Digitized by the Internet Archive in 2011 with funding from LYRASIS members and Sloan Foundation http://www.archive.org/details/correlationinherOOhaye THE CONNECTICUT HOv | ::f ^ AGRICULTURAL EXPERIMENT STATION NEW HAVEN, CONN. BULLETIN 171, MAY, 1912. CORRELATION AND INHERITANCE IN NICOTIANA TABACUM. By H. K. Hayes. TABLE OF CONTENTS. Page Introduction 3 The Material Used 4 The Methods Used 5 Correlation of Parts 6 Inheritance of Characters 7 Family (405 x 400) 8 Family (403 x 401) 10 Family (402 x 405) 19 Summary of Results 26 Interpretation of Results 27 Conclusions 34 Suggestions to the Economic Plant Breeder 35 Literature Cited i 36 Tables 37 The Bulletins of this Station are mailed free to citizens of Connecticut who apply for them, and to others as far as the limited edition permit. CoDnecticBt AiricDltiral Eiperient Statioi. OFFICERS AND STiVKK BOARD OF CONTROL. His Excellency, Simeon E. .Baldwin, Ex Officio, President. Prof. H. W. Conn, Vice Presiden* Middletown George A. Hopson, Secretary Wallingford E. H. Jenkins, Director and Treasurer ........ New Haven J. W. Alsop Avon W. H. Lee . Orange Frank H. Stadtmueller Elmwood James H. Webb Hamden Administration STATION STAFP\ E. H. Jenkins, Ph. D., Director and Treasurer. Miss V. E. Cole, Librarian and Stenographer. Miss L. M. Brautlecht, Bookkeeper and Stenographer. Chemistry. Analytical Laboratory. John Phillips Street, M. S., Chemist in Charge. E. Monroe Bailey, Ph. D., C. B. Morrison, B. S. R. B. Roe, A. B., C. E. Shepard, Assistants. Hugo Lange, Laboratory Helper. V. L. Churchill, Sampling Agetit. Proteid Research. T. B. Osborne, Ph. D., Chetnist in Charge. Miss E. L. Ferry, A. B., Assistant. Miss Luva Francis, Stenographer. Botany. G. P! Clinton, S. D., Botanist. E. M. Stoddard, B. S., Assistant. Miss M. H. Jagger, Seed Analyst. Miss E. B. Whittlesey, Hebarium Assistant. Entomology. W. E. Britton, Ph. D., Entomologist; also State Entomologist. B. H. Walden, B. Agr., D. J. Caffrey, B. S., H. B. Kirk, Assista7its. Miss E. B. Whittlesey, Stenographer. Forestry. Samuel N. Spring, M. F., Forester; also State Forester and State Forest Fire Warden. W. O. Filley, Assistatit State Forester. Miss E. L. Avery, Stenographer. Plant Breeding. H. K. Hayes, M. S., Plant Breeder. C. D. Hubbell, .Assistant. Buildings and Grounds. William Veitch, In Charge. ■^ CORRELATION AND INHERITANCE IN NICOTIANA TABACUM. H. K. Hayes. INTRODUCTION. The objects of this paper are two fold; first, to give some new facts regarding the correlation and inheritance of plant charac- ters in Nicotiana tabacum, second, to show how these facts may be applied by plant breeders to the production of new improved forms. The following facts show that Nicotiana tabacum offers special facilities for the study of the correlation and inheritance of plant characters. 1. There are a large number of different varieties which present easily measured quantitative differences in characters. 2. The Nicotiana tabacum forms are naturally close polli- nated and can be inbred for many years without deterioration. 3. The technique of crossing is very simple and a large number of seeds may be produced by a single cross. 4. The seed is viable for a long time so that a considerable number of generations may be grown on the same field in one year. As tobacco is one of the principal agricultural crops of the United States it is very important that all of the facts regarding the correlation and inheritance of its characters should be known. For the last nine years many attempts have been made to pro- duce improved forms by hybridization without a very definite knowledge of the underlying principles. It is hoped that this paper may be a contribution to this knowledge. 4 INHERITANCE IN NICOTIANA TABACUM. THE MATERIAL USED. The material used for the studies reported in this paper, with the exception of the Broadleaf strain, consisted of types which had been inbred for a number of years and which were uniform to type. These were Havana and Broadleaf, which have been grown in Connecticut for cigar wrappers for many years, and three A^arieties for growing under shade, which had been grown in row selections for a number of years from selfed seed by The Connecticut Agricultural Experiment Station in cooperation with the United States Department of Agriculture. Following is a short description of the forms used in the experiment. Statistical determinations of special characters- are given later. No. 400. Uncle Sam Sumatra. This type proved to be of little practical value for growing under shade because the leaves, when cured, had a papery texture. The number of leaves, counting from the fourth leaf from the bottom to the leaf below the bald sucker*, ranges from seventeen to twenty-five and averages about twent^^-two. No. 401. Broadleaf. A variety which has been cultivated in the open since the early history of the tobacco industry in Connecticut. The number of leaves ranges from sixteen to twenty-two and averages nineteen. The average height is about fifty-five inches and the average leaf area is about 9 sq. dcms. Its leaves are