Heredity 59 (1987) 293-299 The Genetical Society of Great Britain Received 5December1986

Thein vitro genetics of barley (Hordeum vulgare L.): detection and analysis of reciprocal differences for culture response to 2,4-dichlorophenoxyacetic acid

W. Powell and Scottish Crop Research Institute, Pentlandfield, P. D. S. Caligari* Roslin, Midlothian EH25 9RF, Scotland.

The procedures of biometrical genetics have been used to detect and analyse reciprocal differences for the response of immature barley embryos to 2,4-dichlorophenoxyacetic acid in culture. Two genotypes were examined (4082 and R23) which had exhibited large reciprocal differences in culture in previous experiments. The F1, F2 and backcross generations, including full reciprocal families, were produced from these, true-breeding genotypes and evaluated in culture. The experiments confirmed the presence of reciprocal differences. Such differences were subsequently shown to stem from cytoplasmic effects alongside effects traceable to the mother tissue. The implications of these findings for the manipulation of barley genotypes in vitro are discussed.

INTRODUCTION A number of factors may complicate the analyses of genotypic differences and in this paper Genotypicdifferences exist in the response of attention is focused on a relatively neglected area cereals to in vitro culture and it has become of tissue culture response i.e., the possibility of apparent that a thorough understanding of the reciprocal differences. A diallel analysis of in vitro genetical control of growth and differentiation in response in barley indicated the presence of large such systems is necessary for the efficient use of reciprocal differences when immature embryos cell culture technologies in crop improvement. In were cultured on media containing 2, or more, vitro selection (Bright, 1985), exploitation of soma- mg I_I of 2,4-dichlorophenoxyacetic acid clonal variation (Larkin, 1985) and the possibility (Caligari, Powell and Goodall, 1987). The presence of interspecific introgression (Armstrong, of reciprocal differences for culture response has Nakamura and Keller, 1983) rely on the accessibil- also been reported in (Lazar, Collins and ity of commercially desirable cultivars for culture Vian, 1983) and maize (Nesticky et a!., 1983; establishment and plant . A pre- Beckert and Qing, 1984; Tabata and Motoyoshi, liminary evaluation of culture response (Powell 1965). In a more recent study (Tomes and Smith, and Dunwell, 1987) may allow the identification 1985) on the initiation of embryogenic callus from of responsive genotypes and the transfer of desir- maize genotypes, the reciprocal effects observed able controlling culture response may be were attributed to segregating nuclear factors and achieved by conventional methods. Knowledge of explant source rather than cytoplasmic genetic fac- the genetical systems underlying culture response tors. A distinction between maternal and cytoplas- will complement this approach and may even be mic factors is important since the participation of used to predict the response to selection (Mather the cytoplasm would imply that the organellar, and Jinks, 1982) and the probability of obtaining cytoplasmic DNA is involved in the control of desirable genotypes from a given cross (Jinks and culture response. The objective of the present study Pooni, 1976; Caligari, Powell and Jinks, 1985; was to investigate the underlying control of Powell et a!., 1985). differences in culture response to the synthetic *Presentaddress: Department of Agricultural Botany, Plant auxin 2,4-dichiorophenoxyacetic acid (2,4-D). Sciences Laboratories, University of Reading, Whiteknights, This was achieved using biometrical genetical tech- Reading RG6 2AS, U.K. niques which allowed not only the nuclear genetic 294 W. POWELL AND P. D. S. CALIGARI basis to be quantified and analysed but also the family was represented by 5embryosper petri-dish cytoplasmic component. therefore giving a maximum of 30 embryos per family. Examination of the raw data revealed large MATERIALS AND METHODS positive correlations between the estimates of the means and variances. It was therefore considered Experimental material necessary to transform the data. In practice it was found that taking the fourth root (x"4) removed The parental material used in the present study this scalar effect and all subsequent analyses were consisted of the genetically multiply-marked performed on the data transformed in this way spring barley stocks R23 and 4082. R23 possesses a range of morphological markers distributed (see Caligari etal,, 1987). among the seven pairs of homologous chromo- somes with the alleles being in the dominant condi- PRELIMINARYANALYSES tion. 4082 is recessive at the corresponding loci. The parental, F1, F2 and backcross generations Analysesof variance were performed on the trans- including full reciprocal families were produced formed data to test for differences between the 16 in a glasshouse with a minimum 16-hour photo- generation means for the characters: fresh weight period at a temperature of 20°—30°C. Florets were (F.Wt.), dry weight (D.Wt.) and water content emasculated by clipping the lemma and the palea (Wat), and these are presented in table 1. It should and the anthers were removed using fine forceps be noted, however, that the three characters recor- (Pope, 1944). The sixteen families examined in the ded here are not completely independent of one study are given in table 2. In order to avoid prob- another since D.Wt. and Wat make up F.Wt. lems associated with the effects of clipping, all the However, hearing this is mind, the analyses indi- families sampled were clipped at anthesis. cated that there were highly significant differences between families for the three characters scored. Cultureof immature zygotic embryos Furthermore, the interactions between replicates Caryopsescontaining embryos 1-2 mm in length and families were also highly significant (P< were removed from spikes representing each 0.001). The means and standard deviations for the family. In the case of the six-row parent (4082) three characters are given in table 2. These statistics and segregants from crosses with it, only primary may be used to compare the family means as well florets were excised. Caryopses were sterilised for as to detect and test for the presence of reciprocal 15 minutes in 2 per cent sodium hypochlorite with differences. Student's "t-test" was applied and the Tween 20 as a wetting agent. They were then values for "z" from a number of comparisons of washed in sterile distilled water. Embryos were interest are given in table 3. There were significant excised and placed with the scutellum uppermost differences between the parents for the characters on 20 ml of solidified medium (agar 07 per cent) F.Wt. and Wat. The comparison between families at a frequency of five per petri-dish (90 x 15 mm). 3 and 4 was also highly significant for these two The medium employed was based on that of characters indicating that there were large Murashige and Skoog (1962) with the major salts reciprocal differences between the responses of the at half the recommended concentration, 3 per cent F1 hybrids in culture depending on the direction sucrose and 4 mg l_l of 2,4-D. Dishes were sealed in which the cross was made. This reciprocal with paraflim and incubated at 20°C2°Cwith a difference is presented visually in fig. 1; where 16-hour photoperiod produced by white fluores- immature embryo derived callus from the parental cent tubes (30 pEm2s'). genotypes (R23 and 4082) together with the After 28 days in culture, each embryo and reciprocal F1 hybrids are shown. It can be seen attached callus was harvested. After blotting dry that the expression of the F1 derived embryos in with paper tissues the fresh weight (F.Wt.) was culture is clearly dependent on which direction the recorded in mg. Following drying overnight at cross is made. Thus when 4082 is used as the female 35°C,thedry weight (D.Wt.) was recorded and parent the resulting F1 embryos are much subsequently the water content (Wat) calculated. more responsive in culture than those derived from the reciprocal. Apart from the comparison of family 3 with family 1 in table 3, the rest of the Experimentaldesign and data transformation comparisons measure reciprocal effects in the F2's Theexperimental design was randomised complete and backcrosses, and most proved to be significant. blocks with six replicates (petri-dishes). Each The fact that not all of them are significant mdi- IN VITRO GENETICS OF BARLEY: RECIPROCAL DIFFERENCES 295

