TURIiISH TOBACCOS Characteristics and Chemical Composition of Imported Types F. R. DARKIS, E. J. HACKNEY, AND P. 31. GROSS Duke University, Durham, 9.C . The small leaf size of these recent years. T h e chemical analyses of tweiitj -four constituents or tobaccos is obtained by growexcept for a period groups of constituents of imported oriental tobaccos are ing the plants on relatively during the war, over 50 milpresented. The analjses are for tobaccos of karying grades poor soils and by planting a lion pounds of tobaccos of from eighteen areas scattered over four of the main grou the Turkish type lvere inilarge number of plants-40,ing regions of Greece and Turkey; the tobaccos were grown 000 to 70,000-per acre. If ported into the United States in 1937 and 1938. These analjses show significant differfertilizer is used it conimonly annually ( 2 3 ) . These toences in chemical composition among the tobaccos from consists of goat or sheep dung. baccos are normally gron-n in the four regions and among grades. They also show some The growing season after Turkey, Greece, Bulgaria, differences among tobaccos from neighboring areas w ithiri transplanting and during cura n d southern Russia. E x the regions. These differences indicate a basis for the comcept for some regions ill ing is usually dry and warm mercial practice of di*iding oriental tobaccos into type4 Bulgaria, they are grown in in these regions. dnd diwisions of tjpes along geographical lines, and in turn The totiacco is harvested aread relatively close to t h r . reassembling these to obtain a blend of more constant shorcs of the Black, Aegcmi, by priming, the 1eavr.s being composition. idiscussion of chemical analjses as the) arid Mediterranean Seas. picked a few a t a timeas they appear to be related to the commercial usages and evaluaThese tobaccos differ priripen. After priming they are tion of these tobaccos is presented. 41so the chemical marily from domestic tobacstrung by piercing the large analyses are correlated, to the extent possible, with the cos in that their leaf size is end of the midrib with a large existing climatic and soil conditions and the cultural pracsmall and the intensity of needle and pulling the leaf tices in \ o p e in each geographical region. their aroma is greater. onto the string. After the A large proportion o f these strings of leavcs are altobaccos is imported to be incorporated in blends of popular lowed to \yilt for 1-3 (lays in a cool shady place, they are placed brands of blended cigarets to improve their burning quality and in the sun ana air and cured by means of these agents. During aroma. These tobaccos cost the Smerican manufacturer more r d t i n g and curing the leaves change from green to a b r o m or than the average domestic tobaccos. Part of this increased cost yellow color and lose most of their natural moisture content. is the import duty. After the leaves are cured they are stored in a protected place until In view of the importance of this crop to the cigarct industry as they take up moisture again with the coming of damp autumn a whole, and because of the significance that a more detailed weather. They arc then made into temporary bales and dclivknowledge of these tobaccos might hold for the American tobacco ered by the grower t,o the dealer or exporter. farmers, a program o f research was initiated in 1939 at Duke The tobaccos from the top part of the stalk are usually conUniversity t o study these tobaccos. This included an investigasidered to be the best in quality, and those from the basc the tion of the possibility of producing a type of tobacco in the United poorest. From 20-5053 of the production of the plants of most States with properties similar to those of imported Turkish tobaccrops is not imported into the United States because its quality cos. It also included a study of the chemical composition of imdoes not warrant the payment, of import duty. ported types and of domestically grown aromatic tobaccos to The tobacco is delivered by the grower to the dealer or exporter permit comparisons betiveen the t\vo types and to increase our for “manipulation.” This consists of sort,ing,giading, and baling knowledge of each. T o do this numerous samples of tobacco the tobacco in the preferred manner, for shipment and to faciliwhich were grown in Turkey and Greece in 1937 and 1938 were tate fermentation. I n recent years most Turkish tobaccos imobtained from the importers and analyzed. The absence of ported into the United States are baled in the so-called Tongas chemical data in the literature for Turkish tobaccos imported into bale. This consists in placing the loose leaves in bales under this country, and the significance of these data for a large and inipressure and sexing burlap covers securely to the bales. The portant industry, indicate the desirability of publishing this inbales vary in weight from 70 to 125 pounds. formation. The tobacco is usually stored for two or more years before it is These tobaccos are grown in areas surrounding villages and used by the manufacturer. During storage it undergoes fermcntowns in n-hich the people live who produce them. The tobaccos tation. In most instances this begins with the coniing of warm are usually named after one of the main towns in the area. They weather and proceeds for a period of varying length depending may also be known by the name of the seaport from which they upon the temperature conditions and the tobacco. I t ceases x i t h are shipped or the city in which they are prepared for shipment. the arrival of cooler weather. The tobacco may undergo fermenThus, the tobaccos grown in the numerous villages near Serres. tation each summer until it is used. The temperature of the toGreece, are known as Serres tobacco, and those grown in the vilbacco usually exceeds that of its surroundings during fcrmentalages in the general area of Drama, Greece, are known as Drama tion. REVIEW OF LITERATURE tobacco. These tobaccos, as well as others, are shipped out of the port of Cavalla and are often known in the trade as Cavalla toThe literature contains but few data on the chemical coniposibaccos. The name may also arise from the topography of the tion of tobaccos of the Turkish type. Because of the limited region in which they grow; thus Djebel refers to mountainous scope of most o f the investigations recorded in the foreign literacountry, and Yaka refers to the hills in the mountainous regions. ture, they are not comparable, in many instances, to the data prcThe name may also be derived from the name of a section of terrisented in this paper, nor do they aid in the interpretation of these tory such as Souyalassi or Pravi. data.
