DIALLEL ANALYSIS OF WATER-USE EFFICIENCYIN GIANT BERMUDAGRASS CYNODON DACTYLON(L.) (PERS.) VAR. ARIDUS (HARLAN ET DE WET)

Item Type text; Dissertation-Reproduction (electronic)

Authors Lira, Mario de Andrade, 1941-

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this materialis made possible by the University Libraries, University of Arizona.Further transmission, reproduction or presentation (such aspublic display or performance) of protected items is prohibitedexcept with permission of the author.

Download date 30/01/2021 18:00:29

Link to Item http://hdl.handle.net/10150/288319

INFORMATION TO USERS

This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted.

The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction.

1.The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity.

2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame.

3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete.

4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced.

5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received.

Xerox University Microfilms 300 North Zeeb Road Ann Arbor, Michigan 48106

75-4947

LIRA, Mario de Andrade, 1941-DIALLEL ANALYSIS OF WATER-USE EFFICIENCY IN GIANT BERMUDAGRASS CYNODON DACTYLON (L.) PERS. VAR. ARIDUS HARLAN et de WET.

The University of Arizona, Ph.D., 1974 Agronorny

Xerox University Microfilms, Ann Arbor, Michigan 48106

THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED.

DIALLEL ANALYSIS OF WATER-USE EFFICIENCY IN GIANT

BERMUDAGRASS CYNODON DACTYLON (L.) PERS. VAR.

ARIDUS HARLAN et de WET

by

Mario de Andrade Lira

A Dissertation Submitted to the Faculty of the

DEPARTMENT OF AGRONOMY AND PLANT GENETICS

In Partial Fulfillment of the Requirements For the Degree of

DOCTOR OF PHILOSOPHY WITH A MAJOR IN AGRONOMY

In the Graduate College

THE UNIVERSITY OF ARIZONA

19 7 4

THE UNIVERSITY OF ARIZONA

GRADUATE COLLEGE

I hereby recommend that this dissertation prepared under my

direction by Mario de Andrade Lira .

entitled DIALLEL ANALYSIS OF WATER-USE EFFICIENCY IN GIANT BERMUDAGRASS CYNODON DACTYLON (L.) PERS. VAR. ARIDUS HARLAN et de WET

be accepted as fulfilling the dissertation requirement of the

degree of Doctor of Philosophy

Dissertation Director Date *

After inspection of the final copy of the dissertation, the

follovring members of the Final Examination Comniittee concur in

its approval and recommend its acceptance:"''

L, Mm /A 9-6-7H

This approval and acceptance is contingent on the candidate's

adequate performance and defense of this dissertation at the final oral examination. The inclusion of this sheet bound into the library copy of the dissertation is evidence of satisfactory

performance at the final examination.

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

SIGNED:

ACKNOWLEDGMENTS

I wish to express sincere gratitude to Dr. W. R.

Kneebone for his constructive advice and guidance through­

out the graduate program.

Appreciation is due to the graduate committee, Dr.

M. H. Schonhorst, Dr. L. N. Wright, Dr. G. L. Jordan, and

Dr. P. R. Ogden; for guidance, use of growth-chamber, and

assistance in the preparation of the dissertation.

I also wish to express my gratitude to A.I.D., to

the Government of Brazil, and Government of the State of

Pernambuco for financial support during the course of

study; to Drt Kneebone's student helpers for their

cooperation, and to Dr. R. 0. Kuehl, statistician; to

Paul Johnson, computer programmer.

I give special thanks to my wife, Maria, for her

encouragement and assistance during the graduate program.

iii

TABLE OF CONTENTS

Page

LIST OF TABLES V

ABSTRACT viii

INTRODUCTION 1

LITERATURE REVIEW 2

Description of Material 2 The Concept "Water-Use Efficiency" 4 Environmental Factors Associated with w.u.e. . . . 4 Genotypic and Physiological Effects on w.u.e. . . 7 Breeding Forage Crops 8 Heterosis 12 Diallel Cross 13

MATERIAL AND METHODS 17

RESULTS AND DISCUSSION 24

Number of Seed Obtained per Cross and Number of Progenies Studied 24

Yield 28 Transpiration 31 Water-Use Efficiency 35 Environmental Effects 39 Character Associations 40 Heritabilities 42 Combining Ability .... 43

LITERATURE CITED 48

iv

LIST OF TABLES

Table Page

1. Origin of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) clones used as parents in the diallel cross 18

2. Number of seed obtained per crossing bag placed on Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) clones in the field 25

3. Number of progenies studied per cross of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) clones . 27

4. Mean dry matter produced per pot by parents and progenies, Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet fgiant bermudagrass) under growth-chamber conditions 29

5. Mean transpiration losses per pot from parents and progenies, Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass), during 28 days under growth-chamber conditions 32

6. Mean transpiration losses from parents and their crosses, Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass), during 28 days under growth-chamber conditions , . . 33

7. Mean transpiration losses from parents and their grouped reciprocal crosses, Cynodon dactylon (L.) Pers. var, aridus Harlan et de Wet Cgiant bermudagrass), during 28 days under growth-chamber conditions ..... 34

v

vi

Table

8 .

9.

10.

11.

12.

13.

LIST OF TABLES—Continued

Page

Grains of water transpired per gram of harvested dry matter (water-use efficiency) for parents and their crosses of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) 36

Grams of water transpired per gram of harvested dry matter (w.u.e.) for parents and their crosses of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) 37

Grams of water transpired per gram of harvested dry matter (w.u.e.) for parents and their grouped reciprocal crosses of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) .... 38

Mean yield, transpiration, and water-use efficiency (w.u.e.) obtained per run for 174 clones of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) 40

Correlation coefficients between yield, transpiration, and water-use efficiency (w.u.e.) under growth-chamber conditions of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) .... 41

Heritabilities for three characters determined by regression and by diallel analysis of Cynodon dactylon (L.) Pers. var, aridus Harlan et de Wet (giant bermudagrass) 43

Results from analysis of variance for combining ability of a complete diallel with six giant bermudagrass clones, Cynodon dactylon (L.) Pers. var, aridus Harlan et de Wet . 44

vii

LIST OF TABLES—Continued

Table Page

15. Comparison of expected water-use efficiency values of crosses, based on general combining ability, with obtained values in a complete diallel with six giant bermudagrass clones, Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet 46

ABSTRACT

All possible reciprocal crosses were made among six

selected giant bermudagrass, Cynodon dactylon (L.) Pers.

var. aridus Harlan et de Wet, clones. Progenies and parents

were evaluated for forage yield, transpiration, and water-

use efficiency under two moisture regimes and two fertility

levels in growth-chamber studies.

Nitrogen promoted a large increase in yield. This

increase in yield was accompanied by an increase in trans­

piration and an improved water-use efficiency (w.u.e.).

Yield and transpiration were larger under excessive soil

moisture supply than under normal soil moisture supply but

w.u.e. was approximately the same under the two moisture

levels.

Differences among yield means were not significant

but transpiration and w.u.e. showed significant differences

among crosses. These characters were significantly corre­

lated.

Heritabilities determined by regression of mid-

parent and offspring and by regression of female parent and

offspring showed fair agreement. Heritability for w.u.e.

was much higher than that for yield and transpiration.

Reciprocal effects were significant for yield and

transpiration with no other parameters significant for these

viii

two characteristics. Water-use efficiency, on the other

hand, showed significant specific combining ability with

no other parameters significant for this character.

INTRODUCTION

Giant bermudagrass, Cynodon dactylon (L.) Pers. var.

aridus Harlan et de Wet, is adapted worldwide in tropical to

warm temperate regions and has high forage production

potential. It has special forage potential in Arizona.

Arizona is also the major producer of bermudagrass seed for

world trade. Much of the forage and seed production of this

grass is in areas where water is often a scarce and valuable

input commodity. Therefore, bermudagrass can remain pro­

ductive only if it uses water efficiently.

An understanding of the genetic background and mode

of inheritance of water-use efficiency is essential to

effectively develop types of giant bermudagrass that can

fully convert available water to economic production. The

diallel cross technique is a reliable breeding procedure

to obtain desired information.

