STUDIES ON THE FERMENTATION OF TOBACCO '

By JAMES JOHNSON
Horticulturist, Wisconsin Agricultural Experiment Station, and Agent, Division
of Tobacco and Plant Nutrition, Bureau of Plant Industry, United States
Department of Agriculture
INTRODUCTION
After curing, cigar-leaf types of tobacco in particular are allowed to
undergo one or more fairly definite periods of fermentation or *^ sweat-
ing." This process is characterized chiefly by an exchange of gases,
the generation of heat, and a modification of the flavor and aroma of
the leaf. The aging process in tobacco is not in all cases clearly dis-
tinguishable from fermentation, except that the rate of activity in the
latter is more rapid and results in an appreciable liberation and
accumulation of heat and gases. Although the subject of fermentation
has received considerable attention in the past no satisfactory technic
for measuring improvement in quality has been devised, and estimates
of the progress of fermentation are dependent largely upon the opinion
of those experienced in judging tobaccos.
As there is little exact knowledge concerning the nature of the pro-
cess of fermentation, comprehensive investigations from several
different angles will be necessary to establish the facts. The present
investigations, in which the Dewar-flask method was used, are pri-
marily concerned with the possible relationship between micro-
organisms and the changes involved in the fermentation of cigar-leaf
types of tobacco.
EARLIER STUDIES
The chemical changes occurring in tobacco during fermentation
have been given particular attention by several investigators, but
need not be reviewed here. It should be recalled, however, that tobacco
fermentation is generally believed to be an oxidation process, and the
close relation of air to the results secured has been generally recognized.
Analyses show significant decreases in nitrogen compounds, including
nicotine, and other organic substances, accounting for a loss of solid
matter sometimes exceeding 5 percent, during fermentation. The
total loss of weight, including that of free water, may be considerably
higher during the process. On the other hand there is a marked libera-
tion of ammonia and carbon dioxide gases as a consequence of this
activity. The improvement of the aroma, flavor, burn, and other
qualities is to be regarded as of major importance even though not
chemically determinable. Separating the essential and desirable
changes from incidental changes constitutes one of the chief difficulties
of the chemical aspects of the problem.
The previous investigations which are of most interest in relation to
the results secured in the present investigations are those dealing with
the possible relationship of enzymes and micro-organisms to the
fermentation process. There have been certain claims (17) ^ and
' Received for publication Mar, 12, 1934; issued August 1934.
' Reference is made by number (italic) to Literature Cited, p. 159.

Journal of Agricultural Research, Vol. 49, no. 2

Washington, D.C. July 15, 1934
Key no. Wis -6J
(137)
138 Journal of Agricultural Research voi. 49, no. 2

there is some evidence for the contention that, given the proper con-
ditions, fermentative changes may occur in the absence of any enzymic
or microbial activity, but this theory has not received much support.
The problem seems rather to involve the relative importance of the
enzymic and the microbial activities during normal fermentation,
even though these may be regarded only as agencies speeding up the
rate of oxidation and other chemical changes.
As early as 1858 an analogy between tobacco fermentation and
alcoholic fermentation was implied by Koller (10), who added yeast
with the purpose of increasing the rate of tobacco fermentation.
Later, tobacco fermentation was compared with that of silage (4)y
^^brown'' hay (Í), manures (19), etc. The bacterial theory of tobacco
fermentation was most definitely brought forward in 1891 by Suchs-
land (21), who isolated bacteria from sweating tobacco, prepared pure
cultures, and inoculated these back into tobacco. His claim that the
flavor and odor of a specific type of tobacco could be developed in
another type through the use of bacterial cultures has not, however,
been substantiated. Miciol (15), Dávalos (3), Vernhout (22),
Behrens (1), Koning (11), Jörgensen (9), and others soon afterward
isolated organisms from tobacco and in general supported Suchsland's
hypothesis. These workers showed that a variety of organisms were
present, in what were considered large numbers, i.e., as high as
112,500 bacteria and 12,500 fungus spores on 100 cm^ of freshly
fermented leaf.
Suchsland's bacterial theory soon fell into disrepute under a
vigorous though not convincing attack by Loew (12) in 1899. Loew
not only claimed that bacteria were not present in sufficient numbers to
influence fermentation, but that sufficient moisture was not normally
present for their development, and that even if sufficient moisture
were present, bacteria do not find tobacco a congenial medium for
growth.
An even stronger argument against the bacterial hypothesis, how-
ever, was the enzymic theory of tobacco fermentation developed by
Loew and treated in further detail in two succeeding papers (IS, 14)-
This theory ascribes tobacco fermentation to oxidizing enzymes nor-
mally present in all living material, as oxidase, peroxidase, and cata-
lase, the latter being present in dried, cured, and fermented tobacco,
even after several years. The chief role in fermentation was first
ascribed to peroxidase, an enzyme readily identified by its reaction
with tincture of guaiacum in the presence of hydrogen peroxide.
Loew's theory has since been supported mainly by Boeckhout and Ott
de Vries (2) and Jensen (5), but very little new evidence to support or
controvert the enzyme theory has appeared in the literature. Jensen
(^) in 1915 used the Dewar-flask method of study and concluded that
fermentation of leaf tobacco cannot be inhibited through the addition
of chemicals detrimental to micro-organisms. The data presented are
not clear on this point, however, and the variability of the tempera-
tures in the incubator used was such as to render the data of doubtful
value. On the other hand, Jensen suggests the existence of two types
of fermentation, namely, one which proceeds at a moisture content of
20 percent or below and another which requires a higher moisture
content. Recent investigations in Russia, discussed in considerable
detail by Smirnov (20), evidently are in general agreement with
Jensen's conclusions.
 July 15, 1934 Studies on the Fermentation oj Tobacco 139
It is evident from the literature that the determination of the
nature of tobacco fermentation is complicated by the problem of what
constitutes true fermentation and by the variation in the practical
requirements of different types of tobacco.
Studies of a related nature have been conducted by many investiga-
tors on the spontaneous generation of heat in hay, straw, silage, ma-
nures, etc. The purely
chemical, enzymic, and
microbial explanations
have all had stanch
supporters, but it is in-
teresting to note that
recent investigations
support the microbial
theory {16,18), at least
under temperature
and moisture condi-
lions imder which or-
ganisms will iriultiply,
and even discount the
cooperation of en-
zymes {16).
MATERIALS AND
METHODS
At first the present
studies were con-
ducted with the ordi-
nary narrow-mouth 1-
quart thermos bottles.
Later, wide-moutli
flasks were secured
which were easier to
fill and permitted of
handling the tobacco
under fairly satisfac-
tory aseptic and pure-
culture conditions
when desired. This
was accomplished by
first placing the tobac-
co in moisture-proof
cellophane containers,
sterilizing it with heat, riGURE 1.—A simple method for preparinR tobacco .samples for fer-
mentation studies with pure cultures. The steriliz.ed tobacco in
and inoculating it with the cellophane roll (A) may be inoculated at any desired points by
mserting the syringe (B) through the cellophane before the roil is
water suspensions of placed in the wide-mouth Dewar flask (C).
cultures of organisms
by means of a Luer syringe inserted through the cellophane at one oi-
more points. The cellophane containers were at first made the desired
shape and size before being filled with tobacco ; inoculations were then
made by injections at a large number of points (fig. 1).
However, more even distribution of inoculum could be secured by
preparing larger sealed cellophane bags in which the tobacco was held
140 Journal uj Agricultural Research Vol. 4«, no. 2

loosely and into which the inoculum was injected with a syringe and
mixed with the tobacco by turning and agitating it. The tobacco
was then pressed into one end of the bag, which was rolled into form
to fit the wide-mouth Dewar flask (fig. 2).
It was found that the tobacco could be adequately sterilized in these
containers without apparent physical injury to the leaf or any appre-
ciable change in moisture content by placing the roll or bag in a sealed
copper container which was placed in an ordinary steamer. Forty-
five minutes of steaming, during which the tobacco reached a tem-
perature of 80° C, was sufficient to prevent subsequent thermogenesis,
and plating out showed that no organisms were present. In practice.