Table 1Analyses of variance for the 16 families raised from the inbred lines R23 and 4082

F.Wt. D.Wt. Wat Item df MS (102) MS (102) MS (102)

Reps 5 12.08* 1691*** J3.47*** Families 15 5065*** 3012*** 33.49*** Reps x Families 72 (3)t 801*** 607*** 6.36*** Error 350 (34)f 311 249 237 t The number of missing observations. *PS-005—001; ***p <0001. cates that the cause(s) of such effects may not be tions of cytoplasmic determinants from the male simple and warrant further examination. and female gamete to the zygote (Jinks, 1964). This The comparison between families 3 and I is, type of inheritance is through the female parent in fact, the comparison of the higher scoring F1 which contributes by far the major part of the with the better parent and thus tests for heterosis, zygotic cytoplasm. Maternal or mother tissue although it should be noted that it is not formally effects are transient in nature in that they, by independent of the comparisons between I and 2 definition, change with the genotype of the mother. as well as 3 and 4. There are no significant levels A distinction between these two types of inherit- of heterosis observed, although the comparison for ance is important since a cytoplasmic component Wat approaches significance. There is, therefore would suggest that organellar DNA residing in the no evidence, in this study, for superiority of the chloroplasts or mitochondria is influencing in vitro F1 hybrid over the better scoring parent, for the development. It is also important since the success- characters examined. ful manipulation of the character would depend on understanding the underlying modes of GENETICALANALYSES transmission. In order to investigate the genetical reasons for the reciprocal differences, a more inten- Thepresence of reciprocal differences has been sive investigation of the two inbred lines was detected in a number of tissue culture studies and necessary. By producing generations in addition these differences may be due to cytoplasmic to the F1's, genetical models can be fitted which inheritance and/or maternal effects. Cytoplasmic allow the cytoplasmic and maternal effects to be inheritance would result in persistent reciprocal estimated and their adequacy in describing the differences and is the result of unequal contribu- observed differences to be tested.