Objectives were to determine heritabilities, general

combining ability, specific combining ability, and mode of

gene action for the characters yield, transpiration, and

water^use efficiency. These parameters provide a basis for

developing a breeding program aimed at improving these

important plant characteristics.

1

LITERATURE REVIEW

Description of Material

Kneebone (1966) pointed out that Cynodon dactylon

(L.) Pers. (bermudagrass), from its long forgotten origin

in tropical Africa or the Indo-Malaysian area, has spread

around the tropical world. The first deliberate introduc­

tion of bermudagrass into the United States took place in

the eighteenth century (Harlan, 1970).

The genus Cynodon includes nine species: C.

aethiopicus Clayton et Harlan, C. arcuatus J.D. Presl. ex

C.B. Presl., C. barberi Rang, et Tad., C. dactylon (L.)

Pers., C. incompletus Nees, C. nlemfuensis Vanderyst, C.

plectostachyus (K. Schum.) Pilg., C. transvaalensis Burtt-

Davy, and C. X magenissi Hurcombe (Harlan et al., 1970).

Six varieties are assigned to the most widespread species,

C. dactylon. They are dactylon (L.) Pers., afghanicus

Harlan et de Wet, aridus Harlan et de Wet, coursii (Camus)

Harlan et de Wet, elegans Rendle, and polevansii (Stent)

Harlan et de Wet (Harlan et al., 1970). Seed from C.

dactylon var. dactylon and C. dactylon var. aridus,

bermudagrass and giant bermudagrass respectively, are grown

in Arizona. Giant bermudagrass is diploid and the probable

progenitor of bermudagrass which is tetraploid (Kneebone,

2

1973). Harlan and de Wet (1969) first described giant

bermudagrass as Cynodon dactylon var. aridus.

Giant bermudagrass is distributed from South Africa

to Palestine to South India, and introduced in Arizona. It

has a true rhizome in which the tip always stays below the

surface. Leaves are relatively unmodified on stolon tips.

The branching pattern of stolons is alternate with one bud

suppressed. Giant bermudagrass is medium to large size

without turf formation and does not have tissue winter

hardiness (Harlan, de Wet, and Richardson, 1969; Rawal and

Harlan, 1971),

Cynodon dactylon is highly cross pollinated,

without apomixis (Burton, 1947). Another important charac­

teristic, for a breeding program, is that most clones are

highly self-incompatible (Burton and Hart, 1967; Hoff,

1967).

Cynodon dactylon endures variations in soil pH, soil

fertility, soil water content, soil salinity, and grazing

(Kneebone, 1966). The grass can make the most of dry areas

where salt is a problem because it is capable of growing in

environments with 27,000 ppm of salt (Schroder, 1966). The

major bermudagrass problem is poor animal performance

associated with high lignin and silica content of the

forage (Duble, Lancaster, and Holt, 1971).

4

The Concept "Water-Use Efficiency"

The concept and study of water requirement is very

old. Briggs and Shantz (1913) defined water requirement as

"the ratio of water absorbed by a plant during its growth to

the weight of dry matter produced" (p. 7). Agronomists

before and since have studied this important physiological

aspect of plant water relations. Viets (1962) defined

water-use efficiency as the weight of dry matter produced

per unit volume of water used in evapotranspiration. He

also stated that consumptive use is equivalent to evapo­

transpiration; i.e., the sum of the volumes of water used

by plant growth in transpiration and evaporation from the

soil. Teare et al. (1973) defined water-use efficiency

(w.u.e.) as: ratio of dry matter produced per unit area

(P.M.) to evapotranspiration from that area (E.T.).

Dobrenz et al. (1969), on the other hand, reported water-

use efficiency as grams of transpired water per gram of dry

matter produced. As used by Dobrenz et al. (1969) water-

use efficiency and water requirement are synonymous.

Environmental Factors Associated with w.u.e.

Miller and Hunt (1966) stated that a number of

complex factors control water requirement of a plant. This

statement is well supported by an abundant literature

dealing with water requirement and water-use efficiency of

crops. Factors that induce optimum plant growth, in

5

general, are conducive to high efficiency of water-use.

Factors that limit plant growth, on the other hand, are

conducive to low water-use efficiency. Keller (1953)

determined that Dactylis glomerata L. (orchardgrass) plants

high in herbage production were low in water requirement.

Dobrenz et al. (1969) found that dry weight and total

protein were significantly correlated with water-use

efficiency in Panicum antidotale Retz. (blue panicgrass).

Dobrenz, Cole, and Massengale (1971) found a significant

correlation between dry matter production and water-use

efficiency in Medicago sativa L. (alfalfa). Weng and

Quinones (1969) however, stated that, in spite of a sig­

nificant negative correlation between dry matter production

and water requirement, high herbage yield could not be used

with confidence as a guide in selecting for low water re­

quirement in Panicum obtusum H.B.K. (vine mesquite).

High soil fertility induces high water-use effi­

ciency. Hanks and Tanner (1952) stated that water was more

efficiently used by Zea mays L. (corn), Avena sativa L.

Coats), and alfalfa under 70.5 lb per acre of nitrogen than

under 4.5 lb per acre of nitrogen. Burton, Prine, and

Jackson (1957) stated that tropical grasses Cynodon dactylon

(coastal, suwannee, and common bermudagrass), Paspalum

notatum Flugge (Pensacola bahiagrass), and Digitaria

decumbens Stent (pangolagrass) were more efficient in water

use when soil was properly fertilized. Brown (1971)

6

determined that 268 kg per ha of nitrogen increased water-

use efficiency by an average of 56% in Triticum aestivum L.

(wheat).

Management influences water-use efficiency of

plants. Baker and Hunt (1961) stated that Agropyron

desertorum Pisch. ex Link. Schult. (crested wheatgrass)

clipped at two inches was more efficient in use of water

than when clipped at four inches. This difference was

probably due to a greater leaf area of the plants clipped

at four inches. Soil moisture level and soil aeration are

other environmental factors that control water-use effi­

ciency. Khudairi, Wahab, and Thewairi (1962) stated that

Phaseolus mungo L. (green-gram) and Gossypium hirsutum L.

(cotton) grew better when soil moisture was kept between

field capacity and 75% field capacity. Campbell and

Ferguson (1969) studied effects of moisture stress and

aeration on water use by wheat. Pots maintained between

0.2 atm and 1,4 atm had low oxygen diffusion rates leading

to poor aeration. Less wheat was produced and more water

used under poor aeration than under good aeration. Robert,

Joseph, and Stolzy (1972) also found a trend toward greater

water^use by wheat plants with low aeration. Interaction

of low aeration and heat stress sharply reduced root weight

and increased water used per unit of plant tissue dry weight

produced. These effects may be partially interpreted in

terms of heat inducing increased respiration rates of roots,

7

leading to degradation of cytoplasm as a result of induced

oxygen scarcity.

Genotypic and Physiological Effects on w.u.e.

Water-use efficiency is strongly affected by the

environment surrounding a plant. However, water requirement

also shows within and between-species variation. Keller

(.1953) determined that water requirements among 16 genotypes

of orchardgrass were significantly different. Burton et al.

(.1957) stated that coastal and suwanee bermudagrass were

more efficient than common bermudagrass, pensacola bahia-

grass, and pangolagrass. Baker and Hunt (1961) found

significant differences in water requirement among clones

of crested wheatgrass. Hunt (1962) determined, by parent

progeny correlations, that water requirement was a highly

heritable character in the species Elymus junceus Fisch.

(Russian wildrye) and Agropyrum intermedium (Host) Beauv,

(.intermediate wheatgrass) . Miller and Hunt (1966) reviewed

water requirement of plants and its importance to grassland

agriculture. They stated that improvement through breeding

programs can be realized. Wright and Dobrenz (.1970) found

significant differences in water-use efficiency among

selections of Eragrostis curvula (Schrad.) Nees (boer

lovegrass).

Stomate density is a plant factor that controls

transpiration and, consequently, water-use efficiency of a

8

plant (Salisbury and Ross, 1959). However, Dobrenz et al.

(1969) found no significant association between water-use

efficiency and stomate densities for six clones of blue

panicgrass.

