isoenzymatic variation in zea (gramineae) · isoenzymatic variation in zea (gramineae) john f....

Systematic Botany (1984), 9(2): pp. 203-218 C Copyright 1984 by the American Society of Plant Taxonomists

Isoenzymatic Variation in Zea (Gramineae)

JOHN F. DOEBLEY and MAJOR M. GOODMAN

Department of Statistics

CHARLES W. STUBER

Department of Genetics and USDA-ARS, North Carolina State University, Raleigh, North Carolina 27695

ABSTRACT. An average of 14(7-34) plants each for 61 different collections of Zea (maize and its wild relatives, the teosintes) were studied for 12 enzyme systems coded for by 21 loci. Principal component and cluster analyses based on allele frequencies showed Zea can be divided into two major groups: 1) sect. Luxuriantes, including Z. perennis, Z. diploperennis, and Z. luxurians; and 2) sect. Zea (in part), including Z. mays subsp. mays, var. parviglumis, and subsp. mexicana. Zea mays var. huehuetenangensis (Huehuetenango teosinte) is isoenzymatically distinct from both sections, but shows its closest relationship to Z. mays var. parviglumis of sect. Zea. Populations of Z. mays subsp. mexicana and var. parviglumis grade isoenzymatically from one into the other without any clear break, but without any overlap either. Five populations of Z. mays subsp. mays are all isoenzymatically very similar to Z. mays var. parviglumis. The isozyme data are consistent with the theory that Mexican annual teosinte is the ancestor of maize. With the exception of the unusual isoenzymatic nature of Z. mays var. huehuetenangensis, the isoenzymatic data agree with previous studies of tassel morphology and cytoplasm DNAs. A comparison of sympatric populations of maize and teosinte suggests that teosintes are not greatly affected by introgression from maize. Zea mays var. parviglumis and Z. diploperennis have considerable within and among population variation; Z. luxurians has much less. Race Central Plateau of Z. mays subsp. mexicana has greater among population variation than any other taxon in Zea. Zea tends to show greater within population heterozygosity and more variance within species and subspecies than most other plants that have been studied isoenzymatically.

In the research reported below, enzyme electrophoresis was employed to assess phylogenetic relationships, measure the degree of genetic variation, and determine the extent of gene exchange among populations of Zea. This study builds upon three previous studies of isozyme variation in Zea (Senadhira 1976; Masten- broek et al. 1981; Smith et al. 1984), but differs from previous work by including a substan- tially larger sample of the perennial members of the genus and a larger number of loci. Also included are some recently discovered annual populations from Jalisco and Durango, Mexico, that were not previously surveyed.

One purpose of this paper is to examine taxonomic relationships among the populations sampled. These populations occur over an ex- tensive geographic region from Honduras to Chihuahua, Mexico. They are adapted to a broad range of environments, ranging from the seasonally arid savanna of the Mexican Central Plateau to the mesic montane forest of western Guatemala. They include both annuals and pe-

rennials, and, except for a few of the perennial populations, all are diploids. These populations have been treated taxonomically in two different fashions (table 1) by Wilkes (1967) and by Iltis and Doebley (1980). Wilkes recognized two sections: sect. Zea for the cultigen maize (Z. mays) and sect. Euchlaena for the wild taxa, com- monly called teosintes. He recognized two wild species: 1) Z. mexicana, a widespread, diploid (n = 10) annual, and 2) Z. perennis, a tetraploid perennial found only in Jalisco, Mexico. The emphasis in Wilkes's classification, however, was on his division of Z. mexicana into six allopatric races. In agreement with previous authors (Collins 1931; Longley 1941a, b), Wilkes noted that his race Guatemala of southeastern Guatemala is distinct from all other teosintes. He considered race Huehuetenango of western Guatemala to be related to the Mexican annual teosintes. Wilkes also stated that races Chalco, Central Plateau, and Nobogame of central and northern Mexico are maize-introgressed forms of race Balsas of southwestern Mexico. Finally,

203

204 SYSTEMATIC BOTANY [Volume 9

TABLE 1. Two classification systems for Zea.

Iltis and Doebley (1980) and Wilkes (1967) Doebley and Iltis (1980)

Sect. Euchlaena Sect. Luxuriantes Z. diploperennis Iltis, Doebley, & Guzman

Z. perennis (Hitchc.) Reeves & Mangelsdorf Z. perennis Z. mexicana (Schrader) Kuntze

race Guatemala Z. luxurians (Durieu & Ascherson) Bird Sect. Zea

Z. mays subsp. parviglumis Iltis & Doebley race Huehuetenango var. huehuetenangensis Iltis & Doebley race Balsas var. parviglumis Iltis & Doebley

subsp. mexicana (Schrader) Iltis race Chalco race Chalco race Central Plateau race Central Plateau race Nobogame race Nobogame

Sect. Zea Z. mays L. Z. mays subsp. mays

Wilkes noted that there are two forms of race Central Plateau-a robust, weedy type that in- vades maize fields, and a small, wild type that occurs in disturbed habits but does not invade maize fields.

Iltis and Doebley (1980) recognized the following taxa: 1) sect. Luxuriantes, which con- tains three distinct species, Z. perennis, Z. luxurians (=Race Guatemala of Wilkes), and Z. diploperennis, a newly discovered diploid perennial from Jalisco, Mexico, unknown at the time of Wilkes's work; and 2) sect. Zea, with one species, Z. mays, which includes both the cultigen maize and all of Wilkes's races of Z. mexicana except Guatemala. Iltis and Doebley divided Z. mays into three subspecies: 1) subsp. mays (the cultigen, maize); 2) subsp. mexicana, a large-spikeleted form adapted to the arid high elevations (1800-2500 m) of northern and central Mexico, including three of Wilkes's races (Nobogame, Chalco, and Central Plateau); and 3) subsp. parviglumis, a small-spikeleted form adapted to the moister middle elevations (400-1700 m) of southwest Mexico (var. parviglumis) and western Guatemala (var. huehuetenangensis).

Some differences and similarities between the Wilkes and the Iltis and Doebley systems need to be emphasized. Doebley and Iltis (1980) did not believe that maize is genetically or morphologically distinct enough to merit its own section within the genus. Rather, they treated

maize and the annual teosintes of Mexico and western Guatemala as members of not only the same section, but the same species. This was done explicitly to recognize that they regard maize as a cultivated form of Mexican annual teosinte. Second, while Wilkes noted an affinity of race Huehuetenango of western Guate- mala to the Mexican annuals, Iltis and Doebley (1980) were more specific, combining races Huehuetenango and Balsas into Z. mays subsp. parviglumis. Third, like Wilkes (1967), Iltis and Doebley (1980) recognized race Guatemala to be distinct from all other teosintes and, accord- ingly, they treated it as a separate species. However, they noted a relationship between this annual and the perennial teosintes, all of which they combined in sect. Luxuriantes.

Another objective of this paper is to examine briefly the evidence for maize introgression into teosinte. Some authors have argued that such introgression is a major force shaping the morphology and genetics of the various teosintes (Collins 1921; Mangelsdorf 1947; Wilkes 1977; Bird 1978). Other authors have contended that introgression is only a minor, unimportant fac- tor, contributing little to variation among the teosintes (Kato 1976; Doebley 1984b). In this report, five pairs of sympatric maize-teosinte populations are compared isoenzymatically in an effort to provide further evidence concern- ing introgression in Zea.

Finally, the degree of genetic variation ob-

1984] DOEBLEY ET AL.: ZEA 205

served among the teosintes is compared to other plant species for which isozyme data have been published.

