全 文 ：第28卷 第2期 作　物　学　报 V ol. 28, N o. 2
2002 年3月　 196～ 202页 A CTA A GRONOM ICA S IN ICA pp. 196～ 202　M ar. , 2002
Heterosis and Gene Action of Bol l W eight and L in t Percen tage in H igh
Qual ity F iber Property Var ieties in Upland CottonΞ
YUAN You2L u1　ZHAN G T ian2Zhen13 　GUO W ang2Zhen1　PAN J ia2Ju1　R ussell J Kohel2
(1 N ational K ey L aboratory of C rop Genetics & Germ p lasm E nhancem ent, D epartm ent of Genetics & C rop B reed ing , N anj ing A g ricultural
U niversity , N anj ing 210095, China; 2U SDA A R S , S outhern P lain A g riculture R esearch Center, C rop Germ p lasm R esearch U nit, College S tation,
T X 77845, U SA )
Abstract　Tw en ty F 1 com binations crossed among 5 varieties and strain s differen t in fiber p roperties acco rding
to comp lete diallel crossing design w ere used to evaluate the heterosis and gene action of bo ll w eigh t and lin t
percen tage in h igh quality fiber p roperty varieties in U p land Co tton in 1998～ 1999 fo r two year successively at
N an jing. It w as indicated that there ex isted sm all in teractions w ith the environm en tal facto rs w ithout m aternal
effects and the additive gene effect w as in ch ief, attain ing to 51. 2% and 65. 4% respectively fo r bo llw eigh t and
lin t percen tage. T he dom inan t effect w as also in h igher rate, 32. 6% and 16. 8% respectively. T he population
m ean heterosis of bo ll w eigh t and lin t percen tage over the m id2paren talm ean w ere relatively p rom inen t 13. 3%
and 3. 5% respectively in ex trem e sign ificance. How ever, bo ll w eigh t show ed no sign ifican tly surpassing
paren tal F 1 heterosis over the better paren t based on population m ean ( 2. 0% ) ; w h ile the lin t percen tage
exp ressed sign ifican t negative heterosis value ( - 2. 1% ). T he gene actions w ere in confo rm ity w ith the
heterosis exp ression. It w as show n clearly that the F 1 com binations crossed betw een paren ts w ith sim ilar
perfo rm ances had relatively h igh dom inan t effects and sign ifican t positive F 1 surpassing paren tal heterosis (F 1
heterosis over the better paren t) ; w h ile no F 1 com bination crossed betw een the paren ts w ith p rom inen t m utual
difference surpassed the h igher paren t in yield componen ts, w h ich indicated that among those paren ts w ith less
difference and close relationsh ip s, there still ex isted sufficien t genetic variation o r certain m echan ism fo r creating
variation and ach ieving greater advances in breeding. Correlation analyses also indicated that there still ex isted
severely undesirable negative co rrelation betw een yield and fiber p roperties as w ell as the difficulties fo r their
sim ultaneous imp rovem en ts.
Key words　U p land co tton; H igh quality fiber p roperty; Bo llw eigh t; L in t percen tage; Inheritance; H eterosis
CLC num ber: S562　　　D ocum en t code: A
陆地棉优异纤维品系的铃重和衣分的遗传及杂种优势分析
袁有禄1　张天真1　郭旺珍1　潘家驹1　R J Kohel2
(1作物遗传与种质创新国家重点实验室, 南京农业大学遗传育种系, 江苏南京 210095; 2U SDA ARS, Southern P lains A griculture Research
Center, College Station, TX77845, U SA )
摘　要　利用5个具有不同纤维品质性状的品种 (系)配制完全双列杂交组合20个, 通过亲本和 F 1的2年随机区组试验发
现产量性状的铃重和衣分与环境的互作效应小, 不存在母体效应, 并以加性遗传效应为主, 分别占表型方差的51. 2%
和65. 4% ; 显性遗传效应所占的比率也较高, 分别为32. 6% 和16. 8%。铃重和衣分的群体平均优势较大, 分别为13. 3%
和3. 5% , 达到了极显著; 铃重的超亲优势为2. 0% , 不显著; 衣分为显著的负值 (- 2. 1% )。遗传分析与杂种优势结果
一致。具体表现在产量性状上, 亲本相当配制的组合杂合显性较高, 其超亲优势正向显著, 而极值亲本 (差异较大) 所Ξ Foundation item s: N ational N atural Science Foundation of China (39830240)
Biography: Yuan Youlu (1967. 10) ; m ale; Sichuan p rovince; Ph. D. ; P resent address: Cotton Research Institute of CAA S, A nyang Henan,
455112; Research field: Genetics & Cotton B reeding.