Breeding Forage Crops

Selection within a base population followed by

utilization of the selected material for creation of new

populations is the general procedure used in many breeding

programs (Comstock and Robinson, 1952). Breeding pro­

cedures that may be used for forage crop improvement can be

grouped, according to Murphy and Lowe (1966), into 5

categories: natural populations, selection, synthetic

varieties, asexual progenies, and first generation hybrids.

Burton (1973) discussed with some detail the possible uses

of mass selection, recurrent selection, synthetic varieties,

first-generation chance hybrids, self^incompatibility

hybrids, cytoplasmic male-sterile hybrids, and apomictic

hybrid techniques in the improvement of quantitative and

qualitative characters of forage crops. However, as Burton

(1952) earlier pointed out, understanding of hereditary

mechanisms controlling a certain character is of maximum

importance and can change the art of plant breeding to a

science as well as greatly facilitate development of superior

varieties.

9

Kempthorne et al. (1954) stated that variation in

any population can be partitioned into heritable and non-

heritable components. The first could be obtained by the

formula:

(block mean square - error mean square) number in each block

Genetic uniformity within blocks is assumed in this formula.

The nonheritable component, on the other hand, would be

measured by the error mean square. Gardner (1963) stated

that genetic parameters useful to a plant breeder are

additive genetic variance, dominance variance, epistatic

variance, average degree of dominance, genotype by environ­

ment interactions, and genotypic correlations among charac­

ters. Falconer (1960) suggested elimination of genotypic

variance, using clones for example, in calculating variance

due to environmental causes.

Very useful concepts in plant breeding are general

combining ability (g.c.a.) and specific combining ability

(s.c,a.) as introduced by Sprague and Tatum (1942). They

state "The term 'general combining ability' is used to

designate the average performance of a line in hybrid

combination, . . . The term 'specific combining ability' is

used to designate those cases in which certain combinations

do relatively better or worse than would be expected on the

basis of the average performance of the lines involved"

(p. 923). Graumann (1952) reviewed studies with a variety

10

of cross fertilized species demonstrating that combining

ability is heritable. Johnson (1952) and Kehr (1961), among

others, reported that g.c.a. is due primarily to additive

effects of polygenes, s.c.a., on the other hand, is due to

deviations from the additive scheme. In general, as

suggested by Theurer and Elling (1963), a high g.c.a. mean

square suggests rapid progress in plant improvement through

selection. The relative amount of g.c.a. and s.c.a. will

also control the type of breeding program in many other

ways. One example of this control was given by Lonnquist

and Gardner (1961). They suggested use of reciprocal

recurrent selection instead of other methods when s.c.a. is

more important than g.c.a. Sprague (1967) recommended, in

plants that may be reproduced vegetatively, inbreeding when

dominance effects are more important than additive effects.

Allard (1960) pointed out that combining ability in

forage species can be estimated by using open-pollinated

progeny tests, top-cross tests, polycross tests, and single

crosses (including diallel crosses). Frakes and Matheson

(.1973) compared parents, single crosses, open-pollinated

progenies, polycross, and F2 progenies of Festuca arundin-

acea Schreb. (tall fescue) and concluded that all methods

allowed separation of promising parental plants.

Another very useful concept in plant breeding is

heritability. Allard (1960) defined heritability as "The

proportion of observed variability which is due to heredity,

11

the remainder being due to environmental causes. More

strictly, the proportion of observed variability due to

additive effect of genes" (p. 468). The first, broader

concept, is called heritability in the broad sense. The

second, more strict concept, is called heritability in the

narrow sense.

Strickberger (1971) stated that heritability may be

2 estimated as twice the regression coefficient (2b = H ) when

regression of offspring on one parent is used. However,

when the midparent value is used, heritability is con­

sidered equal to the regression coefficient. Smith and

Kinman (1965) stated that the most commonly used expres-

2 sions to estimate heritability appear to be 2b = H and

2 b = H . The first is adequate when the parents are non-

inbred but results in an overestimate of heritability if

the parents are related. Heritability can also be obtained

in other ways such as analysis of variance from clonal data,

from open pollination progenies, and from controlled

crosses, including the diallel cross. Kneebone (1958)

determined a fair agreement for heritability estimates

obtained by variance components from clonal data, variance

components from open-pollination progeny, and parent progeny

regression when studying plant height, plant diameter,

leafness, and protein percentage in the leaves of Andropogon

hallii Hack, (sand bluestem). Usually, as Murphy (1952)

explained, heritabilities calculated by progeny and parent

12

correlation have been variable for yield but have been

higher and of predictive value for characteristics less

influenced by environment such as leaf character and

disease reaction. Hunt (1962) reported that water require­

ment, determined by parent and progeny correlation, was a

highly heritable character in Russian wildrye and inter­

mediate wheatgrass.

Heterosis

Another concept very useful in plant breeding is

heterosis or hybrid vigor. Breese and Hayward (1972)

stated that heterosis may be defined as superiority over

the best parent or over the best available variety. Moll,

Salhuana, and Robinson (1962) stated that heterosis,

expressed as percentage above midparent, increases with

increased genetic diversity among parents. Moutray and

Frakes (1973) stated that crosses between tall fescue clones

of diverse morphology, origin, and anthesis date resulted

in greater heterosis than crosses performed between clones

selected for similarities. Presence or absence of

heterosis will determine to some extent the breeding program

because a hybrid program should be developed only when

heterosis is present and some means of easily obtaining

hybrid seed is available (Breese and Hayward, 1972),

When the end product of a breeding program is a

single plant that may be propagated vegetatively,

13

hybridization may well precede other methods of plant

breeding (Burton, 1947). Potential of this method can be

well illustrated by development of coastal bermudagrass, a

landmark in the annals of forage breeding (Harlan, 1970).

Selection among clones is very effective because it allows

full use of variability present (Dickerson, 1961).

Estimation of genetic parameters of a population

allows a decision about the most efficient improvement

method that can be used. It also allows calculation of the

gain that can be expected from selection. Selection gain

is equal to (k) (sp) (Hn), where k is the selection differen­

tial, sp is the phenotypic standard deviation, and Hn is the

narrow sense heritability (Sleper, Drolsom, and Jorgensen,

1973).

Diallel Cross

A diallel cross may be defined as all possible

combinations of single crosses among a set of parents. It

is a relatively old method of plant breeding and as early

as 1947, Yates (1947) described an analysis of data from all

possible reciprocal crosses among a set of parental lines.

Frandsen (1952) stated that the complete diallel cross is

probably the most significant progeny test. Graumann (1952)

stated that diallel crossing is a valuable technique to

determine both specific and general combining ability.

Johnson (1963), eleven years later, commented that

14

statistical techniques for analysis of diallel-cross data

have been accepted by plant breeders as a very useful method

for rationalizing genetic studies of continuous variation.

Bray (1971) pointed out that the breeder may perform a

partial diallel to estimate variance components, herita-

bility, general combining ability, specific combining

ability, and to select the best parent.

The main disadvantage of the diallel-cross system

is the number of mating combinations required, rendering the

method impracticable if many genotypes are to be tested

(Frandsen, 1952; Corkill, 1956). Hayman (1954) stated that

2 in a diallel cross there are n mating combinations.

Evidently, he included reciprocal crosses as well as selfed

progenies. Le Clerg (1966), on the other hand, excluded

reciprocals and selfs and, consequently, he explained, with

n parents the number of crosses would be obtained by the

formula n(n-l)/2. Griffing (1956) listed and explained four

possible diallel cross models: (1) parents, one set of

F^'s and reciprocals are included; (2) parents and one set

of F^'s are included but reciprocals are not; (3) one set of

F^'s and reciprocals are included but parents are not; and

(4) one set of F^'s but neither parents nor reciprocals.

Interpretation of diallel cross data is valid to

the extent that the population from which inferences are

to be drawn fit the genetic model. Most models assume

(Kempthorne, 1956) that crosses show normal diploid

15

segregation, there are no maternal effects, there is random

epistasis, an arbitrary number of alleles at each locus,

parents have the same coefficient of inbreeding, and there

is no genotype by environment interaction.