1)URIKG
-
1631
1632
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 39, No. 12
Grading Room i n Turkey
Andreadis and T o o k (3),n-orking with Greek ioixic.iw-, 4 i i r w that the nicotine increases from the lower to t h c s interriirdiate leaves on t h e stalk a n d again decreases in the top leave.. In another paper (8)the same authors show t h a t the total nitrogeri content follows the course of the nicotim and that protein nitiugen increases in the leaves from the base to tlie top of the s t d k . T h e work of Pyriki ( f 9 d ) iiidicates t h a t the better grntlcl. of Turkish tobaccos have a lower pH value than the poorer gradci. Vladescu and Diinofte ($4,$5, 26) determined iota1 nitrogen, protein nitrogen, nicotine, soluble carbohydrates, and ash for Turkish, Greek, a n d Bulgarian tobaccos. Their data indicate t h a t the better grades of tobacco possess the lower ash anti nicotine content and t h e higher carbohydrate content. They alw indicate, with some esceptions, a higher total nitrogen content in the poorer grades of tobacco. These authors find the total nitrogen, protein nitrogen, a n d nicotine content of tlie Bulgarian tobaccos t,o be lower t h a n that of the Greek and Tur,liish tobacco, whereas the carbohydrate content of the former is higher. T h e ash content of the Greek tobaccos is somewhat greater than t h a t of tlie other two. Their data show that the Sniyrna tobaccos are lower in content of nitrogenous materials a n d higher in carbohydrarcs t h a n rhe Samsun tobaccos. Also the Djebel tobaccos from both t h e Greek and Bulgarian regions arc lo~verin nicotine contctnt and higher in sugar content than any other t o b 2 C C r J S of the respective regions. Iiadir (f3) discusses the blending of Turliisli t(Jb!i?coSfur cigaret manufacture a n d gives the anal)-& of tobaccos of average grade from many sections of Turlicy. His figures ahon. t h a t the compusition of tobaccos growing in different niitjor regions, such :is Samsun on the Black Sea and Izmir i i i Southern Turlicy, may vary widely and, furthermore, t h a t there may be considernble differences in the coniyosition of tobaccos from differynt areas within the same region. Iioseraif (14) gives some analyses of Turkish tobaccos which lead to similar conclusions as those d r a v a from Kadir's data. METHODS OF ANALYSIS
MOISTURE. Two-gram samples in aluminum dishes iverr dried over concentrated (above 91yG)sulfuric acid at 30" C. f o r 14 days. HYGROSCOPICITY. The dried samples from the moisturc determination \yere placed in a n atmosphere of 72% relative liuniidity at 30" C. for a period of 14 days. The increase in weight on a percentage basis was termed the hygroscopicity.
.
PETROLLL-M ETHERESTR.ICT. Five-gram samples of the ground rriaterial n'ere extracted for 23-24 hours with petroleuni ether (boiling point 30-80" C.) on a Bailey-Kalker extraction ailvias dried to a constant weight a t 95dr.r.o1iu~EXTRACT. The r e d u e from the petroleuni ether estract i~:iscstrartcd 23-21 hours x i t h (35(;) ethyl alcohol on a Baiiey-T\+alkcr extraction apparatus ( 2 7 ) . T h e residue was dricd to constant, weight at 100-102' C. in a n electric oven. STARCH. T h e residue from the alcohol estract was used for the determination of starch by the diastase method with subsequent acid hydrulysis (.4, p. 120). SICOTIXE. The Iieller method as modified by Garner icas usod
!fO, 1 1 ) . PROTEIS SITROGES. Tn-o-gram samples were boiled for 10 minutes Lyith 50 cc. of 0.5:; acetic acid, a n d the mixture was filtered when cool. The residue was washed with hot 0.5yoacetic acid until the filtrate was colorless. T h e nitrogen in the residue was determined by the Iijeldahl-Gunning-Arnold method (4,p. 8). TOTAL SOXVOLITILE ACIDITY.Five g r a m of tobacco and 6 cc. of 6 AI hydrochloric acid \wre mixed into a homogeneous mass, and finely divided neutral pumice stone was later worked into it until a semidry misture was obtained. This mass was estracted with alcohol-iree ether for 40 hours or more in a Soshlet estractor. T h e ether was removed b y the addition of boiling water, and t h e acid-containiug solution was boiled for 5 minutes to remove any volatile acids. The solution was made to a volume of 230 cc. a t rooiii temperature. hliyuots of 10 cc., to irhich 100 cc. of water had been added, were titrated for acidity with 0.1 allrali using phenolphthalein as indicator. Chlorine was determined on other 10-cc. aliquots by the l l o l i r (161method and its acid equivalent subtracted from the alkali titration. The results are expressed iri the number of ec. of 0.1 S alkali requiyed t o neutralize the acidity in 1 grain of tobacco. T k c amino nitrogen, wat oluble nitrogen, total reducing substances, tots1 reducing sugars, and total sugars were dctermined o n a n estract made by extracting 25 grams of t,ol>accowith 376 cc. of water in a Mason ,jar, wit,h the addition of 1 cc. of ehloroforni a t room temperature for 12 t o 14 hours, with occasional shaking for the first hour. The pH was determined on a n extract of the same proportions of tobacco and water but with the omission of the chloroform. T h e extracts were filtered through a linen cloth or a plug of glass wool. +\-
INDUSTRIAL AND ENGINEERING CHEMISTRY
December 194?
HYDROGENIONCOXCESTRATIOK.The Coleman glass electrode was used. The results were espressed in terms of pI3. REDUCINGASD TOTAL SGGARS. Reducing sugars were determined before and after hydrolysis by the Nunson-Kalker method (4,p. 190), and the results expressed in terms of glucose. The estract was clarified with neutral lead acetate. TOT.ALREDUCIXGSUBSTATCES.Total reducing substances were determined on the estract without clarification and after hydrolysis by the hlunson-Walker method (4, page 190). The difference between the total reducing substances and the reducing sugars was called polyphenols'. 1 Wright ( $ 9 ) points out t h a t a n alternative interpreta:ion of the so-called polyphenol content is possible in view of the statements in the literature. Pyriki (19) states t h a t "it is perhaps correct" t o consider t h a t the difference between the total reducing substances and the total sugars are the polyphenols. Our d a t a include values ior both the reducing sugars and total sugars, which makes it possible t o calculate t h e polyphenol values either w a y . After reviewing the confused state oi this matter in the literature we are of the opinion t h a t a more correct value for the content oi phenolic compounds would be obtained if t h e extract n-ere hydrolyzed f i r s to convert all Boluble sugars a n d soluble phenolic compounds into iorms in xhich they would reduce Fehling's solution, and then determine the reducing power before a n d after clarification. If the clarification procedure removed all reducing materials other than soluble reducing sugars. the difference between the t w o values thus obtained for reducing power should more correctly represent the polyphenol content.
