To get these samples was asking a great deal. Even when packed in special boxes glass bulbs are somewhat fragile things for trench life, and the wooden boxes made excellent kindling wood, which was always being sought for. The result is that when the cloud does come along the vacuum bulbs are often conspicuous only by their absence. Even if they are kept whole it is asking rather a lot of a man to take an accurate scientific sample during the excitement of a gas attack which is accompanied by a bombardment by explosive shells and gas shells. For a long time none of the bulbs found their way back to the field laboratory. Eventually one did come, carefully packed in shavings and wadding. I happened to be present when it was brought in, and there was a good deal of excitement at the little prodigal’s return. The bulb was taken out, but under it was found a leaf from a field-service note book, on which was written: “Danger. This bulb was found in a hedge. It seems to have been dropped from an aeroplane and probably contains cholera germs. Fortunately it has not been broken.” The “surprise effect,” which was mentioned above as being the third fresh feature of this new-era gas-cloud attack, took the form of making the attack in the dark and at a time when men were least prepared--that is, just before the morning “Stand to,” the hour before dawn, when all troops in the trenches stand to arms. By making the attack at night, or at any rate in the dark, the boche achieved two objects: First of all, there were better wind conditions for an attack, because the night winds tend to flow down toward the earth and keep the gas cloud low-lying and thick, whereas in the day the sun warms the ground and produces so many upward currents of air that the cloud gets lifted up and dissipated; in the second place it was impossible to see the cloud when it was first liberated, and this reduced the means of detecting the attack to only two--the hissing noise of the gas escaping from the cylinders and the smell of the advanced parts of the cloud. Later on it was known that the best hours for all gas attacks, both cloud and shell bombardment, are in the night; and as a matter of fact practically all gas warfare is now carried out at night, but at that time the significance of this was not grasped, and many of our casualties were due to lack of preparedness, numbers of men being caught “on the hop” and overwhelmed.