MATERIALS AND METHODS

In 1978, Rafael Guzman reported that he had re-discovered Zea perennis, which was thought to be extinct in the wild, at two locations in Jalisco, Mexico, as well as discovered several new populations of annual teosinte in Jalisco (Guzman 1978a, b, 1982). Iltis et al. (1979) de- termined that plants from one of Guzman's localities represented a new diploid perennial (Z. diploperennis). Subsequently, Guzman, Iltis, and their associates discovered three additional sites for Z. diploperennis, all near the previously known locality for this species (Iltis 1980). These collections made by Iltis et al. form the core of materials on which the work reported here was based. These collections were supplemented by other collections of the various annual, diploid teosintes, including many made by T. A. Kato. A population of Z. mays subsp. mexicana from Durango, Mexico, which was recently re-discovered by Doebley (1983b), is also included. Finally, five open-pollinated maize varieties collected from fields in Mexico and Guatemala, where they grew and hybridized with teosinte, were included for a total of 61 collections (table 2).

Most of these 61 collections consist of seed collected in bulk in the wild; however, some collections have been increased under cultivation (table 2). For the most part, we do not know how many plants contributed seed to the bulk collections. In this sense then, the accuracy with which each collection represents its population is unknown.

Seeds were germinated at 22?C in the dark for six days, at which point a 2.5 cm segment of coleoptile was removed, homogenized, and stored as previously described (Cardy et al. 1980; Stuber and Goodman 1983). After removal of the coleoptile sample, plants of 35 of the 61 collections were transferred to greenhouses or field (Florida winter nursery) where they were grown to maturity in order to determine whether they were pure teosinte plants or maize-teosinte hybrids. This is important because all teosintes (including Z. perennis, pers. obs., Doebley) hybridize to some degree with maize in the wild, and it is generally preferable

TABLE 2. List of collections (1-61) of Zea analyzed including taxon name (abbreviation), location data, and number (N) of individuals assayed. # = collections consisting of seed increased under cultivation; * = collections grown in greenhouses or Florida winter garden to check for maize-teosinte hybrids. If hybrids were found, they were not used to calculate gene frequencies.

Z. diploperennis (DIPL). MEXICO. Jalisco. Sierra de Manantlan: 1*, La Ventana, 19'32'N, 104'13'W, 2300 m, 15 Dec 1977, Guzman 777 (N = 10); 2, 2 km E of Las Joyas, 19'35'N, 104'16'W, 1900 m, Nov 1980, Guz- man 1120 (N = 26); 3, location same as collection 2, 6 Jan 1979, Iltis et al. 1250 (N = 12); 4, location same as collection 2, 3 Jan 1980, Iltis et al. 2265A (N = 12); 5*, 0.3 km S of Manantlan, 19'36'N, 104'13'W, 1450 m, 4 Jan 1979, Iltis et al. 1190 (N = 11); 6*, Las Joyas, 19-36'N, 104017'W, 1900 m, 6 Jan 1979, Iltis et al. 1375 (N = 12); 7*, 0.6 km WNW of Rincon de Manantlan, 19036'N, 104013'W, 1520 m, 4 Jan 1979, Iltis et al. 1155 (N = 7).

Z. perennis (PERN). MEXICO. Jalisco. 8#, 1 mi S of Ciudad Guzman railway station, 19042'N, 103029'W, 1520 m, 21 Oct 1921, Collins s.n. (N = 12); 9, 0.2 km W of Piedra Ancha, 19038'N, 103035'W, 2100-2200 m, Oct 1979, Guzman s.n. (N = 12); 10, 1.5 km ESE of los Depositos, 19039'N, 103032'W, 1650 m, Oct 1978, Guz- man s.n. (N = 12); 11, location same as collection 9, 31 Dec 1978, Iltis et al. 1050 (N = 12).

Z. luxurians. Race Guatemala. (LUXS). HONDURAS. Choluteca. 12*#, San Antonio de Padua (see Standley 1950), Galinat 76-2076-B (N = 18). GUATEMALA. Ju- tiapa. 13*, 1.2 km N of El Progreso, 14021'N, 89051'W, 1050 m, Dec 1975, Iltis G-5 (N = 34); 14, 2 km E of El Tablon, 14017'N, 89054'W, 950-1100 m, 31 Dec 1975, Iltis G-27 (N = 12); 15*, 1.6 km E of El Progreso turn- off on the Pan American Hwy., 14021'N, 80050'W, 920 m, 1 Jan 1976, Iltis G-36 (N = 21); 16*, 5 km W of Agua Blanca, 14029'N, 89042'W, 920 m, 1 Jan 1976, Iltis G-38 (N = 23). Chiquimula. 17*, 2 km N of Ipala on road to Chiquimula, 14038'N, 89038'W, 800 m, 2 Jan 1976, Iltis G-42 (N = 21).

Z. mays var. huehuetenangensis. Race Huehuetenan- go (HUET). GUATEMALA. Huehuetenango. 18, 1 km S of San Antonio Huista, 15040'N, 91045'W, 1250 m, 8 Jan 1976, Iltis and Lind G-119 (N = 14); 19, 1.5-2.5 km ENE of San Antonio Huista, 15039'N, 91046'W, 1300-1400 m, 9 Jan 1976, Iltis and Lind G-120 (N = 12).

Z. mays var. parviglumis. Race Balsas (PV-B). MEXICO. Guerrero. 20, 5 km E of Mazatlan, 17030'N, 99030'W, 1100 m, 1972, Beadle s.n. (N = 12); 21, 5 km S of Palo Blanca, 17025'N, 99030'W, 1350 m, 1972, Bea- dle s.n. (N = 12); 22*, El Rincon, 17030'N, 99030'W, 1150 m, Kato 69-14 (N = 12); 23*, km 53 on Iguala- Teloloapan Hwy., 18020'N, 99048'W, 1350 m, Kato 77- 13 (N = 12); 24*, km 75 on Teloloapan-Arcelia Hwy., 18017'N, 100010'W, 1600 m, Wilkes 71-2 (N = 12); 25*,


TABLE 2. Continued.

km 103 on Teloloapan-Arcelia Hwy., 18?23'N, 100'03'W, 1000 m, 24 Nov 1971, Iltis and Cochrane 78 (N = 12). Michoacan. 26*, km 24 on Huetamo-Mo- relia Hwy., 18'25'N, 100?50'W, 800 m, Kato 67-13 (N = 12); 27*, km 127 on Huetamo-Morelia Hwy., 19025'N, 100045'W, 1000 m, Kato 67-15 (N = 12); 28, 1 km S of Tzitzio on road to Huetamo, 19034'N, 100055'W, 1500 m, 6 Dec 1971, Iltis and Cochrane 308 (N = 14); 29*#, km 3 on Tingambato-Tuzantla Hwy., 19018'N, 100018'W, 950 m, Kato 67-22 (N = 12). Mexico. 30*#, km 116 on Toluca-Tejupilco-Lu- vianos Hwy., 19005'N, 100015'W, 1390 m, Kato 67-17 (N = 12); 31*#, km 175 on Toluca-Valle de Bravo-Tin- gambato Hwy., 19027'N, 100010'W, 1110 m, Kato 67- 20 (N= 12).

Z. mays var. parviglumis. Jaliscan populations (PV-J). MExIco. Jalisco. 32, 0.3 km E of Agua Caliente, Mpio. Ejutla, 19059'N, 104003'W, 1300 m, 4 Jan 1978, Cobia- Olmedo s.n. (N= 12); 33, 30 km ESE of Tecalitlan, 19018'N, 103005'W, 1200 m, Guzman and Anaya 32 (N = 16); 34, 8 km (by air) ESE of Casimiro Castillo, 19034'N, 104023'W, 900-1100 m, 14 Dec 1977, Guzman s.n. (N = 12); 35*, Jirosto, 17 km WNW of Purification, 19045'N, 104046'W, 500 m, 11 Jan 1979, Iltis and Nee 1480 (N = 22); 36, El Rodeo, 21 km SE of El Chante, 19033'N, 104003'W, 1200 m, 14 Oct 1982, Iltis et al. 28888 (N = 12); 37*, Cerro la Mesa, 7 km NE of El Palmar, 19057'N, 104004'W, 980 m, Puga 11065 (N = 25).