1633
TOTAL NITROGEN. Kjeldahl-Gunning-Arnold method niodified to include nitrates (4,p. 9) was used. ~ ~ V A T E R - S ONITROGEN. L U B L E Kjeldahl-Gunning-iirnold method (4,p. 8) was used. h m r o SITROGES. Van Slyke method (4, p. 245) was used. S O L ~ B ASH. L E Five-gram samples in tared porcelain crucibles were heated at 250" to 300" C. for 3 to 4 hours, then heated to a temperature of dull redness in a muffle for 14 to l G hours, cooled to room temperature in a desiccator over calcium chloridc, and weighed as total ash. The sand was determined, the difference between the total ash and sand considered t o be soluble ash. SASI), SILICA, OXIDES O F I R O X AND .kLUMINGX, CALcIU31, . I N D These were determined by the officinl methods \IAGTESIrlf. (4,PP. 39-41). Two-gram samples were dissolved i n 30 cc. of PIIOSPHORUS. concentrated nitric acid and 6 cc. of concentrated hydrochloric acid; organic matter was destroyed by boiling and phosphoru! p. 3). iicterniined by the volumetricmethod (4, SGLFUR.The magnesium nitrate method was used (4, p. 45). CHLORITE.This was determined by the official volunietric method (4,pp. 43-4). PoTassruu. This was determined by the Lido-Gladding method (4,p. 42). SAMPLES
TABLE I. REGIONS, AREG, REGIOS Eamsun (Black Sea)
Tows3
OR
T~ILL\GE,>
AREA Djannik
T o w x s ASD VILLAGES Kehyali, Haji, Ismailoglou, Sitma, Suyu, Tash Demir, Baldjali, Yagh Bassan. Harnzslu. Kara Oglan, Iiokdje, Iiisbla, Buyuklu, Iiir
hladen
TekiyKeuy, Tokari Tsinik, P a p a z LIahallesri, Tsinik, Drgherish, Assar Agatch, Okse, Andria Rlah, Tsirakman, Jlyaskeuy Iiertme, Dagh Keuy, Dnz Keuy, Karagol, T a r lan Rlanados, Kavadjik Oyoumdja A d o u l a r , .ilanos, Tchobanli, B'allatch, Dered: jik, Kadi Keuy A k Teke Elifli Kara Keuy Teke Sarmousak, Dedeli: Oreddjik, Ak, Ghuneyi, Elmadjik, Koushlaghan, Kovanlik, Martakala, Dervent, Sourmeli Deressi, D a r Boyaz, ICush Iiayassi, Domdz .ighi Soma, Kirkagac Akhisar, Gordus, 3Ianisa Sindirgi, Yenick, Selendi, Marmara. .ilibeyli: H a r t u , Yayakoy, Suleymanli, Derekoy. Gelenbe LIenemen, Foca, Dikili, Bergama, Ayvalik, Gomec, Jonuzlor. Dernirtas, Acan, Sanakoyu Iiozah, Caniamli, Akcenger, Kasaly, Guzel:
Pidiii
Evgaf
Baira
Smyrna (Aegean Sea)
i m
Akhisar
Foca
h:cnr Y L I Y L
Iamir
Mugla
Eastern Greece
Comontini
Izmir Torbali Buca, Seydikoy Cuniaovasi De;eli, Tarbkli, Kayas, Burno;.a, Dikardas: Urlan, Cesme, Iiaraburun, Salman, .ilcati, Drniaviren selc;hl-' ~ - - ~ ~ hlilas, l I u g l a , Bodrum, Akcaalan T a h k a v a k , I i u r u k o y , Asin, Aydin, Yoran: Akkoy,-Soke, Ula. .%,lacam, Sozkoy, Iiaracahisar, Iiemikler, Kazikli Komotini. Doukatan. Kasmion. Pandrossas.
Xanthi (2nd P a k a ) zolou, Daoutlu, Kipseli, Kiosse Ali Djebel Kechrokambo, Lekani, Dipotumon. Plaiamonia, Ptelea, AIakaklov. I r o n a o u t , Sarnovista. Yasriani. Isqidie Souyalassi Kallithea, Dafndn, Stavroupolis, Sarnitz, Ada, Kourlar., I pnikiiii .~~ ~.. . Drama Drama Doxaton Iioudounia Choristi Kabanbaki,'Fotoliri, Bnriani, Radurista, Tkataldja, T C..i.r n..t w-. l i. Prai-a Dryena, Fteri, Amfipolis, Pravi, Avli, Kipia, Elefthera!, Pangaion, llessoropia, Bostandjil, Sikissiani, Drani tsa Zihna Aliatrari, Sfelinos, Zihna, 'Iholos, Drarsova, Vitasta, Anghista, Rodoliros, Kotsakioi, hIyrkinos, 1-ikissiani, Rahova, Draviskos, Vultsista, Kioupkioi, Hororista Serres Kastri, Kutsos, Rahnianli, Ichinos, Xigrita, Diinitritsi, Serrai, h-eos. Skopar, Homondos. Christoc, Drano\.a, Lacos, Eidirokastron Iiimissis, Karmaroto, T'altero Strymonikon' Staros, Subaskioi Dol-iqta X'Lsmik, Sormou: sakti, Rahovisiia,'Topalian?l Prossotsian Granitis, Gornitsa, Egri-Dere, Procsotsian Rei-ika, Viosatian, Plevna, D r a m a , Koublisrn: Kirlihova Agrinion Agrinion, Stamns, .ingkelokastron, Spoluita Papadatais, Gal-alou, Analipsis, Thernon: Drymonas, Lepena, Seochori ~
__
Agrinion (Testern Greece)
I
Each sample was taken from one individual bale, selected at random from a lame number of bales available from each of the regions considered. The covers were removed from the bale and a section (1/20-1/10) of the bale was removed and taken to the laboratory. P a r t of this section was ground on a TViley mill to pass a 30-mcsh sieve, thoroughly mised, and sealed in a glass jar until used for analysis. The stock of bales available for sclection consisted of bales of grades 1, 2, and 3, grade 1 being the best in quality. The number of bales of grade 3 was greater than the numbw of grades 1 and 2. The tobacco in each of thc bales selected for analysis was prob%hi>-a mixture of leaves from numerous growers from several viilages in the area surrounding thc town for which thc tobacco is named. I t is possible, however, t h a t the tobacco in any bale may have come from a restricted locality. The tobaccos of grade 1 may or may not have been from the same part of any c o ~ i i ~ ~ i i i n ity as those of grades 2 and 3.