Peterson (1971) pointed out that three decisions

are necessary in a diallel cross: (1) number of parents,

(2) number of individuals to be grown per cross, and (3)

number of replicates in the experimental design. He stated

that eight to ten parents are preferred, with few indi­

viduals per cross and only two or three replicates. An

equal number of progenies is preferred since the unequal

number of progenies results in considerable difficulties

in the analysis of variation by an orthogonal analysis of

variance (Kempthorne, 1957). In the case of unequal

progeny numbers, Kempthorne (1957) suggested an analysis of

plot mean;":, ignoring any differences in their reliabilities.

Today, computer models are available that overcome the

difficulties of analysis due to unequal numbers of progeny

(Shaffer and Usanis, 1969).

Numerous diallel crosses have been conducted for

estimation of genetic parameters. Kehr (1961) conducted a

diallel cross among six alfalfa clones and estimated that

the variance component for g.c.a. was more important than

s.c.a, for fall growth habit and rate of recovery; s.c.a.

on the other hand was more important than g.c.a. for forage

yield and spring growth habit. Based on the results, he

16

stated that commercial production of alfalfa crosses to

utilize the s.c.a. of two clones appears to be a possi­

bility. Johnson (1965) evaluated three complete diallels

for alfalfa forage yield and suggested that prior selection

of parental clones for vigor of growth may provide a valid

means of discarding those less desirable in terms of

combining ability. Wilson and Cooper (1969) described a

diallel analysis of photosynthetic rate and related leaf

characters among six contrasting genotypes of Lolium perenne

L. (perennial ryegrass). Fifteen families, represented

by ten plants, and six parents, represented by four plants,

were included in the study. Sleper and Drolsom (1974)

studied a complete diallel cross of Bromus inermis Leyss.

(smooth bromegrass) utilizing a randomized complete block

design with three replications. A comprehensive analysis of

variance is included by these authors. Burton and Hart

(.1967) reported a diallel cross involving six self-

incompatible bermudagrass clones in which four of the

fifteen diallel crosses yielded as much forage as the

vegetatively propagated clones in the test. One clone,

unrelated to the others, showed high g.c.a. for yield. There

is, apparently, no published work about diallel crosses

involving water-use efficiency.

MATERIAL AND METHODS

Research was conducted, during 1972 and 1973, in

the field, greenhouse, and growth chamber of the Agricul­

tural Research Service, U. S. Department of Agriculture

located at the Tucson Plant Material Center, Tucson,

Arizona.

The breeding procedure was a diallel cross, model

number one of Griffing (1956), that included parents, one

set of F^'s, and their reciprocals. Maximum number of

progenies evaluated per cross was ten. The parents, giant

bermudagrass clones originally selected for vigor, were

chosen because they had varied significantly in w,u,e. in

previous studies conducted by W, R. Kneebone. Clones were

Yakima, B 442, B 445, B442 open-pollination selection-one

(B 442 OP^), B 442 open-pollination selection-two (B 442

OP2), B 273, and B 434. The clones varied considerably in

their origins (Table 1).

The crossing procedure was based on assumption of

self ̂-incompatibility (Burton and Hart, 1967; Hoff, 1967).

Thus, selfing was not attempted. The crossing procedure was

to enclose into semitransparent plastic coated bags three

to five inflorescences (as pollen source), sustained in a

vial of water, together with one to three intact in­

florescences (as female) prior to anthesis, All possible

17

18

Table 1. Origin of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) clones used as parents in the diallel cross.

Number Clone

Identification Origin

1 Yakima Yakima, Washington. Received from A. G. Law, Washington State University

2 B 442 Oklahoma State University 8152 (Afghanistan)

3 B 445 P.I. 291 616 (South Africa)

4 B 442 Open pollination progeny B 442, Arizona

5 B 442 0P2 Open pollination progeny B 442, Arizona

6 B 273 Arizona selection from NK 37

7 B 434 Oklahoma State University 983 (Israel)

reciprocal crosses were attempted among the selected

clones. Crosses were made in the field in the summer of

1972 and later with potted plants in the greenhouse. No

record of the number of crosses was kept, but the number of

seeds per cross was determined for the crosses made in the

summer.

Glumes, lemmas, and paleas were removed by rubbing

between rubber surfaces. Caryopses obtained were then

germinated and propagated. Henceforth "seed" and

"caryopsis" will be used interchangeably. Seed were

germinated in a Cleland Model 1000 FAAT germinator, set for

alternating 20 and 30 C temperatures with 16 hours of dark­

ness at 20 C and eight hours of light at 30 C. Relative

humidity within the germinator chamber was maintained at 75

to 85%. Disposable plastic petri dishes (9 cm) were used

as germinator containers. Seed were placed upon two to four

circles of Eaton and Dikeman No. 617 filter paper moistened

with distilled water. Seed that did not germinate within a

few days were perforated with a pin to increase germina­

tion. Seedlings obtained were transplanted to plastic pots

containing approximately 1200 g of potting soil. Plants

were maintained in these pots prior to initiation of the

experiment. In crosses from which more than ten seedlings

were obtained, ten progeny plants each were selected at

random after discarding those with insufficient growth to

allow vegetative multiplication in the water-use efficiency

experiment.

Parents and crosses were evaluated in the growth

chamber for yield, transpiration, and w.u.e. in the summer

and fall of 1973.

Procedures followed in vegetative increase and

testing of the experimental populations were as follows:

1. Stolon cuttings, two nodes each, were placed in

fresh vermiculite for rooting. Stolons required

between 15 and 25 days to root.

20

2. Two to four (depending on the number of rooted

stolons available) well rooted cuttings were trans­

ferred to plastic pots with 2.14 liters capacity

(40 fluid oz). Each pot had a drainage hole

drilled in its side wall near the base. During

establishment the hole was left open. During the

testing period the hole was sealed with a rubber

laboratory cork (size 00). Each pot was filled

with screened pea gravel to bring the weight of the

container, gravel and cork to 200 g, then 1200 g of

air dried screened Mohave soil was added to each

pot.

3. Experimental design was a randomized complete

block with four runs (spaced in time) and two

replications per run of 175 treatments (168 progeny

plants, six parents, one pot without plants to

measure surface evaporation). Each treatment in

each replication consisted of a single pot. Plants

in each replication and each run were genetically

identical to those in other replications and runs.

The growth chamber was set for 14 hours of light

and the temperature was kept between 38.9 C and

21,1 C. The hottest period was between two and four

p.m. and the coolest period between four and six

a.m. The light period was between six a.m. and

eight p.m. with an approximate light intensity of

21

3300 foot candles. The growth chamber setting was

the same for all runs. However, there were

refrigeration compressor difficulties with the

chamber in run II. Data were unreliable and were

discarded from the study.

4. Before transfer to the growth chamber the grasses

were allowed to develop during a variable time (at

least 10 days for run I, at least 50 days for run

III, and at least 87 days for run IV). After that

period the grasses were clipped to the level of the

pot border and 250 ml of ground styrofoam were added

per pot to decrease evaporation from the soil

surface (Dobrenz, Cole, and Joy, 1968). Pots of

run IV were the same as those used in run II and

plants were clipped twice before the beginning of

the experiment. The two replications of run IV also

received one g of ammonium sulfate per pot 10 days

prior to placement into the growth chamber.

5. Immediately after cutting, the pots were moved to

the growth chamber and water consumption was

measured during 28 days. Pots were weighed every

two days. Pots of run I were watered to saturation

point (22% of soil dry weight) when they reached

less than 70% of water held at saturation point

C15,4% of soil dry weight). Pots of runs III and IV

were watered to field capacity (12% of soil dry

22

weight) when they reached less than two-thirds of

water held at field capacity (8% of soil dry

weight).

6. Topgrowth was harvested after 28 days by clipping

at the level of the pot border. Clippings were

oven dried to measure dry matter production (yield,

expressed as g of dry matter per pot).

7. Transpiration per pot was determined by the differ­

ence between the water used by each pot and the

water lost through evaporation from the unplanted

pot (check) in each repetition.

8. Water-use efficiency was calculated as grams of

water transpired per gram of dry matter produced.