Z. mays subsp. mexicana. Race Chalco (MX-C). MEXICO. Mexico. 38*, Colnetza Boyeros, just W of Texcoco, 19031'N, 98054'W, 2200 m, 29 Sep 1981, Doe- bley 479 (N = 10); 39*, Colnetza Boyeros, just W of Texcoco, 19031'N, 98054'W, 2200 m, 29 Sep 1981, Doe- bley 481 (N 9); 40, 5.5 km NNE of Los Reyes Hwy. interchange on road to Texcoco, 19024'N, 98057'W, 2200 m, 29 Sep 1981, Doebley 482 (N = 12); 41, 14 km SE of Chalco on road to Amecameca, 19012'N, 98050'W, 2300 m, 30 Nov 1971, Iltis and Cochrane 178 (N = 12); 42*, 2 km NE of Chalco on road to Ayotla, 19017'N, 98054'W, 2270 m, 23 Sep 1977, Iltis and Doebley 401 (N = 12); 43, location same as collection 40, Oct 1978, Iltis et al. 769 (N = 12). Distrito Federal. 44*, San Mateo, 19010'N, 99015'W, 2300 m, Kato 68-2 (N = 12).

Z. mays subsp. mexicana. Race Central Plateau (MX-P). MEXICO. Durango. 45*, 23 km ENE (by air) of Durango, 20004'N, 104031'W, 2050 m, 1 Oct 1982, Doebley 625 (N = 12). Guanajuato. 46*, Manuel Doblado, 20045'N, 101050'W, 1700 m, Kato 69-2 (N = 12); 47*, Uriangato, 20010'N, 101005'W, 1925 m, Kato 69-7 (N = 12). Jalisco. 48, 0.5 km NNW of La Estan- cia, 21031'N, 101051'W, 2100 m, 30 Sep 1979, Guzman and Perez 110 (N = 12); 49*, 10 km SW of Degollado, 20022'N, 102011'W, 1600-1650 m, 3 Jan 1978, Puga 11066 (N = 31). Michoacan. 50, Quinceo, 6 km (by air) NW of Morelia, 19043'N, 101014'W, 2000 m, 4 Dec 1971, Iltis and Cochrane 276 (N = 12); 51*, Churintzio,

TABLE 2. Continued.

20010'N, 102000'W, 1800 m, Kato 69-3 (N = 12); 52*, Chucandiro, 19050'N, 101015'W, 1800 m, Kato 69-5 (N = 12); 53*, Puruandiro, 21010'N, 101035'W, 2000 m, Kato 69-9 (N = 12); 54*, Patambicho, 19037'N, 101035'W, 2100 m, 18 Nov 1978, G. Prior s.n. (N = 12).

Z. mays subsp. mexicana. Race Nobogame (MX-N). MEXICO. Chihuahua. 55*, Nobogame, 26005'N, 106058'W, 1850 m, Beadle s.n. (N = 15); 56*, Nobo- game, 26005'N, 106058'W, 1910 m, Kato 78-1 (N= 15).

Z. mays subsp. mays (MAIZ). GUATEMALA. Hue- huetenango. 57, location same as collection 19, Iltis and Lind G-121b (N = 12). MExico. 58, location same as collection 45, Doebley 626 (N = 12); 59*, location same as collection 38, Doebley 480 (N = 12); 60, location same as collection 50, Iltis and Cochrane 296 (N = 12); 61, location same as collection 1, 22 Sep 1978, Iltis et al. 436 (N = 12).

to eliminate hybrid plants when calculating allele frequencies. The exclusion of hybrid plants is especially important in assessing relationships as measured by the distribution of rare alleles. Otherwise, the same rare alleles de- rived from maize may be found in unrelated teosinte collections. Such information might be misinterpreted as evidence for a relationship between those teosinte collections. Of the 860 plants analyzed, 526 were checked and 14 of these were hybrids with maize.

Coleoptile extracts of each plant were elec- trophoresed in four separate gel systems (Car- dy et al. 1980; Stuber and Goodman 1983). Gel slices were then stained for 12 different enzyme systems, which produced isozyme bands coded for by 20 loci (table 3). These loci are distributed on at least seven of the 10 maize chromosomes (table 3; Goodman and Stuber 1983). One additional locus, Mmm, which is a modifier locus changing the migration distances for isozymes encoded by Mdhl, Mdh2, and Mdh3, was also studied, making a total of 21 loci. After staining, the gels were scored for all 12 enzyme systems by the first author, and then checked by the second or third authors. Goodman and Stuber (1983) summarized the genetics of the enzymes we have studied.

Principal component analysis is used to assess relationships among the collections. In this analysis, each individual allele represents a character that may have a value from 0 to 1, corresponding to its frequency in a collection.


TABLE 3. Summary of enzyme systems of Zea studied, including chromosomal locations (chromosome number plus L for long arm or S for short arm of the chromosome) and gel types on which enzymes were assayed. A = L-histidine/citric acid pH 5.0; B = L-histidine/citric acid pH 5.7; C = Lithium hydroxide/boric acid pH 8.3; D = L-histidine/citric acid pH 6.5.

Number of loci Chromosomal Gel

Enzyme (abbreviation) assayed location types

Acid phosphatase (Acp) 1 9 B, D Alcohol dehydrogenase (Adh) 1 1L C Catalase (Cat) 1 C,D Endopeptidase (Enp) 1 6L C Esterase (E or Est) 1 3S C fl-glucosidase (Glu) 1 lOL B Glutamate-oxaloacetate transaminase (Got or Aat) 3 3L, 5L, 5S C Isocitrate dehydrogenase (Idh) 1 6L D Malate dehydrogenase (Mdh) 5 8, 6L, 3L, lL, 5S A, B 6-phosphogluconate dehydrogenase (Pgd) 2 6L,3L B, D Phosphoglucomutase (Pgm) 2 1L, 5S A, D Phosphohexose isomerase (Phi or Pgi) 1 1L B, D

Scores for each principal component were cal- culated using the variance-covariance matrix; therefore, the degree to which each allele con- tributes to the discrimination among collections depends on the among-collection variance for that allele. If the variance is large, then the contribution tends to be large. By this method, alleles that are either rare or vary little among collections contribute relatively little to discrimination among collections. It is often helpful when employing principal component analysis to consult a table of distances among collections. This is prudent because graphs of the first few principal components may not accurately represent the actual inter-collection distances, particularly when the percentages of variation represented by those components is small. In the principal component analysis, each individual collection was graphed separately.

In addition to principal component analysis, Nei's coefficient (I) of genetic identity (Nei 1972) and a modified Rogers's distance (Wright 1978; cf. Gower 1972) are presented. The for- mer may vary from 0, for collections that share no alleles in common, to 1, for collections that have the same alleles at the same frequencies. The latter is a simple Euclidean distance, stan- dardized so that it varies from 0, for collections that have the same alleles at the same frequencies, to 1 for collections that share no alleles in common. While these two statistics vary in- versely from 0 to 1, they are not simple invers-

es of one another, and their distributions between 0 and 1 are quite different. Rogers's distance is a more desirable taxonomic measure of the genetic distance among collections than either Nei's genetic identity or his genetic distance (Goodman 1973), especially in cases where hybridization may be important. In such cases, Euclidean representation of parental populations and their derivatives is often helpful (An- derson 1949; Goodman 1967).

To assess variability within collections and species, the following statistics were employed: actual heterozygosity-(H,) - the proportion of loci heterozygous in an individual averaged over all individuals in a collection; expected heterozygosity (H,) - the proportion of heterozygous loci one would find in a collection if the population were at Hardy-Weinberg equilibrium; and gene diversity (HT) - the proportion of heterozygous loci one would find in a taxon if all of its populations formed a single panmictic unit that was at Hardy-Weinberg equilibrium.

