Of course I heard only one side of the conversation, but as near as I could gather Kennedy was asking the inspector to obtain several samples of ink for him.  I had not heard the first part of the conversation, and was considerably surprised when Kennedy hung up the receiver and said: 

“Vandam had the prescription filled again early this morning, and it will soon be in the hands of O’Connor.  I hope I haven’t spoiled things by acting too soon, but I don’t want to run the risk of a double tragedy.”

“Well,” I said, “it is incomprehensible to me.  First I suspected suicide.  Then I suspected murder.  Now I almost suspect a murder and a suicide.  The fact is, I don’t know just what I suspect.  I’m like Dr. Hanson—­floored.  I wonder if Vandam would voluntarily take all the capsules at once in order to be with his wife?”

“One of them alone would be quite sufficient if the ‘ghost’ should take a notion, as I think it will, to walk in the daytime,” replied Craig enigmatically.  “I don’t want to run any chances, as I have said.  I may be wrong in my theory of the case, Walter, so let us not discuss this phase of it until I have gone a step farther and am sure of my ground.  O’Connor’s man will get the capsules before Vandam has a chance to take the first one, anyhow.  The ‘ghost’ had a purpose in that message, for O’Connor tells me that Vandam’s lawyer visited him yesterday and in all probability a new will is being made, perhaps has already been made.”

We breakfasted in silence and later rode down to the office of Dr. Hanson, who greeted us enthusiastically.

“I’ve solved it at last,” he cried, “and it’s easy.”

Kennedy looked gravely over the analysis which Dr. Hanson shoved into his hand, and seemed very much interested in the probable quantity of morphine that must have been taken to yield such an analysis.  The physician had a text-book open on his desk.

“Our old ideas of the infallible test of morphine poisoning are all exploded,” he said, excitedly beginning to read a passage he had marked in the book.

“’I have thought that inequality of the pupils, that is to say, where they are not symmetrically contracted, is proof that a case is not one of narcotism, or morphine poisoning.  But Professor Taylor has recorded a case of morphine poisoning in which the unsymmetrical contraction occurred.’

“There, now, until I happened to run across that in one of the authorities I had supposed the symmetrical contraction of the pupils of the eyes to be the distinguishing symptom of morphine poisoning Professor Kennedy, in my opinion we can, after all, make out our case as one of morphine poisoning.”

“Is that case in the book all you base your opinion on?” asked Craig with excessive politeness.

“Yes, sir,” replied the doctor reluctantly.

“Well,” said Kennedy quietly, “if you will investigate that case quoted from Professor Taylor, you will find that it has been proved that the patient had one glass eye.”