Analysis of the diallel cross was conducted in

accordance with the computer model proposed by Shaffer and

Usanis (1969), Correlations between the characters studied

were also calculated with the aid of the computer. Standard

analyses of variance were made using the average of

progenies in each cross. Analyses of variance were con­

ducted for grouped runs I, III, and IV according to Gomes

C1966). Heritabilities were determined from averages of

all runs, by regression of offspring on mid^parents.

Heritabilities were also calculated from the diallel

analysis:

23

4 sa + 4 sl „ = 2 s b 4 S2 + 4 S2 + S2 + S2 + S2

g s m r e

2 2 2 2 where is broad sense heritability, and S^, Sg, Sm, Sr,

2 Sg are g.c.a., s.c.a., maternal, reciprocal, and error

variance, respectively (Sleper et al., 1973).

RESULTS AND DISCUSSION

Number of Seed Obtained per Cross and Number of Progenies Studied

Number of seed obtained per cross attempted in the

field during the summer of 1972 was variable (Table 2). The

2 X test for the 1:1 hypothesis showed that as female parent,

B 442 OP2 produced significantly more seed than the average

of all crosses and that Yakima produced significantly less

seed than the average. Clone B 434 was not included in the

2 X test, as female, since almost no seed was produced. The

2 X test for homogeneity was also significant meaning that

samples were significantly different. Clones studied also

2 differed significantly as male. The X test for 1:1

hypothesis showed that females produced significantly more

seed when fertilized by the male B 442 OP^ and signifi­

cantly less when fertilized by the males B 434 and B 442.

Average seed from all females (excluding B 434) produced per

field cross bag was 2.15, 4.61, 3.05, 7.86, 2.51, 3.73, 2.04

when fertilized by the males B 434, B 445, B 273, B 442 OP^,

B 442 OPj, Yakima, and B 442, respectively. These results

indicate genetic potential for a breeding program directed

toward improved seed production in giant bermudagrass.

Numbers of progeny plants obtained and studied for

each cross from the crossing program were variable (Table 3).

24

25

Table 2. Number of seed obtained per crossing bag placed on Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) clones in the field.

Female

Parents

Male Bags Seeds/Bag

B 434 B 442 5 .20 B 445 8 .13 B 273 4 .00 B 442 OP;l 3 .15 B 442 0P2 5 .00 Yakima 7 .00

Total 32 .08

B 442 B 434 4 2.75 B 445 4 2.50 B 273 3 4.33 B 442 0PX 4 11.75 B 442 0P2 5 8.60 Yakima 5 4.40

Total 25 5.72

B 445 B 434 5 2.60 B 442 3 4.67 B 273 3 .00 B 442 0PX 3 4.67 B 442 0P2 2 2.50 Yakima 4 8.75

Total 20 3.86

B 273 B 434 2 2.50 B 445 2 6.50 B 442 4 .50 B 442 0P-, 2 2.00 B 442 0P2 2 3.50 Yakima 4 1.25

Total 16 2.71

B 442 OP, B 434 1 1.00 X B 445 1 2.00 B 442 1 6.00 B 273 2 7.50 B 442 0P2 1 .00 Yakima 1 4.00

Total 7 3.41

Table 2.—Continued

26

Parents

Female Male Bags Seeds/Bag

B 442 OP, B 434 4 3,25 B 445 3 6.67 B 442 4 .25 B 273 2 6.00 B 442 OP-l 2 28.00 Yakima 3 4,00

Total 20 8.02

Yakima B 434 9 .78 B 445 5 10.00 B 442 6 .83 B 273 4 .50 B 442 0P-, 4 .75 B 442 0P2 2 .50

Total 30 2,39

27

Table 3. Number of progenies studied per cross of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) clones.

Number of progenies studied

Male parents

Female Parents

Yakima (1)

442 (2)

445 (3)

442 OP-i (4)

442 0P9 (5)

273 (6) Tota!

Yakima (1) — 2 10 1 None 3 17

442 (2) 10 — 3 10 9 7 39

445 C3) 10 7 — 9 10 1 37

442 0P1 (4) None 2 1 — 9 8 20

442 0P2 (5J 1 1 4 10 — 5 21

273 (6) 10 8 10 3 3 — 34

Total 31 10 28 34 31 24 168

28

Clone B 434 was dropped from the program due to lack of

progenies when used as female. The few seed obtained from

this clone failed to germinate.

Yield

Analysis of variance of yield means did not show

significant differences when the runs by treatment interac­

tion mean square was used as the denominator of the F test.

This interaction was the proper denominator due to its being

significant when tested against the error term (Gomes, 1966) .

Progenies averaged only 76% of the production of the parents

(Table 4). Because the parents had originally been selected

for vigor and probably were themselves natural heterotic

hybrids, progeny from them would tend to be less vigorous

and hence lower yielding. Additionally, parents were not

selected on the basis of combining ability but because they

differed in w.u.e. Burton and Hart (1967) reported a

diallel involving six self-incompatible bermudagrass clones

in which only four of the fifteen diallel crosses yielded

as much forage as the vegetatively propagated clones. No

report was found in the literature in which a cross showed

significant positive heterosis when compared to the

vegetatively propagated parent. However, this fact does

not invalidate the search for a breeding combination as

productive as the parent clones because commercial multi­

plication by seed is, usually, more economical than

29

Table 4. Mean dry matter produced per pot by parents and progenies, Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass) under growth-chamber conditions.

Dry matter per pot (g)

Male parent

Female Cross Parental parent 1 2 3 4 5 6 means means

1 — .58 .80 .75 missing .68 .70 .79

2 . 74 — .64 .67 .60 .69 .67 .97

3 .58 .70 — .67 .80 .28 .61 .92

4 missing .48 .35 — .84 .76 .61 .92

5 .37 .52 .66 .59 — .61 .55 .66

6 .78 .79 .63 .71 .69 .72 .79

Cross Means .62 .61 .62 .68 .73 .60 .64 .84

30

vegetative multiplication. Also, on a field scale, the less

vigorous and hence less productive progenies of any given

cross will tend to be eliminated by more vigorous and

productive progenies with a subsequent increase in yield

potential. This hypothesis is unproven and best survival

in the field is not necessarily associated with best yield.

There was a significant correlation (r = .60) between

number of progenies studied per cross and yield of the

cross. A possible explanation, if selfing were possible,

might be that some or all of the progenies of low yield were

selfs instead of true crosses. There was no significant

correlation between yield and seed produced per cross (r =

,02). This lack of correlation indicates that yield is

a character independent of number of seed produced per

bagged cross and that selfing is not a likely explanation

for the correlation yield and number of progenies studied.

Furthermore, most crosses had some progeny with low yields,

the low yield of some progenies being compensated by the

higher yield of other progenies in crosses where several

progenies were studied. For example, the progenies of the

cross 4 by 6 had an average yield of .76 g of dry matter per

pot but one of them produced only .15 g? the cross 2 by 4

had an average yield of .67 g but one progeny plant produced

only .23 g of dry matter per pot. These results were

expected because the parents are heterozygous and the

character yield is dependent upon a large number of genes.

31

Transpiration

Mean transpiration values for parents and their

progeny, with or without grouped reciprocals, showed sig­

nificant differences. Average progeny transpiration was

95% that of the parents. Consequently, transpiration rates

did not reduce as much as yield when compared with parental

values. However, transpiration is closely associated with

yield and this difference in percentage may be fortuitous

(Tables 5, 6, and 7).

Separation of transpiration means for parents and

their progeny, using the Duncan test and the run by treat­

ment interaction standard deviation shows that several

reciprocals differed significantly (Table 6). Cross 3 by 4

was different from 4 by 3 and cross 3 by 6 was different

from 6 by 3. Low transpiration rates of progenies are

significantly correlated with low number of progenies per

cross (r = .51). Transpiration is closely correlated with

yield. The hypothesis presented to explain low yield may be

used to explain low transpiration. Another possible explana­

tion might be a partial incompatibility between the cytoplaan

of clone 4 and the genes of 4 by 3 as well as between the

cytoplasm of 3 and the genes of 3 by 6. This would explain

the small number of progenies obtained as well as the

association of small numbers with low yield and trans­

piration. However, the cytoplasm of clone 4 should be

identical to cytoplasm of clones 5 and 2 (442 OP-^, 442

32

Table 5. Mean transpiration losses per pot from parents and progenies, Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass), during 28 days under growth-chamber conditions.