TAXONOMIC AND PHYLOGENETIC IMPLICATIONS

Figure 1 is a principal component analysis of the collections listed in table 2. The allele frequencies used are summarized in table 4. In discussing this figure, we will defer discussion of Zea mays var. huehuetenangensis (collections 18 and 19), an isoenzymatically unique and most interesting taxon, until after discussion of the


O DIPLOPERENNIS < VAR. HUEHUE A SUBSP MEX.-CHALCO ' SUBSP MAYS O PERENNIS * VAR PARVI.-BALSAS A SUBSP MEX.-CNT PLT. * LUXURIANS < VAR PARVI-JALISCO A SUBSP MEX.-NOBOGAME

2 A49 A56

5 A3 8

1 0 ~~~~1 7 IL 4 1

0 A5 2 5l 1 7 ~~~~~~~~~~~~~40A 48A 1514 18 A 45

H ~~~~~~0 A 42 Z 16 3 4

0 3 3 ~~~~~~~~~~~~* 2 A AA

-2~~ ~ ~~ ~~~ 2 2 I 0 E 3

19* * 0 0 2 1 2 0

C) (Do

- -1 0 611 0 2 7

LU ~~~~~~~~~~~~1Q23, 2 5 -1 ~~~~~~~~~6 30* 2

28, 3 6 <3 2

5 8'~ 3 4<>C37

-2

-2 -10 FIRST COMPONENT

FIG. 1. Principal component analysis of samples of Zea. Numbers refer to collection numbers in table 2.

other taxa. The first component (accounting for 26% of the total variation) divides the genus into two groups: (a) the perennials and Zea luxurians with negative values for this component; and (b) the Mexican annual teosintes and maize with positive values. These two groups corre- spond to sects. Luxuriantes and Zea of Doebley and Iltis (1980). The first principal component also separates Z. perennis from Z. diploperennis. The second component (accounting for 14% of the total variation) separates Z. luxurians with positive values from the perennials with mostly negative values. Further, it separates the large-spikeleted, large-seeded, high-elevation teosintes with generally red hairy leaf sheaths and weak root systems (Z. mays subsp. mexicana) with positive values from the small- spikeleted, small-seeded, middle-elevation teosintes with green glabrous sheaths and strong root systems (Z. mays var. parviglumis) with

negative or small positive values (cf. Doebley 1984a).

Some other aspects of figure 1 need to be emphasized. First, collections of Z. perennis are completely distinct from Z. diploperennis. This agrees with morphology (Iltis et al. 1979) and cytoplasm genome studies (Timothy et al. 1982) for which these taxa are totally distinct. Thus, the isoenzymatic data are consistent with the view that these two perennial taxa are "good" species. Second, the three geographic races of Z. mays subsp. mexicana do not appear to be distinct from one another. This again agrees with morphology (Orozco 1979; Doebley 1983a). Third, populations of Z. mays var. parviglumis from the Valle de Bravo in the western part of the state of Mexico (collections 29, 30, 31) are not distinct from other members of this variety and do not seem to merit separate taxonomic status (cf. Smith et al. 1981, 1982). Fourth,

19841 DOEBLEY ET AL.: ZEA 209

VAR. HUEHUE.

SUBSP. MAYS

VAR PARVI.- BALSAS

VAR. PARVI.-JALISCO

SUBSP. MEX.-CHALCO

SUBSP. MEX.- CNT. PLT.

SUBSP. MEX.-NOBOGAME

LUXURIANS

PERENNIS

DIPLOPERENNIS

0.4 0 3 0.2 0.1

MODIFIED ROGERS' DISTANCE

FIG. 2. Average linkage cluster analysis of species and races of Zea (cf. table 1). Using modified Rogers's distances based on isozyme data (table 5).

Z. mays subsp. mays does not stand alone, as Wilkes's placement of it in a separate section implies, but rather it shows complete overlap with Z. mays var. parviglumis (race Balsas). Oth- er maize races may yield somewhat different results, especially considering the tremendous diversity within this cultigen. However, isozyme analysis of hundreds of Latin American maize collections suggest that maize shows its closest relationship to Z. mays var. parviglumis (Goodman and Stuber unpublished; Smith et al. unpublished). In that maize cannot be dis- tinguished from this teosinte, the isoenzymatic data are consistent with the theory that Mexi- can annual teosinte is the direct ancestor of maize (Galinat 1971; Iltis 1971; Beadle 1972; de Wet and Harlan 1972). Fifth, the perennials, and especially Z. diploperennis, do not show a close relationship to maize or Z. mays subsp. mexicana, as suggested by some authors (Wilkes 1967; Smith et al. 1982).

This last point merits some discussion. Not only does isoenzymatic evidence fail to show a close relationship between the perennials and maize, but so do other data, including tassel morphology (Doebley and Iltis 1980), cytoplas- mic DNAs (Timothy et al. 1979; Sederoff et al. 1981; Timothy et al. 1982), nuclear DNA hybridization (Hake 1981), chromosome knobs (Pasupuleti and Galinat 1982), and seed proteins (Smith and Lester 1980; Mastenbroek et al. 1981). Further, the isozyme data presented here (figs. 1-2), as well as many of the studies cited above, show a loose affinity between Z. luxurians and the perennials.

SUBSP. MAYS

VAR PARVI

VAR HUEHUE

SUBSP. MEX - CHALCO

SUBSP MEX - CNT PLT

SUBSP MEX - NOBOGAME

LUXURIANS

PERENNIS

DIPLOPERENNIS

04 03 02 01

EUCLIDEAN DISTANCE

FIG. 3. Average linkage cluster analysis of species and races of Zea (cf. table 1) based on tassel morphology (data from Doebley 1983a).

Figure 1 also reveals the unusual isoenzymatic nature of some populations of Z. mays var. parviglumis. Specifically, collections 20, 21, 22, and 33 all show a distinctly close relationship to Z. mays subsp. mexicana. Paradoxically, collections 20-22 are from central Guerrero, the southernmost portion of the range of Z. mays var. parviglumis, and the region most distant from subsp. mexicana. Further, these populations are morphologically the most divergent from Z. mays subsp. mexicana (Orozco 1979), and they show no greater cytological similarity to subsp. mexicana than do other populations of var. parviglumis (Kato 1976). The reason for the isoenzymatic similarity of the southern var. parviglumis to subsp. mexicana is not apparent, and this similarity does not seem to be supported by other data.

As discussed thus far, the isoenzymatic data show generally good agreement with both inflorescence morphology (Doebley 1983a) and other data. But now we need to consider Z. mays var. huehuetenangensis (Huehuetenango teosinte). In figure 1, this teosinte has a negative value for the first component, grouping with the teosintes of sect. Luxuriantes and showing a particularly close relationship to Z. perennis. In this figure, it apparently does not portray a close affinity to the Mexican annual teosintes. These relationships are partially cor- rect and partially a mirage because this principal component analysis distorts some inter- population distances while accurately portray- ing others. In table 5, Huehuetenango teosinte is about equidistant from all other groups but


w-~~~ 0000~ 0 000 00 e J0-A0 0 0 0 0e C 0 0 0

Z . 0000oO0oo0 0O0 0oo4 00 00oo o o 0oo

* a) I~~~ 00000000000000000000~~~C 0000000000000 C C C ) ItC)C)C

Lo~

Q = , , o O o o o n M n o0 r- o o o n o o o n on Ln L n 0 r- N n M m m 0 Ln O

C'4 ,$ .. . . . . . . . .. . . . . . . .

I 00

to I W

00 04 C )C )c nC ) C ) C 04r -4 e-e 0CO) 000 -) o0000C)-C) 00 Aa)~~~ <i 00000~~~~~CO0 C C400 N C0000 00 0C)C C - )C)C 000 ' ,0000C)00C00

. n. 0000000000

~~ 0 I E-~~~ 0 c0 0000 c0000c000c0 0 c0 000c0

a) I v ~ 00000000 00000000000 00 00000000

. Q I S 00 0 0 0 0 0 0 00 0 0 0 0 00 0 0 0 0 0 0 0 0 000 00 000

00 0 0 0 c00 000 00 0 000 0c0 0 c00 000 0000 00 000

I t 0 00 0 0 0P00 0 0 0 0 0 00 00 0C0 00 00 0 '000 0 00 0 0 0

a) I0 c 0 c; c0 00c 00c000c; c00c000c00c0

-H t1) I b 0 0 O O 0 00 0 0 0 00 0 0 0 0 00 0 CY 0 0 O0 0 1- C) O O O O 0

Y~~~~~ ; 0 b 0 0 O O O O O I0 O 0 > O O O 0 O O O O N O O 0 r

O O ON O O O O O O

a-4 CZ

C4 I c * . CY c. c. CY . ) . .

