evaporated in a moderate sized capsule over a gas lamp, replenishing it as the water evaporated, till a sufficient quantity had been introduced. It was then evaporated to perfect dryness, finishing the operation in a water bath, and then transferred into a crucible by a platinum spatula, having moistened it barely sufficiently to allow it to be removed without flying. What then yet adhered to the capsule was loosened by adding more distilled water, and rubbing it with the spatula or a feather till it became perfectly loose, then washing it down into a corner of the capsule, again evaporating the water, and transferring the residue; repeating this same operation till the whole residue was effectually removed. It was then heated considerably above the temperature of boiling water. This method is, however, objectionable, for with the utmost care it is impossible to get all the residue removed, and more especially the silica. This same objection applies still more to the method of Prof. SILLIMAN, of evaporating the water in a number of glass flasks, transferring their contents, as they became concentrated, into one, and from this into a small counterpoised capsule, for some precipitate is always deposited during the concentration. The second estimation of 4.421 grains to the gallon was obtained by a separate experiment, performed by a single evaporation in a small capsule, but having only a few hundred grains left of the same lot of water from the analysis; this circumstance renders also this estimation less certain than it otherwise would have been, though the balance employed was of a superior delicacy.

6. The difference of the two analyses in the total amount of solid residue of one (wine) gallon of the water, is 1.08 grains and 1.42 grains. As I heated the residue to a higher temperature, the actual difference is less than this amount, and seems to be due mostly to the amount of organic matter, which I have estimated at only 0.036 grain, while Prof. SILLIMAN's estimation, deducting the carbonic acid, equivalent to all the magnesia, 0.207 (see §7), gives it 1.03 grains, or a difference in organic matter of about 0.99 grain. I am under the impression that my estimation is too low, as I, on the whole from want of time, did not pay as much attention to the organic matter as it deserved, and I have often observed that the organic matter in the Schuylkill water at certain seasons undergoes a change or kind of fermentation by standing in an open vessel, indicated by a deposit of a green film on the bottom of the vessel. As the water analysed by me was filtered, the above difference would be somewhat reduced, if Prof. SILLIMAN did not first filter his, as small animalcules and other very fine organic matters often remain suspended in the

water without settling. Whether the remaining difference must be explained by an actual difference of the amounts of organic matter in the water, at the two different periods, I leave undecided.

7. The difference of the two analyses in the amount of inorganic ingredients, taken as a whole, varies less than the whole residue, including the organic matter. By direct estimation, Prof. SILLIMAN found the fixed ingredients (see Table II) to be 4.26. By calculation from my analysis, for I did not ignite the residue before analyzing it (deducting from the inorganic ingredients the carbonic acid 0.250 grain, of the carbonate of magnesia 0.484 grain), the fixed inorganic ingredients amount to 3.794, making a difference of 0.466 grains to the gallon. Calculated from the same data, in Prof. SILLIMAN's analysis, the difference is still greater, but as this is evidently intended to represent and make up the amount of 4.26 grains obtained by direct weighing of the residue left after ignition (see his Report page xvii), an oversight has been committed in representing it to contain carbonate of magnesia, which, by a red heat, loses its carbonic acid. An error also exists in the summing up of the ingredients (sce Table I, or his Report, page xx), for instead of 4.26, they will be found, by addition, to make only 4.16. This difference (0.1 grain), together with the equivalent carbonic acid (0.207) belonging to the carbonate of magnesia (CO2=0.1815), the chloride of magnesium (CO2=0.0044), and sulphate of magnesia (CO2=0.0207), (for these two latter can certainly not exist at a red heat, together with carbonate of soda), leaves 0.307 grain to be yet accounted for in the analysis.

8. This difference of 0.307 grain, distributed on the different ingredients in proportion to their relative amounts, would raise the amount of carbonate of lime, in Prof. SILLIMAN's analysis to 2.010 grains, and that of the carbonate of magnesia (substituting also instead of the chloride of magnesium and the sulphate of magnesia, their equivalent of carbonate), to 0.521 grain, thus reducing the difference in the two analyses, for the carbonate of lime, to 0.18 grain, and for the carbonate of magnesia, to 0.04 grain, rendering the agreement in the total amount of these two main and most important ingredients, constituting nearly two-thirds of the whole solid residue, complete to within 0.14 grain, to the gallon (see also § 10).

9. The amount of alumina and oxide of iron has not been determined by Prof. SILLIMAN. Their amount is, at all events, but inconsiderable, but some importance attaches itself to it, as, if any phosphoric acid be contained in the water, it must be found in this precipitate. The reason why I omitted to examine it for this sub

R

stance was, that I thought its amount too small to give satisfactory results.

10. In the silica, the apparent difference in the two analyses is considerable. From the method employed, and subsequent experience from other waters, in some of which I have found as much as 50 per cent. of silica, I am convinced that my estimation, if anything, is too low. With all possible care, some of the silica will attach itself so firmly to the vessel during the evaporation, that it becomes. perfectly impossible to remove it completely by rubbing with water, or even with acids. This same cause would act still more to diminish its amount in Prof. SILLIMAN's analysis, where the portion precipitated, during the evaporation in the glass flasks, could not be removed by any mechanical means. That the actual amount of silica in the water analyzed by Prof. S. must have been greater, can easily be proved by his own experiments; for if we, from Table I, add together the carbonate of lime, the magnesia and the silica, they ought, together, to make up all the fixed insoluble ingredients; but in this way we only obtain 2.145 grains, while his direct estimation of the same (see Table III, and his Report, page xvii) yielded 3.69 grains, a difference of 1.54 grains. Allowing a fair proportion of this to be due to the too small estimation of the carbonate of lime and the carbonate of magnesia (see §8), it leaves yet a difference for the silex in the two experiments, which is too great to be supposed to have been committed in that one ingredient, and shows that part of the difference must be due to the separate estimation of the whole insoluble ingredients. The amount of insoluble fixed ingredients calculated in the same manner from my analysis gives it 2.9 grains. The average of the two above numbers of Prof. S. will give it 2.92 grains.

