Geological Survey of the United Kingdom, President of the Section.

ALTHOUGH I have had the honour of presiding over the geologists of the British Association at several previous Meetings since our first gathering at York, now thirty years ago, I have never been called upon to open the business of this Section with an address; this custom having been introduced since I last occupied the geological chair at Glasgow, in 1855.

The addresses of my immediate predecessors, and the last anniversary discourse of the President of the Geological Society of London, have embraced so much of the recent progress of our science in many branches, that it would be superfluous on my part to go over many topics again which have been already well treated.

Thus, it is needless that I should occupy your time by alluding to the engrossing subject of the most recent natural operations with which the geologist has to deal, and which connect his labours with those of the ethnologist. On this head I will only say that, having carefully examined the detrital accumulations forming the ancient banks of the river Somme in France, I am as complete a believer in the commixture in that ancient alluvium of the works of man with the reliquiæ of extinct animals, as their meritorious discoverer, M. Boucher de Perthes, or as their expounders, Prestwich, Lyell, and others. I may, however, express my gratification in learning that our own country is now affording proofs of similar intermixture both in Bedfordshire, Lincolnshire, and other counties; and, possibly, at this Meeting we may have to record additional evidences on this highly interesting topic.

But I pass at once from any consideration of these recent accumulations, and, indeed, of all tertiary rocks; and, as a brief space of time only is at my disposal, Í will now merely lay before you a concise retrospect of the progress which has latterly been made in the development of one great branch of our science. I confine myself then to the consideration of those primeval rocks with which my own researches have for many years been most connected, with a few allusions only to metamorphism, and certain metalliferous productions, &c.

There is, indeed, a peculiar fitness in now dwelling more especially on the ancient rocks, inasmuch as Manchester is surrounded by some of them, whilst, with the exception of certain groups of erratic blocks and drifts, no deposits occur within the reach of short excursions from hence, which are either of secondary or tertiary age. Let us, then, take a retrospective view of the progress which has been made in the classification and delineation of the older rocks since the Association first assembled at York in 1831. At that time, as every old geologist knows, no attempt had been made to unravel the order or characters of the formations which rise from beneath the Old Red Sandstone. In that year Sedgwick was only beginning to make his first inroads into those mountains of North Wales, the intricacies of which he finally so well elaborated, whilst I only brought to that, our earliest assembly, the first fruits of observations in Herefordshire, Brecon, Radnor, and Shropshire, which led me to work out an order which has since been generally adopted.

At that time the terms 'Cambrian,' 'Silurian,' 'Devonian,' and 'Permian,' were not dreamt of; but, acting on the true Baconian principle, their founders and their coadjutors have, after years of toil and comparison, set up such plain landmarks on geological horizons that they have been recognized over many a distant land. Conpare the best map of England of the year 1831, or that of Greenough, which had advanced somewhat upon the admirable original classification of our father, William Smith, and see the striking difference between the then existing knowledge and our present acquirements. It is not too much to say that when the British Association first met, all the region on both sides of the Welsh border, and extending to the Irish Channel on the west, was in a state of dire confusion; whilst in Devonshire and Cornwall many of these rocks, which from their crystalline nature were classed and mapped as among the most ancient in the kingdom, have since been shown to be of no higher antiquity than the Old Red Sandstone of Herefordshire.

As to Scotland, where the ancient rocks abound, though their mineral structure, particularly in those of igneous origin, had necessarily been much developed in the country of Hutton, Playfair, Hall, Jameson, and McCulloch, yet the true age of most of its sedimentary rocks and their relations were unknown. Still less had Ireland, another region mainly paleozoic, received any striking portion of that illustration which has since appeared in the excellent general map of Griffith, and which is now being carried to perfection through the labours of the Geological Survey under my colleague Jukes. If such was our benighted state as regarded the order and characters of the older formations at our first Meeting, great was the advance we had made when at our twelfth Meeting we first assembled at Manchester in 1842 Presiding then as I do now over the Geological Section, I showed in an evening lecture how the paleozoic rocks of Silurian, Devonian, and Carboniferous age, as well as those rocks to which I had assigned the name of Permian, were spread over the vast region of Russia in Europe and the Ural Mountains. What, then, are some of the main additions which have been made to our acquaintance with the older rocks in the British Isles since we last visited Manchester?