-
TERRITORIES
For this work tobaccos from four marketing areas in e:tch the Samsun and the Smyrna tobacco-producing regions of Turkey, nine areas in the hlacedonian and Thracian regions of Greece, and one area in the hgrinion region of Greece were selected. Tobaccos representative of the areas of D j m n i k , Uatien, Evgaf, and Bafra of the Samsun region, and hkhimr, Foca, Izmir, and Mugla of the Smyrna region, were selected as bring typical of those grown in Turkey. Tobaccos representativc of Comontini, S a n t h i (2nd 1-aea), Souylassi, Djebcl, Drama, Prnvi, Zihna, Serres, and Prossotsian in Eastern Greece and Agrinion in Western Greece were selected as being typical of those grown in Grcccc. The geographical locations of these areas arc shown in the accompanying maps (Figures 1 and 2), and some of the towns arid villages located in each area are given in Table I. A11 towns rianicd are not necessarily tobacco-producing centers. The villages from which the tobaccos in the Samsun region were obtained all lie x i t h i n 30 miles of the seacoast except for a scattered villages in the Bafra area. I n the l l a d e n , Djannik, sild Evgaf areas most of the tobacco is gron-n on hilly land, whereas in Bafra part of it is gro1-m on the plain. Thc region in n-hich Smyrna tobacco is grown is much larger, some of thv areas extending inland from the sea as far as 60 to 80 miles. In this region most of the tobacco is grown on hillsides or high tablelands; a part of it, hovxver, is grown on flatland in well drained valleys. f[tw
1634
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
Vol. 39, No. 12
December 1947
I N D U S T R I A L A N D E N G I N E E R IN G C H:E,p I S T R Y
MAP OF GREEK AND
1635
SMYRNAPRODUCING REGIONS AND AREAS FIGURE N0.I
The areas in the Eastern Greek reyiori from which tobaccos were selected lie within 35 miles of the A4egeansea or its arms, \r-ith the exception of parts of the Serres and Prossotsian areas. In general, this territory is mountainous, but much of it is constituted of small valleys. The tobacco-growing land extends from the banks of the streams and small rivers that flow t,hrough the valleys to the sea, into the hills and mountains that border thew valleys. The city of Agrinion is located in the tobacco-growing region of that, name in Western Greece. The tobacco was obtained from an area within a radius of 15 miles of Agrinion. The region is generally mountainous with some flat,land lying w s t of .%grinion. DISCUSSIOT
The data. for twenty-four chemical constituents or properties of each of the samples analyzed are given in Tables I1 and 111. As \Todd be expected, there is rather m-ide variation in some cases between the content of certain constituents in tobaccos of the same grade from year to year and from area to area in each of the regions. The average content of the constituents for the whole group of samples is quite similar for each of the two years. This similarity holds for each grade from year to year. These average d a t a are given in Table 11'. Also the tre.nd from grade to grade is definite when the data on all samples are considered as a whole. This trend, however, fails to hold in many individual cases. The d a t a show that the tobacco from each of the three main regions is quite different chemically. They also shoK that tobaccos from
,
different areas within a spccific region may be quite similar in chemical make-up in some cases and different in others. Information on such points as weather, production, fertilization, and soil fertility, from which a n adequate explanation of these differences could be formulated, is not at hand. The practice of many importers-obt,aining tobaccos from as many areas as possible within a region and then blending the tobaccos from each-indicates that these differences in chemical compositions reflect real differences in the tobaccos and that these are recognized in commercial practice. XITROGESGUS COSSTITUENTS.The data for the individual samples in Tables I1 and I11 show that the cont,t:nt of the nitrogenous components varies over a wide range. On the other hand, the data in Table IV shorn that the average content for these constituents is essentially the same for each of the 1937 and 1938 crops. Isolated cases, however, can be found in Table 111 where the contents of these materials are rather divergent for the two years. For instance, they are highest in the Souyalassi tobaccos of t,he 1938 crop. whereas they are highest in the Coniotini tobaccos of the 1937 crop. With the exception of the protein nitrogen, the average content of these constituents is least in the tobacco of grade 1 and most in that of grade 3. The average protein nitrogen content of the tobacco of grade 3 is greater than that of the tobacco of grades 1 and 2 ; these are essentially the same, with slightly less in grade 2 than in grade 1. This decreased nitrogen content in the tobacco from the upper portion of the stalk (grade 1 is considered to be
1636
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 39, No, 12
KAYAK
mainly from the top of the stalk) is a t variance with the behavior of cigar tobaccos (1) and flue-cured cigaret tobaccos (7) where the nitrogen content is usually maximal in the tobacco grown on the top part of the stalk. However, in respect to grade it is in accordance with the findings for cigaret tobacco ( 7 ) ,the better cigaret tobaccos being of a low and medium nitrogen content. Perhaps this decreased' nitrogen content in the top leaves may be due to the limited supply of nit,rogen available to the plant, this being due in turn to the low level of fertility of the soil and the loss of nitrogen from these leaves to the seed head. The seed head is formed late in the life of the plant and frequently is not removed from the plant. The average content of protein nitrogen in the tobaccos varies but little between the four major regions represented. The average nicotine content of the tobaccos from the Sanisuri, Eastern Greece, and Agrinion regions is similar, varying from 1.29 to 1.34y0. That of the tobacco from the Smyrna regiori, however, is much less, being only 0.96%. The average content of t,otal nitrogen, mater-soluble nitrogen, and amino nitrogen of these tobaccos varies considerably from region to region. The data in Table IV, with the exception of nicotine, s h o n that as one of these constituents varies the others vary; the gradation of these constituents from Samsun to Smyrria through the Greek tobaccos is shon-n in Figure 3. In this figure an arbitrary linear scale was selected for the total nitrogen content of these tobaccoi from !he different geographical regions; the Samsun tobaccu was arbitrarily given a value of 1, Eastern Greek tobaccos a value of 2, the Agrinion tobacco 3, and the Sniyrna 4. If the other nitrogen constituents are plotted against the same arbitrary index, the striking result is found that (with the exception of nicotine) essentially linear relations exist for these constituents. This same index of gradation in tobacco composition from re-
