VOL. XVI.] PHILOSOPHICAL TRANSACTIOBM. 287 



contrivance, and accommodate it more to the literal way of writing with us : 

 and as our way is now possibly brought to the greatest perfection for exactness 

 and expedition, so doubtless must their way of printing, since they can engrave 

 their stamps for a sheet as soon as one of our compositors can set and correct a 

 sheet of our literal character ; and since one man alone will print off 1500 sheets 

 in one day. And though it is generally believed to be much the same with our 

 wooden cuts for printing, yet from some observations, hereafter to be noticed, 

 it seems to be quite different. 



By a Chinese manuscript, having fou^d that the pronunciation had no affinity 

 with the strokes of the character, I conceived it was either a numeral character 

 consisting of numbers, or else a real character, but not a literal, unless it were 

 a literal character of some other language than that by which it was pronounced, 

 whose pronunciation is lost, though the significancy be retained. Since that 

 time I procured from China a dictionary of the court language ; but this whole 

 book consisted only of the Chinese characters, without any inter{)retation or 

 pronunciation ; however, by the help of the pictures of that and a Chinese 

 almanack, I soon found out their characters for numbers, and their way of nume- 

 ration, with the figure and use of their abacus or counting board, for performing 

 the operations of arithmetic ; which I find pretty nearly to agree with that of 

 the ancient Romans, save only, that instead of pins and sliding grooves of the 

 Romans, the Chinese abacus has strings or wires, and beads to slide upon them; 

 and that instead of 4 pins for digits or units, the Chinese has 5 beads ; so that 

 it seems the Chinese abacus was designed for a duodecimal progression, whereas 

 that of the Romans was designed for the decimal. One thing is remarkable in 

 the Chinese, that the places in the abacus lie horizontal, and the first place is 

 next the left hand, which probably was also the first in their old way of reading. 

 Now as the Chinese and Roman abacus do much agree, save only that they 

 proceed contrary ways, so also does their way of expressing numbers by letters 

 or marks, one stroke or line signifying 1 ; two lines, 2 ; three lines, 3 ; a cross, 

 10; two crosses, 20; three crosses, 30; and so on to 100, which they express 

 by a square mark ; and a cross with a stroke added for lOOO. And though the 

 characters are not all the same, yet the order and method of one agrees very 

 nearly with that of the other ; especially if it be allowed, that the primitive way 

 of writing and reading with the Chinese was horizontal, like the European 

 way. 



Having thus discovered their characters for numbers, and their way of nu- 

 meration, I next applied myself to understand something of their language 

 and character. But upon perusing all the accounts I could meet with in books, 

 I found very little satisfaction as to my principal inquiry, which was, first con- 