FKJURE 2.—The eellophane-bag (A, B, C) method permits of a uniform application of inoculum under
aseptic conditions; the tobacco in the cellophane bag is sterilized in a copper, rubber-tube-sealed container
(Ö), without undergoing any change in percentage of moisture.

however, the steaming was allowed to continue for 1 hour, during

which time the tobacco reached a temperature of at least 85°. At
these temperatures peroxidase was destroyed. In other tests peroxi-
dase was found to be destroyed in water extract if heated for 10 min-
utes at 83°, and Loew (12) reports that the enzyme is destroyed at
87° in 3 minutes.
A limited amount of aeration was made available to the tobacco by
various means. In the case of the narrow-mouth bottles a pliable
perforated lead tube extended to the bottom of the flask. In the case
of the cellophane containers, aeration was only provided by puncturing
the cellophane at several points with the syringe or a hot needle and
by not sealing the mouth of the flasks tightly. No other provisions for
aeration were made, but since, as a rule, the tobacco was only loosely
packed and the experiments continued for only 10 days, it is safe to
July 15, 1934 Stvdies on the Fermentation oj Tobacco 141
say that the amount of oxygen available was equal to that in normal
fermentation in large boxes or bulks. At least, preliminary trials
with air aspirated through the bottles did not appreciably increase
thermogenesis, whereas the replacement of the air with nitrogen,
followed by sealing, greatly reduced the thermogenic power.
The tobacco used in the tests was largely of the local variety known
as Havana No. 142, grown on the Wisconsin Experiment Station farm.
This tobacco contained approximately 30 percent moisture, which is
close to normal for Wisconsin tobacco in the bale soon after stripping.
The leaves were first stemmed and cut into strips about one-fourth
inch in width; this made the pack more uniform and easier to
handle. One hundred and fifty grams (about 5% ounces) of this

FiüUKE 3.—One of the constant-temperature incubators and the thermos bottles used in the earlier
experiments.

tobacco was usually used in the 1-quart thermos bottles. The

moisture content was usually raised to 35 percent or more by atomizing
the tobacco with the desired quantity of distilled water. In most
instances the chemicals were applied with the water.
As soon as the bottles were filled they were placed in automatically
regulated constant-temperature chambers. Most of the experiments
were run in a large 30° C. incubator in which the temperature nor-
mally varied less than 1° (fig. 3). More significant results would no
doubt have been secured in some cases by incubating at a somewhat
lower temperature. Readings were taken at 24-hour intervals
(i.e., at 9 a.m.), usually over a period of 10 days. The temperature
increases shown in the tables were then based on the average of the
incubator readings subtracted from the average of the flask readings
over the 10-day period. The maximum increases, which usually
occurred about the fifth day, were considerably higher than the
average. In most cases the incubator temperature recorded was
75317—34 4
142 Journal qf Agricultural Research Vol. 49, no. 2

that of a thermometer inserted in a corresponding empty thermos

bottle.
After the completion of the test in the thermos bottles, rough esti-
mations of ammonia and carbon dioxide were usually made by draw-
ing air from the aeration tube over concentrated hydrochloric acid
and through limewater. These estimations of gaseous products are
to be largely regarded as confirmatory information as to the direction
and degree of activity exhibited. Exact quantitative determinations,
while desirable, did not appear to be justified in the present connec-
tion, and would probably not have influenced the conclusions reached.
The tobacco was then withdrawn from the flask, and a random sample
of approximately one-half gram, or 10 square inches of leaf was
introduced into about 5 cc of sterile water, agitated, and allowed to
stand for about 1 hour before being plated out in nutrient agar or
potato agar. Notes were also made at this time on the color of the

^
y ^^OE ACCO

y
\
/
^ ■^--

I
^^NCUB/VTOR
^^^f^
^^**»*^ ^^ — ».
- "--. /
0 1234567 89 10
DAYS

FIGURE 4.—Thermogenic behavior of tobacco in a thermos bottle held for 10 days at an incubator tempera-
ture of between 30° and 31° C.

water extract, and the odor, as well as on tests for peroxidase with
tincture of guaiacum and hydrogen peroxide. The entire lot of
tobacco was again weighed and then placed in an oven for determina-
tion of dry weight.
EXPERIMENTAL RESULTS
In the present investigation about 350 separate fermentation trials
were made in Dewar flasks. The presentation of all the data secured
seems unwarranted, and consequently only portions will be used to
illustrate the conclusions drawn. The trials were rarely run in
duplicate, preference being given to repeated successive trials, which
were feasible since practically all environmental factors were under
control and the data were based on average temperature readings.
Control flasks, i.e., flasks containing untreated tobacco, were run,
however, in connection with practically all trials.
The typical thermal behavior of such a control may be noted from
the graph shown in figure 4. The tobacco being placed in the flask
at room temperature and the flask then placed in an incubator averag-
ing 30.5° C. (87° F.)^ the temperature rose very rapidly during the
July 15,1934 Studies on the Fermentation of Tobacco 143

first 24 hours as a consequence of both the heat from without and the
thermogenesis from within, which in the presence of sufficient moisture
usually gets under way the first day. The temperature normally
continues on the upgrade in the flasks for 3 to 5 days, after which time
it usually declines gradually. This decline is in part due to a reduc-
tion of the potential heat-producing constituents of the tobacco,
since if this same lot of tobacco is removed and replaced in the flask
with the original moisture content, the temperature curve is lowered
and shortened. However, a gradual reduction in available oxygen
and an accumulation of gases such as ammonia and carbon dioxide
injurious to thermogenesis may also play a role in this decline.
EFFECT OF MOISTURE

The percentage of moisture in Wisconsin tobacco at the time it is

baled on the farm averages approximately 30 percent,^ the variation
being generally between 25 and 40 percent. This tobacco may be
fermented in several different ways. If the leaf is sorted, it is packed
tightly in boxes or placed in bulks, where it goes through one or more
sweats before being packed in boxes. If the leaf is stemming tobacco,
the bales are stacked in large bulks, where they undergo fermentation.
Following a period of aging, usually accompanied by natural drying,
this leaf is moistened heavily before being stemmed, and the strips
are again fermented in large bulks at a moisture content of about
40 percent. The results secured under the various methods of
handling stemming tobacco may naturally be expected to be quite
different, and the results in this type as a whole are apparently not
comparable with those obtained in tobacco types stored at moisture
contents of 20 percent or below, where little or no thermogenic
activity occurs.
In large bulks or under high pressure, tobacco may evidently go
through a normal fermentation process at a relatively low moisture
content (22 to 28 percent). In thermos bottles with only 150 g of
tobacco, loosely packed, no significant thermogenic activity was
obtained at a moisture content of below 30 percent. As the moisture
content is increased above 30 percent, the thermogenic power was
found to increase proportionately (table 1). The rate of this increase
was distinctly raised by increasing the pressure, i.e., placing 300 g
of tobacco in the same volume (fig. 5). However, significant results
by the thermos-bottle method are dependent upon having a moisture
percentage of 35 to 40, which, though comparable with the amount of
moisture present during the practical process of fermenting stemming
tobacco, is not comparable with the usual percentage of moisture in
sorted tobacco. It seems reasonable to assume, however, that within
certain limits higher moisture contents in the thermos bottles for a
time at least only serve to hasten activity of the same character as
proceeds more slowly at lower moisture contents. In order that the
generation of heat may be measurable in the thermos bottles, it must
naturally accumulate at a more rapid rate than that at which it is
lost through the insulating power of the flask. Furthermore, moisture
is actually liberated during normal fermentation, as is suggested by
the term '*sweating'', which no doubt renders the leaf and the environ-
ing air increasingly favorable for additional forms of activity. Taking
3 The moisture percentages as given throughout this paper are computed on the basis of total weight.
144 Journal of Agricultural Research Vol. 49, no. 2

all factors into consideration, limited generalizations with respect

to the similarity of fermentation activity at ranges of from 25 to 40
percent are permissible. The situation in tobacco stored at moisture
contents below approximately 20 percent may be expected to be of

