“In other words, Khalkis wasn’t told the colour of the changed tie. Was it mere accident, then, if he himself had changed from the scheduled green to the red one he later wore―was it mere accident that he picked a red tie from the rack? Yes, that’s possible―for remember that the cravats on the rack in the wardrobe were not arranged by colours―they were mingled in a confusion of colours. But how account for the fact that, whether he picked the red tie by accident or not, he knew―as his subsequent actions proved―that he had picked a red tie?”
Ellery ground his cigarette slowly against the bottom of an ashtray on the desk. “Gentlemen, there is only one way in which Khalkis could have known he was wearing a red tie. And that is―he could distinguish its colour visually―he could see!
“But he was blind, you say?
“And here is the crux of my first series of deductions. For, as Dr. Frost testified and Dr. Wardes corroborated, Georg Khalkis was afflicted by a peculiar type of blindness in which sight might return spontaneously at any time!
“What is the conclusion, then? That last Saturday morning, at least, Mr. Georg Khalkis was no more blind than you or I.”
Ellery smiled. “Questions arise at once. If he could suddenly see after an authentic period of blindness, why didn’t he excitedly inform his household―his sister, Sloane, Demmy, Joan Brett? Why didn’t he telephone his doctor―in fact, why didn’t he inform Dr. Wardes, the eye-specialist then visiting in his house? For only one possible psychological reason: he did not want it known that he could see again; it suited some purpose of his own to continue leading people to believe that he was still blind. What could this purpose have been?”
Ellery paused and drew a deep breath. Knox was leaning forward, his hard eyes unwavering; the others were stiff with attentiveness.
“Let’s leave it there for the moment,” said Ellery quietly, “and tackle the clue of the percolator and the tea-cups.
“Observe the superficial evidence. The tea-things found on the tabouret indicated clearly that three persons had drunk tea. Why doubt it? Three cups showed the usual signs of usage by their dried dregs and the ring-stains just below the rims inside; three dried tea-bags were in evidence and prodding them in fresh water elicited only a weak tea-solution, proving that these bags had actually been employed in making tea; three desiccated, squeezed pieces of lemon were there; and three silver spoons with a cloudy film, indicating use―you see, everything tended to show that three persons had drunk tea. Furthermore, this substantiated what we had already learned; for Khalkis had told Joan Brett on Friday night that he expected two visitors, the two visitors had been seen arriving and entering the study―and this made, with Khalkis himself, three people. Again―superficial corroboration.
“But―and it’s a leviathan “but”, gentlemen―” grinned Ellery, “how superficial the indications were was at once revealed when we looked into the percolator. What did we see there? A percolator, to put it tersely, with too much water. We set about proving our surmise that there was too much water. By draining the water from the percolator we discovered that it filled five cups―the fifth not quite full, to be sure, since we had previously drawn off a tiny sample of the staled water in a vial for later chemical analysis. Five cupfuls, then. Later, when we refilled the percolator with fresh water, we drained off exactly six cupfuls when the tap ran dry. This meant, then, a six-cup percolator―and the stale water had filled five cups. But how was this possible if three cupfuls had been used for tea by Khalkis and his two visitors, as all the superficial signs indicated? According to our test, only one cupful had been taken from the percolator, not three. Does this mean that only a third of a cup of water had been used for each of the three men? Impossible―there was a circular tea-stain around the inner rim of the cup, indicating that each cup had been full. Well, then, was it possible that three cupfuls actually had been drained from the percolator, but that later somebody had added water to the water already in the percolator to make up the difference of the two missing cupfuls? But no―an analysis of the stale water from the little vialful I had taken indicated by a simple chemical test that no fresh water was present in the percolator.
“There was only one conclusion: the water in the percolator was authentic but the evidences on the three cups were not. Some one had deliberately tampered with the tea-things―the cups, the spoons, the lemon―to make it appear that three people had drunk tea. Whoever tampered with the tea-things had made just one mistake―he had used the same cupful of water for each cup, instead of taking three separate cupfuls out of the percolator. But why go to all this trouble to make it appear that three people were there when it was accepted that three were there―from the two visitors and Khalkis’s own instructions? For only one possible reason―emphasis. But if three people were there, why emphasize what is established?
“Only because three people, strange as it seems, were not there.”
He fixed them with the feverish glittering eye of triumph. Someone―Ellery was amused to see that it was Sampson―sighed appreciatively. Pepper was profoundly absorbed in the discourse, and the Inspector was nodding his head sadly. James Knox began to rub his chin.
“You see,” continued Ellery in his sharpest lecture-voice, “if three people had been present and all had drunk tea, there would have been three cupfuls of water missing from the percolator. Suppose now, that all had not drunk―people sometimes refuse such mild refreshment in these days of American prohibition. Very well. What’s wrong in that? Why go through this tortuous rigmarole of making it appear that all had drunk? Again, only to substantiate the accepted belief, fostered by Khalkis himself, please note, that three were present in that study a week ago Friday night―the night Grimshaw was murdered.”
He went on rapidly. “We are therefore faced with this interesting problem: if three were not present, how many were? Well, there might have been more than three: four, five, six, and number of people might have slipped into that study without being seen after Joan Brett admitted the two visitors and went upstairs to tuck the bibulous Alan in his little bed. But, since the number cannot by any means at our disposal be fixed, the theory of more than three leads nowhere. On the other hand, if we examine the theory that there were fewer than three present, we find ourselves on a heated trail.
“It couldn’t have been one, for two were actually seen entering the study. We have shown that, whatever it was, it was not three. Then, according to the only alternative in the second theory―the theory of fewer than three―it must have been two.
“If two people were there, what are our difficulties? We know that Albert Grimshaw was one―he was seen and later identified by Miss Brett. Khalkis himself was, by all the laws of probability, the second of the two. If this is true, then, the man who accompanied Grimshaw into the house―the man “all bundled up”, as Miss Brett described him―must have been Khalkis! But is this possible?”
Ellery lit another cigarette. “It is possible, decidedly. One curious circumstance seems to bear it out. You will recall that when the two visitors entered the study, Miss Brett was not in a position to see into it; in fact, Grimshaw’s companion had shoved her out of the way, as if to prevent her from catching a glimpse of what was―or what was not―in the interior of the room. There may be many explanations for this action, but certainly its implication is in tune with the theory of Khalkis being the companion, for he naturally would not want Miss Brett to look into the study and notice that he was not there when he should have been there . . . . What else? Very well―what are the characteristics of Grimshaw’s companion? Physically he approximated Khalkis’s size and build. That’s one thing. For another, from the incident of Mrs. Simms’ precious puss, Tootsie, Grimshaw’s companion could see. For the cat, perfectly still, lay on a rug before the door and the bundled man checked himself with one foot in the air and then deliberately walked around it; if he were blind, he could not have avoided stepping on the tabby. This checks, too; for from the necktie deductions we have demonstrated that Khalkis was not blind the following morning, but was pretending to be―and we have every reason to postulate the theory that his sight may have returned to him at any time after a week ago Thursday, on the basis of the fact that the last time Di Wardes examined Khalkis’s eyes was on that day―the day before the incident of the two visitors.