Maech 25, 1887.] 



8CIEJSCB. 



297 



Prussia and the empire together would form a 

 fairer basis for comparison with the United States 

 than would the empire alone, because the latter 

 leaves the civil administration almost entirely to the 

 individual states. The comparison with Prussia and 

 the empire together, however, would not be exact, 

 as in Prussia the nation assumes some functions 

 wljich are here left to the states ; but it is safe to say, 

 that, if we could compare with accuracy the expendi- 

 tures for like purposes in Prussia and the empire 

 together and in the United States, it would be found 

 that the proportions in each of war expenditures were 

 nearly the same; and of course, if we consider the 

 productive expenditures of the German states, the 

 percentage of war expenditures will be much smaller 

 than in this country. 



I do not mean to deny Mr. Atkinson's general 

 statement that a larger proportion of expenditures 

 goes for war purposes in Europe than in the United 

 States, nor to underestimate the other burdens which 

 a great standing army imposes, but merely to point 

 out, that, so far as state expenditure for war pur- 

 poses is concerned, the diiference between this and 

 other countries is not so great as we are apt to think, 

 and that in the case of Germany it is doubtful if 

 whatever difference there may be is in our favor. 



Mr. Atkinson also holds ' ' that the revenue of state 

 forests, mines, and other instrumentalities of sub- 

 sistence . . . constitute as true a tax upon the people 

 as if they had been assessed directly on their prop- 

 erty." 



That is a question that ought to be determined be- 

 fore we begin to make comijarisons. If we intend 

 to count profits from lands, mines, and railroads as 

 taxes in Europe, we must do so in this country. 



If the consumer is served equally well and cheaply 

 by a private and public producer, profits are no more 

 a tax in one case than in the other. It would be 

 difficult to convince any one that it makes no dif- 

 ference to the German tax-jDayer whether govern- 

 ments derive from the profits of railroads a revenue 

 sufficient to pay the interest on the public debts, as 

 is the case in the German states, or whether that 

 revenue comes from taxation, provided the railroads 

 are as well managed as they would be if government 

 did not control them. Heney B. Gaednee. 



Johns Hopkins univ., Baltimore, March 21. 



The characteristic curves of composition. 



With regard to Professor Mendenhall's novel paper 

 ■on ' The characteristic curves of composition,' in 

 your issue of March 11 (No. 214), which proposes to 

 Tepresent and compare the orthographical produc- 

 tions of writers by a statistical and graphical method, 

 it seems to me, that, interesting and instructive as 

 Are the results he has reached, they are confined to a 

 range of inquiry too narrow to bring into sufficient 

 relief the personal idiosyncrasies of individual 

 writers, and to a kind of enumeration in which per- 

 sonal peculiarities are too much marked by the par- 

 ticular language in which they write. 



That the characteristic curve is principally con- 

 trolled by the language in which the composition is 

 written, is evident from the comparatively small dif- 

 ference to be found between the various English 

 writers between whom comparison is made, as well 

 as from the marked departure from this general 

 «hape of the English curve to be seen in that of 

 •Oaesar's ' Commentaries.' The curve found for any 



other Latin author would presumably not differ from 

 this one more than the curves of various English 

 writers differ from each other. 



What the general shape of the characteristic curve 

 may be for any writer is determined, then, princi- 

 pally by the language in which he writes. 



It would be interesting to compare several lan- 

 guages with each other, so as to obtain approximately 

 the normal curve for each. An inflected language, 

 like Greek, Latin, or German, will, of necessity, 

 have its normal curve largely affected by the numer- 

 ous letters forming the terminations. Moreover, any 

 tendency toward the formation of compound words, 

 such as Pferdebahnwagon , or toward agglutination, 

 would also have its effect upon the shape of the 

 curve. Such a comparison would doubtless furnish 

 tests on which to build new ai'guments and comjjari- 

 sons respecting the vexed question of Teutonicity, 

 and the like. 



But to return to the point with which I began ; 

 viz., that there are other characteristics of writers 

 equally susceptible of treatment by the statistical 

 and graphical method, in which their personal pecul- 

 iarities differ more widely, and which are therefore 

 more characteristic than the habitual selection and 

 use of long or short words. For example : it seems 

 to me that the length of the sentences employed by 

 a writer is such a peculiarity, and one which, al- 

 though influenced somewhat by the particular lan- 

 guage in which he writes, is nevertheless an expres- 

 sion of his habits, feeling, taste, and individuality to 

 such an extent as to exhibit necessarily some charac- 

 teristics which would distinguish him in a marked 

 manner from other writers. 



The length of the adjective modifiers of substan- 

 tives seems also to be a particular well suited to 

 bring out individual characteristics by a similar 

 enumeration. In this category may be mentioned 

 also the length of the adverbial expressions; the 

 complexity of the verbs ; as well as the character of 

 the vocabulary as regards derivation from Anglo- 

 Saxon, French, Latin, Greek, etc. The list of fit 

 subjects of enumeration can be extended at will. 



It would seem probable that a discussion of the 

 results obtained by the simultaneous application of 

 several of these enumerations would, in any case of 

 disputed authorship, afford decisive tests such as 

 could not be obtained from any one of them singly ; 

 and by its help the person making the investigation 

 could exhibit to the public how weighty the evidence 

 may be on which his judgment is based. 



H. T. Eddy. 

 Cincinnati, March 14. 



Earthquake weather at sea. 



Your European exchanges have no doubt given j'ou 

 so full reports of the recent earthquake in this re- 

 gion, that it would be impossible for me to add any 

 thing that would interest j'ou or your readers. You 

 may be interested, however, to have somewhat in as 

 detail a report of earthquake weather at sea, such 

 was encountered by the steamship Gottardo on its 

 last trip from New York. 



We sailed from New York on the 19th of February, 

 and had disagreeable weather almost from the hour 

 we left Sandy Hook. On Tuesday, the 23d, began a 

 series of storms which kept by us almost constantly 

 until we sighted the African coast outside the Straits 

 of Gibraltar. The disturbance began about 4 p.m. 