Commencing with the oldest strata, I may now assume, from the examination of several associates on whose powers of observation as well as my own I rely, that what I asserted at the Aberdeen Meeting in 1859, as the result of several surveys, and what I first put forth at the Glasgow Meeting of 1855, is substantially true. The stratified gneiss of the north-west coast of the Highlands, and of the large island of Lewis and the outer Hebrides, is the fundamental rock of the British Isles, and the precise equivalent of the Laurentian system of Canada, as described by Sir W. Logan. The establishment of this order, which is so clearly exhibited in great natural sections on the west coast of Sutherland and Ross, is of great importance in giving to the science we cultivate a lower datum-line than we previously possessed, as first propounded by myself before the British Association in 1855*.

*See Report of British Association for 1855 (Glasgow Meeting). At that time I was not aware that the same order was developed on a grand scale in Canada, nor do I now know when that order was there first observed by Sir W. Logan. I then (1855) simply put forward the facts as exhibited on the north-west coast of Scotland; viz. the existence of what I termed a lower or "fundamental gneiss," lying far beneath other gneissose and crystalline strata, containing remains which I even then suggested were of Lower Silurian age. Subsequently, in 1859, when accompanied by Professor Ramsay, I adopted, at his sugges

For hitherto the order of the geological succession, even as seen in the Geological Map of England and Wales or Ireland, as approved by Sir Henry De la Beche and his able coadjutors, Phillips, Ramsay, Jukes, and others, admits no older sediment than the Cambrian of North Wales, whether in its slaty condition in Merioneth and Caernarvon, or in its more altered condition in Anglesea.

The researches in the Highlands have, however, shown that in our own islands, the older palæozoic rocks, properly so called, or those in which the first traces of life have been discovered, do repose, as in the broad regions of the Laurentian Mountains of Canada, upon a grand stratified crystalline foundation, in which both limestones and iron-ores occur.subordinate to gneiss. In Scotland, therefore, these earliest gneissic accumulations are now to be marked on our maps by the Greek letter alpha, as preceding the Roman a, which had been previously applied to the lowest known deposits of England, Wales, and Ireland. Though we must not dogmatise and affirm that these fundamental deposits were in their pristine state absolutely unfurnished with any living things (for Logan and Sterry Hunt, in Canada, have suggested that there they indicate traces of the former life), we may conclude, that in the highly metamorphosed condition in which they are now presented to us in North-western Britain, and associated as they are with much granitic and hornblendic matter, they are, for all purposes of the practical geologist, "azoic rocks." The Cambrian rocks, or second stage in the ascending order as seen reposing on the fundamental gneiss of the north-west of Scotland, are purple and red sandstones and conglomerates forming lofty mountains. These resemble to a great extent portions of the rocks of the same age which are so well known in the Longmynd range of Shropshire, and at Harlech in North Wales, and Bray Head in Ireland.

At Bray Head they have afforded the Oldhamia, possibly an Alga, whilst at the Longmynd, in Shropshire, they have yielded to the researches of Mr. Salter some worm-tracks, and the trace of an obscure crustacean.

The Highland rocks of this age, as well as their equivalents, the Huronian rocks of North America, have as yet afforded no trace whatever of former life. And yet such Cambrian rocks are in parts of the Longmynd, and specially in the lofty mountains of the north-western Highlands, much less metamorphosed than many of the crystalline rocks which lie upon them. Rising in the scale of successive deposits, we find a corresponding rise in the signs of former life on reaching that stage in the earlier slaty and schistose rocks in which animal remains begin clearly to show themselves. Thus, the Primordial Zone of M. Barrande is, according to that eminent man, the oldest fauna of the Silurian Basin in Bohemia*.

In the classification adopted by Sir Henry De la Beche and his associates, the Lingula Flags (the equivalent of the Zone Primordiale of Barrande) are similarly placed at the base of the Silurian system. This Primordial Zone is also classed as the Lowest Silurian by De Verneuil, in Spain; by James Hall, Dale Owen and others, in the United States; and by Sir W. Logan, Sterry Hunt, and Billings, in Canada†. In the last year, M. Barrande has most ably compared the North American Tacotion, the word 'Laurentian,' in compliment to my friend Sir William Logan, who had then worked out the order in Canada, and mapped it on a stupendous scale. I stated, however, at the same time, that if a British synonym was to have been taken, I should have proposed the word 'Lewisian,' from the large island of the Lewis, almost wholly composed of this gneiss.