^
/

/TIGHT PA CK
/ /
/
ê
i
^^OOSE PACK

Ü
2

Y
<
a:

<>
1
ií
25 30 35 40 45 60 55
MOISTURE (PERCENT)

FIGURE 5.—Relation of percentage of moisture and tightness of packing to the thermogenic activity of
tobacco in thermos bottles held at an average temperature of 30.2° C.

quite another nature, as are those in tobacco maintained at excessively

high moisture contents.
TABLE 1.—Relation of ^percentage of moisture in tobacco to its fermentation in
Dewar flasks ^

Average temperature
Moisture
(percent) Ammonia Carbon
dioxide Odor Color of
extract
Incubator Flask Increase

° C. °C.
30.4 30.3 31.0 0.7 + 0 Raw +
32.9 30.3 31.3 1.0 + —do ++
34.2 30.3 31.9 1.6 ++ + Mild ++
37.9 30.3 32.9 2.6 +++ ++ —do +++
41.5 30.3 32.7 2.4 +++ +++ Good +++
50.0 30.3 34.8 4.5 +++ +++ Strong +++
54.3 30.3 35.0 4.7 +++ +++ —do +++
In this and succeeding tables, plus signs indicate relative amounts, i.e., + small, ++ medium, +4-4-

EFFECT OF TEMPERATURE

Under favorable conditions cigar-leaf tobacco is ready for normal

fermentation as soon as it is properly cured. In practice, some
months may elapse between the completion of the curing and the
beginning of fermentation. While this interim is partly a result of
the time required for marketing, sorting, packing, and storage, fer-
July 15, 1934 Studies on the Fermentation of Tobacco 145
mentation does not normally get under way until favorable natural
or artificial temperatures are presented.
Little is definitely known about the most favorable temperature
for fermentation, or regarding the lowest or highest temperatures at
which it occurs. The solution of this problem presents many diffi-
culties. An investigation of the problem by the Dewar-flask method
suggests, however, many points of interest. The results of a series
of trials are shown in table 2. The various temperature control
chambers used could not be maintained regularly at a variation of
less than 1 ° C. ; consequently small temperature increases at the ex-
tremes were difíicult to detect. Evidently, however, little or no
activity occurred at temperatures below 10° C. (50° F.). At about
16° C. (61° F.) fairly evident thermogenesis began, reaching a maxi-
mum at 20° C. (68° F.) in this series. In other trials the indications
were that the maximum of heat production lies closer to 25° C.
(77° F.) (fig. 6). At any rate, spontaneous generation of heat begins

o
o

UJ

g"
<tr / \53pE RCENT MOI 5TURE
UJ
a.
/
/
/ ^^
Z
/
/ J y^ ^
UJ
/ •^-.<.
UJ
\
£E
Ü PERCENT MOISTURE
\
Z
\ \
Ul , ^^ \ \
ce
^ \
UJ
\
0
0 5 10 15 20 25 30 35 40 45 50
TEMPERATURE ('C.)

FIGURE 6.—Relation of temperature of incubation to the thermogenic activity of tobacco at two different
moisture contents.

to drop off at 30° C. (86° F.) and is apparently entirely eliminated

before a temperature of 50° C. (122° F.) is reached. Temperatures
as high as 50° C. are rarely, if ever, reached in fermenting boxes of
leaf tobacco, though in bulk fermentation considerably higher tem-
peratures may be attained {%S), whereas in stemming tobacco the
bulk may be allowed to reach 60° C. (140° F.). For cigar wrapper
and binder tobacco, however, it is now generally believed that tem-
peratures above 50° C. are injurious to the quality of the leaf. Nor-
mally the bulk is therefore turned or the leaf transferred to boxes
before this temperature is reached. The curve of the temperature
rise in such a bulk is typically such as is shown in figure 7, in which
it will be seen that following a rapid rise of temperature, the rate is
diminished considerably beyond 43° C. (110° F.) and, in the fourth,
fifth, and sixth rebulkings, drops off at increasingly lower tempera-
tures, the primary reason perhaps being the lowered moisture content.
In practice, bulks are now usually turned only once or twice before
being boxed.
146 Journal of Agricultural Research Vol. 49, no. 2

TABLE 2.—Relation of temperature to the fermentation of tobacco in Dewar flasks ^

Average temperature
Moisture Ammonia Carbon Color of Peroxi-
(percent) dioxide Odor extract dase
Incubator Flask Increase

° C. ° C. ° C.
39.3 9.8 9.3 2-0.5 0 0 Raw ++ ++
38.3 16.6 17.8 1.3 ++ ++ do ++ +++
39.1 20.0 23.7 3.7 ++ +++ Good +++ +++
37.8 25.4 28.0 2.6 +++ +++ Strong +++ +++
37.0 27.0 28.4 1.4 +++ +++ do +++ +++
37.3 36.6 38.4 1.8 +++ +++ do +++ +++
37.0 44.1 44.9 .8 4- 0 Mild +++ +++
38.0 48.5 48.3 2-.2 0 0 do +++ +
I See footnote, table 1. 2 Decrease.

Judging from the results obtained with the thermos bottles, fer-
mentation apparently does not proceed much below 18° C. (65° F.)

70

FIGURE
10
m^ 20 30 40
DAYS
50 60 70

7.—Generation of heat in a large bulk of cigar-leaf tobacco following repeated rebulking.

or above 45° C. (113° F.). Several trials with tobacco ''incubated''

in a moist condition at temperatures between 45°-55° C. (113°-131°
80

F.) for 20 days yielded little or no evidence of normal fermentative

changes. On the other hand, there is some chemical evidence that
heating tobacco directly may bring about changes resembling certain
of those due to true fermentation, although evidently such heating
cannot replace fermentation. Apparently, it is not the actual tem-
perature reached in fermentation which is significant, but the actual
increase in temperature above the surrounding atmosphere, since
this increase represents the intensity of the process. Taking all
facts into consideration, it will be seen that problems of considerable
practical importance arise in this connection. Apparently fermenting
rooms should preferably be maintained at approximately 22° C.
(72° F.), and temperatures occurring in the fermenting leaf above
July 15, 1934 Studies on the Fermentation oj Tobacco 147
45° C. are less likely to be effective than are lower temperatures. A
30-degree increase in temperature between 75° and 105° F. should,
for example, be considerably more effective than a 30-degree rise in
temperature between 100° and 130° F. Hence, bulkings which result
in excessive accumulation of heat may actually retard true fermenta-
tive changes.
EFFECT OF HEAT AND ANTISEPTICS

