tf;?^ting the buknin(} quality ok tobacco. -SU 



rolled ([uite tit;litl\ . rirst on the t\)nii and thi'ii on the j^lass. At tlio 

 betiiiinini;- of tlu> pioe-es.s of lolliiio, the oxtroiuo outer ccrnci of the 

 base of the wrapi)er is attached to the overlappiiiu' j)()rtioii with a hit 

 of eio^ar paste, and at the end of the operation the tii) of the \vrai)i)er 

 is attached to the i-eceiver by the same means. A number of sumph>s 

 to be tested are thus wrapped on the forms and set aside until they 

 have dried out properly. The pluif in the end of the form is then 

 withdrawn and the rubber band causes the walls of the latter to col- 

 lapse, so that it can be easily withdrawn from the receiver. This 

 leaves the sample of wrapper securely attached to the glass tulie, 

 and in exactly the same form it would have on a cigar, 'i'he tube 

 carrying the sample to t)e tested is connected with the remainder of 

 the apparatus, shown in figure ; . the construction of which will be 

 understood without further exphnuition. The current of air is fur- 

 nished by means of an ordinary filter i)ump, and its rate can be con- 

 trolled with sullicient accuracy by measuring the How of water through 

 the pump. Th(^ end of the wrapper is ignited wilh a flat gas flame, 

 and the evenness of the buiMiand the portion consumed liefore it ceases 

 to o'low are carefullv noted. Our method of lecoi'ding the results is 

 to grade each sample on a scale of ten. tjoth with reference to the 

 evenness of the l)urn and the tire-holding ca})acity. Of course, stand- 

 ards in these tests are pundy arl)itrarv, as the results are only intended 

 to be comparative. I'nder the conditions l:iid down for the experi- 

 ment, wrappers having markedly good burning (pialities will bui'ii up 

 completely and evenly with only one lighting, and these are given a 

 grade of 10. 



For the purpose of comparing the results obtained b}' this method 

 with those given by the cigar test with reference to the evenness of the 

 burn, a number of leaves were selected from dith^rent types of wrap- 

 per tobacco. One half of each leaf was used f( r wrapping a cigar and 

 the second half was wrapped on the form for testing, as has just l)een 

 described. There was a decided lack of agreement in the results 

 obtained b}^ the two methods when onlj^ one t^'pe of filler was used in 

 making the cigars. It was found that frequently a wrapper that 

 graded onh^ 5 or 6 on a scale of 10 in what may be called the ",foi"m 

 test" would burn quite evenl}" on the cigar, whereas another wrapper 

 grading as high as 9 in this test would show an uneven burn on the 

 cigar. A good illustration of this pomt is found in a wrapper which 

 was scored 10, 9, 10, respectively, in three experiments Avith the form 

 test and gave a fire-holding capacity of 65 seconds b}^ the old method 

 of Nessler. On one type of filler this wrapper gave a very uneven 

 burn, but when smoked on a lighter filler the burn was perfectl3' satis- 

 factory. These results, then, seem to emphasize the fact that, although 

 the final judgment as to the burning qualities of a wrapper which has 

 shown up w^ell in the preliminary tests must be based on the smoking 



100— IV 



