"The manner in which it was sought to prove the correctness of this theory appeared to me at the time (and still does so) as most weak and fallacious, and certain experiments I made only strengthened that opinion. However, scientific people accepted it as proof.
"In making this experiment schoolboys were seated in rows at different measured distances from a map of Mars, which they were told to copy. The map showed all the well-known dark patches and markings, but no fine lines. About the places where some of those lines should have been, dots, curls, wisps, &c., were inserted at irregular distances, and not always exactly where the lines should have been shown. The inevitable result was that the boys who were too far away to see clearly saw these small markings as continuous straight lines, and so drew them. In the circumstances they could not do otherwise; for if sufficient marks were inserted nearly in alignment, they would necessarily produce the effect of lines.
"These drawings were then acclaimed as proving that the lines seen on Mars were only discrete markings viewed from beyond the distance of clear seeing, and that the network of lines seen and drawn by so many skilled and careful observers of Mars had no actual existence upon the planet. Thus all their work was completely discredited.
"Experiments like these could not possibly prove any such thing, because it would be easy to insert in a map various markings which, when viewed from a distance, would appear to form almost any design that one might choose to depict. Any desired effect might thus be obtained; and I have seen many pictures so formed in which the illusion was perfect. When viewed from a distance each appeared to be a picture of something entirely different from what was seen when it was viewed from a near standpoint.
"The linear illusion could not arise from a mere multiplicity of faint scattered markings, but all the more conspicuous markings must be in alignment. It seems impossible to imagine that so many hundreds of lines on Mars could thus fortuitously be formed by illusion, and every line be connected to some definite point at each end.
"To argue that because illusory lines can be formed as in these experiments proves that the Martian lines are also illusions is claiming far too much. For instance, if I drew what was actually a map of South Africa, and was so seen at close quarters, yet in consequence of the insertion of numerous small marks and shadings formed a portrait of Lord Blank when viewed from a distance, it would be very far indeed from proving that every map of South Africa was a portrait of the noble lord, or that his portraits were all maps of South Africa.
"Moreover, as I myself saw, some of the boys were so unskilled that they had not even drawn correctly the outlines of the dark patches about which there was no dispute.
"It is obvious that such erroneous and unreliable work as this could not be regarded as evidence upon which truly scientific argument could be founded for the purpose of deciding such a contentious question; yet mainly upon this very slender and unreliable evidence meetings of two of our leading astronomical associations endorsed the illusion theory, and for a long time it held the field.
"M. Flammarion made some similar experiments in Paris, and even inserted spaced dots along the sites of canal lines on the map put up as a copy, yet not one boy drew a canal. M. Flammarion evidently was rather too sparing with his dots and marks.
"A long series of experiments was carefully carried out by Professor Lowell and his colleagues, from which it was deduced that if in any line on Mars there was a gap of sixteen miles in length, our present telescopes would suffice to discover it. It is most improbable that in so many hundreds of lines, several of which are over two thousand miles in length, there would not be numerous gaps over sixteen miles long if the lines were made up of separate markings.