If tobacco is heated to a sufficiently high temperature in dry or moist

heat, both the microbial ñora and the enzymes present are destroyed.
If such heated tobacco is placed in a thermos bottle in a moist con-
dition, heat generation will still occur, though after some delay. This
is true microbial thermogenesis, resulting from reinfestation with
micro-organisms. If, however, the technic is modified so as to insure
the introduction of heat-sterile tobacco into the container and the
maintenance of aseptic conditions, thermogenesis is completely
prevented, and the leaf is essentially preserved and unchanged
(table 3). However, it is evidently not possible by the use of heat to
destroy the microbial flora without destroying or injuring the plant
enzymes or other constituents of the plant cells which may be con-
cerned with normal fermentation, since the thermal inactivation points
of most forms of living matter often too closely overlap.
It seemed quite conceivable, however, that certam chemicals,
especially those of an antiseptic nature, might exert a selective action
between microbial activity and the activity inherent in plant cells.
Johnson and Murwin (^), for example, found in testing various dis-
infectants for tobacco seed, that while corrosive sublimate under
certain conditions completely checked germination, silver nitrate
under similar conditions was not injurious to the germination of the
seed, a function evidently dependent upon enzymic activity. Both
chemicals, however, were about equally destructive to micro-
organisms. Chloroform, toluene, and acetone are usually conceded
to be effective germicides, but with the additional property of not
being particularly destructive to enzymes or harmful to their activity
at low concentrations.
TABLE 3.—Effect of heat applied to moist tobacco in air-tight copper containers on
subsequent behavior in Dewar flasks when reinfestation with micro-organisms is
prevented ^

Average
Heat treatment of 150 g of tempera- Ammonia Carbon Peroxi- Bacteria Fungi
tobacco ture in- dioxide dase
crease

°C.
None . - - -- 4.3 +++ ++ H-++ +++ ++
20 minutes, to 45° C _ _ _ 2.8 +++ ++ +++ ++ ++
30 minutes, to 67 ° C_ __ ___ 2.0 + ++ + + +
40 minutes, to 74° C ._ _ . .2 + 0 0 + +
45 minutes, to 80° C. 2-.1 0 0 0 0 0
60 minutes, to 85° C_-- _ _-. 2 -.2 0 0 0 0 0

1 See footnote 1 table 1. 2 Decrease.

According to Loew (14), many salts and bases are not harmful to
catalase activity, and some may be beneficial. Mercuric chloride
however, was found to be distinctly harmful, as was formaldehyde in
high concentrations (1 to 5). Chloroform was not found to be
148 Journal of Agricultural Research Vol. 49, no. 2

destructive. While the data on this subject are not entirely clear
in differentiating between destruction and inactivation, nevertheless
the report by Jensen (5) that mercuric chloride, formol (formalde-
hyde) and chloroform do not prevent tobacco fermentation has been
regarded by some as strong evidence in support of Loew's theory of
enzymic fermentation, and good evidence against the microbial theory.
The results of the experiments here reported, in which the thermos-
bottle method was used, have been quite contradictory to those of
Jensen. Not only do mercuric chloride and chloroform effectively
prevent thermogenesis, but acetone, toluene, beta-napthol, and other
antiseptics at low concentrations also prevent thermogenic activity
(table 4). The data secured with formaldehyde were less certain.
Under the conditions of the experiments, both enzymes and microbes
were apparently inactivated but not necessarily destroyed by many
antiseptics. Chemicals other than antiseptics, such as ethylene
chlorhydrin, acids, and bases, were tested with the purpose of dis-
covering possible stimulatory action on thermogenic activity, but no
conclusive evidence in this direction was secured.
TABLE 4.—Effect of antiseptics and other chemicals on tobacco fermentation in
Dewar flasks ^

Average temperature Car- Staphylo-

Mois- Am- bon Perox- cocci to
Treatment of 150 g tobacco ture Incu- monia diox- Odor idase the
In- square
bator Flask crease ide inch«

Per-
cent °(7. Number
None ! 37.0 30.3 33.7 3.4 +++0 +++0 Mild.._ +++ 100,000
Acetone, 5 cc 37.0 30.3 30.7 .4 raw +++ 120
Toluene, 5 cc 37.0 30.3 30.7 .4 0 0 _-do__. +++ 0
Chloroform, 1 cc 37.8 30.3 30.6 .2 + + —do__.
Formalin, 2 cc 51.1 30.0 31.7 1.7 +++0 ++0 —do-_. +++ 500
Mercuric chloride, 1 g 37.1 30.5 30.8 .3 ,-do__. ++ 0
Silver nitrate, 1 g 37.1 30.2 31.7 1.5 + +++ —do-_- +++ 0
Silver nitrate, 3 g 41.3 30.2 31.6 1.4 + +++ _-do-_. +++ 0
Ethylene chlorhydrin, 1 cc. 43.8 27.7 30.3 2.6 +++ +++ —do___
Strong- +++ 20,000
None - 43.8 27.7 32.4 4.7 +++ +++ -_.do__. ++ 50,000
Potassium hydroxide, 2.5 g-- 38.5 30.5 33.4 2.9 +++ +++ ++ 100,000
Oxalic acid, 0.5 g 41.0 30.3 34.4 4.1 +++ ++ .._do-_. +++ 10.000

1 See footnote 1, table 1. 2 Estimated.

The results secured with silver nitrate are of particular interest as

indicative of a differential effect on microbial and enzymic activity.
Under the conditions of the experiments, silver nitrate reduces ther-
mogenesis to about one-half that of the untreated controls. The chem-
ical, however, is quite as effective as the other antiseptics used in reduc-
ing microbial activity, as indicated by the freedom of the treated
tobacco from surface fungus growth and from the results of plating
out. Varying the quantity of silver nitrate from 0.25 to 4.0 g to
150 g of tobacco did not appear to affect the type of result secured.
While the bacteria are evidently kuled by the silver nitrate, many
fungus spores may survive, raising some doubt as to the efficiency of
the silver nitrate in checking fungus activity. The proof of their
elimination by suver nitrate therefore must be sought by other means.
This may evidently be accomplished by first destroying both the
enzymes and the microbes in the tobacco by heat. Such tobacco will
soon develop strong thermogenic activity as a consequence of micro-
July 15,1934 Studies on the Fermentation oj Tobacco 149
bial development resulting from natural or artificial reinfestation.
If, however, such heated tobacco is in addition treated with silver
nitrate, no thermogenesis follows (table 5). It seems logical to con-
clude, therefore, that silver nitrate effectively checks microbial
activity and that the thermogenesis occurring in unheated tobacco
with süver nitrate is not due to microbial action, but is evidently due
to some constituents (presumably enzymes) of the tobacco leaf itself,
which are not harmfully affected by silver nitrate. The evidence
suggests, therefore, that thermogenesis in tobacco may be brought
about by enzymes alone, by enzymes in combination with micro-
organisms, or by micro-organisms alone.
TABLE 5.—Relation of silver nitrate to fermentative activity of tobacco in Dewar

Silver Average
Treatment of nitrate Artificial inoculation tempera-
tobacco, 150 g ture in- Peroxidase Bacteria Fungi
used
crease

Grams
Not sterilized _/ 0 None _-- 3.7 +++ +++ +
do 1.5 +++ 0
I1 ""'doI"""""""I-^
4.6
2-1
0
0
H-
0
+++
0
Sterilized 1 Aspergillus flavus — 0 0 0 +
1 Aspergillus ochraceus .3 0 0 +
1 Aspergillus terreus 0 0 0 +
0 4.1 0 0 +++
1 See footnote 1, table 1. 2 Decrease.

