ANTHROPOMETKIC INVESTIGATION IN GREAT BRITAIN AND IRELAND. 395 



The chief instruments used are Flower's craniometer, Martin's 

 traveller's aiithri)pometer, Pearson's headspanner. 



The records of the Anthropometric Committee of tlio Philosophical 

 Society extend over many years, and they were used about 15 or 20 

 years ago by Dr. Venn in a paper entitled ' Cambridge Anthropometry ' 

 ('Jour.' Anth. Tnst.) (see also Horton-Smith in 'Nature' about 1895). 

 Measurements for crania and on living persons are published in the 

 'Journal' of the Anthropological Institute and in the ' Journal ' of the 

 Camb. Phil. Society. (See a pamphlet on the 'Anatomical Museum at 

 Cambridge,' by W. L. H. Duckworth.) 



The observations used by Dr. Venn enabled him to compare, in respect 

 of physical development, men who in their final examinations took 

 1st, 2nd, and 3rd classes I'espectively. Otherwise, detection of racial 

 differences has been the chief object of research. 



The Camb. Phil. Soc. maintains an a.ssistant who has been instructed 

 how to make the observations : this assistant does not appear to have 

 attended any .special course of instruction on the principles of anthro- 

 pometry. 



The measurements ni.nde in connection with the Philosophical Society 

 have been but little utilised. At the University Laboratory, instruction 

 has been given to about 300 students in the last four years, and about 

 two per cent, of these have up to the present contributed to our know- 

 ledge of general anthropology. As the great bulk of these students are 

 still pursuing their medical studies, it is still early to pronounce on the 

 final results of the instruction given, for naturally medical students have 

 but little time for the pursuit of such researches as are the subject of 

 inquiry here. 



Galton, Francis, F.R.S., refers to the list of papers published by him 

 in the Royal Society Catalogue of Scientific Papers. He says : ' The 

 conclusions to which a great many and various experimental inquiries 

 have led one are a distrust of statistical results unless the data are 

 collected under conditions that (1) wltolly exclude bias and (2) the occa- 

 sional presence of large disturbing injluences. T do not think that laxity 

 in measurement matters much, so long as laxity does not lead to error in 

 otw. direction ; in fact, I know that a vast deal of effort is wasted in 

 minuteness of measurement. In speaking of bias, I mean, not only 

 personal (often unconscious) bias, but any influence that gives a one-sided 

 direction to the results. An instrument that has a sensible index error, 

 which is not allowed for, gives a bias to the results. In every new pro- 

 posed inquiry, great pains and much consideration should be given before 

 beginning it, to be sure that the plan is not vitiated by unsuspected 

 causes of error.' 



Mr. Galton considers that 400 cases, of whicli the two sets of 200 

 agree, are quite sufficient for any statistical investigation. The probable 

 error of the result should always be calculated and given. He higjdy 

 recommends the methods of the higher statistics as so greatly advanced 

 by Professor Karl Pearson. 



Garson, J. G., M.D., has described his method of measurement and 

 instruments in ' Notes and Queries on Anthropology,' published by the 

 Anthropological Institute. He uses callipers, steel tape, set squares, aiid 

 sliding rules of various kinds. He has designed a combination instrument 

 called the ' Traveller's Anthropometer ' by whicli all tlje usual measurer 

 ments car^ be made. 