Table 2The means () and standard deviations (S.D.) for each of the 16 families for three chharacters

F.Wt. D.Wt. Wat Family S.D. t S.D. x S.D.

1 P, (4082) 0569 0028 0353 0033 0531 0018 2p2 (R23) 0385 0018 0230 0009 0362 0014 3F, (P, x P2) 0657 0061 0379 0045 0621 0048 4F, (P2xP1) 0458 0016 0277 0013 0438 0012 5F2 [(P1 x P2) x (P1 x P2)] 0508 0039 0311 0019 0484 0034 6F2 [(P2xP1)x(P2xP)] 0597 0026 0351 0023 0571 0021 7F2 [(P1 x P2) x (P2 x P,)] 0.439 0050 0295 0014 0489 0039 8F2 [(P2 x P1) x (P1 x P2)} 0589 0059 0412 0055 0529 0041 9B1 [P1 x (P1x P2)] 0707 0059 0455 0048 0646 0049 10B, [(P1 x P2) x P,] 0516 0035 0288 0-015 0500 0032 11B, [P1 x(P2 x P1)] 0-727 0032 0444 0022 0687 0032 12B, [(P2 x P1) x P1] 0836 0023 0610 0051 0713 0020 13B2 [P2x(P1xP2fl 0457 0059 0289 0037 0427 0045 14B2 [(P1 x P2) x P2] 0559 0053 0366 0045 0-509 0039 15B2 [P2x(P2xP,)] 0-399 0-020 0237 0.007 0383 0019 16B2 [(P2 x P1) x P2] 0662 0065 0446 0068 0586 0043 296 W. POWELL AND P. D. S. CALIGARI

Figure 1 Response of immature barleyembryos to medium containing 4mg Iof 2,4-D. The two parents (P1= 4082and P2 =R23) areshown together with the reciprocal F1's.

The expectations for the 16 families produced [cd] and[C5]respectively.Parameters which rep- from the F1, F2 and backcrosses, including all pos- resent the maternal components are also tabulated sible reciprocals, are given in table 4. The nuclear for two aspects: [d,], to represent the maternal genetic contributions to the progeny means may effects traceable to the additive genetic component be specified by the parameters: m, [di, [h], [i], [j] and [h,] that reflecting the dominance effects of and[/1,whichprovide estimates of the main effects the maternal genotype. and first-order interactions of these families The model can be fitted to the observed data (Mather and Jinks, 1982). The contribution of the using weighted least squares with the inverse of cytoplasm is represented by [ci while the interac- the appropriate variance providing the weight for tions between it and the additive, [di, and domi- each family mean. The significance of each of the nance, [h], nuclear determined effects are given as parameters estimated and the adequacy of the model can then be tested in the standard way (Mather and Jinks, 1982). The results of such Table 3 Student's "1-test" between the means of some of the model fitting are given in table 5 for the three 16 families characters scored. Models were considered to he Family comparison F.Wt. OWl. Wat satisfactory when the chi-square testing the good- ness of fit was non-significant, and all the estimates I vs. 2 4.60*** 3.27*** 5.02*** of the parameters were themselves significantly 3vs. 4 3.855*5 2.26* 4.02*** different from zero. As can be seen from the table, 3vs. 1 159 050 191 the three characters give significant estimates for 5vs. 6 177 1'OO 188 7vs. 8 240 236 075 all the nuclear genetic parameters, apart from the 9vs. 10 3.27*5 3.4g*** 269 additive genetic one. The lack of an additive 11 vs. 12 2.52* 3.37*5* 063 genetic component, [d], while the additive by addi- 13vs. 14 141 127 132 tive, [ii,andthe additive by dominance, [11,inter- 15 vs. 16 4.82*** 4.06*5* 4,345*5 actions are significant, would indicate that there * P = 005—001; P = 001 —0001; P <0001. is dispersion of the alleles for increasing and IN VITRO GENETICS OF BARLEY: RECIPROCAL DIFFERENCES 297