gion to region also gives surprisingly good linear relations for the carbohydrate type of constituents. This is shown by the curves for carbohydrate constituents in Figure 3. The juxtaposition of the curves for nitrogenous constituents and carbohydrate constituents in Figure 3 emphasizes again the relations found in previous work as characteristic of the flue-cured type of tobacco :6, 7 , 8 ) . These may be stated briefly as follows: ( a ) If one nitrogen constituent is high the others are high; ( b ) if nitrogenous constituents are high, the sugar constituents are low; and (c) if total nonvolatile acids are high the carbohydrate constituents are low. The large amount of data presented in Tables 11, 111, and IV (much of Tyhich will not be mentioned specifically in !he discussion) indicates quite clearly that the Turkish tobaccos show similar trends in their chemical composit'ion. This indicates that those factors influencing growth, which originate from the soil and climatic differences existing in the middle East region where these tobaccos were grown, operate in essentially the same manner as those influences which determine the tlifferences in flue-cured types from different regions (6). If this principle is valid, as it seems to be, it, is of considerable importance, as it will permit application of the extensive knowledge already available (6, 7 , 8)regarding the effects of varying cultural and fertilization practices on tobaccos of the flue-cured type, and lead to more intelligent procedures in producing tobaccos of t,he Turkish type. This knowledge has already been applied in field work now in progress. Thus it has been possible to reduce greatly the amount of phosphorus availabl? to the plant so as to delay flower formation. This results in a elover maturing plant, which in turn gives the grorrer a longer period over which to spread the manual work required in harvesting. The percentage of total nitrogenwhich is water soluble is greatest
December 1947
INDUSTRIAL AND ENGINEERING CHEMISTRY
1637
1638
INDUSTRIAL AND ENGINEERING CHEMISTRY
FIGURE
N0.3
T O T A L NITROGEN
O WATER SOLUBLE NITRO(
.‘‘w G
PROTEIN NITROGEN
A AMINO NITROGEN
2’
N
/ ’
,.d TOTAL /’ _’’
ACIDS ALCOHOL LXTRACJ;20
0 TOTAL S U G A R S
I COT I NE
ircoiitc,rit may be affected considerably by the weather conditions ( 7 ) which prevail 3lioPtly before harvest. The absence of weather data precludes any adequate discussion or valid coirelations in this case. The data in Table IV show that tobacco of the 1937 crop na5 O.55Yc higher in the extiact than that of the 1938 crop, and that grade 1 contains 1% less of the extract than grades 2 a n d 3 . They also show that the tobacco from the Sambun region contains the least amount of the extract, whereas that from the Greek regions contains the greatest amount. The data in Table I1 show that the extiact content i5 consistently less in the tobaccos of the 1938 crop from the Smyrna region. This w u l d tend t o indicate that dryer, hotter weather prevailed in 1937 in this region. For the Samsun
for the grades within a specific area of the region. The data in Table I11 show a general definite trerid tubvard a greater content of the extract in the 1937 crop lor the EasternGrcektobaccos. They also show that the tobacco from the Djebel and Souyalassi areas contains much less of the extract than the tobaccos from any other area of the region. Table I11 shows that the content of the extract varies from 2.82 to 8.76e; for individual samples. These variations among samples are much greater for the t,obaccos from the Eastern Greek region than for those from the Samsun and Sniyrna regions, which probably iiidicates that either the weather conditions were more variable or the grading practice was less efficient in the Eastern Greek region. Other factors being equal, the smaller the petroleum ether estract the better the burning quality and the poorer the aromatic qualities. If this is correct, the data tend t o lend validity to the contention t h a t the Samsun tobaccos are most desirable for general blending qualities, while the Smyrna and many of the Eastern Greek tobaccos are desired if a greater intensity of aroma is required in the blend. ClRBOHYDRATE Ah‘D A C I D COSSTITCEXTS. Tables 11 and 111 show the wide range in content of carbohydrate material found in thede tobaccos; Figure 3 s h o w the relation of other carbohydrate constituents to one another when they are plotted in an arbitrary manner, as was done for the nitrogenous constituents. As the sugar type materials increased the acids decreased. The alcohol extract contains a conglomerate mixture of materials. The main part, however, consists of sugar, acid, protein, gum, and resinlike substances. The data in Table I V and Figures 3 and 4 show that in general the larger the amount of alcohol-solubIe material found in a tobacco, the larger the amount of soluble m g a r materials and the smaller the aniount of nitrogenous materials it contains. Figure 4 also indicates, as has been observed previously for the Hue-cured type of cigaret tobacco (?‘), t h a t as the soluble sugar content increases the total acid content decreases. I n the ease of the alcohol extract, sugar, starch, and polyphenols, Figures 3 and 4 show t h a t as one of these constituents varies, each of the others do also. Unpublished work of this laboratory indicates t h a t the aromatic principle of tobacco of this type is included in the alcohol extract. The specific constituents from which the aroma arises, however, have not been identified. T h e difference between the total sugars and reducing sugars is a measure of the disaccharides present in the tobacco. I n general, the greater the total sugar content the greater the content of disaccharides. The gradation of the sugar-containing constituents from Smyrna t o Samsun and from grade 1 t o grade 3 is clearly shown in
December 1947
INDUSTRIAL AND ENGINEERING CHEMISTRY
