is, that the surface water of lakes is purer than that taken from some depth. Nos. 3, 8, and 10, also 2, 6, and 7, also 5 and 9, offer illustrations of this remark. The rivers also gather impurities as they draw near their mouths, the water from high up the stream being purest. It will also be seen that the number and proportion of the ingredients vary in the same water. These results were not a little interesting to the writer and his excellent assistant, Mr. Hunt, when after our labors were closed, we were for the first time supplied with the names of the various sources from whence the waters on which we had labored were obtained.

The specific gravities of these waters were taken with a delicate balance, and with every precaution as regards temperature, &c. A glass bottle with a perforated stopper was used for the purpose. The results were in some cases anomalous, giving for some of the waters a density less than that of pure water. But these differences are so infinitesimal as to be easily accounted for by the gaseous matters which these waters contain. The actual differences from the density of pure water at 60°, afforded even by the most dense of the waters examined in this research, affect only the third place of decimals. Thus No. 11 has 87-811 parts of solid matter in a hundred thousand parts of water, or nearly one per centum, yet its gravity differs from pure recently boiled water distilled from glass by only. We should not look therefore for a very appreciable increase of gravity in a water containing only three parts of solid matter in a hundred thousand. Table III shows the specific gravity.

Number of Water,
Specific Gravity,

Number of Water,
Specific Gravity,

| 1·000541 | 1-000060 | 1-0000894 | 1-000016 | 1-000118 | 0-999842

| 1·000062 | 0-999924 | 1-000090 | 0-999967 | 0·001126

The water No. 11 (from a well in Boston) proved to be highly saline. One standard quarter gallon of the water was boiled with the previous addition of a known quantity of pure anhydrous carbonate of soda, which threw down, during the progress of the ebullition, all the earths and other bases present, except the soda, converting them of course into carbonates. Then by the well known methods of analysis, the various ingredients were separately determined, and the soda quantitatively. I give subjoined the actual quantities obtained for one gallon of the water, without attempting to reconstruct the order of arrangement which we may suppose they had in nature. We found

Both the silica and phosphate of alumina in the analysis of No. 11 were separately analyzed after their weight was recorded, to ascertain their entire purity and freedom from other matters. We also made a series of experiments to determine the action of these waters on metallic lead. The result of these trials was that every water, except No. 2, (Croton,)acted more or less on the lead. Those interested will find these results tabulated in the report before quoted.

In conclusion we may remark, that all natural waters may in a certain sense be properly called mineral waters, as they must each possess a specific and peculiar character dependent on the nature and amount of solid matters which they contain, and this must depend ultimately on the geological structure of the country where they are found. It is curious and instructive to see that even those waters which we consider the purest, contain, in a notable quantity, matters which are absolutely indispensable to satisfy the demands of the vegetable world, (and ultimately the animal also;) and when we remember the vast amount of evaporation from the expanded leaves of a full grown forest tree during a single summer day-can we any longer be at fault for a cause sufficient to account in the most satisfactory manner, for the various inorganic constituents of plants? It cannot be doubted that natural waters act a most important part in conveying into the upward current of organic life those inanimate elements, which, from their constant presence in plants, must be of primary importance, although we are at a loss to explain the mode of their action.*

Yale College Laboratory, July 15, 1846.

*The foregoing analyses were made by authority of the city of Boston, preparatory to selecting one of the sources to supply that metropolis. Long Pond (Nos. 5 and 9) was finally chosen.

ART. XXIII.-Description of a remarkable fossil Echinoderm, from the Limestone Formation of St. Louis, Missouri; by J. G. NORWOOD, M. D., and D. D. OWEN, M. D.

In the winter of 1844-5, during a visit to St. Louis, our search after organic remains in that vicinity, was rewarded by the discovery of a slab containing three specimens of a magnificent fossil belonging to the class Radiata, one of which is here represented.