?~~~~0 0] 0U 0 0 Col Cl Ln C4 l C>l 1- O 0 0 O O r- 0 C1 O O 00 0 O) O) O O

0

A ?I OO OO O00 N OO OO OO OOOOON O~ O O %%f O 0

0 - O L

CN 4

Q tb4 q > . . . . . . . . + .. . I . . . . r-

co r4 co i

< a)O I > 000000 000t0000000400000O040000~000N000

cir

tt t X I a) iz 4

I I 0000 0 00000000000 00000 00 00000CO00000 a) *

9 =I R oOoOoOOoXO?OOOwo O OO O O O O O O OO OO OO OO O O ,,- 003~ ...............0000.............

_ 2; I 00 0- 0 .................. _ e IX OO oo oo 0o0oo 0oo 0oo 0oo 0oo0oo0o 0o00o00o00o00o0 Q :t X

-~ 000000?00?000000o OO - o oo oo U -oo oo o o oo oo I X o~0 o0 0 0 ~ ~ 0 00 0 0 - c~ 00 oC

.ZCI > )H

S 0=

Q iIINIDooooiooo se

0O 0e'4e O r O 0v4D0,0 0O0O C0fL -4 4 <YQ I 0000e'4-400Lfe0 0oo0o000 oo'o0000ro00o00 -o H?.................. *~~O 00 00 00 0 00 0 0 00000000000000 oo ooo ooo ooo ooo

n I ,OOOOXnO OXOOOOc 00 0 000 0 0 0uE00 000O0o00O00000

SX I R * I.**.** ***..** . I oooooooooooo -

S~~~~~r- N .-4 I t O( LD O O. Lo 4 O O N O. O~ .O 4 .O rD O + O~ O~ 00 C O O~ O 0 ON Cs O >~~~~ ~~ ~ ~ ~ I * .% * . >>>..>.>>> I }, ~ ~ 0 -l 0%O X ~

1W 0 9 oa)oosonooo

o w . sooo noooooooooo e u U> . . . . . . . . .


I N O O O X O O O O O b O O O O O 00 0 0 "o Un O O O

I ~~ 0000000000000000000~0000 00 00 0C)00 0 0 0

,I 1 000000000c000c00 0 C00 00,000000o0-co 40jo I4 >

I 0000 0 o o o o o o o o oo

'? ..000000.0000000000000000000000000

o o o0o0o0o0o0o0o00s o0o000o0o0o0o0o0o o0o0o o o o o o o CN o N o o

I ~~~~~~~~~~~~~~~~ ~Lnr - ~ 0

0

O 0e O O- O O O OO O' O O O OOOOO O O CY 00 O

z 0 0 00o o o o o o o 0 o o o 0 0 o 0 C) o C) o 0 0 0 o 0 o 0 0 o 0

,~~~~ CY) o o o o o CY) o

c) co c c) c co C) C) C) o o -o o o o o

ooooo

p4

~~ 00000000~~~~0 000 0 0 N00 0 0CD0 0 0 0 00-0 0 000

* O C4OOC)OOOOOOO O O Oo o o o oo o o o

' Z 0 0 0 0 0 0 0ONO00 oo O OOC ) )C ) )C % 0O OO O

co Cy c c o c c co c co oo o o o o o o

o o C, o o4

o o %o

o 0

o C0

o C)

:2:~~~~~ o o oN o o o o o C) C) o,o C) o o o o o o o o o o o o ooo

o o 0 00 0 ) C C o o 00 0 0 00o o 0 04 0 0 0 0 0 4 0 o o 00 C)

?I O 0000000000 0 00 0 o 0000 CO o 00 O 0O4 O O O OC) 0 00 O

I CO O O 00 N O O o o oo O O O O O O O O O N O O O C) 00 0 r O 00 0r0

'- o - of o~ oo o - o o o o

-o o-- o o o n o 4

o on of -o o o

C)r4 00000000000000000 00000000000 0 0 0 0

I NO O O e4 O O O O O O OOcO

IJ2 ,X OI DOXOOOOO I ~ ~ ~ ~o P~ .. . .. . . .. . . .. . . .. . .

I;O OO OOO OOO 00 000 OOO OOON I

ZO @ OO OO XO OO O _ O 000O Oo O OO I 40000o

0 Iz . . . . . . . .. . . . . . . . . V 00

N 09w NBv

I N sOOOOC 000 00 00 000000000000000000000000

I 1 O O O O00 0 0 0 0 0e 0 00 O OOO LOOO 0OO O0 OOO 0

I S 000000000000000000000000000000000

IX .................................

00 0 > -0 0 0 0 00 00 0 0 0 0 00 OO 0 OOC 0 0000 00

0C")e~ O OO -0 0000 N10 Cl0 000C'lOOOOO - ~0C0 0 O

I 00000 00 00 00 0 0 00 00 00 000

PC m sOXx NOObc 000 O0O0O000 00000000000 00 0 0 ctI O Oe 0 O D 110 0000oOOOOOoOOOOOXOOOt

I 00000 000 000000000~00000000000000

I t 000000000000000000000000000000000

I X OOOOO O O O O O O0O0 O 0C 00C~0 0 0 0 000 O OO - I ; 00 00000000000O000000000000 00 0

IX O 00000000000000000000000000000000

I~~~~C t O~ O~ O .O 00 0~ 0 C>O N O .O O~ 0D O N> 00 0. rl N Or O Ot N Or Or .O 01-

IW a < I * * < I * . cn I) Z- i N N N( X D O

O 0 0


N o o o o o ooo o o 000 000 0 00000000000000 0 0 0 0 00

?4 .......oo

0000000

,z oooooooooooooooooooooooooooooooo

0 ~ ~ ~ ~ ~ ~ ~ ~ ~ 00 0

I 0 OD 0 0 0 0 0 O O O O O O O O 0 0 0 0 0 0t 0 0 0 O OOO 0

I 000 O O O0 0 O O 4 D O O 0000 0 0 00 0 0 0 0 0 0 0 0 0 000

I c 0 0 0 000000000 N 00 D 0eL0000000e 00 %0 0000i

,to o oooo 4 o o oooooo

a) 0 0 0 0 0 0 0 0 0 0 0000 0000

z4 L o o o o o o

o o o oooooooo

I O o O 000000000000000 0 0 0 0 0 0 O O O

. . . .................... . . . .

cn X O oooo DOOO DOO OOOOOOO 00 N O 0 0 0 0

,P . . . . . . . . . . . . . . . . . . . . . . . . . . . : o o o oY o o o CD CD CD CD CD CD CD o oooooo

I

X O O 4 O00000000000000000 0 O O N 00 0 0 00 0 00 0 0 00O

I I O H I Xc

rI c

'0 00 000 - 00 0 0 J0 0 0 ' O O 0

PA .... . .. . .. . .. . .. . .. .