PRODUCTION OF AMMONIA

The most characteristic odor in rooms where large amounts of

actively fermenting tobacco in bulks are being processed, is that of
ammonia gas. According to Loew (Í5), this ammonia results from
the destruction of the nitrogenous compounds by oxidizing enzymes
and is not a product of putrefaction. The data secured in the present
experiments repeatedly indicate the formation of large amounts of
ammonia in the absence of oxidizing enzymes in the case of heated
peroxidase-free tobacco. Ammonia production seems in all cases to
be closely correlated with microbial activity and with thermogenesis.
However, the determinations of ammonia here reported are only
estimations made on the basis of the amount of fuming produced
when air from the aeration tube was drawn over concentrated hydro-
chloric acid. More accurate determinations of correlation between
enzymic activity and ammonia production must be made before it
can be definitely concluded that peroxidase, catalase, or other enzymes
are not also instrumental in the formation of ammonia in fermenting
tobacco.
Ammonia production is only faintly discernible at low moisture
percentages, where thermogenesis is not evident, but rises rapidly in
direct proportion to the increase in moisture. At such low tempera-
tures as 12° C. (54° F.) or below, ammonia production either is not
determinable by the method used or is entirely absent even though
slight thermogenesis has occurred. Its production then gradually
rises up to about 25° C. (77° F.), and falls off again gradually to
44° C. (111° F.), with zero production again at 48° C. (118° F.)
(table 2).
150 Journal oj Agricultural Research voi. 49, no. 2

Mercuric chloride, acetone, and toluene may entirely prevent am-

monia production, although measurable quantities developed in the
presence of chloroform, beta-napthol, and Ceresan, in cases where
thermogenesis was practically nil. Silver nitrate apparently checks
ammonia production more effectively than it does thermogenesis.
Other chemical agents used of a nonantiseptic nature had no depress-
ing effect on the formation of ammonia.
PRODUCTION OF CARBON DIOXIDE

The formation of carbon dioxide is a result of oxidation and should

therefore be closely correlated with the fermentation process. The
data secured on this subject are again only estimations based upon
drawing the atmosphere from the thermos bottles through limewater.
As with the formation of ammonia, carbon dioxide formation is fairly
closely correlated with thermogenesis, although it may often appar-
ently be absent under conditions which permit some formation of
ammonia. Its production is completely checked at low percentages
of moisture and at the extremes of temperature. Most antiseptics
usually completely prevent its formation, but silver nitrate usually
reduced but did not check its production.
OTHER FERMENTATIVE CHANGES

The desired results of fermentation are those which bring about the
best flavor and aroma in tobacco. While the term ''aroma'ls some-
times used in referring to leaf tobacco, it should preferably be reserved
for cigars or manufactured tobacco and particularly for the smoking
quality. The term ^'odor^' will therefore generally be used for the
changes which occur during fermentation. Chemically, there may be
no particular relation between the odor and the aroma, although in
practice a desirable odor is believed to bring about a desirable aroma.
The lack of any standard of measurement of odor, either qualitative
or quantitative, makes it a difficult problem to attack. The ability
of some persons to judge odor by the sense of smell is, to be sure, often
remarkably developed, and, following much and varied experience,
may be, for all practical purposes, sufficiently reliable for fine dis-
tinctions. In the present experiments it has been necessary to con-
fine the estimations to rather wide differences in odor. The designa-
tion ^*strong'' odor may be quite unsatisfactory, however, since it
probably more correctly refers to the odor of ammonia, which may
actually mask other odors of greater significance. The terms ''raw''
and ''none", "mild" and "good" are more rehable. "Sweet" or
"pleasant" odors were rather unusual types, secured at the higher
temperatures, and may not have been of a true fermentative nature.
No fermentative odors are produced in the thermos bottles in the
period of 10 days unless the moisture content is above about 32 per-
cent, even though some heat production and some gases may be
evident. At a 35-percent moisture content, however, a fair develop-
ment of odor occurs at 30° C. (86° F.) in 10 days and apparently
increases rapidly in direct proportion to the moisture present, up to a
45-percent moisture content or more, in which case a very strong but
undesirable odor occurs in considerably less than 10 days. With a
moisture content of 35 to 40 percent, no apparent fermentative odor
develops below 18°, in 10 days. At temperatures beyond 20°, the
July 15,1934 Studies 071 the Fermentation oj Tobacco 151

characteristic odor develops rapidly, reaching the maximum probably

between 25° and 35°. Beyond 40°, the strength of the odor is strik-
ingly decreased, and a sweet and more pleasant odor, sometimes
perfumelike, results at temperatures up to 48°. At still higher tem-
peratures it is doubtful whether any true fermentative odors are
discernible.
Such antiseptics as were found to prevent active thermogenesis
(mercuric chloride, chloroform, acetone, toluene, beta-napthol, etc.)
also prevented the formation of any desirable or undesirable fermenta-
tive odors. Such chemicals as potassium hydroxide, oxalic acid,
sodium carbonate, and eth^^l chlorhydrin were without particular
effect. Silver nitrate, while it permitted some thermogenic activity,
checked the formation of fermentative odors quite effectively. Heat-
ing the tobacco to 80° C.,or above, is in itself productive of some change
in odor which may be confused with the odor of fermentation. How-
ever, it seemed apparent from a limited number of trials that the
destruction of peroxidase by heat does not prevent the subsequent
production of a fairly typical fermentation odor, when proper con-
ditions for microbial activity are again brought about. Under such
conditions there appears, however, to be more likelihood of certain
fungi or other undesirable organisms developing excessively, with the
resultant production of **wild^' fermentative changes.
The color of a water extract of tobacco is of some value in com-
paring the relative degree to which the fermentative changes have
progressed, although of little value in comparing different lots of
tobacco with each other. As the fermentative activity increases, or
the time is prolonged, the extract becomes increasingly darker in
color; this condition is generally closely correlated with thermogenesis.
ENZYMIC RELATIONS

Although the presence of a number of different enzymes may be

demonstrated in cured tobacco, it has not yet been conclusively shown
that any particular enzyme, or these enzymes as a group, are
responsible for fermentation in whole or in part. For the present,
therefore, it may be satisfactory to select one of the easily deter-
minable enzymes in cured tobacco as a general criterion of probable
relation of enzymes to the experimental results secured. Peroxidase,
as determined with tincture of guaiacum in the presence of hydrogen
peroxide, was therefore chosen for this purpose. Unfortunately, this
test does not distinguish between the presence of peroxidase in the
active and in the inactive state; consequently, since it is normally
present in tobacco, the significance of the present tests is largely
limited to cases in which this enzyme is destroyed and hence elimin-
ated as a factor in fermentation.
Peroxidase, as compared with living matter in general, is fairly
resistant to unfavorable conditions. None of the antiseptics at the
strengths used destroyed the j)eroxidase reaction with guaiacum.
Since certain of these same antiseptics prevented thermogenesis, it
must be assumed either that peroxidase is not connected with thermo-
genesis or that the enzyme was rendered inactive by the antiseptic.
The latter assumption is quite logical.
In water extract from cured tobacco leaves peroxidase is destroyed
at approximately 83° C, at a 10-minute exposure. When 150 g lots
152 Journal of Agricultural Research Vol. 49, no. 2

of moist tobacco are placed in a dry-air oven for 24 hours, it normally

required a temperature of 75° to 80° to destroy the peroxidase. In
the moist condition peroxidase withstood a temperature of 49° for
10 days, though it was decidedly weakened by this treatment.
Exposure to a temperature of 54° for 10 days in the moist condition
destroyed peroxidase completely. It is therefore very unlikely that
any thermogenesis resulting from peroxidase can occur above 50° C.
(122° F.).
The behavior of enzymes and micro-organisms toward heat and
antiseptics, as well as toward environmental conditions in general,
is such that it becomes difficult to satisfactorily separate their activity
in this manner in tobacco fermentation. Tobacco catalase, however,
appears to be an exception. The optimum temperature for the activ-
ity of tobacco catalase as measured by the evolution of oxygen gas
40