drooping in habit. No. 402. Havana. Another Connecticut out-door variety, which averages about twenty leaves per plant, with a range of from sixteen to twent}— five. The average height is fifty-six inches and average leaf area 7 sq. dcms. Its leaves are more erect than the Broadleaf and droop slightly at the tip. No. 403. Small-leafed Sumatra. This type was introduced for shade purposes, but did not prove so satisfactory as the Cuban. It averages about twenty- * The "bald sucker" is a farm name for the last sucker or flowering stem on the top of the plant which has no true leaves. THE METHODS USED. 5 •seven leaves per plant with a range from twenty-three to thirty- one. The average height is about seventy-six inches and the average leaf area is about 3 sq. dcms. The leaves are erect in habit. No. 4-05. Cuban. This type is now used for growing under shade in Connecticut, ■over two thousand acres being raised in the valley in 1911. It has a range of from sixteen to twenty-five leaves and averages about twenty. The average leaf area is about 5 sq. dcms. and average height about sixty-five inches. THE METHODS USED. As shown by the descriptions, each type has been given a number. A cross between No. 405, Cuban, and No. 402, Havana, has been written (405x402), the female parent coming first. Whenever later generations have been grown they have been noted by further numbers, as 402-1, (405 x 402)-l, which denote respectively the second generation of Havana and the second generation of the cross between Cuban female and Havana male. The tobacco flower is naturally arranged for self-fertilization. If inbred seed is desired it is only necessary to cover the flower cluster with a Manila paper bag; the 12 lb. size having been found to be most satisfactory for this purpose. It is, however, •advisable to take off all but about twenty of the seed pods, as these will produce an abundance of seed. The technical work in crossing two varieties of tobacco is very simple. The corolla is split up one side, before the blossom •opens, and the stamens are removed. Pollen from another variety (taken from its stamens by means of a scalpel or other sharp instrument) is applied to the pistil of the variety from which the stamens have been removed. Those blossoms not used in crossing are removed and the seed-head covered with a Manila paper bag. The following characters were studied with reference to correlation and inheritance. 1. Number of leaves per plant. The number was counted from the fourth leaf from the bottom of the plant to the leaf just below the bald sucker at the top, which gives about .the number that is usually harvested. 6 INHERITANCE IN NICOTIANA TABACUM. 2. Height of plant measured from the ground to the last leaf counted. 3. Average area of leaves. After the plants had reached maturity, tracings of the fourth leaf from the bottom, the middle leaf and the last leaf below the bald sucker were made on smooth paper and each was given a series number. The area of each tracing was determined with a planimeter which gives an experimental error of only about 5 sq. centimeters per leaf. The term "average area of leaf" is the average area of these three separate leaves. 4. Average length of midrib, which is the average length of the three leaves used for the area measurements. 5. Average width of leaf, taken in the same manner as the length measurements. The data, with one exception, which will be mentioned later, were all taken in a uniform manner by the author and Air. C. D. Hubbell. The planimeter measurements were made by Mr. Hubbell, who has given much efficient assistance in this work. We wish also to express our thanks to Dr. E. M. East for much helpful advice and cooperation. CORRELATION OF PARTS. The question of correlation between parts is of great im- portance when applying the principal of selection to improve- ment of plants. In our work the usual correlation table has been used and the coefficient of correlation determined. The coefficient of correlation shows the degree of mutual relation, between the characters in question. It it is low (i. e., much below 0.50) it indicates that they do not depend very much upon each other; if high, it indicates that they are closely related and when it rises to unity it shows that both characters depend upon the same cause and are inherited together. If two genes are located in the same chromosome as supposed by Emerson ( :11) they could be inherited together but not depend on the same cause. Two types were used to study the correlation between parts. Correlation tables of the results are given at the end of this paper. For convenience in discussing results, the dift'erent coefficients of correlation are here grouped in tabular form. INHERITANCE OF CHARACTERS. TABLE I. COMPARISON OF CORRELATION COEFFICIENTS. No. Correlation between no. of leaves and hght. of plant Correlation between no. of leaves and aver, leaf area Correlation between length and breadth of leaf No. 401 Broadleaf No. 403 Sumatra (403 X 401 )F, (403x401)-lF2 (403 X 401)-4F2 +.368 ±.048 +.631 ±.033 + .406 ±.046 + .342 ±.058 + .408 ±.036 -.165 ±.054 -.008 ±.055 -.226 ±.052 -.124 ±.065 -.076 ±.043 + .684 ±.029 + .497 ±.041 + .818±.018 + .737 ±.030 + .761 ±.018 The above table shows that the crosses between the Nos. 401 and 403 have not apparently affected the mutual relation- ship of the different characters studied. Thus, while there is a positive correlation between the number of leaves per plant and total plant height, this correlation as a rule is somewhat less than +0.5 in our tests. One might expect some correlation between the height and number of leaves because the former is the combined length of the internodes, and the number of internodes depends on the number of leaves. But the corre- lation is not very large and shows no very close relation between height and number of leaves. There is a small negative correlation of leaf area and number of leaves but the relation between the two is so small as to have no practical value. That is, number of leaves is not a character distinctly opposed to leaf area. The high correlation between length and breadth of leaf indicates that both are very closely related, that is, that both are dependent on the same cause or series of causes in inheritance. INHERITANCE OF CHARACTERS While all of the characters studied show fluctuating varia- bility they are very differently affected by environment. The most uniform character of all was the number of leaves per plant, which was little affected unless the conditions were so unfavorable as to greatly stunt or dwarf the growth of the plant, as appears in the following table. Each of these four selections was grown at Forest Hills, Massachusetts, Bloomfield, Connecticut, and New Haven, Connecticut, from seeds of a single plant. The Forest Hills plants were grown and the 8 INHERITANCE IN NICOTIANA TABACUM. data taken by Dr. E. M. East. The calculated mean is used to determine the value of the selection. TABLE II. NUMBER OF LEAVES PER PLANT. Selection Forest Hills Bloomfield New Haven Average Mean 1 2 3 4 25.8 ±.091 30.8±.115 25.3 ±.085 25.8 ±.091 25.7 ±.081 29.6 ±.078 25.2 ±.074 27.4 ±.079 25.2 ±.077 30.7^.090 24.7 ±.073 26.7 ±.078 25 . 6 30.4 25.1 26.6 The field at Forest Hills was fairly fertile but in a region where tobacco is never grown commercially. Bloomfield is in the center of the tobacco-growing region and the soil is per- fectly adapted to it and heavily fertilized. The soil at New Haven is a thin, poor, sandy loam only moderately dressed with manure and chemicals. The means of the different selections compared with the average mean show a variation of only ±0.8 leaves, and as only about one hundred plants were counted for each determination the results seem very uniform. Three crosses have been studied as to inheritance of charac- ters and for convenience each will be considered separately. Family {If-Oo x JfOO) Ctthan x Uncle Sam Sumatra. This cross was made in 1907, the reciprocal Fi generations and the parents were grown in 1908, and the F- generations of the crosses and parents in 1909. In both years the crosses and parents grew on the same plot under shade and therefore under uniform conditions. The data given in Table 3. show the range of variation, the number of plants studied and the usual statistical determinations which consist of the Mean, A, the Standard Deviation, S. D., and the Coefficient of Varia- bility, C. V. The plants were not topped as uniformly in 1908 as with the later generations and therefore the number of leaves per plant cannot be accurately compared with the numbers in later years. This 3"ear's work, however, shows that there was no increase in variability as determined b}' C. V. due to \\ie crossing. >• o 6. 08 ±.276 5. 51 ±.215 6 . 77 ± . 234 6. 97 ±.271 4. 62 ±.186 5. 64 ±.218 8.10±.315 8.56±.333 d in 1 . 03 ± . 047 1 . 08 ± . 042 1 . 28 ± . 044 1 . 52 ± . 059 0.85 ±.029 1.01 ±.039 1 . 53 ± . 060 1.66 ±.065 < 'C Z~. Tl — -^ O '^f 1— 1 c; ic -* X -t- o 00 ai coccroooo ^ ^ 41 ^ ^' 4l' M M OcOOCXJ^OlCi'*' cOOCO^OOt^OOO r-^T-lCqT-lr-HT-Hr-H 13 o ccocoooo r-1 L-^ C; LO '^ l-O >0 O ■Si M tfi u : ; ; CO : : irH • ■ -T^ ■ • .,-( CO (M • ■ CO ■ • • y3 .-H 'Xi • • CO t^ CO ■ • •cooo CO • • -H o • (M rH CO • • (M --< O Ol • 1-H CO CO T-H t^ C5 o C2 -r! r^ !M > < a w J t-l fa m o < oi H e g D Z (K -. o z o LOco-to-^'-t'oc^iT^ic OGcaso^^oocoi> > (NC^J«-HC0fN^r-<^iO(M i^i'-fi'-H'-H'-H'-H'-n'-n'-H" L- (NoooioioaiiOTfico ■o CO ic o Lo 03 00 ai m 00 00iOI>00i-(^H-tcOC0T-H lOt^cOlOiOC^lC^tMOcO QOOOOOOOr-o « -H -f] -H -H -H -H -H ■« -H -H W 03COC0 10 000TtlTj i-o oo o 00 CO ci G3 CO oq (N c-* t> -# in rH ,-H CO (M =■' 1 'XiC !N Tf CO O rt og 1 T— < CO 1 coco CO-*COCOi-H C^l 1 ^,-1 CO^i-M ■^ 1 CO^ OOC33COI> m (M 1 !N(M •woqco CD i COLO' O Ol t^ ^ .-H lO IM 1 ^Ol CO w 1 lO I CO t^ CO CO o^ooc^ !N (N O) 0.110 1^ .-H cq < 1 !-< CO -*< CO 00 o o IM • rt TfHCOOrHCO COIMCO O I •■-i(MO'* CO 00 • • ■coc^ • iMrtca t^ 1 • ■C003 ■ O3OJ00 • o 1 ■ Ol ■ w (_) 1 0) -; ^^ u • 05 ■ O ■ O Tf .-^ IC rt 0) -rt ■ CT— oq ■ c^ ca C^ M (N J e U S i O.