Correspondence author: E2m ail: co tton@m ail. njau. edu. cn; Fax: 02524395307
Received on: 2001204224, A ccep ted on: 2001208207

配组合没有超过高亲的。这表明亲本差异小、亲源关系较近的亲本中仍然存在足够的遗传变异或某种机制以创造变异
使育种取得更大的进展。相关分析表明了仍然存在严重的品质与产量的负相关, 遗传改良的难度较大。
关键词　陆地棉; 优质纤维; 铃重; 衣分; 遗传; 杂种优势
　　U p to date, the yield increase in o rdinary co tton
varieties seem s no t very p rom inen t, w hereas hybrid
co tton show ed rather p rom ising po ten tial fo r yield
increasing [ 1～ 5 ], especially the comm ercial cultivation
of transgen ic B t in sect2resistan t hybrid co tton
p romoted the utilization of co tton heterosis to a new
stage[ 6 ].
In the w ake of successes in breeding large
amoun t of m aterialsw ith excellen t fiber quality [ 7 ] , to
realize the genetic speciality of these m aterials and the
po ten tiality of their hybrid vigo r m ay be conducive to
enfo rcing the tactics fo r breeding desirable varieties
and hybrid com binations. In th is research, comp lete
diallel2crossing com binations among U p land co tton
varieties w ith differen t fiber strength w ere analyzed
to realize the inheritance of the h igh quality fiber
p roperties and the po ten tiality fo r imp roving yield
and fiber p roperties sim ultaneously. It w as found
that the bo ll w eigh t and lin t percen tage of these
paren ts w ere all d iversified from each o ther, besides,
the inheritance and heterosis fo r these two characters
w ere sign ifican tly differen t from those of fiber
p roperties. In th is paper, the results of studying the
bo ll w eigh t and lin t percen tage of these m aterials
w ere discussed more emphatically.
1　M a ter ia ls andM ethods
1. 1　M a ter ia l
F ive varieties o r strain s, TM 21 ( TM 21 ) ,
PD 6992 ( PD 69 ) , M D 51ne (M D 51 ) , H S427210
(H S42) and 7235 differen t in fiber strength from
each o ther (T able 1) w ere used. 7235 w as supp lied
by Industrial C rop Institu te, J iangsu A cadem y of
A gricultural Sciences and w as further pedigree2
selected after being in troduced in our labo rato ry.
TM 21, PD 6992, M D 51ne, and H S427210 w ere all
supp lied by D r. Kohel in the Southern P lain
A gricultural R esearch Cen ter, Co llege Station,
T exas U SA. T he genes relevan t to the excellen t
fibers in the paren tal m aterials used w ere differen t
from each o ther. Tw en ty com binations w ere
p roduced acco rding to the comp lete diallel2crossing
model in the W ild Co tton P lan ting Garden of Co tton
R esearch Institu te ( CR I ) , CAA S, in H ainan
p rovince in the w in ter 1997 and w ere p lan ted at the
J iangpu Cotton B reeding Station, N an jing
A gricultural U n iversity (NAU ) in the summ er
1998. Fo r scien tific analyses the cross com binations
w ere divided in to differen t types as: h igh×h igh (H×
H ) , h igh × low (H ×L ) , and low × low (L ×L )
acco rding to fiber strength. T he paren tsw ere all self2
po llinated fo r keep ing their genetic purity.
1. 2　Exper im en ta l m ethod
F ive paren ts and 20 F 1 com binations w ere
p lan ted in J iangpu Cotton B reeding Station of NAU
in 1998 and 1999 in random ized block s w ith 3
rep lications, double row s in a p lo t, each row as long
as 5 m , 0. 8 m apart betw een two adjo in ing row s,
and 15 p lan ts in each row. Seeds w ere sow n in the
nursery on A p ril 15 and A p ril 11, and the seedlings
transp lan ted to the experim en tal field on M ay 13 and
M ay 17 in 1998 and 1999. O ther field m an ipulations
w ere sam e as those exercised in the comm ercial
fields. T he clim atic conditions w ere no rm al in 1998,
but w ith more rainy days and low er temperature
during later developm en tal stage in 1999, w h ich
caused adverse influences on the fiber developm en t.
T he 50 norm ally opened bo lls w ere p icked fo r
assaying the bo ll w eigh t (gram s of seed co tton per
bo ll) , lin t percen tage and fiber p roperty characters.