30

2 20

10

AT I I I IM0 10 20 30
TEMPERATURE (''c.)
40 50 60

FIGURE 8.- -Relation of temperature to the activity of catalase as measured by the evolution of oxygen from
hydrogen peroxide.

with hydrogen peroxide was found to be close to 45° C, with the

maximum at perhaps 58° and the minimum below 0° (fig. 8). Accord-
ing to determinations of others, the activity of catalase between 10°
and 40° is more or less constant, and inactivation occurs at 55°.
While the higher temperatures for catalase activity correlate roughly
with all the fermentative changes and their decline at the upper limits,
this correlation is no less evident for microbial activity.
On the other hand, there seems to be very little correlation between
catalase activity and fermentative changes at the lower end of the
temperature scale. If catalase is, for instance, approximately as
active at 15° C. as at 25°, thermogenesis and other changes in tobacco
should be quite comparable at these temperatures; whereas they are
as a matter of fact strikingly different and more closely correlated
with microbial behavior at such temperatures (fig. 9.).
Evidently it is necessary, however, to develop other forms of
technic to distinguish enzymic from microbial activity in the case
of active thermogenesis in tobacco. Theoretically, the destruction
of all enzymes present by adequate heating, followed by the reintro-
July 15, 1934 Studies on the Fermentation of Tobacco 153
duction of enzymes which may be concerned, together with the main-
tenance of aseptic conditions, should furnish conclusive proof of
their connection. However, such evidence may, to be sure, lack
definiteness except in the case of positive results. Using the cello-
phane-bag method for securing aseptic conditions, experiments of the
above nature were performed in the following manner: One thousand
grams of cured tobacco was extracted with 2,000 cc of water and
yielded 1,300 cc of extract. This extract was filtered through a
bacteria-proof porcelain filter, into approximately 50 cc lots. The
filtered extract was stored for several days and remained sterile, as
was shown by plating out from the separate lots. Unfiltered control
extract soon became clouded with growth of organisms. These
extracts yielded strong peroxidase reactions at dilutions as high as

KiouRE 9.—Relation of temperature of incubation of tobacco in Dewar flasks for the 10 days to the
development of micro-organisms; A, 9.4° C; B, 14.8°; C, 19.9°; D, 24.4°; E, 27..')°; F, 30.7°; G, 35.5°;
H, 40.2°.

1 to 1,000. One-hundred-and-fifty-gram lots of sterilized, peroxidase-

free tobacco in cellophane bags were then treated uniformly with
these extracts in varying amounts ranging from 40 to 80 cc, the tobacco
in these lots being thoroughly wetted by the larger amounts of extract
applied. The bags were then rolled and introduced into the thermos
bottles. Adequate controls were also prepared. In some cases con-
tamination with organisms occurred as a consequence of providing
for aeration, and the usual thermogenic activity resulted. However,
in cases where contamination did not occur, no thermogenic activity
developed, ammonia and CO2 were hardly perceptible, and no changes
in aroma were evident (table 6). Peroxidase and other enzymes or
leaf constituents which are water-soluble were present, as was shown
by tests at the end of the experiments, but the quantity of enzymes
was evidently below normal. However, it is not unlikely that certain
other conditions or activators necessary for cnzymic activity were
lacking; the results, being negative, are not regarded as conclusive.
The method, however, offers possibilities of further development.
154 Journal of Agricultural Research Vol. 49, no. 2

TABLE 6.—Effect of applying varying amounts of filter-sterile extract containing

peroxidase to sterile tobacco in Dewar flasks ^

Filter-
Treatment of 150 sterile Average
tobacco Moisture temperature Ammonia Carbon Peroxidase Fungi
g of tobacco extract dioxide
increase
added

cc Percent ° C.
40 36.3 2-0.1 + 0 0 0
60 43.2 .1 + 0 + 0
Sterilized 80 46.0 2 -.1 + + + 0
3 60 43.8
43.1
4.1 +++ +++ + ++
I 0 0 0 0 0
Not sterilized 0 35.0 3.3 +++ +++ +++ +
I See footnote 1, table 1. 2 Decrease. 3 Accidental contamination.

MICROBIAL RELATIONS

Many difficulties also lie in the way of furnishing conclusive

evidence of any essential relation between micro-organisms and the
fermentation of tobacco, and until the limits of fermentation are
more clearly defined than at present, generalizations on this subject
are premature. For those types of tobacco in which normal fermen-
tation may be said to proceed at percentages of moisture that entirely
preclude the development of micro-organisms, negative proof of
microbial relationship should be quite simple. On the other hand, in
the fermentation of tobacco in which the activity of micro-organisms
may be readily demonstrated, it still remains to be shown that they
play an essential rather than an incidental role. Such proof may be
rendered difficult for several reasons, particularly the possibility that
microbial activity under the conditions of the experiment may replace
in whole or in part other agencies which may normally be responsible
for fermentation.
The existence of a variety of micro-organisms in relatively high
numbers in stored and fermenting tobacco has been repeatedly
shown by different investigators. That certain species may find
tobacco a favorable substratum for growth under certain conditions
is amply demonstrated by the occurrence of such maladies as black
rot, musts, and molds. The regularity and frequency with which
still other species, which evidently cause no injury, occur suggest
that these may exert some influence on fermenting and stored tobacco,
although this influence may be admittedly harmful if allowed to
progress too far.
In the present investigations over 350 platings have been made
from the Dewar-flask-fermented lots, and from tobacco from various
other sources, including widely separated tobacco-growing districts.
The regularity of occurrence and the numbers of certain species of
organisms were in some instances very striking. On the other hand,
the relative infrequency of the occurrence of the known injurious
forms, such as Aspergillus niger (7) and Oospora nicotianae in these
plates suggests the limited degree of infestation necessary to produce
subsequent extensive changes in tobacco under conditions favorable
for the development of the particular species concerned. Two large,
white-colony forms of bacteria of undetermined species were com-
monly present, but these were not regarded as significant for the
double reason that no marked increase in numbers occurred during
July 15, 1934 Studies on the Fermentation of Tobacco 155
fermentation and because they were evidently unable to induce ther-
niogenic activity in tobacco when introduced in pure culture.
On the basis of almost universal association with stored and fermenting
cigar tobacco, a tiny coccus form seemed worthy of more consideration
(fig. 10). This organism has regularly been secured on nutrient
agar from fermenting tobacco from widely separated sources (table 7),
often in numbers indicating 300,000 to 500,000 organisms to the
square inch of leaf surface. Pi-eliniinary studies on this organism in

FiciURE 10.—I.arüC niinilicrs of a Staphylococcun species seciircrl in dilution plate.s from fermenting tobacco,
and tlie relative size of these colonies as compared with three colonies of large bacteria.

