410 MR. c. o'neill's experiments and 



threads. These experiments, being repeated many times, 

 gave him for 



Louisiana cotton 2^ grammes. 



Jumel „ 4-^ „ 



Georgian „ long staple 3| „ 



Georgian ,, short staple 4-^ „ 



It will be found that these results are considerably below 

 the truth ; for this method, which seems ingenious at first 

 sight, is fundamentally fallacious, because a spun yarn or 

 thread made from short hairs like cotton is not like a 

 bundle of wires, the whole lengths of which may coincide, 

 but, on the contrary, is composed of lengths which theo- 

 retically never coincide at all, and practically very few do 

 coincide in their whole lengths : but if a very short grip 

 was taken, and if the cotton hairs were uniform in strength 

 throughout their length, the results might come tolerably 

 near ; but cotton has not a uniform strength in all parts, 

 and a number of experiments show that this method can 

 never give more than about one-third of the mean strengths 

 of the hairs in a yarn or thread, and generally very much 

 less. 



After many tedious trials, I found out a very simple 

 and ef&cacious way of securing the ends of the cotton hairs. 

 I take some tough thin paper, like letter-copying paper or 

 good tissue paper, and put a thin layer of gum on one side, 

 and, when dry, cut it into slips about half an inch broad ; 

 I bend some of the thinnest iron wire into triangles having 

 sides of about half an inch, and cutting the gummed paper 

 into lengths of about an inch, double a length of it over 

 one side of the triangle, with the gummed side inwards, 

 and just moisten it at the bend for a short distance with a 

 cameFs-hair pencil, press together, and so secure it on the 

 triangle, leaving, however, the ends free, and forming an 

 open flap of a quarter of an inch or more in length. The 

 cotton hair about to be tested is brought by the forceps 