' 555.....................................................oooooooo~oooooooooooooooooooo

I .. ................................................... o oo oo0oo00o00Uno0oo0n oo0oo00o00o000 oo0oo0oo0

H> .$. X .X

I) O 00 0 0 0 0000 0 0 0 0 0 0 000 000sO OOOO000DCsOOO I , O OO00OO00 OO000 O O0000 OO0 OO o00 00 0 00 0000000

w I * . 000000000 X X* t 00O00 *00 *00 *00O O O OO * * I O 0 0 0 0 0O O00OO00O00OO00O00OO00o00Oo00O00oo00O00oO

I ' OOOOOOO4x I O O OO0 OO0 OO0o00 OO0O0000o00o00o o0 u0o oo0o00oo0

< 0000I ) O~L~ I O~~~ O O ~~O OC N Ot cOLr O O O Eqt 0* 0 N

X~~ ~ ~ O- I q Ooou Eoooooooooooooooooo<c


o N m f)0NLf)00000L N " m No N m ooo ooo t N t ON ctl oq ~C oq oq W% C

0000000 m 0 mN m00 N oq oq oq a\ m

ts N a\ N \0 a\ o on

\ 0o a\' m N N o \

0 N oX N N Noo oo ooa'ol

0 6C o oo o Lo ooo 00

N N N a\ x oA

m N \ ' a\ \ 00 N a\

N 00 N N a O /N N N Oi )

N N o0 bo~ LO

_ N \o NN ' " 1^ 14 o

Y

N N N N N o oo/ o a N N

C OCONOO 0 0/ 0/-0L0

L N) 0 N O) 0O N 6 - NN

o N N1 C)00 0N0 666oo m N N /oo/oo m \o

0 4 00N a 0N\ Z00 N e

r- 0 COO0--4a\ 0 N LO O

X u66 o 6oooooooN

Ut ,~~~~~~~~~~~~~(

64

o0 a\N m o oo . o

1 ? N o too soo 4 e e o b1 .0

CZ

LO~~~~~~~~~~~~~~C

CZ ,C

t I I

0 = 7 oooo

164 Wn In I

Nuo / ec N tce tsi ds ts ts Os

:~~~~~~~~~c Y ol o~ o \oo o~ o o~ o

shows its closest relationships to Z. mays var. parviglumis (race Balsas), and not to Z. perennis, as indicated by figure 1. In an average linkage cluster analysis based on the distances in table 5, Huehuetenango teosinte is distinct from all other taxa (fig. 2). Thus, in regard to Huehue- tenango teosinte, the isoenzymatic data reported here show that it is very different from all other Zea, but closest to Z. mays var. parviglumis.

This result does not agree well with either male inflorescence morphology (Iltis and Doe- bley 1980) or with other morphological studies (Wilkes 1967; Smith et al. 1981), all of which showed a much closer relationship between the two varieties of subsp. parviglumis. Similarly, it disagrees with cytoplasm DNA analysis (Tim- othy et al. 1979), nuclear DNA hybridization studies (Hake 1981), and seed protein studies (Smith and Lester 1980; Mastenbroek et al. 1981), all of which showed a very close relationship between Huehuetenango teosinte and the other teosintes of sect. Zea. Cytological studies (Longley 1941a, b; Kato 1976; Smith et al. 1982) have shown Huehuetenango teosinte to be distinct from the Mexican annuals, but these data also indicated a closer relationship between Z. luxurians and Huehuetenango than is supported by isozymes (table 5).

What then should be the taxonomic status of Huehuetenango teosinte? Isoenzymatic data and cytology both suggest that this teosinte be treated as a separate species. Morphology, seed proteins, and cytoplasm genomes suggest that it be considered conspecific with the Mexican annual teosintes and maize. Given that there is little morphological basis to support separate species status for this teosinte, this alternative seems taxonomically indefensible. Nor can we justify treating it as a mere race within Zea mexicana, or as a variety within Z. mays subsp. parviglumis. The best solution may be to treat it as a genetically distinct subspecies within the highly polymorphic Z. mays.

Isoenzymes and tassel morphology can be directly compared by presenting an average linkage cluster analysis for tassel morphology (fig. 3) similar to the one presented for isozyme data (fig. 2). Figure 3 shows generally good agreement with isozyme data (fig. 2), with two ex- ceptions. First, as discussed above, Huehuete- nango teosinte clusters close to var. parviglumis for tassel data but does not do so for isozyme data. The morphological similarity of these two


teosintes may result in part from convergent evolution. Both vars. huehuetenangensis and parviglumis occupy seasonally moist (120-200 cm rainfall annually), middle elevation (400-1700 m) sites, and both have small seeds and spike- lets and generally green, glabrous leaf sheaths. Second, maize shows a close relationship to var. parviglumis for the isozyme data but appears much more distinct for tassel morphology. The morphological differences between maize and teosinte probably arose during the domestica- tion process, while isozyme profiles would probably be largely unaffected by artificial se- lection.

Finally, we should note that the tetraploid, Zea perennis, showed no fixed heterozygosity at any of the loci we examined. This is consistent with the hypothesized autopolyploid origin of this species (Shaver 1962).

MAIZE INTROGRESSION INTO TEOSINTE

Zea has been cited as a model example of reciprocal introgression between a crop and its wild relatives (de Wet and Harlan 1972; Heiser 1973). Nevertheless, the evidence for introgression of maize into teosinte has been largely cir- cumstantial, and a rigorous consideration of this evidence failed to support the hypothesis that teosinte morphology has been noticeably al- tered by maize introgression (Doebley 1984b). The isoenzymatic data presented here offer another opportunity to assess the degree of maize introgression into teosinte.

This study includes five sets of teosinte and maize populations that grew together in the same field. The maize collections (57-61) were obtained from fields with Z. mays var. huehuetenangensis (19), Z. mays subsp. mexicana (45, 38, 50), and Z. diploperennis (1), respectively. At all five localities, hybrids between maize and teosinte were observed.

Figure 1 shows that not one of these teosinte samples has a particularly close relationship to its corresponding maize population. Thus, it seems that these five teosinte populations maintain genetic distinctiveness from their sympatric maize population for traits other than those essential to their survival in the wild. Of the five pairs, teosinte collection 50 shows the closest relationship to its corresponding maize collection (60). However, this teosinte population (as judged by field observations) contained many F, hybrids with maize, and we did not

check our sample to remove hybrid plants. Thus, the slight closeness of collection 50 to the maize collections may only be the effect of including hybrid plants when calculating the allele frequencies. This might also explain why this collection of Z. mays subsp. mexicana shows a closer relationship to Z. mays var. parviglumis than do other subsp. mexicana collections.

One of the most interesting results here con- cerns the three of the maize populations that were sympatric with Z. mays subsp. mexicana. This subspecies was considered by Wilkes (1967) to be a maize-introgressed form of Z. mays var. parviglumis (Balsas teosinte) and by Doebley and Iltis (1980) to be the most maize-like teosinte. The isozyme data do not support either conclu- sion. Zea mays subsp. mexicana does not appear to be a maize-introgressed form of var. parviglumis, nor is it genetically the most maize-like teosinte. Rather, figures 1 and 2 suggest that var. parviglumis is isoenzymatically the most maize-like form. It is difficult to attribute the maize-like isozyme profile of var. parviglumis to maize introgression, simply because hybridization in the Balsas river drainage, where this variety occurs, is rare (Wilkes 1977). In three separate trips to the Balsas ragion, the first author found only a single maize-teosinte hybrid among hundreds of plants checked. Further, var. parviglumis, as noted by Wilkes (1977), is a wild plant that occurs mostly in open wood- lands, prairies, and roadsides. Thus, paradoxically, we seem to find the robust maize-like plants of Z. mays subsp. mexicana, which occur almost exclusively as weeds in or near corn- fields, are isoenzymatically distinct from sympatric maize populations; and the more slen- der, wild plants of Z. mays var. parviglumis, which rarely occur in maize fields and grow mostly in wild places, are isoenzymatically very similar to maize. This unexpected result merits further studies with a larger sample of materials.

Wilkes (1967, 1977) suggested that introgression accounted for the differences between the large, maize-like and small, weedy forms of race Central Plateau. Our collections 45, 49, and 52 are the small weedy type, and collections 46, 50, and 54 are the large maize-like form. Col- lections 46 and 50 show greater affinity to maize, and hybrids were common in these collections, offering some support for Wilkes's suggestions.


0.7 -

A

0.6 *

a. 0.5 -

0.4

0.1 0.2 0.3

FIG. 4. Graph of proportion of polymorphic loci per population (PLP) and mean expected heterozygosity (HF) for taxa of Zea. Symbols as in figure 1 except for circled star, which is an average for outcrossing species (Gottlieb 1981).