-^O^O^'Hrt^rt 0 osoioicioooc^c^c; ;^ o s D ooooommmoo ^C D S 1 fe-^t-r-r g o oooooo 2 3 £; TTTfTf-^Tfirtf m^P5^ X X X X « X -1 - 1 CO CO CO CO CO CO CO CO rt -H Q O J-, Q Q Q 1 1 '^l -*^ Tf ■* Sl-SS-2!- ■* s < o 2 z :=> o 1^ H P TOTt<(NOTC^O-fCC^- TOTOb-TO^'^OOTO^O > 01 TO (M -# C^ ^ rt .- 0 T ^-H-H-H-ti'-H-H-H-JI-H .0 1 CZ; [^ C2 I^ rH IC ?.) ic 00 --S C5 6 oc 32 ci TO fi 1 ,-lrt t^ -t C --f< l> CO (M 0 M C-1 t~ TO Lt — t^ I- t^ C-. TO 0 d ^ C-l rt ■>! .-H 0 C 0 TO ?.! -H-H-H-H-H-H-H-H-H-H 0 30 0 cc Tf< X i> 0 c^ 0 iC 0 00 -* 0 !N rt< 0 C^) TO Tt>LOTO-*-*i3 0^I>0 i-H ^ 05 !N 0 -* L'^ i--; -1 0 ■ C TO r- 0 LO 0 0 TO t^ f~ C-] TO !N TO M — rt rt rjH c^^ < U -H -H 41 4m ^^ -H -H -H 1 -^TOS-st'XtCiCrtrtO loOLOoioC^tN-HTOO t> lo Lo -* i> i> :^ t^ 0 -n "3 1 Ot^OOO^TOOt^^ 4J I 10 0 0 0 "-T 0 TO C: 0 -^ 1 rtrtrt^^--t-c„c^ H j ^ ^^ ^ 0 1 • ■ (N r - • 0 1 • • TtflMi-i 03 1 • ■ rtOt^ < X 1 • ■ ID 1 ^ • t^,-H 0 X 1 •TOTfrt TO 1 0 ■ ATOXIC 0 X 1 ^ ■ cir^ TO 0 1 IM • TO t^ TO t^ X 1 TO • -#ICX tH 1—1 r~ 1 i> • X - 1 z 01 C^-H TOX 0 02 1 .fO^ ^r- OX02ON w 0 1 TO(M TOTOTOi-H IM --0 1 ^OTOO30rt50T-it^ H 0 1 IM TftTO r-l — i-ltM f^ 1 t^ T*< 10 0 TO i-H 1-0 0 -* a ■n 1 -*IM (^ TO 1 0 1 ^ot^-H . • -a 0 C^ rt • .rt T)H t~ 1 (M -«t^ • ■ -rHX ^ 1 I— 1 T— 1 >-H • -xx 0 ^H • • TO 1—1 1 • • 1— ( • -TOM • • cq ??l • -oto lO 1 • - ^ • c 1 ^ 2::sz;3:^::::::::: 0 > g ro 0 cj ci m 0 01 05 C3 C3 0 rH^^^^^^^«^ B 1 M > < H O < Ui pi < QD 02 c; TjH t^ O OiC O C/J t^ t^ t^ o to > ' ' ,J " ! 41 -H -H -H -H : o C-. O (M "* Ci ■ LI CD t> O I> • t- CT. t-^ Ol Cr --I >-! ^ C^l iM (M O -* t^ »0 (M CO "TT* LO CO OOOOO • Q -H -H -H -H -fi : c/i t^ o CO CO ^ • lO i^ i-H (M 1-1 • O^ -Hn-H^ >-H CO c-i O oci CO 05 CO CZ) •* coooo • <■ -H -H -ti -H 4] : CO loiooo-* • iM CO CO ICCO • CO GO CO lO LO r< -4 ^ OOOt-rH . -1 LC O IC o ■* • C-l .-H T-H T-H .-H C3 . Tt^ .^ ... 1 '"' CO T-H ^ C-l Ci < > < ^ I> - Pi o t^ t- 1— 1 CJ o fa f" t/; 02 ■ coTfi ,-1 ■ CO CO S M R > oC' • OCSCOt> 00 • coi-i t/2 . fa I> • T-H ^ cot^co • CO"* --iiM '"' <; «g CO • O0CD{M (M • g< ■* CO t^ • Z fa lO (M CO (M ai a: • o CO(M t^ • r/) t/: < ^ C2 • lO ■* t^ • J CO • (M -^ • U o . . CO O . .(MO . (M 41 1 4J ^' ^" U CO 00 03 CO GO ^ '^ Ol-H ttl fa < CO (M iM LQ Cd w T— 1 T-H T-H T-H T-H J Q < C5 CO CC T-H T:t^ O Ot-h OCOt^ Oi Ot-hOt-hO Q 4i ^ ^ 41 ^ OJ CO t^ T-H lo -^ O O CO . lO ■ ■ • < fa > a J - a H s fa -*! ■ t^ O O IM Q J iM • lOM^ T-H u ^ z fa o K Q 1— I T-H -*l -tH CO Ol CO < en CM -* CO COt^ . fa > H Z P oo "0 CO t^ T^ ^ <; T-H CO CO-^ (M T-H fa o (/2 z <; >o o • lo CO CO ,_, C3 • (M CO o D H p c^ T-H ■ ■ T-H (M T— ( IM • • S , G 5-< -^ 03 IS OOOt-h T-( 5 ^H T-H T-H T-^ T-H D O o; cr:^ cft c;i o >- ^ rt rt ^ ^ Z w p o w fa d g gooo ^ h; ^ "^i ^ 2 mPQ X X X -■ - CO CO CO 2;:z!ooo O O -^ -gH ^ -^ TJ< ^^-^.^ > t-- lO 0 0 02 01>CO(M 0 CO 00 CO ic CO 41 41 41 41 41 GO CO GO 00 CO GO 0 ^ Ol t^ t^O^OD ^ C2 Q 2. 27 ±.088 4.69±.183 3. 39 ±.132 4. 22 ±.195 3. 69 ±.117 < 28. 8 ±.125 48 . 7 ± . 258 40.0 ±.187 37 . 4 ± . 275 37 . 9 ± . ] 60 O 10 LQ IC 0 '^ ^ ^ ^ ,^ ^1 CLASS CENTERS IN CM. FOR AVERAGE LENGTH OF MIDRIB OF LEAF CO CO - § Tt* I> T— 1 »c CO 00 00 CO CO ^H ,—1 (M CO lOt^ .-I'^.-HCO 02 CO 10 ceo 'CO LO CO GO CO ^ cocoes H < S. V. OO (M t^ LO CO -* CO T-i ■ ■ (N --H lO ^ • rt COOCO CO 'ti lO CD (M O t^ (M O O CD t^ »0 T-H (M ^ ^ (M T— 1 lO CD rN 1— 1 O ^ CO ^ !M CO CO (M O CDTt< (M ^ OO COCO'* COO CO 1— < t^ CD -*i CO 02 00 CO ^ ^ (M COrH 00 CDCO -H -H -ti 4i d Tfl ,-H 00 1> coo^ CO l:^I> t^Tt< T^ !>•* T-H OC 00 lOO; r-l T-H 1— 1 i-H CO Q -H -H ^^ -H c/i OCOO CO 00 02 Oi(M "* CO-^ 05 ^ ooooi CO -H t^TH (N Y\ ^ t^ , — 1 1-H T— 1 1—1 03 0J t^ IM 00 lO ^^ CO lO CC CO .co^ 00 z fa " o CO CO Tt< (M ;A aj Gj lO to ,^ >oo 6 r^ cj -^ Tt< ^ 405, C 402, H (402 X (402 X 22 INHERITANCE IN NICOTIANA TABACUM. 405 Cuban was 65.4 ±.264, of No. 402 Havana was 56.5 ±.218, while the Fi had a Mean for height of 65. 5 ±.270. This, how- ever is no doubt due to increased vigor from crossing and has no significance in inheritance. The variability was a very little larger for the Fi generation than for either parent and well within the probable error of the determinations. The F2 generation showed an increased variability and a considerable degree of correlation between height of plant and number of leaves. Thus, a correlation table gives a correla- tion coefficient of +.786 ±.023 for number of leaves and plant height. As Table XI proves, selection No. 405 Cuban has a smaller leaf than No. 402 Havana, although the range of variation is very much the same. The Fi generation showed the same range of variation as the parent No. 402, although the Mean was lower for the cross. The variability in Fi was also greater than that of either parent, but within the probable errors. The F2 generation produced some leaves with as small average size as the smaller parent and some leaves which averaged larger than either parent. The variability as determined by C. V. was also materially greater than that of the parents. The correlation coefficient for the average area of leaves and number of leaves per plant was —.092 ±.048, which shows conclusively that there is very little correlation between number of leaves and