He began with the waiter at Charenton. Very skilfully he approached the subject, and, painting a moving picture of an innocent man falsely accused of murder, he gradually enlisted the man’s sympathy. Then he appealed to his cupidity, promising him a liberal reward for information that would save his client, and finally he soothed his fears by promising that in no case should any statement he might make get him into trouble. The waiter, who seemed a quiet, honest man, was perfectly open, and readily replied to all La Touche’s questions, but except on one point he stoutly adhered to his previous statement to Lefarge. M. Boirac--whom he identified unhesitatingly from a photograph--had lunched in the café about 1.30, and had then telephoned to two separate places--he had heard the two numbers asked for. As before, he made the reservation that he was not certain of the day of the week, his impression having been that it was Monday and not Tuesday, but he stated that in this he might easily be mistaken. There was no shaking his evidence, and La Touche was strongly of the opinion that the man was speaking the truth. But as well as repeating his statement to Lefarge, the waiter added one item of information that seemed important. Asked if he could not recall either of the numbers demanded, he now said he recollected the last two figures of one of them. They were 45. They caught his attention because they were the café’s own telephone number--Charenton 45. He could not recall either the previous figures of the number nor yet the division. He had intended to tell this to Lefarge, but being somewhat upset by the detective’s call, the point had slipped his memory, and it was only when thinking the matter over afterwards it had occurred to him. For La Touche to look up the telephone directory was the work of a few seconds. The number of Boirac’s house in the Avenue de l’Alma did not suit, but when he looked up the Pump Construction Office he found it was Nord 745. Here was fresh confirmation. It was obvious the waiter could not have invented his tale, and La Touche left utterly convinced that Boirac had indeed lunched at the café and sent the messages. As he was returning to the city it occurred to him that perhaps the waiter’s impression was really correct and that Boirac had been in the café on Monday afternoon instead of Tuesday. How was this point to be ascertained? He recollected how Lefarge had settled it. He had interviewed the persons to whom Boirac had spoken, the butler and the head clerk, and both were certain of that date. La Touche decided he must follow Lefarge’s example.