THE ARCHEOLOGICAL HISTORY OF NEW YORK 



43 



valuable facts in the study of cultural stages, and enable us to trace 

 and perhaps identify the factors influencing cultural growth, trans- 

 formation or decadence. Lists are not difficult to tabulate if a com- 

 plete collection has been made, or if a good catalog record has been 

 kept. It is when -we begin to estimate the relative frequency of 

 material that errors are most likely to occur. This is because no 

 system has been generally agreed upon by archeologists and col- 

 lectors. 



In considering methods of comparison, we find that there is one V- 

 general form that may be used with advantage, for statistical informa- / 

 tion and for ascertaining percentages. 



By this we enumerate all classes of articles from a site or region, 

 and divide the total by the number of thousands or fraction of 

 thousands. Thus we obtain a factor for determining the num- 

 ber of given objects that occur in a thousand of all kinds. We 

 also do two other things: (a) provide the means of estimating the 

 probabilities of finding other similar objects; (b) provide the means 

 for describing the relative frequency of the article. Thus, if among 

 2500 artifacts found in a site there are 15 bone fishhooks, we may 

 divide 15 by 2.5 (the number of thousands) and as a result 

 determine that 6 fishhooks occur to every 1000 specimens found. 

 Thus we may state that .6 per cent of all articles collected are fish- 

 hooks. On the other hand, among the 2500 specimens there may be 

 250 hammerstones of all classes. This would give hammerstones a 

 frequency of 10 per cent. 



As an example of how to lay out a specific site frequency table 

 let us take a certain number of representative articles from the Rich- 

 mond Mills site, where about 2500 specimens are enumerated in 

 the census made by Mr Dewey. 



By this means we can name an article by its specific cultural fre^ 

 quency and state it as i or 10 per cent as the facts bring out, and 

 arbitrarv terms need not be used. 