leaf area and that these two characters are inher- ited independently. Tables XII and XIII show that the difference of the parents in size characters of the leaf is chiefly a difference in average length, as the average width of leaves of both parents is very nearly the same. Table XII shows that the range of variability of the Fo genera- tion for length of midrib is as great as the combined variability of the Fi generation and the parents. One variate had a smaller average length of leaf than the lower class of the lower parent, and seven variates had a longer average midrib than the larger parent. The abrupt ending of the parent classes, however, and the fact that a larger number of variates occurred in Fo than in the parents makes it probable that the variability for average OixJH c» CO CD CC CO c^ O COt^ Ol > -H -H -H -fl coo^oioi O r- o ^ iM CD t^ 00 lO T-HT-H^tM z <: w 00 CO CD CD ^CDCD CO S" OOOO o -^ < -H -H -H -H X CD CO CD LO CSl 1> '^ cn (M O lO CO CO CD -*! ^" — ^ 7) 'ra O OO (M m O to LO LC O O H r-H ^ I— 1 ^ (M b ! 1 ! T— 1 o M h-i O ■< Pi w > o fa ^ '. '. '. ^ X > ^^ -J [I, O 1— I ; ! ;oo o < 02 • 00 <-o ^ OS S <; (J t/5 Q M . > 00 • ic i^ CO o <: • CO c-1 CO <" W i> Ci ^ CR »0 Z^ ^ CO^ > ^ fa 75 O <; -^OICDIM fa o lo^ ^ CO z IQ OOCOt^ (M o iC(M (MCO en < TjH CD-- X w J w fa hJ o pq < H o z w J M O < fa o iz; o ceo G5t^ OOGO ^(N (M (N CO"* > -H -H -H -H CJ O L': O cq M^ CO ?1 (M t^t-CCIN '"' ^ C-. lOlM O (M COCO ^^ T-H r— < T— t d •H 4< -H -tj * m r^ (M coco O COTfi (M (M CO CO "O ■rti CO T— t^ 0 00 01-^ ^ rHr-l(M < •H -1] -H -H O iC (M CO ^ , rt O OO (M o lO >-0 O Oi H """" i '. ; ! t^ ,_, • lOiM —1 H lO • T-H T— i O w 00 •r^ ^ ^ ^ • CO rt CO o o • CO (M oc < Tt< • lO"* ^ >. x; CI lO rH OO iQ ^ coTt< CO o S f^ fa C55 (MTtoo :z ^ a3-*Tt< Offi X X - -(M (M SSco O O T^ Tt< TT ^--^-^ Tti lO O lO OOt-H ^(^ CO "* ^ lo > ^ ^ ^ 41 lO ■* CO --H U oo»o-* o C500CO t^COTtH ^ t- (X)00 (M OOOt-h Q ^^ -H -H ^^ CO 1.98 2.14 2.17 3.49 ai ooc o O T-H C <( > < w ^ t^ C^] O iM o CTj (M C»t- 00 CO »i 00 «:> CO CO i^ Cx^ ^ 1-H ^ lococo •z M U C/3 iC lO 03 (M -H Cfi r-H i-H < ►J O (M ■ • -^ fc =^C!^^ c c!^_ Co cS (M (N o .£) >00 ^ ^ cj "^ "^ um X X - - CO O iCOl o o O O -^ ^ j Tt<'^--^C^ 26 INHERITANCE IN NICOTIANA TABACUM. length of midrib is no greater for F2 than for the combined parent and Fi generations. Table XIII shows an increased variability in F2 for average leaf width, although the parents have nearly the same Alean and range of variability. This, we believe, is due to correlation between width and length of leaf* and also explains w^hy leaves yielded by F2 were larger than those of the larger parent. This explanation seems logical in view of the fact that there were no leaves with a smaller average area than that of the smaller parent. SUMMARY OF RESULTS. A. Correlation of Characters. 1. In the two types studied and in the first and second generation of crosses between them there was a positive corre- lation between number of leaves and height of plant although in all but one case this was less than +0.5. 2. The number of leaves and average leaf area showed only a slight negative correlation, i. e., a large number of leaves was associated with a slightly smaller average leaf area. 3. There was a distinct plus correlation between length and width of leaf, i. e., the longer leaves were on the average also the broader ones. B. Inheritance of Characters. 1. The characters studied showed very different fluctuating variabilities due to environment. The most uniform charac- ter, in this respect, was number of leaves per plant, which was little affected unless the conditions of growth greatly stunted or dwarfed the plant. 2. Reciprocal crosses are equal within the limits of fluctuat- ing variability. 3. The Fi generation is intermediate in the characters studied, being as a rule somewhat larger than the average of the^ parents. All characters studied except the number of leaves per plant showed added vigor. 4. The Fi generation is no more variable than the parents, the variability of Fi being found slightly greater than the average * The correlation between length and width of leaf as determined by the correlation coefficient proved to be +.814 ±.016. INTERPRETATION OF RESULTS. 27 of the parents in six cases and less in five cases. This result agrees with Johannsen's (:07) observation. 5. Different variates in Fi give similar results m F2, showing that the variation in Fi is fluctuating variation due to environ- ment and is of no germinal value. 6. The F2 generation is more variable than the parents. When sufficient numbers of variates were studied the F2 showed a range of variation equal to the combined range of the parents and Fi. 7. In the two crosses studied there was only a small negative correlation between average leaf area and number of leaves per plant. This indicates that leaf number and average leaf area are inherited independently; therefore we can combine the desirable leaf size characters of one variety with the number of leaves of another form. 8. The results show some variation in the correlation be- tween height of plant and number of leaves. Thus, the corre- lation coefficients of the two Fo generations of the cross between (403x401) were +.342±.058 and +.408±.036, while in the F2 of the cross between (402 x 405) the correlation coefficient was -1-.814±.016. 