1. 3　Sta tistica l m ethod
T he average values per p lan t w ere used as un its
fo r Statistical analyses. T he additive2dom inance2
m aternal genetic model ( ADM ) and additive2
dom inance genetic model (AD ) w ere analyzed by
M INQU E ( 1 ) m ethod [ 8～ 9 ] fo r estim ating variance
and covariance as w ell as fo r further calculating the
ratio s of genetic variance ( V g ) over pheno typ ic
7912期　　YUAN You2L u et al. 　H eterosis and Gene A ction of Bo llW eigh t and L in t⋯⋯　　　　

Table 1　The f iber properties of f ive parents
Parents
F iber strength
(cN ötex)
1998 1999
M icronaire
　
1998 1999
F iber length
(mm )
1998 1999
U niform ity
(% )
1998 1999
E longation
(% )
1998 1999
TM 21 20. 77 20. 83 5. 13 4. 53 30. 33 30. 37 49. 30 48. 20 5. 97 5. 33
PD 6992 23. 63 23. 57 4. 77 3. 83 33. 43 32. 97 47. 00 45. 87 5. 27 5. 23
MD 51ne 21. 80 20. 07 4. 90 4. 00 30. 93 30. 23 49. 07 46. 00 5. 83 5. 23
HS427210 26. 20 25. 80 4. 90 4. 57 31. 97 33. 10 49. 17 48. 77 5. 07 4. 30
7235 28. 50 29. 50 4. 13 3. 70 34. 03 35. 50 45. 77 46. 70 5. 13 4. 20
variance (V p ). T he adjusted unbiased p rediction
m ethod (AU P) [ 8 ] w as adop ted to estim ate the genetic
effects, geno typ ic values, heterosis over the m id2
paren ts o r better paren t (BP ) based on population
m ean. T he Jackkn ife re2samp ling m ethod [ 10 ] w as
used to calculate the standard erro rs of the estim ated
values fo r t2test to assay the sign ificance of
differences among p roperties. T he m ethods fo r
calculating heterosis w ere as fo llow s:
H eterosis over m id2paren t based on population
m ean (m id2paren t population m ean heterosis) :
Hpm (F 1) = (F 1- M P) öΛ;
H eterosis over better paren t based on population
m ean (surpassing paren tal heterosis) :
Hpb (F 1) = (F 1- BP) öΛ;
In above fo rm ulas, F 1 w as rep resen ted as F 1
geno typ ic value, M P as the average of bo th paren t′s
geno typ ic values; BP as the better paren ts geno typ ic
value, and Λas the arithm etic m ean of all the paren ts
and F 1 s in the diallel2crossing.
2　Results
2. 1　The var iance ana lysis of bol l we ight and l in t
percen tage
T he bo ll w eigh t and lin t percen tage of paren ts
and F 1 com binations in the diallel2crossing design
experim en t fo r 1998 and 1999 w ere listed in T able 2.
Excep t the PD 69, ( 7235 × PD 69) F 1, ( TM 21 ×
M D 51) F 1, the bo ll w eigh t of o ther paren ts and F 1
com binations in 1999 w as heavier by 0. 43 g on
average than those in 1998 at ex trem ely sign ifican t
level ( t2test, P = 2. 07×10- 8 ). T he bo ll w eigh t of
TM 21, PD 269 and M D 51 w as relatively heavier,
5. 92g, 5. 71 g and 5. 88 g, respectively; w h ile that
of H S42 and 7235 w as relatively ligh ter, 4. 01 g and
4. 12 g, respectively. T en F 1 com binations obviously
surpassed the better paren t in bo ll w eigh t, show ing
p rom inen t surpassing paren tal heterosis. T he lin t
percen tages of all the paren ts and F 1 s in 1999 w ere
low er than in 1998 by 2. 42% on the average at
ex trem ely sign ifican t difference level ( t2test, P= 1. 6
×10- 13). Among 5 paren ts, H S42 had the h ighest
lin t percen tage, 38% ; w h ile those of 7235 and PD 69
w ere the nex t 33. 16% and 32. 98% , respectively;
those of TM 21 and M D 251 w ere relatively low ,
31. 26% and 30. 67% respectively. T he lin t
percen tages of most F 1 com binations w ere
app rox im ate to those of the h igher valued paren t. F 1 s
from the recip rocal crosses show ed sm all difference
bo th in bo llw eigh t and lin t percen tage.