Specific isozyme data suggest that some individual plants could contain maize germ- plasm. A few plants of Z. luxurians (collection 13) had the allele Glul-7, a common maize vari- ant that is unknown in sect. Luxuriantes, except for these few plants. One plant of Z. diploperennis (collection 1) had alleles Enpl-8 and Pgdl-3.8, both common in maize, but unknown, except for this plant, in sect. Luxu- riantes. Interestingly, the two loci at which the alleles occur are tightly linked (5 map units apart) on chromosome 6 (Goodman et al. 1980). Because this plant was homozygous for the usual isozymes of Z. diploperennis at other loci, and because it was a morphologically typical specimen of its species, the data suggest that a segment of chromosome 6 from maize has been successfully incorporated into Z. diploperennis.

ESTIMATES OF GENETIC VARIATION

In this section, we assess the degree of variability within the teosintes, as well as compare the isoenzymatic data for teosinte with that of other species. In figure 4, the proportion of polymorphic loci per population (PLP) is graphed against average expected heterozygosity (I-s). In calculating PLP, a locus was considered polymorphic if its most common allele had a frequency of less than 99% (Gottlieb 1981). Both PLP and HR are measures of within population variation.

In figure 4, it is not surprising that Balsas teosinte has the greatest within population

TABLE 6. Mean actual heterozygosity (H,), mean expected heterozygosity (H3), and gene diversity (HT)

for species, subspecies, and races of Zea. * = un- weighted averages.

Taxon H, Hf HT

Z. diploperennis 0.183 0.231 0.292 Z. luxurians 0.072 0.110 0.155 Z. mays subsp. parviglumis

var. heuhuetenangensis 0.127 0.158 0.173 var. parviglumis (Balsas) 0.233 0.261 0.311 Jaliscan populations 0.127 0.137 0.218

Z. mays subsp. mexicana Race Chalco 0.231 0.234 0.279 Race Central Plateau 0.215 0.239 0.336 Race Nobogame 0.182 0.215 0.247

Z. mays subsp. mays 0.215 0.229 0.311

variation, because this teosinte occurs in large populations in which such variation could be maintained. It is surprising that Z. diploperennis, Chalco, and Nobogame teosintes show considerable variability, because all have restricted geographic distributions, and the latter two occur only in relatively small populations. Hue- huetenango teosinte shows less variation than those just discussed, which seems reasonable, given that it is a narrow endemic. Zea luxurians shows little variation, yet there is no obvious reason why this should be so. The size and structure of its populations are not greatly different from Chalco teosinte or Z. diploperennis. Race Central Plateau- shows slightly less within population variation than other annual Mexi- can teosintes.

In figure 4, each Zea taxon possesses greater variation than the average outcrossing species. While this may be true, it may also, in part, reflect our use of two different gel systems for many enzyme systems. This method may have enabled the detection of heterogeneity that re- mained hidden in other studies. The values for expected heterozygosity (Hr) are, consistently, slightly higher than the observed heterozygosity (HI) (table 6). The difference between expected and observed levels of heterozygosity for the genus overall is highly significant (P < 0.0005).

Gene diversity (HT) provides a measure of overall variation in a taxon (table 6). Race Cen- tral Plateau has the greatest value for HT, followed by maize and Balsas teosinte. Zea diploperennis and Chalco teosinte have slightly less,


and race Nobogame slightly less than this. The two Guatemalan teosintes (Z. luxurians and Z. mays var. huehuetenangensis) have the least overall variation of any Zea. Mean genetic identity (I) among populations of the same taxon is a measure of the similarity of the populations. Small values for I indicate considerable diver- gence among populations. As measured by I, populations of Central Plateau teosinte show the weakest affinity to one another, followed distantly by Z. diploperennis and races Balsas and Chalco (table 5). Maize, Z. luxurians, Nobogame teosinte and Z. mays var. huehuetenangensis populations appear to be genetically more uni- form. Populations of the races and species of Zea tend to show weaker affinity to one another than do populations of other plant species (cf. Gottlieb 1981).

Considering all four measures of variation (Hr, PLP, HT, I), the following conclusions can be drawn. (1) Zea luxurians and Z. mays var. huehuetenangensis have little variation as compared to other teosintes. (2) Zea diploperennis and Balsas teosinte have a great deal of variation, which is equitably distributed among their populations. (3) Chalco and Nobogame teosintes and maize have a moderate amount of variation within populations, but their populations retain a considerable degree of genetic similarity. (4) Central Plateau teosinte has substantial variation, both within and among populations. This may suggest that its populations are heterogeneous and isolated from one another. (5) In general, Zea species and races seem to show greater variation both within and among populations than other species. How- ever, this may be, in part, an artifact of different methods employed here and elsewhere.

We have not included Zea perennis in this discussion of diversity because it is a tetraploid, and thus, not directly comparable to the diploids. However, it maintains substantial diversity, as can be seen from the spread of its populations on figure 1. Also, the average number of alleles per sample of Z. perennis was 34.3, compared to 38.2 for Z. diploperennis, 35.7 for Z. mays subsp. mexicana, 30.5 for Z. luxurians, 43.7 for maize, and 46.7 for Balsas teosinte.

SUMMARY

The isozyme data presented here and elsewhere augment our understanding of taxonomic and evolutionary relationship in Zea.

These data, when considered with other information, show Zea luxurians, Z. perennis, and Z. diploperennis to be distinct from other Zea and from one another. At the same time, these three species show a certain degree of affinity to one another, and thus, may be placed into sect. Lux- uriantes. The annual teosintes of Mexico and maize have a close relationship to one another. Within the Mexican annual teosintes, isozyme data show close correspondence with ecologi- cal and morphological observations and separate the large-spikeleted, arid, upland form (subsp. mexicana) from the smaller-spikeleted, more mesic, middle-elevation populations (var. parviglumis or Balsas). Of these two forms of Mexican annual teosinte, the middle-elevation type shows the greatest isoenzymatic similarity to maize. This is somewhat surprising because the upland forms have been generally regard- ed as the more maize-like in gross morphology. Annual teosinte from western Guatemala (Z. mays var. huehuetenangensis) is distinct from other Zea; however, it shows its closest affinity to Z. mays var. parviglumis from the Balsas river drainage.

In regard to introgression, our data suggest that some populations of race Central Plateau could be contaminated by maize. Further, a few plants of Z. luxurians and Z. diploperennis had isozymes otherwise typical of maize. This could be interpreted as evidence for low-level introgression from maize.

Finally, the levels of variation within and among populations in Zea taxa varies consid- erably. Zea luxurians and Z. mays var. huehuetenangensis seem to have the least variation, while Central Plateau and Balsas teosintes and Z. diploperennis seem to have the greatest. In general, Zea taxa seem to have more variation than most other plant species for which isozyme data are available.

ACKNOWLEDGMENTS. Paper No. 8697 of the Jour- nal Series of the North Carolina Agricultural Re- search Service, Raleigh. This investigation was supported in part by NIH Research Grant No. GM 11546 from the National Institute of General Medical Sci- ences of the U.S.A. Our special thanks go to H. H. Iltis, R. Guzman M., and T. A. Kato Y., the gleanings of whose hands formed the core of our research materials. Drs. R. R. Sederoff, H. T. Stalker, and D. H. Timothy made many helpful comments on this manuscript. Present address for J. F. Doebley is: Dept. of Genetics, North Carolina State University, Ra- leigh, North Carolina 27695.


LITERATURE CITED

ANDERSON, EDGAR. 1949. Introgressive hybridization. New York: John Wiley and Sons, Inc.

BEADLE, G. W. 1972. The mystery of maize. Field Mus. Natl. Hist. Bull. 43(10):2-11.

BIRD, R. McK. 1978. A name change for Central American teosinte. Taxon 27:361-363.