11. It has already been seen that some great mistake must exist in Prof. HORSFORD's estimations of the amounts of solid ingredients in the different waters, making the Schuylkill water contain 35 grains of solid residue to the gallon. From the foregoing, it appears that his estimation of the relative proportions of fixed and insoluble matter is equally erroneous. The main ingredients of the fixed residue of the Schuylkill water are formed of carbonates of lime and magnesia, with some silica, constituting the insoluble and greater part of it (according to the above analysis, at least more than 50 per cent.). How Prof. HORSFORD can make it to be only 12 per cent., (0.0239 out of 0.1975, see the Boston Water Commissioners' Report to the City Council, page 33) is utterly inconceivable, more especially as it is evident that he did not heat the residues to the expulsion of the car

bonic acid from the lime, or that he means, by "insoluble after ignition," insoluble in acids; for he states expressly, that these amounts, "insoluble after ignition," "are, for the most part, carbonate of lime."

12. The amount of chlorohydric acid, and its equivalent of chloride of sodium, is almost identical in both analyses, the latter varying only 0.006 grain to the gallon.

13. The difference in the amount of sulphuric acid, though small by itself, is nevertheless considerable, when compared with its quantity. As estimated by me, from a portion of the residue (the same that was used for the estimation of the chlorohydric acid), it amounts to 0.302 grain per gallon. Prof. SILLIMAN estimated its amount by precipitation of the water during boiling, without previous concentration, to be 0.0245 grain per gallon, making a difference of 0.277 grain per gallon, and being less than one-tenth of the amount found by me. Calculated from the analysis in Table I, the amount is a little different (0.0376). It is therefore uncertain whether Prof. S. also estimated it from a part of the solid residue, as he makes no mention of it. If not, I should consider his method likely to give a too small result, for as the gallon, or 58,330 grain, can contain only .03 of a grain, the water contains only 15400 of its weight of sulphuric acid, which is rather too great a dilution for forming and collecting, completely, a precipitate, of however insoluble a nature it may be.

14. The amount of alkali in my analysis is 0.448 grain to the gallon, in Prof. S.'s 1.041 grains to the gallon, making a difference of 0.593 grain to the gallon. Prof. SILLIMAN gives the whole as soda, not mentioning whether he tested it for potassa. Though I did not estimate its quantity, I convinced myself of its existence. Prof. S. also considers all the carbonate of soda to have been formed from the decomposition of crenates and nitrates, but I am confident that the water contains free alkaline carbonates, and that if it be evaporated to dryness in a water bath, and again treated with a small portion of water and filtered, the concentrated solution, or the residue obtained from it, will show an effervescence with an acid; an experiment which, though uncertain whether performed with the residue from the Schuylkill water, I have performed with similar waters from the neighbourhood of Philadelphia. The slight alkaline reaction which I found on litmus paper seemed also to indicate the existence of free alkaline carbonates.

15. In regard to the existence of nitrates in the Schuylkill water, it is an important observation of Prof. S., that by heating the residue, a

distinct deflagration took place; though this would undoubtedly tend strongly to infer their existence, still it would be well to confirm this by collecting or observing the nitrous oxide gas given off at the same time. To me no such deflagration occurred; perhaps I did not heat the residue to a sufficiently high temperature, though I heated it considerably above the boiling point of water, and I am under the impression that it began to turn slightly brown on the edges.

16. It will thus be seen that, as a whole, the two analyses confirm each other considerably, and I believe that the following results may be derived from them in regard to the composition of the Schuylkill water, and the differences in it at the two different periods, having three years between them.

The Schuylkill water, when settled clear, is a water of superior purity.

Its amount of chlorine and sulphuric acid is too small to produce any sensible reaction under most circumstances.

Its main characters it derives from its carbonates, more especially the earthy carbonates, which, if they were present in larger quantities, would give it the character of a hard water, and

In regard to a difference of composition, at the two different periods, that,

1. The amount of sulphuric acid varied less than 0.277 grain to the gallon (§13).

2. Chloride of sodium, less than 0.006 grain (§12).

3. The earthy carbonates less than 0.4, perhaps not more than 0.14 (§8).

4. The silica less than 0.32 grain (§10).

5. Organic matter less than 1 grain to the gallon (§6).

6. Total amount of solid residue less than 1.4 grains to the gallon.

Dr. GIBBON, Dr. HARE and Prof. JOHNSON, also entered into some explanations relative to the character and variability of the Schuylkill water.

After the conclusion of these remarks, Prof. HENRY made a further brief explanation of the objects of the Smithsonian Institution. After which, the Association adjourned, to meet this afternoon at 4 o'clock.

Monday, September 25, 4 P. M.

The Association met agreeably to adjournment. President, WM. C. REDFIELD, Esq., in the chair.

The Standing Committee made a report recommending the appoint