9. There was found a large positive correlation between length and breadth of leaf, which indicates that the inheritance of these characters depends on the same cause or series of causes. INTERPRETATION OF RESULTS. When Mendel's law was rediscovered in 1900 it was generally believed that it applied only to a few isolated cases of inheri- tance and many apparent exceptions were cited. By a better understanding of the complexity of the facts or by simple extensions of the Mendelian notation, most of these apparent exceptions have, one by one, been shown to follow the law. The inheritance of morphological characters, i. e., form characters such as size of stalk and leaf, shape of leaf, etc., which show fluctuating variability, has been considered by many to be an exception to the Mendelian rule. By fluctuating varia- bility, as used in this paper, we mean the quantitative fluctua- tions of characters, which are due solely to environmental conditions, soil, climate, etc. While such fluctuations have 28 INHERITANCE IN NICOTIANA TABACUM. no value in inheritance they make more difficult the correct interpretation of experimental data. Mendel's principal discovery — the segregation of potential characters in the germ cells of hybrids and their chance recom- bination in later generations — has given a logical explanation, at any rate, to the facts which we now have. Whether all characters can eventually be shown to be Mendelian is of course not certain. The results above given are statements of the actual behavior of tobacco plants under careful observation. The interpre- tation of these results which follows is an expression of opinion. For the characters studied there is a much greater range of variation in F2 than in Fi. In the light of our present knowl- edge segregation seems to be the best interpretation oj this fact. While we have no data regarding the F3 generations of these crosses, we have no doubt but that some of the F2 t\^pes will breed true. Our reasons for this belief are based upon some unpublished results of the study of the inheritance of number of leaves per plant of a tobacco cross, which show that in gener- ations later than F2 both intermediates and extremes may breed true. How then may the tacts be explained? The first Mendelian interpretation of variation that is ap- parently continuous, known to the writer, was made by East (:10). This assumes that the parent plants, for the character in question, differ in more than one separateh^ inherited unit or gene. Each of these independent, interchangeable units, allelomorphic to its own absence, is capable of adding to the character, and the heterozygous condition of an}^ unit is half the homozygous condition. There are cases of color inheritance which can only be explained by the presence of two or more separately inherited characters in the reproductive cells. Thus, Nilsson-Ehle (:09) found in one case two definite, independently inherited characters for blackness of glumes in oats, although glume blackness in other crosses behaved as a simple Mendelian mono-hybrid. Many crosses were made between wheat varieties having red and white seeds and in all but one of these the F2 generation gave the ordinary three-to-one ratio. But a cross between an old red seeded wheat from the north of Sweden and a white variety produced only red seeds in a total progeny of 78 F2 INTERPRETATION OF RESULTS. 29 plants. The expectancy for F2 if the parents differed in three characters for red would be 63 reds to 1 white. The progeny in F3 of these 78 F2 plants gave ratios which proved that he was dealing with three separately inherited characters for red. East (:10) found that in certain cases there were two indis- tinguishable independent yellow colors in the endosperm of maize. Some evidence was also received of three independent red colors in the pericarp and two colors in the aleurone cells. East and Ha^^es (:11), in a study of inheritance in maize, gave complete results of a number of observed crosses between yellow and white varieties, which behaved as if there were two separately inherited characters for yellow color in the endosperm of maize, either of which could produce the yellow color. Other crosses were mentioned, between yellow and non-yellow (white) families, which behaved as simple mono-hybrids. Data were given of a number of flint-dent crosses, one of which in F) gave about one pure ear in every sixteen, while one cross gave an indication of a higher ratio. Crosses between families which showed quantitative differences in morphological characters showed wide ranges of variability in F2 nearly equal to the combined range of the parents. Emerson (:10) found that crosses between races of plants which differ in sizes and shapes have increased variability in F2 as compared with the parent or Fi