* I learn, however, that in Bohemia Dr. Fritsch has recently discovered strata lying beneath the mass of the Primordial Zone of Barrande, and in rocks hitherto considered azoic the fossil burrows of annelide animals similar to those of our own Longmynd.

In completing at his own cost a geological survey of Spain, in which he has been occupied for several years, and in the carrying out of which he has determined the width of the sedimentary rocks of the Peninsula (including the Primordial Silurian Zone, discovered by that zealous explorer, M. Casiano de Prado), M. de Verneuil has in the last few months chiefly examined the eastern part of the kingdom where few of the older paleozoic rocks exist. I am, however, informed by him, that Upper Silurian rocks with Cardiola interrupta, identical with those of France and Bohemia, occur along the southern flanks of the Pyrenees, and also re-occur in the Sierra Morena, in strata that overlie the great mass of Lower Silurian rocks as formerly described by M. Casiano de Prado and himself. The southern face of the Pyrenees, he further informs me, is specially marked by the display of mural masses of Carboniferous strata, which, succeeding the Devonian rocks, are not arranged in basin-shape, but stand out in vertical or highly inclined positions, and are followed by 1861. 7

nic group of Emmons with his own primordial Silurian fauna of Bohemia and other parts of Europe; and although that sound palæontologist, Mr. James Hall. has not hitherto quite coincided with M. Barrande in some details†, it is evident that the primordial fauna occurs in many parts of North America. And as the tree order of succession has been ascertained, we now know that the Taconic groups of the same age as the lower Wisconsin beds described by Dale Owen, with their Paradoxides, Dikelocephalus, &c., as well as of the lower portion of the Quebec rocks, with their Conocephalus, Axionellus, &c., described by Logan and Billings. Of the crystalline schists of Massachusetts, containing the noble specimen of Paradorides described by W. Rogers, and of the Vermont beds, with their Oleni, it follows that the Primordial Silurian Zone of Barrande (the lower Lingula-flags of Britain) is largely represented in North America, however it may occupy an inverted position in some cases, and in others be altered into crystalline rocks.

In determining this question due regard has been had to the great convulsions, inversions, and breaks to which these ancient rocks of North America have been subjected, as described by Professors Henry and W. Rogers.

In an able review of this subject, Mr. Sterry Hunt thus expresses himself:"We regard the whole Quebec group, with its underlying primordial shales, as the greatly developed representatives of the Potsdam and Calciferous groups (with part of that of Chazy), and the true base of the Silurian system..... The Quebec group, with its underlying shales," this author adds (and he expresses the opinion of Sir W. Logan), "is no other than the Taconic system of Emmons;" which is thus, by these authors, as well as Mr. James Hall, shown to be the natural base of the Silurian rocks in America, as Barrande and De Verneuil have proved it-to be on the continent of Europe.

In our own country a valuable enlargement of our acquaintance with the relations of the primordial zone to the overlying members of the Silurian rocks has been made through the personal examination of Mr. Salter, aided by the independent discoveries of organic remains by MM. Homfray and Ashe, of Tremadoc.

It has thus been ascertained that the lower member only of the deposit, which has been hitherto merged under the name of Lingula-flags, can be considered the equivalent of the primordial zone of Bohemia. In North Wales that zone has hitherto been mainly characterized by Lingula and the crustaceous Olenus and Paradoxides. Certain additions having been made to these fossils, Mr. Salter finds that of the whole there are five genera peculiar to the lower zone, and seven which pass upwards from it into the next overlying band or the Tremadoc slate. But the over lying Tremadoc slate, hitherto also grouped with the Lingula-flags, is, through its numerous fossils (many of them of recent discovery), demonstrated to constitate a true lower member of the Llandeilo formation. For, among the trilobites, the wellknown Llandeilo forms of Asaphus and Ogygia range upwards from the very base of these slates. Again, seven or eight other genera of trilobites, which appear here for the first time, are associated with genera of mollusks and encrinites which have lived through the whole Silurian series. Such, for example, are the genera Calymene, Illanus, among crustaceans; the Lingula, Orthis, Bellerophon, and Conularia among mollusks, together with encrinites, corals, and that telling Silurian zoophyte, the Graptolite. By this proof of the community of fossil types, as well as by a clear lithological passage of the beds, these Tremadoc slates are thus shown to be indissolubly connected with the Llandeilo and other Silurian formations above them; whilst, although they also pass down conformably into the zone primordiale, the latter is characterized by the linguloid shells (Lingwella, Salter) and by the genera Olemus, Parado.rides, and Dikelocephalus, which most charseterize it in Britain as in other regions.

extensive conglomerates and marls of triassic age, and these by deposits charged with fessils of the Lias.