A cco rding to the results analyzed by ADM and
AD genetic model[ 9 ] , the residual variances of bo ll
w eigh t and lin t percen tage w ere 11. 8% and 15. 5%
respectively, the in teraction variances of the additive
and dom inance effect w ith environm en tal facto rs
w ere very sm all, on ly 0. 3%～ 2. 7% , w h ich
indicated that bo th bo ll w eigh t and lin t percen tage
w ere m ain ly con tro lled by genetic facto rs. Bo th bo ll
w eigh t and lin t percen tage show ed no m aternal
effects, but additive genetic effect w as in ch ief,
amoun ting to 51. 2% and 65. 4% , respectively. Bo th
bo ll w eigh t and lin t percen tage show ed, how ever,
relatively h igh dom inance effects at 32. 6% and
16. 8% respectively, bo th at ex trem ely sign ifican t
level by T 2test (T able 3). In general, bo th bo llw eigh t
and lin t percen tage, sam e as fiber strength and length
show ed h igh heritability and additive genetic effects
in the m ain and less in teraction w ith environm en tal
891　　　　　　　　　　　　　　　　　　　　　作　　物 　　学　　报　　　　　　　 　　　　　　　　　　　　28卷

Table 2　The bol l we ight and l in t percentage of parents and 20 F1 combinations
Parent
TM 1b
1998 1999
PD 69b
1998 1999
MD 51b
1998 1999
HS42b
1998 1999
7235b
1998 1999
Boll TM 1a 5. 77 6. 07 6. 10 6. 83 5. 70 5. 58 5. 57 5. 87 5. 37 5. 99
w eigh t PD 69a 6. 17 6. 67 5. 77 5. 65 6. 10 6. 68 5. 10 5. 76 4. 93 5. 54
(g) MD51a 5. 73 6. 34 6. 13 6. 84 5. 87 5. 90 5. 73 6. 26 5. 47 5. 99
HS42a 5. 70 6. 43 5. 37 5. 48 5. 77 6. 23 3. 60 4. 43 4. 53 5. 14
7235a 5. 50 5. 92 5. 50 5. 49 5. 67 5. 93 4. 57 4. 92 3. 83 4. 40
L int TM 1a 32. 17 30. 36 33. 63 31. 18 32. 51 31. 43 34. 63 32. 99 35. 07 32. 44
percentage PD 69a 33. 69 30. 70 34. 84 31. 12 33. 64 30. 50 39. 08 37. 13 37. 15 34. 55
(% ) MD51a 31. 90 31. 07 34. 25 31. 63 31. 89 29. 45 35. 48 33. 17 35. 16 32. 88
HS42a 35. 03 32. 80 38. 09 37. 78 35. 46 32. 04 39. 09 36. 91 38. 80 36. 13
7235a 34. 59 30. 94 36. 75 33. 68 33. 98 31. 64 39. 02 35. 96 34. 49 31. 82
　Note: a2m aternal parent; b2paternal parent
Table 3　The ratios between estimated var iance and phenotypic var iance for bol l we ight and l in t percentage
Character VA öV P VD öV P VA EöV P VDEöV P VM öV aP VM EöV aP Residual
Bollw eigh t (g) 0. 512± 0. 326± 0. 017± 0. 027± 0. 003± 0. 034± 0. 118±
0. 0533 3 0. 0453 3 0. 0053 0. 0133 0. 002 0. 019 0. 0323 3
L int percentage, 0. 654± 0. 168± 0. 003± 0. 020± 0. 036± 0 0. 155±
(% ) 0. 0463 3 0. 0383 3 0. 0013 0. 013 0. 026 0. 0383 3
　Notes: a2ADM model analysis; V P2phenotyp ic variance; VA 2additive variance; VD 2dom inance variance; VA E2in teraction variance betw een additive
and environm ental effects; VDE2in teraction variance betw een dom inance and environm ental effects; VM 2m aternal genetic variance; VM E2
in teraction variance betw een m aternal and environm ental effects; Residual2p roportion of the residual variance
　3 , 3 3 2significance at P= 0. 05 and P= 0. 01 level respectively.
Table 4　The dom inance genetic effect of bol l we ight and l in t percentage
Cross
Bollw eigh t (g)
D DE1 DE2
L int percentage (% )
D DE1 DE2
TM 1×PD69 0. 44±0. 113 - 0. 02±0. 05 0. 28±0. 12 - 0. 56±0. 30 - 0. 13±0. 16 - 1. 24±0. 96
TM 1×MD 51 - 0. 36±0. 20 - 0. 04±0. 11 - 0. 22±0. 24 0. 40±0. 29 - 0. 31±0. 50 1. 68±0. 633
TM 1×HS42 0. 52±0. 063 3 0. 13±0. 07 0. 19±0. 043 3 - 1. 00±0. 313 - 0. 42±0. 33 - 0. 21±0. 27
TM 1×7235 0. 30±0. 063 3 0. 04±0. 06 - 0. 13±0. 21 0. 32±0. 18 0. 35±0. 42 - 0. 42±0. 23
PD69×MD 51 0. 41±0. 053 3 - 0. 06±0. 03 0. 52±0. 193 - 0. 24±0. 23 0. 07±0. 14 - 1. 81±1. 68
PD 69×HS42 0. 06±0. 08 0. 03±0. 04 - 0. 03±0. 05 2. 40±0. 443 3 0. 24±0. 14 2. 67±1. 09
PD 69×7235 0. 01±0. 07 0. 04±0. 03 - 0. 19±0. 18 1. 23±0. 55 0. 51±0. 55 1. 01±0. 85
MD 51×HS42 0. 62±0. 063 3 0. 11±0. 09 0. 20±0. 053 - 0. 69±0. 263 0. 06±0. 21 - 1. 58±0. 68
MD51×7235 0. 40±0. 103 0. 06±0. 09 0. 02±0. 08 0. 77±0. 46 - 0. 36±0. 33 1. 50±1. 02
HS42×7235 0. 10±0. 04 0. 03±0. 06 - 0. 02±0. 04 1. 32±0. 223 3 0. 70±0. 48 0. 22±0. 36
　Note: D 2dom inance effect; DE12in teraction effect betw een dom inance and environm ent 1,
DE22in teraction effect betw een dom inance and environm ent 2
facto rs, but they w ere characterized as rather h igh
positive dom inan t effects fo r most of F 1 s (T able 4) as
compared w ith fiber strength (published elsew here).
