43 * Major-General Roy’s. ‘Account of the 
centers of the pipes terminating the bafe, reduced to the level 
of the loweft, or that at Hampton Poor-houle, in the tempe- 
rature of 63°, being that of the brafs fcale when the lengths 
of the deal rods were laid off. All this, however, fuppofes 
three things to be abfolutely certain : firft, that the expanfion 
of the rods has been accurately eftimated; fecondly, that no 
error has arifen from the butting of the rods again ft each 
other, in order to bring them into contact ; and, thirdly, that 
no miftake of any kind has been committed in the execution. 
When we come to give the true length of the bafe, as ulti- 
mately afcertained by means of the glafs rods, it will appear* 
that one or more of thefe three have a&ually taken) place; 
although it is moft probable, that only the two firft fources of 
error have contributed their (hare of the total difference be- 
tween the two refults. But the difcuftion of this point muff 
be deferred for the p relent ; and I fhall now finifti the fubjedt 
of the expanfion of the deal rods, by mentioning two other 
companions of them, which ferve to Ihew ftiil more obvioufly, 
how improper they are for very accurate meafurement! 
It has already been remarked, that the laft week of July 
was fo wet as to occalion a total fufpenlion of the operations on 
Hounflow- Heath. On the 26th of that month, at 8 h. A,M. 
the temperature being then 63°, the rods were compared with 
the ftandard, and found to exceed it, at a medium, one-fifteenth 
part of an inch. Now, if we fuppofe the whole bafe to have 
been meafured with the rods in that ftate, the difference would 
have amounted to more than yi feet, exclufive of what the 
ftandard itfelf might have altered from its original length. 
The other comparifon was made at Spring-Grove, in the 
beginning of September, after our operations on the heath had 
been finilhed, and the deal rods with their apparatus depofited 
under 
*7 