Figures 3 and 4, and the data in Table I\' show the average content to be essentially the same for each of the tn-o years. The average content of thcse materials for each of the areas in the Samsun and Smyrna regions varies but little. I n the case of the Eastern Greek region, however, a division occurs; the tobaccos from the western part or those from the Serres, Drama, Zihna, Pravi, and Prossotsian areas contain less of these constituents than do those from the eastern part. This again supports the practice of dividing geographically the tobaccos from this region into two classes, Cavalla and Xanthi, as is usually done in the trade. -Again, as in t h e case of the nitrogenous nxitcrials, the tobacco of the S a n t h i area is not similar in carbohvdi,atc composition to tobaccos from neighboring areas. The average starch content of the tobaccos from thc Smyrria i~r:.ion is g r c a ~ e rthan that of thc tobaccos from tlic otlicr regions, t1i:it of tlic Smisun region being least. The starch content is gwatest in grade 1 tobacco and least in grade 3. The tobacco of the 1!G7 crop contained more starch than that of the 1938 crop. The starch content of tobacco from each area in the Samsun region is I[JTV, that of each area in the Sniyrna region is high, and that of the areas in the Eastern Greck region is rather variable. In general, it is liighcr in those tobaccos grown in the castcrn part of thc rcpion I t i h general11 conccdcd (9, 28) thkit the starch content 01 f i ~ A i l yharvcswd tobacco is high \vI~il(?the solublc sugar content is ion-, that t h e soluble sugars arc formed by the conversion of sr:ti~cIi t o sugars, and that the cured tobaccos are higher in sugar and lo\vcr in starch content; unpublished work on curing in this laboratory confirms this. Air-cured tobaccos, when cured over a prolonged period, are low in soluble sugar content (5, 121, whereas flue-cured tobaccos, when cured rapidly, are high in solublc sugar content (6). It vas pointed out previously that the 1937 growing and curing srason was probably dryer and warmer than the 1938 season, and that the Smyrna region is generally much dryer and warmer than the Samsun region. Thus i t would be expected that the tobacco would cure much faster in the former region and that it probably cured faster during the 1937 season than during the 1938 season. If such were the case it is logical to attribute the variation in content of starch and sugars of these tobaccos, a t least in part, to the differencein length of t,he period required for curing. I n the hot, dry Smyrna region the length of time in curing during which the tobaccos contained sufficient moisture to allow the starch to be converted to soluble sugars was relatively short; this resulted in a rapid and incomplete conversion of starch, as n-1~11 as a short period of life in the green tissue during which the sugars could be lost by respiratio?. This would result in both higher starch and higher sugar contents, as found in the Smyrna tobaccos. The color of Smyrna tobaccos usually contains some green, which indicates rapid drying and a n incomplete loss or con3 version of chlorophyll. I n the Samsun region the climate is cooler and more moist. The period of curing is probably much longer, and the tobacco loses moisture more slowly. Consequently the period during which ' the conversion of starch t o sugars can take k place, and also the period during which the loss W of sugar (by respiration) can occur, is longer. U Therefore, i t would be expected, as found, t h a t a W a cured tobacco of less starch and sugar would I result. I n contrast to the Smyrna tobaccos, the Samsun tobaccos, which are usually some shade of brown or red, contain no traces of green. The total reducing substances were determined and the polyphenols calculated by 0 LADE obtaining the difference between the total N0.3
1-
1639
reducing substances and reducing sugars. The polyphenol coefficient (62),in g r a m of polyphenol per 100 grams of total reducing substances, was calculated for each sample. The average values are given in Table IV. Smuck (dg) claims t h a t the polyphenol content is related to flavor and aroma, and the polyphenol coefficient is related t o color and quality; the greater the polyphenol content the better thc flavor and the greater thc aromatic properties; and the higher thc polyphenol cocfficicnt the darker the color. The polyphenol content of the tobaccos studied here is least in the Samsun tobaccos and greatest in the Sniyrna and Eastern Greek tobaccos. Xniong thc gradcs the polyphenol content is largest in grade 1 and least in grade 3. The polyphcnol coefficient. is highest in the Samsun tobaccos n-hich werct darkest and ]cast in the S n i y i a tobaccos which \ve:'c the lightest in color. Total acidity content, expressed in terms of cc. of 0.1 .Y alkali required to neutralize the acid in I gram of tohucco, is, on tlic average, just the reverse of the alcoliol ( > \ i L a c and i sugars in rclative magnitude (Figurc 3). Thi. Sarnsun tob:tccos are lowest in extract and .sugar content illid highest in acids, IT-hcrcas the Snirrna totmccos are highcst in extract and sugar content an11 lon-c~stin acids. This relation in conncction with flue-cured tob:iccos 1i:is b w n pointcd out previously ( 7 , 8). The work of Richards (20) would indicate that organic acids lire by-products of respiration as carbohydrates are broken down. If this br truc, a part of t h r increased acid content of t,he Samsun t o b a c c o may be duc to this proccss ivhich results in a decreased sugar content. The average acid content of tobacco of grades 1 and 2 is less than t h a t of grade 3 (Figure 1). This agrees with the obscrvafion of Piatnitzki (17, 18) who worked with Russian tobaccos of the aromatic type. Hc concluded t h a t the content of organic acids benrs an inverse relation to quality as judged by tobacco classification. The acid content of the tobacco from each of the areas of the Samsun region is essentially the same. This is true also for each of the areas of the Smyrna region. I n the Eastern Greek region the same geographical division t h a t seems t o be correlated with differences in nitrogenous and sugar constituents holds, those tobaccos grown in the western part of the region being higher in acids than those grown in the eastern part. The hydrogen ion concentration is expressed as pH. The average p H value is largest for Smyrna tobaccos and least for the East'ern Greek tobaccos. It is less for the tobaccos in grades 1 and 2 than it is for those of grade 3. This agrees with thc ob-
FIGURE NO. 4 *TOTAL NITROGEN
0 WATER SOLUBLE NITROC G PROTEIN NITROGEN d TOTAL ACIDS
A AMINO NITROGEN
0 TOTAL SUGARS
C STA?CH
I
GRADt NO 2'
CIR
I
N .