2. 2　Heterosis Ana lysis of Bol l W e ight and L in t
Percen tage
F rom T able 5, bo ll w eigh t and lin t percen tage
show ed relatively h igh m id2paren t population m ean
heterosis, 13. 3% and 3. 5% respectively at
ex trem ely sign ifican t level. Bo ll w eigh t show ed
surpassing paren tal heterosis 2% , at in sign ifican t
level; w h ile lin t percen tage show ed negative
surpassing paren tal heterosis at sign ifican t level
( - 2. 1% ). A s regards to the differen t F 1
com binations, excep t TM 21×M D 51, all the o ther
n ine F 1 com binations show ed sign ifican tly positive
m id2paren t population m ean heterosis fo r bo ll w eigh t
( 8. 9%～ 21. 8% ) , w h ich m igh t be relevan t to the
h igh ly positive dom inan t effect (T able 4) ; th ree F 1
com binations ( TM 21 × PD 69, PD 69 ×M D 51 and
H S42×7235) show ed sign ifican tly positive surpassing
9912期　　YUAN You2L u et al. 　H eterosis and Gene A ction of Bo llW eigh t and L in t⋯⋯　　　　

paren t heterosis fo r bo ll w eigh t (11. 1%～ 13. 8% ) ;
w h ile the difference betw een bo th paren ts of these
com binations w ere all sm all, the ratio s of bo th
paren ts in each com bination w ere 1. 02～ 1. 04. T he
com bination of TM 21 × 7235 show ed sign ifican tly
negative surpassing paren tal heterosis(- 3% ) fo r bo ll
w eigh t, the ratio of bo ll w eigh ts betw een bo th
paren ts w as 1. 44, m uch h igher than that above
m en tioned.
Six F 1 com binations show ed sign ifican tly positive
m id2paren t population m ean heterosis fo r lin t
percen tage (2. 5%～ 8. 8% ) , w h ich m igh t be
sim ilarly relevan t to the h igh positive dom inance
effects ( T able 5 ). T h ree F 1 com binations show ed
surpassing paren tal heterosis. T he surpassing
paren tal heterosis of TM 21 ×H S42 and M D 51 ×
H S42 w ere w ith sign ifican tly negative values
- 13. 7% and - 13. 1% , respectively; w h ile PD 69
×7235. Show ed sign ifican tly positive value 7. 8%.
In the 4 F 1 com binations among w h ich the bo ll
w eigh t o r lin t percen tage all show ed sign ifican tly
positive surpassing paren tal heterosis, the paren tal
values in each com bination w ere m utually
co rresponding, the ratio s betw een two paren t in each
com bination w ere in the range of 1. 01～ 1. 04, these
com binations consisted of such types as H ×H , L ×
L , and M H×M H. How ever, the com binationsw ith
paren ts w ith po larized values, i. e. bo th paren ts had
large difference from each o ther, such as TM 1 ×
7235 fo r bo ll w eigh t, TM 1 ×H S42 and M D 51 ×
H S42 fo r lin t percen tage, the paren tal ratio s in each
com binations w ere larger as 1. 22～ 1. 44, the
population surpassing paren tal heteriso s w ere allw ith
ex trem ely sign ifican t negative values ( T able 6 ).
T herefo re, as regards yielding traits, the F 1
com bination w ith bo th paren ts m utually
co rresponding in values show ed rather h igh
heterozygous dom inance and sign ifican tly positive
surpassing paren tal heterosis, w hereas no F 1
com bination w ith m utually po larized paren ts (w ith
large difference in values betw een them ) show ed
surpassing paren tal heterosis (w ith heterosis h igher
than the better paren t).