Female parent

Transpired water per pot (g)

Male parent Cross Parental means means

1 — — 287 368 365 missing 359 345 327

2 354 — 297 318 292 320 316 374

3 255 293 — 298 372 153 274 373

4 missing 223 164 — 419 338 286 392

5 209 288 322 302 — 319 288 281

6 379 388 321 324 371 — 357 383

Cross means 299 296 294 321 364 298 312 355

33

Table 6. Mean transpiration losses from parents and their crosses, Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass), during 28 days under growth-chamber conditions.

Parents and crosses

Means (g/pot)

Parents and crosses

Means (g/pot)

4 x 5 419 a* 6 X 3 321 abed

4 392 ab 2 X 6 320 abed

6 x 2 388 ab 5 X 6 319 abed

6 383 abc 2 X 4 318 abed

6 x 1 379 abc 5 X 4 302 abed

2 374 abc 3 X 4 298 abed

3 373 abc 2 X 3 297 abed

3 x 5 372 abc 3 X 2 293 abed

6 x 5 371 abc 2 X 5 292 abed

1 X 3 368 abc 5 X 2 288 bede

1 X 4 365 abc 1 X 2 287 bede

1 JX 6 359 abc 5 281 bcdef

2 x 1 354 abc 3 X 1 255 cdef

4 x 6 338 abed 4 X 2 223 def

1 327 abed 5 X 1 209 def

6 x 4 324 abed 4 X 3 164 ef

5 x 3 3 2 2 a b e d 3 x 6 1 5 3 f

*Means with the same letter are not statistically different (p = .05) according to Duncan's Multiple Range Test, using the interaction run x treatment standard deviation.

34

Table 7. Mean transpiration losses from parents and their grouped reciprocal crosses, Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass), during 28 days under growth-chamber conditions.

Means Parents and crosses (g/pot)

4 392 a* 6 383 ab 2 374 ab 3 373 ab

1 X 6, 6 X 1 369 ab 1 X 4, missing 365 ab 4 x 5, 5 X 4 361 ab 2 x 6, 6 X 2 354 abc 3 x 5, 5 X 3 347 abed 5 x 6, 6 X 5 346 abed 4 x 6, 6 X 4 331 abed

1 327 abede 1 X 2, 2 X 1 321 abede 1 X 3, 3 X 1 312 abede 2 x 3, 3 X 2 296 abede 2 x 5, 5 X 2 291 abede

5 281 abede 2 x 4, 4 X 2 271 bede 3 x 6, 6 X 3 237 cde 3 x 4, 4 X 3 231 de missing, 5 X 1 209 e

•Means with same letter are not statistically different (p = .05) according to Duncan's Multiple Range Test, using the interaction run x treatment standard deviation.

35

442) but 5 and 2 are compatible with 3. This would refute

this hypothesis.

The combined reciprocal crosses 3 by 6 and 6 by 3

as well as 3 by 4 and 4 by 3 (Table 7) had significantly

lower transpiration levels than either of their parents

(63 and 60% that of their respective parental averages).

These crosses therefore showed significant heterosis.

Yield levels for these crosses, however, averaged 53 and 55%

those of the respective parental averages (Table 4) and

although a significant reduction in transpiration might be

wished for, it should not be at the expense of yield.

Water-Use Efficiency

Progenies averaged 15% less efficient than parents

(535 vs. 467, Table 8). Separation of means was done by

the Duncan test using the error term standard deviation when

reciprocal crosses were not grouped (Table 9) and using the

runs by treatments interaction when the means of reciprocal

crosses were grouped (Table 10). The separation of means of

parents and crosses (grouped or not, Tables 9 and 10) showed

that the cross 3 by 2 was significantly more efficient than

the crosses 6 by 3 and 3 by 6. Crosses 3 by 6 and 6 by 3,

in spite of heterotic low transpiration, were very ineffi­

cient in water use. Cross 3 by 2 associated a relatively

low transpiration rate with a relatively good yield. It is

of practical importance that the most efficient cross (2 by

36

Table 8. Grams of water transpired per gram of harvested dry matter (water-use efficiency) for parents and their crosses of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass).

Water-use efficiency

Male parent Female Cross Parental parent 1 2 3 4 5 6 means means

1 — 513 486 552 missing 563 529 453

2 521 — 519 522 525 506 519 430

3 530 448 — 483 519 635 523 447

4 missing 552 517 — 543 515 532 451

5 588 558 546 561 — 544 559 503

6 516 513 618 518 573 548 520

Cross means 539 517 537 527 540 553 535 467

37

Table 9. Grams of water transpired per gram of harvested dry matter (w.u.e.) for parents and their crosses of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass).

Parents and Means Parents and Means crosses (w.u.e.) crosses (w.u.e.)

2 430 a* 6 520 abc

3 447 ab 2 X 1 521 abc

3 x 2 448 ab 2 X 4 522 abc

4 451 ab 2 X 5 525 abc

1 453 ab 3 X 1 530 abc

3 x 4 483 abc 4 X 5 543 abc

1 X 3 486 abc 5 X 6 544 abc

5 503 abc 5 X 3 546 abc

2 x 6 506 abc 1 X 4 552 abc

6 x 2 513 abc 4 X 2 552 abc

1 X 2 513 abc 5 X 2 558 abc

4 x 6 515 abc 5 X 4 561 abc

6 x 1 516 abc 1 X 6 563 abc

4 x 3 517 abc 6 X 5 573 abc

6 x 4 518 abc 5 X 1 588 be

2 x 3 519 abc 6 X 3 618 c

3 x 5 519 abc 3 X 6 635 c

*Means with same letter are not statistically dif­ferent (p = .05) according to Duncan's Multiple Range Test, using the error term standard deviation.

38

Table 10. Grams of water transpired per gram of harvested dry matter (w.u.e.) for parents and their grouped reciprocal crosses of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass).

Means Parents and crosses (w.u.e.)

2 430 a* 3 447 ab 4 451 abc 1 453 abc

2 x 3 3 X 2 483 abed 3 x 4 4 X 3 500 abed

5 503 abed 1 X 3 3 X 1 508 abed 2 x 6 6 X 2 509 abed 4 x 6 6 X 4 517 abede 1 X 2 2 X 1 517 abede

6 520 abede 3 x 5 5 X 3 533 abede 2 x 4 4 X 2 537 abede 1 X 6 6 X 1 540 abede 2 x 5 5 X 2 542 bede 1 X 4 missing 552 bede 4 x 5 5 X 4 552 bede 5 x 6 6 X 5 559 cde missing, 5 x 1 588 de 3 x 6, 6 X 3 627 e

*Means with same letter are not statistically dif­ferent (p = .05) according to Duncan's Multiple Range Test, using the interaction run x treatment standard deviation.

39

3, 3 by 2) was obtained from the most efficient parents

(3 and 2). Contrary to yield and transpiration, there was

no significant correlation between w.u.e. and number of

progenies per cross (r = .36).

Environmental Effects

Variances for average yield, transpiration, and

w.u.e. per run were significantly greater than the error

term variance and the run by treatment interaction variance

(Table 11). The Duncan test, however, did not allow separa­

tion of yield means. Nitrogen, applied in run IV, promoted

a large increase in yield. The increase in yield was

accompanied by an increase in transpiration and an improved

w.u.e. The results are in accord with previous work (Hanks

and Tanner, 1952; Burton et al., 1957; Brown, 1971). Yield

and transpiration were larger under excessive soil moisture

supply (run I) than under normal soil moisture supply (run

III). However, w.u.e. was approximately the same under the

two moisture regimes. Therefore, both stresses, deficient

N and excessive moisture supply, had comparable effects on

the w.u.e. Campbell and Ferguson (1969) reported that pots

maintained between .2 atm and 1.4 atm had low oxygen dif­

fusion rates leading to poor aeration. Bermudagrass showed

consequently ability to grow under poor aeration. The

ability can be of great advantage in certain environments,

40

Table 11. Mean yield, transpiration, and water-use efficiency (w.u.e.) obtained per run for 174 clones of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass).