cooperation with the Wisconsin Department of Bacteriology indicate

that it belongs to the Staphylococcui^ group. While the regularity of
occurrence and the niunerousness of this organism in fermenting
tobacco suggest some relation to the fermentation process, it also
failed to induce thermogenic activity in tobacco imder the conditions
of the experiments.
The fungi encountered in stored and fermenting tobacco are nor-
mally fairly common species in nature, though little is known about
what significance varieties or strains of these fungi may have in the
present connection. Aspergillus flavus Lk., A. terreus Thoni, and
Pénicillium brevicaule Sacc. were commonly found in fermenting
cigar-leaf tobacco and were used in connection with the present
156 Journal of Agricultural Research Vol. 49, no. 2

studies. A. ochraceus Wilhelm, a form growing abundantly on a

sample of flue-cured tobacco, was included in the studies, as was
A, niger Tiegh., the causal organism of black rot of fermenting
tobacco. These species were either determined or verified by Charles
Thorn, of the Bureau of Chemistry and Soils, United States Depart-
ment of Agriculture, from cultures submitted. In contrast with the
bacteria, these fungi were found to be very active thermogenically
on tobacco and consequently should be given more consideration as
possible factors in fermentation, whereas previous students of the
microbial hypothesis considered the bacteria of major consequence.
TABLE 7.—Fermentation results in Dewar flasks with tobacco from various cigar-
tohacco growing districts i

Average tempera- Estimated cocci

ture Car- to square inch
Am-
Source of tobacco and year Mois- bon Odor
Variety mo-
grown ture nia di-
Incu- Flask In- oxide Con-
In flasks
bator crease trol

Per- Num-
cent ° C. ° C. ° C. Number ber
Ohio (1931) Dutch.. 39.0 23.9 26.9 3.0 +++ +++ Strong. 100,000 50,000
Do Zimmer. 38.3 23.9 26.4 2.6 +++ +++ .-.do—. 150,000 80,000
Connecticut (1930). Havana 38.5 23.9 25.2 1.3 + + Mild.-. 100,000 1,000
seed.
Connecticut (1931) ...do..... 35.0 23.9 26.6 2.7 +++ +++ —do—.
Good-. 100,000 200
Massachusetts (1931) ...do.-... 38.0 23.9 26.1 2.2 +++ +++ 80,000 0
Canada (1929) ...do-.... 44.2 23.9 27.2 3.3 ++ +++ —do-.- 100,000 3,000
Wisconsin southern (1931) _ .-.do-—, 27.3 23.9 23.9 0 + 0 Raw--. 30 0
Wisconsin northern (1931). .-.do-—, 33.0 23.9 26.7 2.8 + Slight- 100,000 30
Pennsylvania (1931) .--do—. 41.2 23.9 29.0 5.1 +++ +++ Strong . 150,000 40,OOD

1 See footnote 1, table 1.

TABLE 8.—Effect of pure cultures of various organisms on sterilized tobacco in

Dewar flasks ^

Average
tempera- Ammo- Carbon Perox- Color of
Inoculum Moisture ture in- dioxide idase extract
nia
crease

Percent ' C.
Aspergillus niger 37.0 4.2 ++ + 0
Aspergillus ochraceus... 36.1 3.6 +++ + 0 +
Aspergillus terreus.. 37.7 6.1 +++ ++ 0 +++
Aspergillus flavus 37.8 2.8 +++ + 0 +++
Pénicillium brevicaule... 40.0 2.7 +++ +++ 0 +++
Pénicillium baiiolum— 37.5 .3 ++ 0 0 +
Oospora sp 39.1 .6 ++ 0 0
Bacterium sp. 1 38.7 2 . .2 0 0 0 +
Bacterium sp. 2 40.1 .1 0 0 0 +
Siaphylococcus sp 39.0 .1 0 0 0 +
Not inoculated; control 38.1 0 + 0 0 +
Not sterilized; control.. 39.8 .. 1 +++ +++ +++ ++
1 See footnote 1, table 1. 2 Decrease.

By using the cellophane bags in the Dewar flasks, it was possible to

first eliminate all possible agencies of fermentation by heating the
tobacco for the required length of time. The above-mentioned
organisms carried in pure culture could then be subsequently intro-
duced for comparative studies. The bacteria, as has been noted,
caused no appreciable thermogenic activity in tobacco, although the
Staphylococcus form developed in large numbers. Neither did the
July 15,1934 Studies on the Fermentation of Tobacco 157

bacteria affect any other measurable changes in the leaf under the
conditions of the tests. Certain fungi, however, not only usually
induced a striking generation of heat, and a resultant production of
ammonia and carbon dioxide, but developed odors and other charac-
teristics fairly typical of fermenting tobacco (table 8). Other fungi
evidently were capable of httle or no influence of a fermentative
nature. Detailed studies are required, however, before satisfactory
conclusions can be drawn in regard to the comparative activity and
actual influence of these and other fungi on the fermentation process
under both experimental and commercial conditions.
DISCUSSION OF RESULTS
As understood in the present investigation, the fermentation of
cigar-leaf tobacco is an oxidation process resulting in the formation
of carbon dioxide and the generation of heat, together with a general
improvement in the desired qualities of the tobacco, usually accom-
panied by the production of ammonia. The influence on quality
may be of various kinds, but in general the process eventually removes
a ^^green'' or raw taste and odor and develops an aroma. The
immediate result may often be the formation of such amounts of
ammonia and other products as to render the leaf excessively strong,
and the benefits of subsequent storage and aging are in part at least
due to the gradual loss of such objectionable substances.
The oxidation process may no doubt be brought about in different
ways or at least at markedly different rates, depending upon the en-
vironmental conditions. The results secured in the present studies
have not eliminated any of the hypotheses held at present as to the
nature of the causative agent in fermentation but have rather added
some support to both the enzymic and microbial hypotheses. The
latter hypothesis has been in danger of being totally discarded from
consideration without all the evidence being taken into consideration,
Loew's {12, 13) conclusions, for instance, are particularly weak with
respect to microbial behavior in tobacco, though his evidence for
'^ oxidizing enzymation" is convincing. Evidently the chief basis for
denying any essential relationship between fermentation and micro-
bial activity is the low moisture content present in tobacco types which
are "reordered'' or redried before being placed in storage. Flue-
cured tobacco, in particular, is much more susceptible to fungus de-
velopment than is cigar tobacco, and unless stored in a condition
relatively dry as compared with storage conditions for cigar-leaf
tobacco, usually becomes overgrown with fungi to a damaging extent.
However, neither the results nor benefits of fermentation as contrasted
with those of aging are clearly defined in redried tobacco, and it is
quite conceivable that a slow oxidation, entirely independent of micro-
organisms, may proceed in such tobaccos. On the other hand,
millions of pounds of stemming tobacco and cigar tobacco are regularly
stored or processed in such a manner as to render microbial develop-
ment certain. It is not unlikely that this procedure has some essential
relation to the desired results. Whether this procedure is to be re-
garded as of the nature of fermentation or special processing may be
open to question.
While it has been claimed by Jensen iß) and by Smirnov {20), on
the basis of trials with antiseptics, that fermentation may proceed
75317—34 5
158 Journal of Agricultural Research voi. 49, no. 2