2. 3 　The correla tion s between bol l we ight, l in t
percen tage and f iber properties
Correlation analyses(T able 7) indicated that bo ll
w eigh t show ed ex trem ely sign ifican t negative
co rrelation w ith fiber length and strength, w ith
geno typ ic co rrelation coefficien ts - 0. 47 and
- 0. 74, respectively, w hereas the lin t percen tage,
just on the con trary, show ed ex trem ely sign ifican t
positive co rrelation w ith these fiber p roperties , w h ich
Table 5　M id-parent population mean heterosis and surpassing parenta l heterosis of bol l we ight and l in t percentage
Character
Hpm (F1)
A verage Range + N - N
Hpb (F1)
A verage Range + N - N
Bollw eigh t (g) 0. 1333 3 - 0. 007～ 0. 218 9 (9) 1 0. 02 - 0. 057～ 0. 138 5 (3) 5 (1)
L int percentage (% ) 0. 0353 3 - 0. 028～ 0. 088 8 (6) 2 - 0. 0213 - 0. 137～ 0. 078 5 (1) 5 (2)
　Note: Hpm (F1) : A verage m id2parent population m ean heterosis in F1; 　Hpb (F1) : surpassing parental heterosis in F1;
+ N : num ber of com binations w ith positive heterosis; 　 () : num ber of com binations attaining to positive and negative heterosis;
- N : num ber of com binations w ith negative heterosis
Table 6　The parenta l difference and dom inance effects of combinations with sign if icant
heterosis over m id-parent or the better parent based on population mean
Character Com bination Type P1- P2 P1öP2 D ij Hpm (F1) Hpb (F1)
Bollw eigh t TM 1×7235 H×L 1. 80 1. 44 0. 303 3 0. 1373 3 - 0. 0303
g PD 69×MD 51 H×H 0. 17 1. 03 0. 413 3 0. 1333 3 0. 1163 3
TM 1×PD 69 H×H 0. 21 1. 04 0. 443 0. 1313 3 0. 1113 3
HS42×7235 L×L 0. 11 1. 02 0. 10 0. 1543 3 0. 1383 3
L int percentage% TM 1×HS42 L×H 6. 74 1. 22 - 1. 003 - 0. 028 - 0. 1373 3
MD51×HS42 L×H 7. 33 1. 24 - 0. 693 - 0. 010 - 0. 1313 3
PD 69×7235 M H×M H 0. 18 1. 01 1. 23 0. 0793 0. 0783
TM 1×MD 51 L×L 0. 59 1. 02 0. 40 0. 028 0. 017
　D ij: dom inance effect
002　　　　　　　　　　　　　　　　　　　　　作　　物 　　学　　报　　　　　　　 　　　　　　　　　　　　28卷

Table 7　The correlation between bol l we ight, l in t percentage and f iber properties
Character L ength U niform ity Strength E longation M icronaire Boll w eigh t
Boll w eigh t - 0. 39±0. 033 3 a - 0. 04±0. 09 - 0. 65±0. 043 3 0. 49±0. 063 3 0. 18±0. 073 2
- 0. 47±0. 043 3 b - 0. 11±0. 12 - 0. 74±0. 053 3 0. 63±0. 073 3 0. 23±0. 093 2
L int percentage % 0. 43±0. 043 3 0. 21±0. 11 0. 55±0. 043 3 - 0. 49±0. 063 3 0. 17±0. 07 - 0. 53±0. 043 3
0. 55±0. 053 3 0. 34±0. 17 0. 65±0. 063 3 - 0. 61±0. 103 3 0. 13±0. 09 - 0. 59±0. 043 3
　Note: a2phenotyp ic correlation; 　　b2genotyp ic correlation
kep t in confo rm ity w ith the ex trem ely negative
co rrelation heterosis betw een bo ll w eigh t and lin t
percen tage (- 0. 59). In th is experim en t, w e found
that bo ll w eigh t show ed sign ifican t heterosis
exp ression, w h ich suggested there still ex isted
severely negative co rrelation betw een fiber p roperty
and yields.
3　D iscussion
3. 1　The inher itance of bol l we ight and l in t
percen tage of excel len t f iber var ieties
A lthough the five paren ts used in th is research
p ro ject w ere sign ifican tly differen t in fiber p roperties
( length, strength and m icronaire ) and yielding
componen ts (bo ll w eigh t and lin t percen tage) , yet
the inheritance patterns of yield componen t w ere
apparen tly differen t from those of fiber p roperties.