1640
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 39, No. 12
os
0.
L 1 1
h
t-h
00
0
00 0
P--
m
00
0
hh
Fa
u
m-
cu
m h
m
ID
00 0
E
0 *
i.
w
2 +
December 1947
INDUSTRIAL AND ENGINEERING CHEMISTRY
1641
scrvation of Pyriki (19.4) t h a t the better grades of Turkish tobacco have a lower pII value than the pooror grades. , I S H Ah-D 1 I I S E R A L C O S S T I T I - E S T ? .
T h e work of Darkii and collaborators ( 7 , 8) with flue-cured tobaccos s1ion.s that the mineral constituents in soniv tobaccos are localized in certain parrs of the plant and that thc water supply available to it affects the amount of ash constituents takeu fmm the soil by the plant. T h e interpretations placed on tlie data for the mineral constituents in the follon-ing discussion are subject t o error because of the absence oi rainfall data, yield data, soil analyses, and analytical d a t a for all parts of the plant. The total ash data. obtained are included here in order to afford a complete picture of the mineral constituent of these tobaccos. I n the absence of the complementary d a t a on weather a n d soil a n adequate interpretation of the significance of the mineral content of these tobaccos is not possible a t this time. However, esperiments are in progress wherein the nutrition of t h e plant is controlled a n d h o r n and the whole of t h e plants produced is being subjected to analysis. Complete weather and soil d a t a are available. When the complete d a t a from these esperiments become available it is believed that it will be possible t o give a t least a partial interpretation of the significance of the different trends in the mineral content of the tobaccos of the various types discussed in this paper. The folloiving condensed figures show the ! d e range over which mineral constituents vary: soluble ash 9.30 to 16.74, silica 0.47 to 3.12, calcium oxide 2.74 t o 6.37, magnesium oxide 0.61 to AllOa 0.34 to 1.37, 1.26, potassium oside 1.96 t o 5.00, Fe& P20s0.44 to 0.96, chlorine 0.15 t o 1.34, and sulfur 0.25 to 0.58y0. The average ash content as well as the average content of silica, calcium, magnesium, and potassium is slightly greater in the tobacco of the 1938 crop than i l ~that of the 1937 crop. This would probably be espected ( 7 , 8)if the 1937 season was dryer than the 1938 season, as suggested. The average content of soluble ash, magnesium, and potassium is somen-hat higher in the tobaccos from the Samsun region than i t is in those from the Smyrna a n d Greek regions. The average silica content is highest in those tobaccos grown in the Smyrna region, the average calcium content highest in the Greek tobaccos. T h e higher calcium content of the lat,ter may be due to the high calcium content of the soils on which t h e tobacco x a s grovin. The soils of Macedonia originate from the limestone formations of which the Rhodope mountains are composed ( 1 5 ) . T h e calcium contents of the tobaccos of each area in the Samsun region are about equal. This same relation holds for the tobaccos of each area in the Smyrna region; this indicates t h a t soils of areas within each region have about a n equivalent content of available calcium. T h e calcium content of the tobaccos from the different areas of the Eastern Greek region is quite variable and seems not to be correlated directly with geographical locations. It probably reflects, however, the content of available calcium in the soil of each area. T h e average content of potassium in the tobacco varies rather widely among the areas within each region a s well as among the regions. This is typical of the potassium content of tobaccos ( 7 , s ) .T h e generally increased potassium content of the Samsun tobaccos, however, would indicate a larger content of available potassium in the soils of t h a t region.
+
Workers Grading Tobacco in Greece
The generally increased ash content of the tobaccos from the Samsun region is probably due, in part, to the increased loss of carbohydrate during the curing of these tobaccos, which in turn xould give a n apparent increase in content of mineral materials. T h e tobaccos of grade 1 are lowest and those of grade 3 highest in average soluble ash, calcium, magnesium, and potassium content. These differences in relation to grade would be expected to follow this pattern because of the increased percentage of cartiohydrate materials in the tobaccos of better grade. I n general the content of iron arid aluminum in most of these tobaccos is high ( S ) , that of the tobaccos from the Smyrna region bring the highest. This would indicate t h a t the soils used for the production of these tobaccos were high in available iron or aluminum. hiany of these soils are red (15),a color indicating a high iron content. The phosphorus content of the Samsun tobaccos is greatest, a n d t h a t of rhc Smyrna tobaccos smallest. I n general, the phosphorus content tends t o be low; this would indicate soils of low available phosphorus content. The average content of chlorine is high. This nould be expected because of the fact that the soils on which the tobaccos are generally grown are fertilized by excrement of sheep and goats, which furnishes considerable chlorine t o the soil. T h e average sulfur content of these tobaccos is low (8). As the plant will take u p much more sulfur than these tobaccos contain, it is logical to assume t h a t the content of sulfur is low in the soils on n-hich the tobaccos Xvere grown. Little sulfur is added to the soil by thp escrement of the sheep a n d goats. The hygroscopicity, or the ability of a tobacco t o take up water, is of importance in determining the suitability of a tobacco for blending purposes. Therefore, the amount of water that these tobaccos xould take up in an atmosphere of 72c0 relative humidity, after being dried over concentrated sulfuric acid, was determined. 'This is purely an empirical procedure which gives relative results only. These vary from 12.21 to 17.73y0 for the acid-dried tobacco. T h e hygroscopicity does not seem t o be directly correlated with any one or any specific group of tlie chemical constituents determined. I n the foregoing discus3ion i t has been assumed t h a t these tobaccos did not possess characteristics of a n y innate nature t h a t would cause the plants from each region a n d area t o develop according t o different physiological patterns, in respect t o the final chemical make-up of their physical structure. If such is the case,
1 N D U S T R I A I, A N D E N G I N E E R IN G C H E M I S T R Y
1642
the assumption could probably be made that seed of tobacco from each area and region, if planted in the same locality, grown, and cured under similar conditions, should give a product rather similar in chemical composition to t h a t of each of the others. Incomplete and inconclusive unpublished work of the w i t e r s indicates that characteristics of an innate nature do esist between Samsun and Sniyrna tobaccos. If such is the case these characteristics probably originated as a result of the selection of natural crosses by the growers during the period, since tobacco was first introduced into Turkey shortly after the discovery of .Imerica. T h e presence of innate characteristics may cast much doubt on the validity of any of the theoretical speculations offered. They do not, however, change the fact t h a t differencesamong thechcmical compositions of these tobaccos exist. Therefore, the practice of the trade in attempting to get tobaccos of each type to blend in making blended products is justified. The data presented here shorv t h a t the tobaccos of tlic Turliish type may vary within wide limits in chemical make-up and t h a t the tobacco from any area may not be of constant chemical composition from year to year. I t also s h o w that the tobaccos of a given main region tend to be dissimilar in chemical composition from those from othcr regions. The similar data obtaiued for the 1937 and 1938 crops, TT-hcn the analyees of all samples of thc specific crops are averaged, indicate t h a t an experienced tobacco blender could maintain a blend of rather constant chemical composition, if a sufficient supply of tobacco from s e v c d crops from many areas of the different regions \vas available. LITERATURE CITED (1) Andcrson, P. J., S ~ ~ a i i h a c lT. i , R., and Stieet, 0 . I:,, (’orin. .1gr. Ecpt. Sta. Btil2. 422, 22 (1939). ( 2 ) Andreadis, T. B. and Toole, E. J., C‘hem. Zrritr., 1934, I, 3141. (3) Ahdreadis,T. B., and Toole E. J., %. C-ntwsuch. Leiiensm.. 68, 431-6 (1934).
