The spatial resolution of image data tends to constrain the horizontal leng
th scale of genetic hypotheses that are addressable by those data. No 'simp
le' formula exists when image resolution is sufficient to test a given geom
orphic process, which is dependent on what characteristics are diagnostic o
f the particular process. Genetic hypotheses should be formulated along the
lines of the "multiple working hypotheses" concept as described in a class
ic paper by Chamberlin [J, Geol, 5 (1897) 837]. An essential element of a v
iable working hypothesis is a clear indication of the characteristics predi
cted by, or a consequence of, the hypothesis. An untestable hypothesis is n
ot an effective working hypothesis. The history of the study of lunar sinuo
us rilles is outlined as an illustration of the influence of image resoluti
on and the formulation of genetic hypotheses on the subsequent advancement
of understanding of the problem. Sinuous rilles on Venus and Mars, and cont
roversial sinuous ridges on Mars are also reviewed. In the lunar case, the
three-order-of-magnitude improvement in spatial resolution provided by Luna
r Orbiter photographs over Earth-based telescopic photographs did nor resul
t in definitive examination and elimination of published hypotheses for the
formation of sinuous rilles. Topographic data obtained from cartographical
ly controlled Apollo orbital photographs, along with important observations
and samples obtained by the astronauts on the lunar surface, did test and
exclude several hypotheses. The formulation of a genetic hypothesis, includ
ing testable consequences of that hypothesis, is a greater determinant of i
ts ultimate utility to the scientific community than is the image resolutio
n available at any,given time. Published by Elsevier Science B.V.