CARDY, B. J., C. W. STUBER, and M. M. GOODMAN. 1980. Techniques for starch gel electrophoresis of enzymes from maize (Zea mays L.). North Car- olina State Univ. Institute of Statistics mimeo ser. no. 1317.

COLLINS, G. N. 1921. Teosinte in Mexico. J. Hered. 12:339-350.

1931. The phylogeny of maize. Bull. Torrey Bot. Club 57:199-210.

DE WET, J. M. J. and J. R. HARLAN. 1972. Origin of maize: The tripartite hypothesis. Euphytica 21: 271-279.

DOEBLEY, J. F. 1983a. The maize and teosinte male inflorescence: A numerical taxonomic study. Ann. Missouri Bot. Gard. 70:32-70.

. 1983b. A brief note on the rediscovery of Durango teosinte. Maize Genetics Cooperation Newsletter 57:127-128.

1984a. The taxonomy and evolution of Trip- sacum and teosinte, the closest relatives of maize. Pp. 15-28 in Proceedings of international maize virus disease colloquium and workshop, eds. D. T. Gordon et al. Wooster, Ohio: Ohio Agric. Res. Devel. Cent. In press.

1984b. Maize introgression into teosinte-a reappraisal. Ann. Missouri Bot. Gard. In press.

and H. H. ILTIS. 1980. Taxonomy of Zea I. Subgeneric classification with key to taxa. Amer. J. Bot. 67:982-993.

GALINAT, W. C. 1971. The origin of maize. Annual Rev. Genet. 5:447-478.

GOODMAN, M. M. 1967. The identification of hybrid plants in segregating populations. Evolution 21: 334-340.

1973. Genetic distances: Measuring dissim- ilarity among populations. Yearbook Physical Anthropology 17:1-38.

and C. W. STUBER. 1983. Isozymes of maize. Pp. 1-33 in Isozymes in plant genetics and breeding, eds. S. D. Tanksley and T. J. Orton. Amsterdam: Elsevier Scientific Publ.

f , K. NEWTON, and H. H. WEISSINGER. 1980. Linkage relationships of 19 enzyme loci in maize. Genetics 96:697-710.

GOTTLIEB, L. D. 1981. Electrophoretic evidence and plant populations. Progress in Phytochemistry 7: 1-46.

GOWER, J. C. 1972. Measures of taxonomic distance and their analysis. Pp. 371-380 in The assessment of population affinities in man, eds. J. S. Weiner and J. Huizinga. London: Oxford Univ. Press.

GUZMAN M., R. 1978a. Una neuva localidad para el teosinte Zea perennis y primer reporte de Zea mexicana para Jalisco. Bol. Informativo Instit. Bot., Univ. Guadalajara 6:9-10.

. 1978b. Redescubrimiento de Zea perennis (Gramineae). Phytologia 38:177.

. 1982. El teosinte en Jalisco: su distribucion y ecologia. Ingeniero Agronomo Tesis, School of Agriculture, Univ. Guadalajara, Jalisco.

HAKE, S. C. 1981. The genome of Zea mays: Its organization and homology to related grasses. Ph.D. dissertation, Washington Univ., St. Louis.

HEISER, C. 1973. Introgression re-examined. Bot. Rev. (Lancaster) 39:347-366.

ILTIS, H. H. 1971. The maize mystique-A reappraisal of the origin of corn. [4 pp. mimeo manuscript abstract of a lecture given at the Corn Con- ference, Univ. Illinois, Urbana, in 1969, and Iowa State Univ., Ames, in 1970. Botany Dept., Univ. Wisconsin, Madison.]

1980. The 3rd University of Wisconsin-Uni- versity of Guadalajara teosinte expedition to the Sierra de Manantlan, Jalisco, Mexico: December 28, 1979, to January 21, 1980-Background, pre- liminary results and commentary on nature preservation in Mexico. [60 pp. offset report, Dept. Botany, Univ. Wisconsin, Madison.]

and J. F. DOEBLEY. 1980. Taxonomy of Zea (Gramineae) II. Subspecific categories in the Zea mays complex and a generic synopsis. Amer. J. Bot. 67:994-1004.

f , R. GUZMAN M., and B. PAZY. 1979. Zea diploperennis (Gramineae): A new teosinte from Mexico. Science 203:186-188. 1

KATO Y., T. A. 1976. Cytological studies of maize. Mass. Agric. Exper. Station Bull No. 635.

LONGLEY, A. E. 1941a.* Chromosome morphology in maize and its relatives. Bot. Rev. (Lancaster) 7: 262-289.

. 1941b. Knob positions on teosinte chromosomes. J. Agric. Res. 62:401-413.

MANGELSDORF, P. C. 1947. The origin and evolution of maize. Pp. 161-207 in Advances in genetics, vol. 1. New York: Academic Press.

MASTENBROEK, INEKE, C. E. COHEN, and J. M. J. de Wet. 1981. Seed protein and seedling isozyme pat- terns of Zea mays and its closest relatives. Bio- chem. Syst. Ecol. 9:179-183.

NEI, M. 1972. Genetic distance between populations. Amer. Naturalist 106:283-292.

OROZCO, J. L. 1979. Interrelacion de poblaciones de teocintle anual Mexicano. M.S. thesis, Colegio de Postgraduados, Chapingo, Mexico.

PASUPULETI, C. V. and W. C. GALINAT. 1982. Zea diploperennis I. Its chromosomes and comparative cytology. J. Hered. 73:168-170.

SEDEROFF, R. R., C. S. LEVINGS, D. H. TIMOTHY, and W. W. L. Hu. 1981. Evolution of DNA sequence


organization in mitochondrial genomes of Zea. Proc. Natl. Acad. U.S.A. 78:5953-5957.

SENADHIRA, D. 1976. Genetic variation in corn and its relatives. Ph.D. thesis, Univ. California, Da- vis.

SHAVER, D. L. 1962. A study of meiosis in perennial teosinte, in tetraploid maize and in their tetraploid hybrids. Caryologia 15:43-57.

SMITH, J. S. C., M. M. GOODMAN, and T. A. KATO Y. 1982. Variation within teosinte II. Numerical analysis of chromosome knob data. Econ. Bot. 36: 100-112.

- , , and R. N. LESTER. 1981. Variation within teosinte. I. Numerical analysis of morphological data. Econ. Bot. 35:187-203.

- , , and C. W. STUBER. 1984. Variation within teosinte III. Numerical analysis of allozyme data. Econ. Bot. 38:97-113.

and R. N. LESTER. 1980. Biochemical system- atics and evolution of Zea, Tripsacum and related genera. Econ. Bot. 34:201-218.

STANDLEY, P. C. 1950. Teosinte in Honduras. Ceiba 1:58-61.

STUBER, C. W. and M. M. GOODMAN. 1983. Allozyme genotypes for popular and historically important inbred lines of corn. U. S. Agricultural Re- search Service, Southern Series no. 16.

TIMOTHY, D. H., C. S. LEVINGS, D. R. PRING, M. F. CONDE, and J. L. KERMICLE. 1979. Organelle DNA variation and systematic relationships in the genus Zea: Teosinte. Proc. Natl. Acad. U.S.A. 76:4220-4224.

f , A. K. WEISSINGER, R. R. SEDEROFF, M. M. GOODMAN, and W. W. L. Hu. 1982. Com- parisons of mitochondrial DNAs of Zea mays L. and its relatives. Bot. Soc. Amer. Misc. Publ. No. 162:47-48. (Abstract).

WILKES, H. G. 1967. Teosinte: The closest relative of maize. Cambridge: Bussey Inst., Harvard Univ.

1977. Hybridization of maize and teosinte, in Mexico and Guatemala and the improvement of maize. Econ. Bot. 31:254-293.

WRIGHT, S. 1978. Variability within and among nat- ural populations. Evolution and the genetics of populations, vol. 4. Chicago: Univ. Chicago Press.

isoenzymatic variation in zea (gramineae) · isoenzymatic variation in zea (gramineae) john f....

Documents