forms. His data were on maize, bean and gourd crosses. Shull (:11), in a study of defective inheritance ratios in Bursa hybrids, gave results which indicate the presence of two genes, each of which is independently responsible for the Bursa- pastoris-type of capsule. The Heegeri-type appeared only when both genes were absent. The only change which it is necessary to make in the inter- pretation of Nilsson-Ehle and East for inheritance of color characters, in order to have the hypothesis fit the facts for inheritance of fluctuating plant characters, is to suppose the heterozygous condition for each character to be only half the homozygous condition. Thus the Fi condition for any character is a blend between the parent types, instead of being like one .or the other parent forms as is the case where complete domi- nance is the rule. 30 INHERITANCE IN NICOTIANA TABACUM. In a discussion of the explanation of results received from crossing certain Linum forms, Miss Tammes (:11) uses a similar interpretation and gives an excellent discussion of this hypothe- sis. The number of individuals studied by Miss Tammes for the different generations is very small. Table XIV gives the theoretical expectation for the F2 genera- tion when the above hypothesis is used. The first column of this table shows the number of units or genes in which the P. or parent forms differ. For any case this number may be represented by n. The second column gives the numerical proportion of the different forms until the parent form is reached. The parent form is represented by P. in the table. These classes are the coefficients in the binominal expansion where the exponent is twice the number of characters; for four characters the condition would be represented by (a-Fb)", the coefficients of this expansion giving the numerical results given in the table for four characters. The third column gives the number of individuals which must be studied in order to have an even chan e of receiving some individuals in each class'. This number is equal to 4^^ where n equals the number of unit characters in which the parents differ. The fourth column gives the number of homozygous individ- uals which may be expected in each case. This number equals 2i^. The fifth column gives the per cent, of homoz\^gous in- dividuals which may be expected in each case. In order to understand this complex class of results we will discuss a specific case. Suppose, for example, we are dealing with number of leaves per plant in tobacco crosses and that both parents of a certain cross are pure for the same basal condition of twenty leaves per plant and that one -parent has in addition some inherited properties which result in a produc- tion of twenty-six leaves per plant. Let us suppose this con- dition due to three interchangeable, allelomorphic character pairs, each inherited separately, and that the heteroz^'gous condition is half the homozygous condition. If we follow the usual Mendelian method and represent the presence of our three characters by A, B and C, and their absences by a, b and c, we get a condition in Fi of AaBbCc, or 23 leaves. X pq < .. , "2 O (M , ^J txO 3 IC (M CO C >;^'d IC (M ^ lO u S-r CJ o > O »0 (M CO CO i-H ^ 6^ lO (M T-H f^i.5 S^ ^ O rt O M 3 c N-rca C^ "* GO CO (M -^ ^ CO ^ o c jz!'- CTj rt ■* CO ^ CO ^ CO ^ CO LO (N c; iM O O Z^ c •r-l ^0, -flnS Uj -f^S § s ^1 o ^C.00 - 1 c ^PhO 2S (M C55 ie "-^ T-H T}^ Q 1^ (--J O C5 .^ -^ (M t- -rJ "o O O (M -^ o. CO g g o g '-p -A.^ 2 § 2 § o •^ (M t^ o 1-. f-yn O LO rHflnCO ^ (M Ci Oh ^^ T-H -^ rf CJ LO O r-HpLnOO ^ W z -OhS § -0.^ ^PLh C t:? t« •-^.ti ° s CD o g 2 gQ 3 No. of Leaves 24 2.5 26 27 28 29 30 31 .... 1 2 3 2 3 9 15 26 1 2 .. 16 22 7 .... 3 6 16 1 10 3 1 1 . . . . M ^ 9 100 39 2 2 3 13 23 46 28 27 8 150 No. of Leaves. A. =28.3 ±.082 S. D. = 1.49 ±.058 Coef . Cor. = Aver. Area of Leaves. A. =3.23 ±.031 S. D. =0.57 ±.022 -.008 ±.055 TABLE XXII. CORRELATION BETWEEN NO. OF LEAVES AND AVERAGE AREA OF LEAVES OF NO. (403 X 401), SUMATRA X BROADLEAF, Fi. No. of Leaves. 19 21 22 23 24 25 26 ^S 4 1 2 2 5 ??Q 5 4 13 7 5 3 32 u . 6 1 6 13 13 9 4 46 1 2 4 10 16 11 44 U)c 8 1 6 8 2 2 19 r' 9 1 2 1 4 2 3 21 47 38 29 10 150 No. of Leaves. A. =23.6 ±.072 S. D. = 1.30 ±.051 Coef. Cor. Aver. Area of Leaves. A. =6.35 ±.062 S. D. =1.13 ±.044 -.226 ±.052 CORRELATION TABLES. 41 TABLE XXIIL CORRELATION BETWEEN NO. OF LEAVES AND AVERAGE AREA OF LEAVES OF (403 X 401)-1, SUMATRA X BROADLEAF, Fo. No. of Leaves. can <; cr C/2 19 20 21 22 23 24 25 26 27 28 30 3 4 5 .. 1 .. . . ..112 116 2 6 5 . . 1 . . 4 2 4 4 4 1 '2 '1 'i 1 4 .. 6 7 8 .. 2 .. 6 .. 2 2 .. ..Ill 4 3 1 4 8 4 2 2 2 2 3 1.. q .. i 1 8 10 11 19 16 18 11 6 6 1 i 2 24 29 32 13 6 Ji^ 107 No. of Leaves. A. =23.8 ±.146 S. D. = 2.24 ±.103 Coef. Cor. Aver. Area of Leaves. A. =5.58 ±.080 S. D.= 1.23 ±.057 -.124 ±.065 TABLE XXIV. CORRELATION BETWEEN NO. OF LEAVES AND AVERAGE AREA OF LEAVES OF (403 X 401)-4, SUMATRA X BROADLEAF, F2. !> No. of Leaves. 18 19 20 21 22 23 24 25 26 27 28 29 3 4 5 .. 1 . . 1 2 . . 1 5 6 10 1 6 12 18 18 1 7 6 17 16 2 15 6 6 1 . . 2 2 . . 1 2 .. 11 6 3 12 8 3 1 1 3 .. 