*The Silurian classification was proposed by me in 1835, and in the following year (1836) Dr. Emmons suggested that his black shale rocks, which he called Taconic, were older than any I described.

Nor are the writings of the Professors W. B. and H. D. Rogers in unison with the opinions of the authors here cited.

In the last edition of 'Siluria' the distinction was drawn between the lower and upper Lingula-flags, but the fauna of the latter is now much enlarged.

I take this opportunity, however, of reiterating the opinion I have expressed in my work 'Siluria,' that to whatever extent the primordial zone of Barrande be distinguished by peculiar fossils in any given tract from the prevalent Lower Silurian types, there exists no valid ground for differing from Barrande, De Verneuil, Logan, James Hall, and others, by separating this rudimentary fauna from that of the great Silurian series of life of which stratigraphically it constitutes the conformable base. And if in Europe but few genera be yet found which are common to this lower zone and the Llandeilo formation (though the Agnostus and Orthis are common to it and all the Silurian strata), we may not unreasonably attribute the circumstance to the fact that the primordial zone of no one country contains more than a very limited number of distinct forms. May we not, therefore, infer that in the sequel other fossil links, similar to those which are now known to connect the Lower and Upper Silurian series-which I myself at one time supposed to be sharply separated by their organic remains-will be brought to light, and will then zoologically connect the primordial zone with the overlying strata into which it graduates? Let us recollect that a few years only have elapsed since M. de Verneuil was criticised for inserting, in his Table of the Paleozoic Fauna of North America, a number of species as being common to the Lower and Upper Silurian. But now the view of the eminent French Academician has been completely sus tained by the discovery in the strata of Anticosti, as worked out by Mr. Billings under the direction of Sir W. Logan, of a group of fossils intermediate in character between those of the Hudson River and Clinton formations, or, in other words, be tween Lower and Upper Silurian rocks. In like manner, a similar interlacing seems already to have been found in North America between the Quebec group, with its primordial fossils, and the Trenton deposits, which are, as is well known, of the Llandeilo age.

I have thus spoken out upon the fitness of adhering to the classifications decided upon by Sir Henry De la Beche and his associates long before I had any relation to the Geological Survey, and which places the whole of the Lingula-flags of Wales as the natural base of the Silurian rocks. For English geologists should remember that this arrangement is not merely the issue of the view I have long maintained, but is also the matured opinion of those geologists in foreign countries and in our colonies who have not only zealously elaborated the necessary details, but who have also had the opportunities of making the widest comparisons.

On the continent of Europe an interesting addition has been made to our ac quaintance with the fauna of one of the older beds of the Lower Silurian rocks, or the Obolus greensand of St. Petersburg*, by our eminent associate, Ehrenberg. He has described and figured† four genera and ten species of microscopic Pteropods, one of which he names Panderella Silurica; the generic name being in honour of the distinguished Russian palæontologist, Pander, who collected them. It is well to remark, that as the very grains of this Lower Silurian greensand seem to be in great part made up of these minute organisms, so we recognize, in one of the oldest strata in which animal life has been detected, organisms of the same nature, and not less abundant than those which constitute the deep sea-bottoms of the existing Mediterranean and other seas.

Before I quit the consideration of the older paleozoic rocks, I must remind you that it is through the discovery, by Mr. C. Peach, of certain fossils of Lower Silurian age in the limestones of Sutherland, combined with the order of the strata, observed in the year 1827 by Professor Sedgwick and myself, that the true age of the largest and overlying masses of the crystalline rocks of the Highlands has been fixed. The fossils of the Sutherland limestone are not indeed strictly those of the Lower Silurian of England and Wales, but are analogous to those of the calciferous sand-rock of North America. The Maclurea is indeed known in the Silurian limestone of the south of Scotland; but the Ophileta and other forms are not found until we reach the horizon of North America. Now, these fossils refer the zone of the Highland limestone and associated quartz-rocks to that portion of the Lower Silurian which forms the natural base of the Trenton series of North America, or the lower part of the Llandeilo formation of Britain. The intermediate formation

* See 'Russia and the Ural Mountains.'

† Monats-Bericht d. König. Akad, der Wiss. Berlin, 18 April, 1861.