Table 4The expectations for the family means in terms of the genetic parameters for nuclear and cytoplasmic components as well as maternal effects

Nuclear Cytoplasmic Maternal , Family m [d] [h] [i] [/] [1] [c] [ca] [Chl [d,,] [h,] — 1P1—4082 1 1 — 1 — — —1 ---1 — 1 2 P2—R23 1 —1 — 1 — — 1 —1 — —1 3 F1 (P1xP2) 1 — 1 — — 1 —1 — —1 1 — 4 F (P2xP1) 1 — 1 — — I I — 1 —1 — 5 F2[(P1xP2)x(P1xP2)] 1 — 0-5 — — 0-25 —1 — —0-5 — 1 6 F2 [(P2xP1)x(P2xP1)] 1 — 0-5 — — 0-25 1 — 0-5 — 1 7 F2 [(P1xP2)x(P2xP1)] I — 0-5 — — 0-25 —1 — —05 — I 8 F2[(P2xP1)x(P1xP2)] 1 — 0•5 — — 0-25 1 — 05 — 1 9 B1 [P1 x(P1 x P2)] 1 0-5 0-5 025 025 025 —1 —05 —0-5 1 10 B1 [(P1 x P2) x P1] 1 0-5 0-5 0-25 0-25 0-25 —1 —0-5 —05 — 11 B1 [P1 x(P2x P1)] 1 0-5 0-5 0-25 0-25 025 —1 —0-5 —0-5 1 — 12 B1 [(P2x P1) x Pj 1 0-5 0-5 0-25 0-25 0-25 1 0-5 0-5 — 13 B2 [P2 x (P1 x P2)] 1 —0-5 0-5 0-25 —025 0-25 1 —0-5 0-5 —1 14 B2 [(P1 x P2) x P2] 1 —0-5 0-5 0-25 —0-25 0-25 —1 0-5 —0-5 — 1 15 B2 [P2x(P2x Pt)] 1 —0-5 0-5 0-25 —0-25 0-25 1 —0•5 0-5 —1 — 16 B2 [(P2x P1)x P2] 1 —0-5 0-5 0-25 —0-25 0-25 1 —0-5 0-5 — decreasing effect between the two parents (Jay- selection directly affecting in vitro response, it is asekara and Jinks, 1976). conceivable that the differences in these characters The magnitudes and signs of [hi and [1] are may be reflections of other morphological char- not influenced by the way the alleles are distributed acters expressed in normal growth and develop- between the parents of the cross and hence these ment, and hence be subject to natural or artificial two components may be used to classify the pre- selection. In any case, the past history of selection dominant type of epistasis (non-allelic interac- and its effect could only be judged satisfactorily tion), as described by Mather and Jinks (1982). In with a larger, and hence more representative, all three cases the [h] increments of the majority sample of genotypes. It should also be noted that of loci are positive while the [1] increments for the two genotypes used were chosen as represent- pairs of loci are negative. Epistasis is therefore ing extreme responses, yet they appear to show mainly of the duplicate kind. This form of epistatis gene dispersion which at first sight appears para- is generally associated with characters which have doxical but, in fact, reflects the complex inherit- been exposed to directional selection whereas the ance of these traits involving, as they do, reciprocal dispersion of genes leading to a non-significant differences. Thus the parents have been chosen, as estimate of [d] would suggest stabilizing selection. it turns out, more on the basis of their non-nuclear Although it is difficult to envisage a history of past effects.

