Objections of M. de la lande. 149 
is, that fome few fpots differ from all the reft, or from the 
multitude, and are not like theie excavations in the fun. Such 
cafes or exceptions will not lurely warrant the conclufion, that 
no Ipot can be an excavation. This would be to reverie all the 
rules of a juft induction, by oppoiing to an irrefragable general 
argument, the force of one extremely limited and feeble. 
But notwithftanding thefe few inftances where the umbra is 
not found to change, when we confider how perfectly all fpots 
referable one another in their moft ftriking features, there na- 
turally arifes fome preiuraption for all under that defcription 
we have given partaking of one common nature ; and for this 
only difference in the phenomena depending upon fomething 
which does not neceflarily imply a complete generical dii- 
tinclion. 
It comes therefore to be inquired, how far fpots, which 
when near the middle of tire difk appear equal and fimilar in 
all tilings, may yet differ from one another confidered as exca- 
vations, or as poffefting the third dimenlion of depth, and how 
far the peculiar circumftances by which they may difagree can 
contribute to make fome ref ft this change of the umbra, when 
near the limb, much more than others. 
In order to this, fuppofe two fpots which occupy a fpace 
upon the fun correfponding to the equal arches GD (fig. x.) ; 
and let GM, DM, be drawn fo as to coincide with the plane 
of the excavation in fuch cafe. The breadth of the nucleus 
being commonly equal to that of the furrounding umbra, if 
the bafe MD of the triangle GDM conceived reftilineal be 
divided in L, fo as ML : LD :: MD : DG ; and if through 
L be drawn LS parallel to DG, then will DGSL be the lection 
of two fpots having this condition, and which as to fenfe 
would, when far away from the limb, be equal in all apparent 