1 . . i i 6 7 8 13 5 6 3 1 2 1 . . 1 1 . . 1 . . 5 16 30 50 52 41 22 15 7 1 1 1 No. of Leaves. A. =22.0 ±.083 S. D. = 2.24 ±.103 Coef. Cor. .076: 9 47 79 72 27 7 241 Aver. Area of Leaves. A. =5.34 ±.048 S. D. =1.11 ±.035 :.043 42 INHERITANCE IN NICOTIANA TABACUM. TABLE XXV. CORRELATION BETWEEN AVERAGE WIDTH OF LEAVES AND AVERAGE LENGTH OF MIDRIB OF NO. 401, BROADLEAF. Aver. Width of Leaves in Cms. h- 1 "^ < 18 21 24 27 30 33 8Q 4 1 5 42 45 48 4 7 . 27 8 1 .. . 7 25 5 . . . 12 34 38 51 54 57 '. i 18 14 1 . 6 9 .. . 3 3 i 33 16 7 60 68 12 1. 1 4 1 8 44 57 34 5 2 150 Aver. Width of Leaves. A. =23.8 ±.164 S. D. = 2.97±.116 Coef. Cor. Aver. Length of Leaves. A. =48.7 ±.258 S. D. = 4.69 ±.183 +.684 ±.029 TABLE XXVI. CORRELATION BETWEEN AVERAGE WIDTH OF LEAVES AND AVERAGE LENGTH OF MIDRIB OF NO. 403, SUMATRA. Aver. Width of Leaves in Cms. 12 15 18 21 24 5 5 27 14 33 7 1 30 2 50 18 2 33 2 10 36 i 21 90 35 4 ^o M > < Aver. Width of Leaves. Aver. Length of Leaves. A. =15.4 ±.111 A. =28.8 ±.125 S. D. = 2.03 ± 069 S. D. = 2.27 ±.088 Coef. Cor. = +.497 ±.041 CORRELATION TABLES. 45 TABLE XXVIL CORRELATION BETWEEN AVERAGE WIDTH OF LEAVES AND AVERAGE LENGTH OF MIDRIB OF NO. (403 X 401), SUMATRA X BROADLEAF. Fi. Aver. Width of Leaves in Cms. L5 18 21 24 27 30 88 1 4 2 1 6 ^u 36 22 4 26 X^ 39 42 13 34 3 50 SoP 26 20 46 S ^ 45 2 14 1 17 48 3 1 4 < 5 37 66 40 2 150 Aver. Width of Leaves. A. =20.9 ±.138 S. D. = 2.51 ±.098 Coef. Cor. Aver. Length of Leaves. A. =40.0 ±.187 S. D. =' 3.39 ±.132 + .818 ±.018 TABLE XXVIIL CORRELATION BETWEEN AVERAGE WIDTH OF LEAVES AND AVERAGE LENGTH OF MIDRIB OF (403 X 401)-1, SUMATRA X BROADLEAF, F2. Aver. Width of Leaves in Cms. ^O < 12 15 18 21 24 27 30 7 33 1 11 5 36 4 19 3 39 1 11 20 1 42 5 5 4 i 45 1 2 2 48 2 2 1 23 41 32 9 1 7 17 26 33 15 5 _^ 107 Aver. Width of Leaves. A. =18.8 ±.186 S. D.= 2.85 ±.131 Coef. Cor. = +.737 Aver. Length of Leaves. A. =37.4 ±.275 S. D. = 4.22 ±.195 t.030 44 INHERITANCE IN NICOTIANA TABACUM. TABLE XXIX. •CORRELATION BETWEEN AVERAGE WIDTH OF LEAVES AND AVERAGE LENGTH OF MIDRIB OF (403 X 401)-4, SUMATRA X BROADLEAF, F2. Aver. Width of Leaves in Cms. ^O 12 15 18 21 24 27 1 1 30 2 6 1 9 33 . . 19 14 33 36- 7 52 3 36 1 62 39 46 83 42 7 25 5 37 45 4S 1 8 1 6 15 1 2 33 121 73 12 i 241 Aver. Width of Leaves. Aver. Length of Leaves. A. =18.8 ±.102 A. =37.9 ±.160 S. D. = 2.34 ±.074 S. D. = 3.69 ±.117 Coef. Cor. = +.761 ±.018 TABLE XXX. CORRELATION BETWEEN NO. OF LEAVES AND HEIGHT OF PLANT OF (402 X 405)-l, HAVANA X CUBAN, Fz- No. of Leaves. 1415 161718 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 41 2 . . 1 1 44 1 . . ' 47 . . 2 1 . . 1 1 . . 50 53 56 59 62 . . 2 1 . . 1 . . 1 1 .. 3 3 6 3 3 .... 23533 4 3 5 2 4 7 .... 1 1 2 2 .. 1 3 2 1 1 5 3 3 2 1 65 68 1 1 1 9 2 2 3 8 2 3 4 3 2 1 1 42 3 2.. 1 1 5 4 2 1 2 1 1 1.... 2 2.. 1.... 1.. 1.. 2 . . . . 1 1 71 74 77 80 83 1 . . 1 . . . .' 1 86 1 1 .. 3 4 8 8 2018 30 24 25 17 16 5 4 3 1 1 1 1 2 1 No. of Leaves. A. =20.9 ±.161 S. D. = 3.31 ±.114 Coef. Cor. Height of Plants. A. =62.9 ±.449 S. D. = 9.23 ±.318 +.786 ±.023 4 1 5 6 22 19 19 27 28 15 12 17 8 4 3 2 192 CORRELATION TABLES. 45- TABLE XX XL CORRELATION BETWEEN NO. OF LEAVES AND AVERAGE AREA OF LEAVES OF (402 X 405)-l, HAVANA X CUBAN, F2. No. of Leaves. o o u > < 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 3 4 5 6 1 12 2 '1 '1 1 '2 3 2 5 1 .. 1 .. 6 3 3 1 . . 1 .... 2 . . 1 1 1 . . 3 6 2 2 2.. 1.. 1 5 2 2 3 . . 1 2 . . . . 1 . . 1 . . 7 8 9 10 11 1112 2 4 9 .. 12 2 4 3 4 .... 2 . . 444 . . 1 1 1 . . . . 2 1 6 8 2 6 1.... 1 1 4 4 4 2.. 2 1.... 2 4 1 2 1 1 1 1 1 . . 1 . . 2 1 . 1? 3 4 8 8 20 18 30 24 25 17 16 5 4 3 1 1 1 1 2 1 , 3 10 32 31 45 33 24 9 4 1 192 No. of Leaves. A. =20.9 ±.161 S. D. = 3.31 =t. 114 Coef. Cor. = Average Area of Leaves. A. = 6.96 ±.085 S. D. = 1.76 ±.061 .092 ±.048 TABLE XXXII. CORRELATION BETWEEN AVERAGE WIDTH OF LEAVES AND AVERAGE LENGTH OF MIDRIB OF (402 X 405)-l, HAVANA X CUBAN, F2. Aver. Width of Leaves in Cms. 27 30 33 36 39 42 45 48 51 54 12 15 18 21 24 27 30 1 11 .53.. . 3 13 2 . 2 14 11 . . . 5 22 8 .. .. ... 1 20 7 21 6 . . 15 11 1 5 6 .. 2 3 2 1 11 37 62 52 26 3 Aver. Width of Leaves. A. =21.8 ±.170 S. D. = 3.49 ±.121 Coef. Cor. 8 18 28 35 48 34 11 _J 192 Aver. Length of Leaves. A. =43.2 ±.247 S. D. = 5.28 ±.182 + .814±.016 PLATE O O^ • rj i- o 1^ '-' ^-^ P. C 0) tfl U .—I 0) o oS PLATE II. a. Average middle leaf of No. 402, Havana at left, of No. 405, Cuban at right and Fi in center. b. Some Fo middle leaves of cross between No. 402, Havana and No. 405, Cuban. PLATE III. to I > D 03 O pj en in «-t-J ^ cd rt o .^ o '^ o fe S' CD ^ '_, A s a^ (J T-H O 13 'd ^ c 11 1=1 rt y=! _rt 03 M P^ w o 03 > 01 o K rt fi 0) (-> (UTS rO D rrt < 0) 1—1 T3 1 rn Pi 1 o ;3 u-3 CJ r{ ffi & M-1 o o PLATE IV. p m OS'S , ^ c/3 b)D > p; >r>^ ■1) ??, R tn O^ ' O u o a cu ^ ClJ r- •s^ s o ,rtc^ o ^ 2 ^ ^ CQ 5 ^ ;-( 03 t/i r/l 01 ci; b/1 > m o3 t-i <1J , 1 0) > l-l 1— 1 ON < '"' ^J <1) cj ^. > 03 u fl> C^ ^1 O O * rn <1) ■M > o3 o ,^ PLATE V. M o C 0) n) < S ^ R R : 7 S '■„/ i„^ \„." v.../ i V.J University of Connecticut Libraries 39153029221142