Vol. 39, No. 12
(4j Assoc. Officiald g r . Chein.. Methods of .Inalysis, 2nd ed., 1925. ( 5 ) Bailey, C. F., and Petre. A. W., IXD.EX. CHEM.,29, 11 (1937). (6) Darkis, F. R., Dixon, L. F., and Gross. P. hf., Ihid., 27, 1152 (1935).
(7) Darkis, F.
R.,Disor~,L. Y.,ITolf, I:. A . , and Groa5, P. 11..Ibid.,
28, 1214 (193G). ( 8 ) Ilkl., 29, 1030 (1937). (9) l ~ r n l l k e n b u r g IT. , c;., .4ilCU/L (10) G u n e r , K.IT., U. d . Dept.
11
Entymoi., 6, 309 (1916:i. Bur. Plant I i i d . B u l l . 102, liY
I..,
(1907).
(11) (12)
Garner, W.IT., r.S.L)ept. .igr., Tech. 1 3 ~ 1 1 414, . 3 1 (1934) Garner. K.\I-., Baron, C. W., and Bowling, T. D., IXD. Ex(,,.
(,‘HEX., 26, 970 (1934). (18) Kadir, Gulteliin, I d ~ i s a r l a rTutun Institusit R a p o r l a r i , S o . 2 . .-lug. 1939. I141 Koscmif, E r o l , Tt~tunlcriinkBlill. 115, 2S-4S (July 1041). (1 I Laureiit, AI., “lleniorial des iiianufactures de L’etat tahnc~-alluiiietes,” J-01. 3 . P a r t 1 (1895). [ici) i r o i l r , I;., A ~ ~ R .97,335 , (1~56). i 17) Piatnitzki, S. AI., State Inst. Tobacco Inuest. (U.S.S.K.) Bull. 38 (1027j. (1s) Piatnitxki, l I . , ILid., Bull. 5 1 (1929). j 19) Pyriki, Constatitin, Z . Cnlersuch Lebensm., 73, 199 ( 1 9 3 ) . [1$lA1)Pj-riki, C o n ~ t n n r i n .ICid., . 77, 157 (1039,. ( 2 0 ) llichnids, 11. SI.: C’aiiicgielnst. 11-ash. Pub., 209 (1915). (,21I Siiiii.nov, A I., niicl Izvoakihov, V. P., S t n t e ’ l n s t . Tobacco Inced. (r,7.S.S.R.)Bull. 7 1 (1930). dniucli, A , , Irist. Tohacco Iucest., Krasiiodar LT.S.S.R. BidZ. 33
c.
(1!)?7J, (?;-err of metal foil and paper has been employed for drying the tissue. The interrelation of the physical phenomena of moisture sorption and the practical problems involved in its removal by vacuum pumping systems are presented. Drying may be supplanted at high temperatures by what appears to be a decomposition of some paper constituent.
P
APER exposed to room conditions noimally contains a t least 570 water sorbed throughout its structure. The
exact amount to be found on any particular type of paper depends on the relative humidity of the surrounding atmosphere and also on the previous history of the paper sample, the temperature, and the pulp composition (4,5, 14,15, 16, 22, 26, SO, 3 f ) . This moisture exerts a marked effect on the physical properties of papers, although its exact role is not known (1, 2, 3, 6, 9). The serious decrease of electrical insulating quality with increasing moisture content is of importance in the numerous uses of dielectric paper (7, 8, 11, 24, S I ) . Tissue paper (23, 27) used in the manufacture of rolled electrical capacitors was examined in the present investigation.
I:vacuation under heat has been the method used to dry radio parts containing paper. This process presents several problemb, but the two major factors considered are the extent of the drying that wsults from various times of treatment under factory conditions \\ ith varying loads in commercial vacuum tanks, and the extent of damage attending exposure to elevated temperatures. The removal of water from materials with large surface area is usually achieved by lo^ ering the relative humidity or partial pressure of water in the gas surrounding the sorptive material The effectiveness of various reductions in humidity has not been tho~oughlystudied hut will probably depend upon temperature and the partial pressure of water vapor. Inert gas pressure should be of importance only as it affects the rate of drying (11, 13). I n vacuum drying these same considerations hold. The rate oi water removal and the limiting desiccation attained depend upon the temperature and the atmosphere within the system. The vacuum tank atmosphere, in turn, is dependent upon the load, the leaks in the system, and the pump speed. I n order t o obtain data relative t o paper drying and decomposition, sorption isotherms for kraft paper a t the full range of temperatures were studied by successively evacuating and reestablishing equilibrium within a large volume connected to paper samples through a large vacuum stopcock. Details of this