A minute inspection of the specimens, induces us to believe that they are gigantic Echinoderms, closely allied to the Echinideans; especially to the Echinus. It is true that in some parts of their anatomy, they differ from Goldfuss's description of that di

vision of the order Echinodermata; yet, it is thought these aberrant characters are hardly sufficient to entitle them to be classed as a distinct group.

With the Echinideans or Echinides of Goldfuss, they agree in three essential characters, viz.: They are composed of ten fields (area); five broad (area majores), and five smaller (area ambulacrorum); and these are made up of little plates, disposed in Pores or holes run in vertical rows up and down the small field (area ambulacrorum). There is, moreover, every evidence that the os inferum and the anus were central, as in the genus Echinus.

rows.

In the following, more trivial characters, they differ from Goldfuss's description of this group.

The plates of which the area majores consist, are mostly sixinstead of five-angled, and are far more numerous than the elementary plates of which Goldfuss's genera are made up.

The plates of the area majores are arranged, not in single or double rows, but in many rows, varying from seven or eight at the widest part, to five or six at the top and bottom.

The plates constituting the area ambulacrorum are, doubtless, also more numerous, for, though rather obscurely marked, a close inspection shows that there are two kinds, viz., a double vertical row of elongated hexagonal plates, interlocking and connected laterally, on either side, with three rows of smaller, four-angled, irregularly rhomboidal plates, (very like the eschars on some spe

Fig. 2.

cies of Lepidodendron,) making in all, eight rows of plates, arranged as exhibited in fig. 2; while the corresponding area of Echinides, according to Goldfuss's description, are composed of but two rows. Each plate is perforated by two holes; so that each ambulacrum consists of eight double rows of pores.

No evidence of any kind has been obtained to show whether these bodies were pedunculated; but there is strong presumptive evidence that they were not. This is rendered probable, 1st, from their near approximation to the true Echinides; 2d, from their gigantic dimensions; 3d, because no stems have yet been found connected with any of the specimens; 4th, because pedunculated Echinoderms of such vast dimensions, would require a much larger stem than is exhibited by any portions of columns

hitherto discovered in this rock; 5th, all the Echinoderms at all approaching the size of this fossil are free bodies; 6th, the absence of a pelvis or circular rim for articulation to a stem; 7th, the presence of depressions in the position, and apparently for performing the offices of the anus and os inferum.

To sum up, then, with a connected description of the fossil, as far as the specimens in our possession will enable us to do so:

Body ovoid or nearly spherical, free; os inferum central; anus central and above; area ten, five large (area majores), five small (area ambulacrorum); plates of the area majores mostly 6-sided, in many rows; those of the area ambulacrorum of two kindsone set are elongated hexagons, disposed in double vertical rows in the centre of the area, which is elevated into a prominent ridge, along the summit of which the interlocking serrated suture of this double row is situated; the other set smaller, irregularly rhomboidal, and running in oblique rows on either side of the former. Each plate of the area ambulacrorum is pierced by two holes; these are central in the rhomboidal plates, but in the hexagonal plates, are situated near the angle furthest from the before mentioned central suture. Each ambulacrum is thus constituted of eight alternating double rows of pores.

Figure 3, is an outline of a restored representation of the fossil reduced. We propose as an appropriate name for it, Melonites multipora, on account of its resemblance, in general outline, to some species of melon, and the great number of pores in the ambulacrum.

This fossil is known amongst the quarrymen, as the "coltsfoot." By reference to

Fig. 3.

fig. 1, it will be seen that it bears considerable resemblance to the impress of the frog of a horse's foot. This is worthy of note, since it confirms, in a most remarkable manner, an assertion made in a former number of this Journal, (1st Ser. vol. xliii, No. 1, p. 17,) that those unacquainted with the science of geology, frequently mistake for fossil footprints, what are, in fact, moulds of shells, or merely casual appearances. It also warns the geologist how cautious he should be in investigating reported footmarks in solid rock.

The discovery of this fossil is peculiarly interesting, not only on account of its gigantic dimensions, measuring, as it lies on the