Characters

Runs

Characters I III IV

Yield (g/pot) .59 a* .35 a 1.15 a

Transpiration (g/pot) 302 e 184 f 469 g

w.u.e. 569 m 565 m 456 n

*Means bearing the same letter are not statistically different (p = .05) according to Duncan's Multiple Range Test.

such as the humid tropics, where the soil may remain water-

saturated during part of the rainy season.

Character Associations

Correlations among values of transpiration, yield,

and w.u.e. indicate that with an increase in yield there

was an increase in transpiration but a decrease in water

used to produce a unit of dry matter (Table 12). Plants

were most efficient in water use when transpiring at rela­

tively high levels. Results are in accord with previous

work (Keller, 1953; Dobrenz et al., 1969, 1971).

Correlations between transpiration and yield as well

as between yield and w.u.e. were uniform from run to run and

41

Table 12. Correlation coefficients between yield, trans­piration, and water-use efficiency (w.u.e.) under growth-chamber conditions of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass).

Correlations (r)

Characters

Run Rep Transp.-Yield Transp.-w.u.e. Yield-w.u.e.

1 1 . 92** -.14 -.37**

1 2 . 92** -.26** -.45**

3 1 .82** .16 -.39**

3 2 .67** .35** -.44**

4 1 .88** -.13 -.44**

4 2 .85** .11 -.39**

Combined .93** -.31** -.50**

from replicate to replicate within runs. Correlations

between transpiration and w.u.e. were not uniform from run

to run and from replicate to replicate. Transpiration and

yield as well as yield and w.u.e. showed the same type of

association under different fertility leyels and watering

regimes but the transpiration and w.u.e, association varied

with environment. The results suggest that yield can be

used with confidence to select for w.u.e, but transpiration

can not be used as a selection criterion for w.u.e.

42

Heritabilities

Heritabilities determined by regression of mid-

parent and offspring and by regression of female parent and

offspring showed acceptable agreement between the two methods

(Table 13). Heritabilities calculated from the diallel

analysis were smaller for w.u.e. than heritabilities

calculated by regression. For yield and transpiration/

results were essentially identical in all three methods of

calculation. The three methods of calculation provided the

same conclusion. Heritability was higher for w.u.e. than

for transpiration and yield. The results indicate that the

proportion of observed environmental variation, when compared

to the variability due to heredity, is higher for trans­

piration and yield than for w.u.e. Results obtained are

well supported by the literature. Murphy (.1952) stated that

heritabilities calculated by parent-progeny correlation have

been variable for yield. Hunt (1962) reported that water

requirement, determined by parent and progeny correlation,

was a highly heritable character in the species Russian

wildrye and intermediate wheat grass. Parents with high

w.u.e, can be expected to produce progenies that are more

efficient in water use than progenies of low water-efficient

plants. Parental performance for yield and transpiration

were not predictive of progeny performance and progeny tests

would be an essential part of breeding programs for these

characters.

43

Table 13. Heritabilities for three characters determined by regression and by diallel analysis of Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet (giant bermudagrass).

Characters

Middle parent-offspring

Female-offspring

Diallel analysis

Characters byx H2(%) V H2(%) H2 (%)

w.u.e. •

00

00

88 .41 81 18

Transpiration -.04 00 -.06 00 6

Yield .17 17 .04 8 4

The significant correlation between dry matter

production and w.u.e. suggests that yield improvement will

improve w.u.e. However, the higher heritability of w.u.e.

suggests that the measurement of the ratio transpiration and

yield (w.u.e.) allows a more predictable gain through selec­

tion than either character considered individually.

Combining Ability

Reciprocal effects for yield and transpiration were

significant with no other parameters significant for these

two characters (Table 14). The interaction run by treatment

may have partially invalidated the diallel analysis for

combining ability for yield and transpiration (Kempthorne,

1956), Possible explanations for significant reciprocal

effects have been already discussed. The character w.u.e.,

44

Table 14. Results from analysis of variance for combining ability of a complete diallel with six giant bermudagrass clones, Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet.

Statistics

Characters and significance

Yield Transpiration w.u.e.

General combining ability N.S. N.S. N.S.

Specific combining ability N.S. N.S. **

Maternal effects N.S. N.S. N.S.

Reciprocal effects ** ** N.S.

N.S. = Not significant.

** = Significant at 1%.

in contrast to transpiration and yield, did not show

significant reciprocal effects (Table 14) but did show

significant s.c.a. This significant s.c.a. points out that

certain crosses did relatively better or worse than pre­

dicted by the g.c.a. of the clones involved. The cross

5 by 1 required 588 g of water to produce a gram of dry

matter but it should have required, taking into considera­

tion the g.c.a. of clones 5 (B 442 0P2) and 2 (Yakima), 542

g of water to produce a gram of dry matter. The cross 2 by

3 (grouped with 3 by 2) required 484 g but should have

required 524 g, the cross 3 by 6 (grouped with 6 by 3)

required 627 g but should have required 540 g.

45

Comparisons between expected w.u.e. values, based on the

g.c.a. of parents, and values actually obtained, are

2 presented in Table 15. The X test for the 1:1 hypothesis

shows that the only significant difference between observed

and expected value was in the cross 3 by 6 (grouped with

2 6 by 3). The X test for homogeneity was not significant.

Data show that, in spite of significant s.c.a.,

g.c.a. allows separation of most promising crosses. The

crosses 1 by 2, 1 by 3, 1 by 4, 2 by 3, 2 by 4, 2 by 5,

2 by 6, and 3 by 4, showed better w.u.e. expectation, had

519 as average w.u.e. The other seven crosses, with worse

w.u.e. expectation, had a mean w.u.e. of 559. Data obtained

suggest the possibility of using g.c.a. to separate the best

crosses as far as w.u.e. is concerned. The final per­

formance of the cross must be the definite guide for

selection due to significant s.c.a.

Presence of a significant s.c.a., as well as the

presence of crosses as efficient in w.u.e. as the parents,

associated with an easy method of obtaining good quality

seed, suggests that the use of first-generation self

incompatibility hybrids is an acceptable breeding technique

to improve w.u.e. of giant bermudagrass. The cross 442 by

445 was very promising and should be field tested. The

relative amount of s.c.a. in relation to g.c,a. suggests

that reciprocal recurrent-selection might be efficient

(Lonnquist and Gardner, 1961). If inbreeding were possible,

46

Table 15. Comparison of expected water-use efficiency values of crosses, based on general combining ability, with obtained values in a complete diallel with six giant bermudagrass clones, Cynodon dactylon (L.) Pers. var. aridus Harlan et de Wet.

Water-use efficiency values

2 Crosses Expected Obtained X 1:1

1 X 2, 2 X 1 526 517 .0776

1 X 3, 3 X 1 532 508 .5538

1 X 4, missing 532 552 .3690

missing f/ 5 X 1 542 588 1.8724

1 X 6, 6 X 1 542 540 .0036

2 x 3, 3 X. 2 524 484 1.5872

2 x 4, 4 X 2 524 537 .1592

2 x 5, 5 X 2 534 542 .0594

2 x 6, 6 X 2 534 510 .5516

3 x 4, 4 X 3 530 500 .8736

3 x 5, 5 X 3 540 533 .0456

3 x 6, 6 X 3 540 627 6.4858*

4 x 5, 5 X 4 540 552 .1318

4 x 6, 6 X 4 538 517 .4180

5 x 6, 6 X 5 550 559 .0730

2 X for homogeneity = 13.2616 N.S.

selection among inbreds might improve hybrid performance

because dominance effects (s.c.a.) were more important than

additive effects (Sprague, 1967).

LITERATURE CITED

Allard, R. W. 1960. Principles of plant breeding. John Wiley and Sons, Inc. 485 p.

Baker, J. N., and 0. J. Hunt. 1961. Effects of clipping treatments and clonal differences in water require­ment of grasses. J. Range Management 14:216-219.

Bray, R. A. 1971. Quantitative evaluation of the circulant partial diallel cross. Heredity 27: 189-202.

Breese, E. L., and M. D. Hayward. 1972. The genetic basis of present breeding methods in forage crops. Euphytica 21:324:336.

