fUj.RGPffqP
7 U-7 esearcft
2 ndDevelopment
cautpe
TO Dr. S.R. Massey FROM T. Hirji/G.A. Few
REFTH/GAF/CAP/61 DATE 7th November, 1978
PROJECT 12.01.23
O1 WP ANALYSIS OF AGRINION AFI/4 AND THE
FLUE-CURED
PORTION OF STATE EXPRESS BLEND
INTRODUCTION
The object of this experiment was to identify the
difference
between Agrinion AFI/4 and the flue-cured portion
of the State Express
blend. For this purpose, strips being used for the
manufacture of
commercial State Upress were obtained from the
factory and filter
tipped cigarettes were manufactured in the
Development Hall. The
specifications of cuts per inch, weight, paper,
filter etc. were as
per the commercial product.
The cigarettes.produced were as follows:
Cigarette Code L86
Blend: State Express blend WITHOUT AF1/4.
Average p.d.: 11.6 C.W.G. Range: 10.3-13.0 C.W.G.
Average total weight: 1.074 g.
Cigarette Code L87
Blend: 100% AFIA.
Average p.d.: 11.5 C.W.G. Range: 10.0-11.9 C.W.G.
Average total weight: 1.074 g.
A sample of commercial State Express was also
collected from the
factory and this was coded L88. The idea of
introducing the third
cigarette was to try and identify the contribution
of AFI/4 to the
State Express. However, it must be remembered that
L86 and L87 were
manufactured in Group R. & D. Centre and L88 was
manufactured in the
factory. Thus any differences between L88 and L86
or L87 may be
attributed to the differences in the manufacturing
conditions and
processes adopted in the two sites.
EXPERIMENTAL PROCEDURE
In order to be able to study the differences
between the three
blends and the three products, it was decided to
carry out OIWP analysis
on S, and S2 smoke for each of the three
cigarettes L86, L87 and L88.
Two replications were carried out for each, giving
3 (codes)
x 2 (SI and S2) x 2 (replications) = 12 O/WP runs.
BATCo document for Province of BritiSh Columbia 1
November 1999


-2- fo"% -&=
Dr. S.R. Massey 7.11.78 HE1?.!gA_-z
COLM
All the analyses were carried out by the same
operator using
one machine. With 2 O/WP runs per day being
practicable, the smokings
and the analyses were carried out over a period of
6 days. The order
of analyses was randomised in such a way that one
replicate of each
was carried out in the morning and one in the
afternoon.
STATISTICAL ANALYSIS
The structure of the data collected in this
described by the following diagram:
5 f
TYPE
CODE L87 J_
A
REPLICATIONS 2, 1 2_ 1
L86
A
For each type (Sl and S2) one-way analysis of
variance was carried
out using raw data for each of the peaks. The
Student-Newman-Keul's
test (SNK) was carried out to test the differences
between (1,86, L87),
(1,86, L88) and(L87, L88). The data was analysed
using SPSS package
(sub-program ONEWAY, RANGES = SNK) and 95% level
of significance was
used for SNK test.
The inclusion of L88 in the above tests produced
some anomalies in
the results and at present it is not clear whether
the difference in the
manufacturing conditions would completely account
for the anomalies.
In order to avoid confusion, the results with
respect to the comparison
between L86 (100% flue-cured) and L87 (1002 AF1/4)
have been extracted
from the full analysis and are summarised below:
I L86 is significantly greater than L87 for both
S, and S2 in
the-following peaks:
Peak 8
10
11 Methyl Ethyl Ketone
15 2,3 Butandione
19
23
26
27 2,4 Pentadienal
29 Pyridine
33
34 2,6 Dimethyl Pyridine
44 Cyclo pent-2-en-l-one
47 3 ethyl pyridine
75 Nicotine
experiment is best
BATCo document for Province of British Columbia 1
November 1999


. -3- 0. i R
Dr. S.R. Massey 7.11.78 Ud"
calre
II For the following peaks, L86 is significantly
higher than
L87 for S2 deliveries, and for S, deliveries L86
is higher
but not significant:
Peak 22 Butanitrile
30 Cyclopentanone
31 2 Methyl Cyclopentanone
.40
43 2,3-Dimethyl Pyrazine
46 2 Methyl cylcopent-2-en-l-one
59
65 Acetophenone
III For the following peaks L86 is significantly
higher than
L87 for Sj; for S2 the direction of the difference
is the
same but the difference is not significant:
Peak 9 Acetone
20 2-Butenitrile
24 N-methyl pyrrole
39 3 or 4-methyl pyridine
52 Furfural
53 3-vinyl pyridine
56 Pyrrole
IV For S2, L87 is significantly higher than L86 in
the following
peaks; the direction of the difference is
maintained in SI:
Peak 71 1,2-dimethyl pyrrole-5-aldehyde
73
77 Phenylacetonitrile
V For peak 41 (2,5-Dimethyl Pyridine), L87 is
greater than L86
for S, but this difference is not significant for
S2.
Ta,fe 1 presents the mean peak areas for L86 and
L87 for the peaks
listed above.
BATCo document for Province of BritiSh Columbia I
November 1999