—_—- 
1903.] LOWELL—THE CARTOUCHES OF MARS. 309 
just passed, 375 complete ones in all. They date from January 21 
to July 26, inclusive, and were divided by months as follows : 
ANALY) Sete e causa rae crernc tee hares ee 18 
HC DEUALY eee nett eee ete ed ee La Tate se mE 48 
MMT A Lay SS yee pay He er MG Cay, BION se Rien eee 49 
BNE ats erst cpen one Ses pA ON SYNE BeGNe tee Saas rater ea 70 + I unfinished. 
IVAN Ss ch ewbyafee (car's wepskaroyetowices gels lcke-o jake miu ats ReMttee 63 
Jie SD aa eRe ho CAS cia Re ae eri ees 70 
“Fall UE Se Ne ee er sieleceietshap tala 57 -+ 1 unfinished. 
375 + 2 unfinished. 
Sketches of particular parts are not included in the list, as being 
unfit for comparison purposes. 
The principle I adopted in making the drawings was that of 
momentary representation. My object in each was not so much an 
exhaustive map as an instantaneous photograph. From ten to 
twenty minutes only was the time allotted to each. .In that period 
the shift of the longitudes is not enough substantially to change the 
degree of visibility of a marking and thus to make of the drawings a 
composite picture. 
Eighty-five canals were examined for presence or absence in 
these drawings. The average number of times a canal might have 
been seen, had it been sufficiently conspicuous, proved to be about 
one hundred. The number of times it actually was seen varied 
with the particular canal, some canals being but rarely detected, 
others being almost continuously visible. From the above it 
follows that eight thousand five hundred separate examinations for 
the visibility or non-visibility of the canals had to be made in all ; 
an undertaking of some length, but adding proportionately to the 
trustworthiness of the result. 
For getting the percentage visibility of a canal at any presen- 
tation it seemed on the whole best to consider all three of the above 
pair of zones together, or, in other words, the percentage of visi- 
bility within 60° of the central meridian, limited as above described 
toward the terminator. Any other pair of zones might have been 
used with equal correctness, but the greater number of determin- 
ations got from considering all three together commended itself for 
its increased accuracy. 
The percentages thus obtained’ proved sufficiently suggestive, 
even before any corrections had been applied. To give them, 