Table 5Significant parameters in the model necessary to describe the differences between the families for the three characters

F.Wt. D.Wt. Wat

m 0-2360-095 0-1570-073 0-3070-085 [d] — — — [hI 0-896±0-227 0-564±0- 176 0-6230-0204 [i] 0-378±0-096 0-282±0-083 0-2210-085 Li] 0-192±0-060 0-1230-051 0-162±0-054 [1] — 0-572 0-141 — 0-386 0-115 — 0-400 0-128 [c] 0-0780-013 0-0590-011 0-0530-012 [cs] 0-1350-029 0-146±0-030 0-080±0-026 [ch] — — — [d,,,] 0-179±0-020 0-124± 0-016 0-145±0-017 [h,,j — — — Xsj 6-8ns 11-On.s 5-On.s 298 W. POWELL AND P. D. S. CALIGARI There is evidence in table 5 for the presence of reported for callus growth in alfalfa (Keyes and a cytoplasmic effect [ci as well as its interaction, Bingham, 1979) and in Nicotiana (Keyes et a!., [Cd],withthe additive genetic component. There 1981). it has been recommended that hybrid vigour is also a significant maternal, [drn], component exhibited in cell and tissue culture may be of value detected. It is interesting to note from the signs of in the identification of heterokaryons following the estimates of [c]and[dr,] and the signs somatic fusion. Smith, Rao and Combatti (1976) attributed to them in the model (table 4) that they have demonstrated that partial selection for are, in fact, opposing one another in the two somatic fusion products of N. glauca and N. parents, with 4082 showing the increasing effect of Iangsdorfii may be based on the more the maternal component but a decreasing one for rapid growth of hybrid colonies. However, the the cytoplasm while R23 shows, of course, the main emphasis must be on the need for a better reverse. It would thus seem that as far as nuclear understanding of the genetical basis for the genes as well as reciprocal differences (of the two expression of the character. It will then be possible types) the parents used are not extremes, although to predict the level of heterosis likely to occur and 4082 is very responsive compared with other lines to be clear about its underlying causation. This tested (Caligari et a!., 1987). will then allow an assessment of its potential. At present, advances in the application of the potential myriad of tissue culture techniques to DISCUSSION commercially important crop species are ham- Theprocedures of biometrical genetics have been pered by the inability to handle a sufficiently wide used to elucidate the inheritance in response of range of genetic material. A better understanding immature barley embryos to in vitro culture. Such of the underlying genetical determination control- procedures showed that nuclear genes were in- ling in vitro response is therefore critical if the volved and allowed the detection and separation techniques are going to be developed as useful of cytoplasmic and maternal effects. Indeed, they additions to those traditionally available for crop allowed all the components involved to be quan- improvement. The multiply-marked stocks used in tified within a single model, its adequacy to be this study (R23 and 4082) differed in their in vitro tested statistically and the significance of the para- response to 2,4-D. Both nuclear and cytoplasmic meters to be assessed. Thus the methods outlined factors were clearly implicated in the control of by Jinks, Perkins and Gregory (1972) for the analy- culture response. It would therefore appear that sis of reciprocal differences have allowed the the inheritance pattern, although complex, would response to be described although it is clearly still be capable of manipulation in conventional complex in showing an array of underlying deter- crossing programmes. minants. Mathias, Fukui and Law (1986) using cytoplas- Acknowledgements We are grateful to V. Goodall for technical mic substitution lines demonstrated the import- help and to Dr W.T.B. Thomas for the use of some computer ance of cytoplasmic and nuclear-cytoplasmic inter- programmes. actions on tissue culture response in wheat. Calli were found to differ in their growth rates, their REFERENCES sensitivity to 2,4-D and their ability to organize shoot primordia. These authors do not, however, ARMSTRONG, K. C..NAKAMURA, C. AND KELLER, W. A. 1983. rule out the possibility that the cytoplasm may be instability in tissue culture regenerants of (Trilicosecale Wittmack) cv. "Welsh" from 6 influencing the original explant and not operating months old callus culture. 7 Pflanzenzuch, 91, 233—245. directly on culture response. In the present study UE(KFRT. M. AND QING,C. M. 1984.Resultsof a diallel trial on barley it is shown that both the cytoplasm and and a breeding experiment for in vitro aptitude in maize. the maternal or mother tissue significantly effects Theor. AppI. Genet., 6X, 247—251. culture response. The hexaploid nature of culti- RRIGIIT.S W .1. 1985.Selectionin vitro. In Bright, S. W. J. and vated wheat has also allowed Jones, M. J. K. (eds.) Cereal Tissue and Cell Culture, Martinus Nijhoff/Dr. W. Junk, pp. 232-260. engineering techniques (Mathias and Fukui, 1986) (.'ALIGARI,P.D.S.,POWELL,W. ANDJINKS,J. L.1985. Theuse to identify the role of specific on ofdoubled haploids in barely breeding. 2. An assessment tissue culture response in wheat. The substitution of univariate cross prediction methods. Heredity, 54, 353— of the "Capelle" 4B chromosome resulted in an 358. increase in cell proliferation and shoot regeneration. CALIGARI, P. 0. S., POWELL, W. ANI) 0000ALL, v., 1987. The in vitro genetics of barley (Hordeum vulgare L.): Genetical Although the present study failed to detect analysisof immature embryo response to 2,4-dichloro- significant positive heterosis it has previously been phenoxyacetic acid. Heredity. 59, 285-292. IN VITRO GENETICS OF BARLEY: RECIPROCAL DIFFERENCES 299