in the absence of microbial activity, the data presented are not con-
vincing on this point. Evidently the selective influence of various
antiseptics on enzymes and microbes under different circumstances
needs more intensive study. Silver nitrate appears to possess this
differential property more completely than any of the other antiseptics
used, and, on the basis of the results secured with this chemical, it
seems evident that fermentation may proceed in the absence of micro-
bial activity. On the other hand, fermentative changes of an ap-
parently normal character may evidently be brought about by various
fungi in tobacco heated sufficiently to destroy all other forms of
activity. Bacteria, which have previously been generally believed
to be concerned in this process, were not found to be capable of de-
veloping thermogenic activity.
SUMMARY
Microbial and enzymic factors which may be concerned with the
fermentation of cigar-leaf tobacco were investigated by the Dewar-
flask method. Modifications of technic were used which permitted
of operations under aseptic conditions and inoculations with pure
cultures of micro-organisms.
With 150 g of tobacco, average temperature increases in Dewar
flasks over a 10-day period ran as high as 5.6° C. In general, the
generation of heat (thermogenesis) was directly proportional to the
percentage of moisture present, a minimum of about 30 percent being
necessary to secure measurable activity under the conditions of the
experiments.
The highest thermogenic increases occurred at incubator tempera-
tures of about 20° to 25° C, very little if any activity developing at
temperatures below 10° C. (50° F.) or above 45° C. (113° F.). This
experimental maximum temperature is considerably lower than the
temperature often allowed in practice by the bulk-fermentation
method, and suggests the possibility of obtaining better results at
more moderate temperatures.
Chloroform, mercuric chloride, acetone, toluene, or beta-napthol
may almost completely check thermogenesis. These antiseptics
check microbial activity, and although they do not destroy peroxi-
dase and other enzymes, they apparently cause inactivation. Silver
nitrate, on the other hand, reduces thermogenesis under simüar con-
ditions to only about one-half that which normally obtains. It is
believed that süver nitrate prevents microbial activity without being
harmful to the action of enzymes under the conditions of the ex-
periments.
Heating the tobacco to a sufficientlyhigh temperature will completely
check thermogenesis or any other expression fermentation, provided
aseptic conditions are maintained. Treating such tobacco asepti-
cally with porcelain-filtered extract from unheated tobacco con-
taining peroxidase and other enzymes failed to induce thermogenesis.
Three species of bacteria commonly occurring on tobacco failed to
induce thermogenic activity in heated tobacco, but several fungi iso-
lated from tobacco were very efficient in this respect, yielding results
thermogenically and otherwise comparable to normal fermentation.
The evidence appears to justify the conclusion that micro-organ-
isms, especially fungi, may play a role in the fermentation or in the
July 15,1934 Studies OU the Fermentation qf Tobacco 159

processing of certain types of cigar-leaf tobacco, although such or-

ganisms may not necessarily be essential to fermentation.
LITERATURE CITED
(1) BEHRENS, J.
1896. DIE BEZIEHUNGEN DER MIKROORGANISMEN ZUM TABAKSBAU UND
ZUR TABAKFABRIKATION. Centbl. Bakt. [etc.] (II) 2: 514-527,
540-545.
(2) BoECKHouT, F. W. J., and VBIES, J. J. OTT DE
1909. ÜBER TABAKSFERMENTATION. Centbl. Bakt. [etc] (II) 24:
496-511.
(3) DXVALOS, J. U. ^ . ^^ ^.
1892. NOTA SOBRE LA FERMENTACIÓN DEL TABACO. Cromca MedlCO-
Quirurgica Habana no. 15. [Abstract in Centbl. Bakt. [etc.]
(I) 13: 390-392, 1893.]
(4) FESCA, M.
1888 UEBER KULTUR, BEHANDLUNG UND ZUSAMMENSETZUNG JAPANISCHER
TABAKE. Landw. Jahrb. 17: [329]-272,
(5) JENSEN, H. n +KI t» i +
1908. UEBER DIE NATUR DER TABAKFERMENTATION. Centbl^ Bakt.
[etc.] (II) 21: 469-483, illus.

1915. PROEVEN OVER TABAKSFERMENTATION IN ''DEWAR'SCHE VATEN''.

Proofsta.. Vorstenland. Tabak [Dutch East Indies], Meded.
12: 22-38.
(7) JOHNSON, J.
1914. BLACK ROT, SHED BURN, AND STEM ROT OF TOBACCO. WlS. Agr.
Expt. sta. Research Bull. 32, pp. [63]-86, illus.
(8) and MURWIN, H. F.
1925. EXPERIMENTS ON THE CONTROL OF WILDFIRE OF TOBACCO. WlS.
Agr. Expt. Sta. Research Bull. 62, 34 pp., illus.
(9) JöRGENSEN, A. p. C. T. , K X • j
1925. MICRO-ORGANISMS AND FERMENTATION. Ed. 5, reset, revised
throughout ..., 467 pp., illus. London.
(10) KOLLER, J. B. C.
1858. DEB TABAK IN NATURWISSENSCHAFTLICHER, LANDWIRTSCHAFT-
LICHER U. TECHNISCHER BEZIEHUNG ... 134 pp. Augsburg.
(11) KONING, C. J.
1900. DER TABAK; STUDIEN üBER SEINE KULTUR UND BIOLOGIE. 8D pp.,
illus. Amsterdam and Leipzig.
(12) LOEW, O. TT CI T-k 4.
1899. CURING AND FERMENTATION OF CIGAR LEAF TOBACCO. U.b. Ucpt.
Agr. Rept. 59, 34 pp.
(13) TTQ
1900. PHYSIOLOGICAL STUDIES OF CONNECTICUT LEAF TOBACCO. U.Ö.
Dept. Agr. Rept. 65, 57 pp.
(14)
1901. CATALASE, A NEW ENZYME OF GENERAL OCCURRENCE, WITH SPECIAL
REFERENCE TO THE TOBACCO PLANT. U.S. Dept. Agr. Rept. 68,
47 pp., illus.
(15) MicioL.
1891. NOTE SUR LES VÉGÉTATIONS QUI SE DEVELOPPENT PENDANT LA
FERMENTATION DU TABAC. Mem. Maufr. Etat, Tabacs—Allu-
mettes 2(2): [182]-191.
(16) MiEHE, H.
1931 THE PRESENT STATUS OF THE PROBLEM OF THE SPONTANEOUS
HEATING OF HAY. Fifth lutematl. Bot. Cong., Cambridge,
Rept. Proc. 1930: 45-57.
(17) NESSLER, J. ^ 1 CI
1867. DER TABAK, SEINE BESTANDTHEILE UND SEINE BEHANDLUNG. iDi
pp. Mannheim.
(18) NORMAN, A. G.
1930. THE BIOLOGICAL DECOMPOSITION OF PLANT MATERIALS. PART III.
PHYSIOLOGICAL STUDIES ON SOME CELLULOSE-DECOMPOSING
FUNGI. Ann. Appl. Biol. 17: 575-613, illus.
160 Journal of Agricultural Research voi. 49, no. 2

(19) SCHLOESING, T. H.
1888-89. SUR LA FERMENTATION EN MASSES DU TABAC POUR POUDRE.
Mem. Manfr. État, Tabacs—Allumettes 1: [514]-552, 1888:
2: 119-136, 192-210, 1889.
(20) SMIRNOV, A. I.
1933. THE PHYSIOLOGICO-BIOCHEMICAL PRINCIPLES OF TOBACCO CURING
AND FERMENTATION. 507 pp., ülus. Krasnodar. (Tabako-
vedenie t. 3).
(21) SUCHSLAND, E.
1891. UEBER TABAKSFERMENTATION. Ber. Deut. But. Gesell. 9: 79-81.
(22) VERNHOUT, J. H.
1899. RAPPORT OVER HET BACTERIOLOGISCH VAN GEFERMENTIERDE
TABAK. Ber. Lands Plantentuin 34. [Abstract in Centbl.
Bakt. [etc.] (II) 4: 778-779.J
(23) WHITNEY, M., and MEANS, T. H.
1899. TEMPERATURE CHANGES IN FERMENTING PILES OF CIGAR-LEAF
TOBACCO. U.S. Dept. Agr. Rept. 60,28 pp., illus.