If we reduce these figures to elementary form and include the minor constituents which are frequently found in rocks, and which in this laboratory are often estimated, the two averages compare as follows:

As the old mean represents an attempt to measure the composition of the entire solid crust of the earth, and so includes an allowance for the carbon in the limestones, the two columns are not strictly compa rable. They are, however, corroborative of each other, and show that within reasonable limits the statistical method is applicable to the problem under consideration. For the arguments upon which the discussion is based the original paper should be consulted. The distribution of the rarer elements has also been elaborately discussed by Vogt,' and their percentages may be regarded as small corrections to be applied to the table at some future time.

By a similar statistical process I have attempted to ascertain something with regard to the relative abundance of the more important rock-forming minerals. Nearly 500 analyses of igneous rocks were

1 Zeitsch. Prakt. Geologie, 1898, pp. 225, 314, 377, 413; and 1899, pp. 10, 274.

discussed, and the subjoined percentages, which are probably nothing more than rough approximations to the truth, were obtained:

The less frequent minerals make up the remaining 6 per cent. The computation, although it is by no means conclusive, is perhaps a little more satisfactory than any similar estimate which has so far been made.

For computing the average composition of the sedimentary rocks the existing analyses of individual samples are inadequate. They are too few and too incomplete to yield any conclusions of value. Attempts have been made to partly use the data, as, for example, by Joly; and it seems probable, therefore, that better material will not be without interest or scientific value.

Some five years ago, at the request of Mr. G. K. Gilbert, a series of composite analyses of sedimentary rocks was made in this laboratory. Many samples were mixed into one uniform sample, from which, by a single analysis, an average composition was determined. The material was selected and the samples were prepared by Mr. Gilbert, assisted by Mr. G. W. Stose, and the analyses were made by Dr. H. N. Stokes. The data obtained may be tabulated as follows:

A. Composite analysis of 27 Mesozoic and Cenozoic shales. Each individual shaie was taken in amount roughly proportional to the mass of the formation which it represented.

B. Composite analysis of 51 Paleozoic shales, weighted as in the former case.

C. General average of A and B, giving them, respectively, weights as 3 to 5. This average represents 78 rocks.

D. Composite analysis of 253 sandstones, about one gramme of each being taken in preparing the average sample.

E. Composite analysis of 371 sandstones used for building purposes. Equal weights taken.

F. Composite analysis of 345 limestones, equal weights being taken.

G. Composite analysis of 498 limestones used for building purposes, equal weights taken.

1 An estimate of the geological age of the earth: Sci. Trans. Royal Dublin Soc., vol. 7, p. 23, 1899.

These analyses may be used for a variety of purposes. For example, they can help in tracing the change from an average igneous rock to an average sediment. They suggest something as to the characteristic. features which distinguish a good building stone from other limestones and sandstones. They are applicable to the discussion of a variety of large theoretical problems, like that chosen by Professor Joly. These considerations alone justify their publication here.

In the former edition of this bulletin (No. 148) a chapter upon analytical methods, by Dr. W. F. Hillebrand, was included. For that chapter there has been a large separate demand, and for that reason it will be expanded into a work of greater detail and issued as a distinct bulletin. It will be noticed that to Dr. Hillebrand many of the best analyses in this compilation are due, and a full statement of methods, embodying his experience, will be of the utmost value.

During the preparation of this bulletin much assistance was rendered by the petrographers and geologists connected with the Survey, especially with reference to analyses hitherto unpublished. In each case credit has been given for the data thus added. Twenty-eight analyses Bull. 168-2

of rocks from Montana, executed by or under the direction of Prof. L. V. Pirsson, of Yale University, and having been made in connection with regular Survey work, are included in the tabulations. With this exception all of the analyses given were made in the Survey laboratories. To those executed in the laboratory at Washington "record numbers" are attached, which serve to identify them on the record books of the Division of Chemistry. Of the abbreviations used for bibliographic reference only three need explanation, and they refer to the official publications of the Survey. "Ann." for Annual Report, "Mon." for Monograph, and "Bull." for Bulletin are the three in question. The others relate to well-known journals, and are familiar to all geologists. The letters P. R. C., following the description of a rock, refer to the Petrographic Reference Collection of the Survey, and are followed by the number assigned to the rock in that series.

ANALYSES.

IGNEOUS AND CRYSTALLINE ROCKS.

MAINE.

1. ROCKS FROM AROOSTOOK COUNTY.

Described by H. E. Gregory in Bull. 165. Analyses by W. F. Hillebrand, record No. 1795.

A. Quartz-trachyte (bostonite), Quoggy Joe Mountain, Presque Isle Township. Contains quartz, orthoclase, albite, and magnetite, with siderite, kaolin, and chlorite.

B. Teschenite, Mapleton Township. In dikes cutting shales. Contains andesine, augite, biotite, apatite, analcite, and magnetite.

C. Andesite, Edmunds Hill, Chapman Township. Contains labradorite, orthoclase, pyroxene, apatite, and magnetite.

D. Calciferous sandstone, New Sweden Township. Contains calcite, alkali, feldspar, quartz, magnetite, muscovite, and siderite. Included here because studied as one of the group.

Traces of lithia present in all. F and Cl not looked for.