Bo ll w eigh t and lin t percen tage did no t show
m aternal effect, w h ile additive genetic effectsw ere in
ch ief w ith rate as h igh as 51. 2% and 65. 4% ,
respectively, w h ich w as in confo rm ity w ith the
inheritance of fiber length and strength. How ever,
the rate of dom inance genetic effect of bo ll w eigh t
and lin t percen tage w ere as h igh as 32. 6% and
16. 8% , respectively, w h ile that of strength and
length w ere on ly 3. 8% and 6. 0% , respectively,
that of m icronaire w as also as low as 11. 5%
(published elsew here ). It confirm ed our p revious
result that the inheritance of fiber p roperties w as
ch iefly con tro lled by additive effects, less by
dom inance effects; w hereas among yielding
componen ts, bo ll w eigh t and lin t percen tage w ere
con tro lled by additive genetic effects in ch ief and by
relatively h igh dom inance effects as w ell. T he
in teraction effect w ith the environm en tal facto r w as
less on bo th aspects. T hese results w ere in
acco rdance w ith m ajo rity results by fo rm er
researches.
3. 2　Heterosis performance of y ield ing perfor-
mances under d ifferen t f iber property backgrounds
Boll w eigh t and lin t percen tage show ed rather
p rom inen t m id2paren t population m ean heterosis,
13. 3% and 3. 5% , respectively, attain ing to
ex trem ely positive sign ifican t level in most
com binations. T he surpassing paren tal heterosis of
bo llw eigh t w as insign ifican t 2. 0% ; w h ile that of lin t
percen tage w as negative sign ifican tly - 2. 1%.
T hese results of th is experim en t w ere also in
confo rm ity w ith those in curren t researches. In th is
experim en t, bo ll w eigh t and lin t percen tage show ed
sign ifican t m id2paren t population m ean heterosis,
w h ile fiber p roperties show ed less m id2paren t
population m ean heterosis, on ly m icronaire show ed
relatively h igh value ( 3. 2% ). T he sign ificance in
m id2paren t population m ean heterosis m ay keep close
relationsh ip w ith the dom inance effect of these traits
and w as in confo rm ity w ith the fo rm erly m en tioned
genetic analytical results.
T here ex isted rather large difference among fiber
p roperties and yielding componen t among the
paren tal varieties used in th is research. T he hybrids
could be divided in to differen t types of crossing
com binations such as h igh×h igh, h igh×low and low
× low etc. In 15 F 1 s (H ×L ) w ith the m ax im um
difference betw een two paren ts fo r yield componen ts
and fiber p roperties, negatively surpassing paren tal
heterosis w ere p roduced, in 12 of them the heterosis
attained to ex trem ely sign ifican t level, w h ich m ean t
that none of the F 1 com binations w ith sign ifican t
paren tal differences gave positive surpassing paren tal
heterosis E igh t F 1 com binations w ith in sign ifican t
difference betw een two paren ts (H ×H , L ×L , and
M H ×M H ) fo r the fiber p roperties show ed no
sign ifican t m id2paren t population m ean and
1022期　　YUAN You2L u et al. 　H eterosis and Gene A ction of Bo llW eigh t and L in t⋯⋯　　　　

surpassing paren tal heterosis. How ever, fo r bo ll
w eigh t and lin t percen tage, 5 F 1 com binations w ith
in sign ifican t difference betw een two paren ts (H×H ,
L ×L , and M H ×M H ) show ed positive dom inance
effect, positive m id2paren t population m ean and
surpassing paren tal heterosis 4 of them attain ing to
sign ifican t o r ex trem ely sign ifican t levels. T hese
results indicated apparen tly that F 1 com binationsw ith
sim ilar paren tal values usually p roduced relatively
strong heterosis fo r yield componen ts. T he results
w ere in confo rm ity w ith those of heterosis exp ressed
in F 2 repo rted by T ang et al. (1993).
Curren tly w ith regard to the relationsh ip
betw een paren tal diversity and F 1 heterosisM eredith
( 1998 ) concluded that paren tal coefficien ts and
molecular m arkers could scarcely be utilized in
distinguish ing the paren ts of excellen t F 1
com binations. T he co rrelation coefficien t betw een
genetic distance (calculated from R FL P data ) and
heterosis w as 0. 08, w h ile that betw een paren tal
coefficien t and heterosis w as also low 0. 05 on ly, all
show ed very insign ifican t co rrelation [ 18 ] , V an
E sbroeck (1998) considered that genetic distance did
no t con tribute m uch to the varietal imp rovem en t,
w h ile excellen t varieties m igh t be o riginated from the
paren ts w ith clo se relationsh ip s[ 19 ].