JAYASEKARA, N. F. M. AND JINKS, J. L. 1976. Effects of gene MATHIAS, R. J., FUKUI, K. AND LAW, C. N. 1986. Cytoplasmic dispersion on estimates of components of generation means effects on the tissue culture response of wheat (Triticum and variances. Heredity, 36, 3 1—40. aestivum) callus. Theor. App!. Genet, 72, 70—75. JINKS, J. L. 1964. Extrachromosomal Inheritance, Prentice-Hall, MURASHIGE, T. AND SKOOG, F. 1962. A revised medium for London. rapid growth and bioassays with tobacco tissue cultures, JINKS, J. L., PERKINS, J. M. AND GREGORY, S. R. 1972. The Physiol. Plant., 15, 473-497. analysis and interpretation of differences between NESTICKY, M., NOVAK, F. J., PIOVARCI, A. AND DOLEZELOVA, reciprocal crosses of Nicotiana rustica varieties. Heredity, KS. 1983. Genetic analysis of callus growth in maize (Zea 28, 363-377. mays L.) in vitro. Z. Pflantzenzucht., 91, 322-328. JINKS, J. L. AND POONI, H. s. 1976. Predicting the properties POPE, M. N. 1944. Some notes on techniques on barley breeding. of recombinant inbred lines derived by single seed descent. J. Hered., 35, 99-111. Heredity, 36, 253—266. POWELL, W., CALIGARI, P. D. S., McNJCOL, J. W. AND JINKS, KEYES, G. J. AND BINGHAM, F. T. 1979. Heterosis and J. L. 1985. The use of doubled haploids in barley. 3. An effects on the growth of alfalfa callus. Crop Sci., 19,473—476. assessment of multivariate cross prediction methods. KEYES, 0. J., DEATON, W. R., COLLINS, 0. B. AND LEGG, P. D. Heredity, 55, 249-254. 1981. Hybrid vigour in callus tissue cultures and seedlings POWELL, W. AND DUNWELL, J. M. 1987. In vitro genetics of of Nicotiana tabacum L., J. Hered., 72, 172-174. barley (Hordeum vulgare L.) 1. Response of immature LARKIN, p,j. 1985In vitro culture and cereal breeding. In embryos to 2,4-dichlorophenozyacetic acid. Heredity, (in Bright, S. W. J. and Jones, J. J. K. (eds) Cereal Tissue and press). Cell Culture,MartinusNijhoff/Dr. W. Junk, pp. 274-296. SMITH, H. I-I., RAO, K. N. AND COMBATrI, M. C. 1976. Inter- LAZAR. M. 0., COLLINS G. B. AND VIAN, W. F. 1983. Genetic specific hybridisation by fusion in Nicotiana: and environmental effects on the growth and differentiation confirmation and extension. J. Hered., 67, 123—128. of wheat cultures. J. Hered., 74, 353—357. TABATA, M. AND MOTOYOSHI, F. 1965. Hereditary control of MATHER,K. ANDJINKS. J. L. 1982. Biometrical Genetics, 3rd callus formation in maize endosperm cultured in Vitro. edition. Chapman and Hall, London. JapanJ. Genet., 40, 343-355. MATHIAS,R.J. AND FUKUI, K. 1986. The effect of specific TOMES, D. T. AND SMITH, 0. s. 1985. The effect of parental chromosome and cytoplasm substitutions on the tissue genotype on initiation of embryogenic callus from elite culture response in wheat (Triticumcestivum) callus. Theor. maize (Zea mays L.) germplasm. Theor. App!. Genet., 70, App!. Gener., 71, 797—800. 505-509.