244 OUR HOUSEHOLD INSECTS 



proportion to the length of time which the gnat has 

 occupied in actually drawing blood," and argues, perhaps 

 somewhat inconsequently, that this indicates the con- 

 stant outpouring of some sort of poisonous fluid during 

 the blood-sucking process. But notwithstanding this, 

 he was unable to detect any channel for the conveyance 

 of poison into the wound. And, moreover, it is difficult 

 to conceive of a double flow of liquid poison downwards 

 and blood upwards as taking place simultaneously 

 within the narrow compass of the proboscis of a gnat 

 or mosquito. Or, again, if the movements were not 

 simultaneous, but a downflow of poison were followed 

 by an updraught of blood, it would seem that the 

 greater part of the poison would be sucked out of the 

 wound almost as soon as it was instilled, and that, 

 therefore, it could hardly exercise much influence upon 

 surrounding tissues. Humboldt, who was a firm believer 

 in the poisonous nature of the bite, considered this 

 sucking out of the poison to be the explanation of the 

 painlessness of some wounds. His experience was 

 almost the reverse of that of Mr. Dimmock, as detailed 

 above. He affirmed that if the insect were allowed to 

 suck to satiety, no swelling took place, and no pain was 

 left behind, and considered that when pain was produced 

 it resulted from the hasty interruption of the process of 

 sucking, since then the last infused poison would not be 

 able to be withdrawn. He experimented with one of 

 the most virulent species, allowing it gently to settle on 

 the back of his hand, and reports of it : "I observed 

 that the pain, though violent in the beginning, diminishes 

 in proportion as the insect continues to suck, and ceases 

 altogether when it voluntarily flies away." The follow- 

 ing experiment, however, seems to throw some doubt on 

 the poison theory altogether. He says, "I wounded 