It w as indicated by all these researches that
among paren ts w ith less diversities w h ile w ith clo se
relationsh ip s, there still ex ist sufficien t genetic
variation o r certain m echan ism that m ay be
conductive to creating variation fo r the advances in
breeding. P resen t Q TL researches revealed that
among morpho logically alike paren ts do ex ist large
amoun t of Q TL w ith diversified effects and opposite
in direction [ 20 ]
References
[ 1 ] 　Deussen H. Imp roved cotton fiber p roperties: the textile
industry′s key to success in global competition. In
(P roceedings from cotton fiber cellulose: structure, function,
and utilization conference. ) M emphis, TN ; N ational Cotton
Council of Am erican, 1992, 43～ 64
[ 2 ]　 M acdonald A G. F iber needs for textile consum er. In
( P roceedings of the beltw ide co tton p roduction research
conference. )M emphis, TN : N ational co tton council, 1997,
1235
[ 3 ] 　Xiang S2K (项时康) , Yu N (余楠) , Hu Y2C (胡育昌).
D iscussion on the current situation of co tton quality in China.
A cta Gossy p ii S inica (棉花学报) , 1999, 11: 1～ 10
[ 4 ] 　Chaudhry M R. Cotton yield stagnating. (The Int. Cotton
A dvisory Comm ittee Recorder. ) 1997, XV (2) : 2～ 3
[ 5 ] 　Chaudhry M R. Comm ercial co tton hybrids (The Int. Cotton
A dvisory Comm ittee Recorder. ) 1997, XV (2) : 3～ 14
[ 6 ] 　Yuan Y2L (袁有禄) , J ing S2R (靖深蓉) , X ing C2Z (邢朝
柱). A Review on research and utilization of hybrid vigor of
co tton. China Cotton (中国棉花) , 2000, 27: 2～ 5
[ 7 ] 　Q ian S2Y (钱思颖) , Huang J2Q (黄骏麒) , Peng Y2J (彭跃
进 ). Studies on the hybrid of G. hirsutum L. and G.
anom alum W aw r & Peyr. and app lication in breeding. S ci.
A g ri. S inica (中国农业科学) , 1992, 25: 44～ 51
[ 8 ] 　Zhu J (朱军). M ethods of p redicting genotype value and
heterosis for offsp ring of hybrids. J . B iom ath (生物数学学
报) , 1993, 8: 32～ 44
[ 9 ] 　Zhu J (朱军) , J i D 2F (季道藩) , Xu F2H (许馥华). M ixed
model app roaches for analyzing intra2cultivar heterosis in crop.
A cta Genetica S inica (遗传学报) , 1993, 20: 262～ 271
[ 10 ] 　M iller R G. The jackknife: a review. B iom etrika, 1974, 61: 1
～ 15
[ 11 ] 　Yuan Y2L (袁有禄) , Zhang T2Z (张天真) , J ing S2R (靖深
蓉). Studies on utilization of second generation co tton hybrids
betw een the recessive glandless varieties and the dom inant
glandless lines. A cta Gossy p ii S inica (棉花学报) , 1999, 11: 11
～ 19
[ 12 ] 　M eredith W R. Yield and fiber quality potential fo r second2
generation co tton hybrids C rop S ci, 1990, 30: 1045～ 1048
[ 13 ] 　M iller P A , L ee J A. Heterosis and com bining ability in
varietal top crosses of U p land cotton Gossy p lum hirsutum L.
C rop S ci, 1965, 5: 646～ 649
[ 14 ] 　M eredith W R J r, B ridge R R. Heterosis and gene action in
co tton, Gossy p ium hirsutum L. C rop S ci, 1972, 12: 304～ 310
[ 15 ] 　M eredith W R J r. Relative perform ance of F1 and F2 hybrids
from doubled hap lo ids and their parent varieties in U p land
Cotton, Gossy p ium hirsutum L. C rop S ci, 1970, 10: 295～ 298
[ 16 ] 　M ay O L. Genetic variation in fiber p roperty. In: Basa (ed. )
(Cotton F ibers: developm ental bio logy, quality imp rovem ent
and textile p rocessing N ew York ). Food P roducts P ress,
1999, 183～ 229
[ 17 ] 　Tang B, Jenkins J N , M cCarty J C. F2 hybrids of host p lant
germp lasm and cotton cultivars: É heterosis and com bing
ability for lint yield and yield components. C rop S ci, 1993,
33: 700～ 705
[ 18 ] 　M eridith W R J r, B row n. J S. Heterosis and com bing ability
of co ttons originating from different regions of the U nited
States. J Cotton S ci, 1998, 2: 77～ 84
[ 19 ] 　E sbroeck G V , Bowm an D T. Cotton germp lasm diversity and
its importance to cultivar developm ent. J Cotton S ci. 1998,
2: 121～ 129
[ 20 ] 　De V icente M C, Tanksley S D. Q TL analysis of
transgressive segregation in an interspecific tom ato cross.
Genetics, 1993, 134: 585～ 596
202　　　　　　　　　　　　　　　　　　　　　作　　物 　　学　　报　　　　　　　 　　　　　　　　　　　　28卷