Briggs, J. L., and H. L. Shantz. 1913. The water require­ments of plants. I. Investigation in the Great Plains in 1910 and 1911. USDA Bur. Plant Ind. Bull. 284. 49 p.

Brown, P, L. 1971. Water use and soil water depletion by dryland winter wheat as affected by nitrogen fertilization. Agron. J. 63:43-46.

Burton, G. W. 1947. Breeding bermudagrass for the south­eastern United States. Agron. J. 39:551-569.

Burton, G. W. 1952. Quantitative inheritance in grasses. Proc. 6th Int. Grassland Congr. 1:277-283.

Burton, G. W. 1973. Breeding better forages to help feed man and preserve and enhance the environment. Bio. Sci. 23:705-710.

Burton, G. W., and R. H. Hart. 1967. Use of self-incompatibility to produce commercial seed-propagated Fi bermudagrass hybrids. Crop Sci, 7: 324-327.

Burton, G. W., G. M. Prine, and J. E, Jackson. 1957. Studies of drought tolerance and water use of several southern grasses. Agron. J. 49:498-503.

48

49

Campbell, C. A., and W. S. Ferguson. 1969. Influence of air temperature, light intensity, soil moisture stress and soil aeration on moisture use by wheat. Canad. J. Plant Sci. 49:129-137.

Comstock, R. E., and H. F. Robinson. 1952. Genetic parameters, their estimation and significance. Proc. 6th Int. Grassland Congr. 1:284-291.

Corkill, L. 1956. The basis of synthetic strains of cross-pollinated grasses. Proc. 6th Int. Grassland Congr. 1:306-313.

Dickerson, G. E. 1961. Biological interpretation of the genetic parameters of populations. Statistical Genetics and Plant Breeding. NAS-NRC Pub. 98 2:225-252.

Dobrenz, A. K., D. F. Cole, and R. J. Joy. 1968. Com­parison of materials for reducing evapotranspiration of soil moisture in water efficiency studies. Agron. J. 60:446.

Dobrenz, A. K., D. F. Cole, and M. A. Massengale. 1971. Yield components and leaf characteristics associ­ated with the water requirement of alfalfa. Crop Sci. 11:124-125.

Dobrenz, A. K., L. N. Wright, M. A. Massengale, and W. R. Kneebone. 1969. Water-use efficiency and its association with several characteristics of blue panicgrass (Panicum antidotale Retz.) clones. Crop Sci. 14:213-216.

Duble, R. L., J. A. Lancaster, and E. C. Holt. 1971. Forage characteristics limiting animal performance on warm-season grasses. Agron. J. 63:795-798.

Falconer, D. S. 1960. Introduction to quantitative genetics. Oliver and Boyd, Glasgow. 365 p,

Frakes, R. V., and K. I. Matheson. 1973. Progeny testing for forage yield in tall fescue, Festuca arundinacea Schreb, Crop Sci. 13:293-295.

Frandsen, K. J. 1952. Theoretical aspects of cross­breeding systems for forage plants. Proc, 6th Int. Grassland Congr. 1:306-313,

50

Gardner, C. 0. 1963. Estimates of genetic parameters in cross-fertilizing plants and their implications in p'iant breeding. Statistical Genetics and Plant Breeding. NAS-NRC Pub. 982:225-252.

Gomes, F. P. 1966. Curso de Estatistica Experimental. Univ. S. Paulo. Escola Superior de Agricultura "Luiz de Queiroz." Piracicaba. 404 p.

Graumann, H. 0. 1952. The polycross method of breeding in relation to synthetic varieties and recurrent selec­tion of new clones. Proc. 6th Int. Grassland Congr. 1:314-319.

Griffing, B. 1956. Concept of general and specific combining ability in relation to diallel crossing systems. Australian J. Biol. Sci. 9:463-493.

Hanks, R. ,J., and C. B. Tanner. 1952. Water consumption by plants as influenced by soil fertility. Agron. J. 44:98-100.

Harlan, J. R. 1970. Cynodon species and their value for grazing and hay. Herbage Abs. 40:2a3-238.

Harlan, J. R., and J. M. M. de Wet. 1969. Source of variation in Cynodon dactylon (L.) Pers. Crop Sci. 9:774-778.

Harlan, J. R., J. J. J. de Wet, W. W. Huffine, and J. R. Deakin. 1970. A guide to the species of Cynodon CGramineae). Oklahoma S. U. Agric. Exp. Sta. Bull. B-673. 37 p.

Harlan, J, R., J. M. J. de Wet, and W. L. Richardson. 1969. Hybridization studies with species of Cynodon from east Africa and Malagasy. Amer. J. of Bot. 56: 944-950.

Hayman, B. I. 1954. The theory and analysis of diallel crosses. Genetics 39:789-869.

Hoff, B. J. 1967. The cytology and fertility of bermuda-grass, Cynodon dactylon (L.) Pers. in Arizona. Dissertation Abst. 28:2227-B.

51

Hunt, O. J. 1962. Water requirement of selected genotypes of Elymus junceus Fisch. and Agropyrum intermedium (Host) Beauv. and their parent-progeny relation­ships. Crop Sci. 2:97-99.

Johnson, I. J. 1952. Evaluating breeding materials for combining ability. Proc. 6th Int. Grassland Congr. 1:327-334.

Johnson, I. J. 1965. Heterosis in alfalfa. Proc. 9th Int. Grassland Congr. 1:124-129.

Johnson, L. P. V. 1963. Applications of the diallel-cross techniques to plant breeding. Statistical Genetics and Plant Breeding. NAS-NRC Pub. 982:561-578.

Kehr, W. R. 1961. General and specific combining ability for four agronomic traits in a diallel series among six alfalfa clones. Crop Sci. 1:53-55.

Keller, W. 1953. Water requirement of selected genotypes of orchardgrass, Dactylis glomerata L. Agron. J. 45:622-625.

Kempthorne, 0. 1956. The theory of the diallel cross. Genetics 41:451-460.

Kempthorne, 0. 1957. An introduction to genetic statistics. John Wiley and Sons, New York.

Kempthorne, 0., T. A. Bancroft, J. M, Gowen, and J. L. Lush. 1954. Statistics and mathematics in biology. The Iowa College Press, Ames. 632 p.

Khudairi, A. K., A. S. A, Wahab, and A. J. Thewairi. 1962. Plant response to different soil moisture levels. Arid Zone Res. XVI. Plant-water relationships in arid and semi-arid conditions. Proc. of the Madrid Symposium. UNESCO:265-269.

Kneebone, W. R. 1958. Heritabilities in sand bluestem CAndropogon hallii Hack.) as estimated from parental clones and their open-pollination progenies. Agron, J. 50:459-461.

Kneebone, W. R. 1966. Bermuda grass—worldly, wily, wonderful weed. Econ. Bot, 20:94-97,

Kneebone, W. R. 1973. Breeding seeded varieties of bermudagrass for turfgrass use. Proc. Scotts Turfgrass Res. Conf. 4:149-155.

Le Clerg, E. L. 1966. Significance of experimental design in plant breeding. Plant Breeding. The Iowa S. U. Press, Ames. 243-313.

Lonnquist, J. H., and C. 0. Gardner. 1961. Heterosis in intervarietal crosses in maize and its implication in breeding behavior. Crop Sci. 1:179-183.

Miller, D. G., and 0. J. Hunt. 1966. Water requirement of plants and its importance to grassland agriculture. U. Wyoming Agr. Expt. Sta. Res. J. 3:11 p.

Moll, R. W., W. S. Salhuana, and H. F. Robinson. 1962. Heterosis and genetic diversity in variety crosses in maize. Crop Sci. 2:197-198.

Moutray, J. B., Jr., and R. V. Frakes. 1973. Effects of genetic diversity on heterosis in tall fescue. Crop Sci. 13:1-4.

Murphy, R. P. 1952. Comparison of different types of progeny in evaluating breeding behavior. Proc. 6th • Int. Grassland Congr. 1:320-326.

.Murphy, R. P., and C. C. Lowe. 1966. Methods of breeding perennial forage species of the temperate zones in the United States. Proc. 10th Int. Grassland Congr. 607-613.

Peterson, D. G. 1971. The estimation of heritability and degree of dominance from a diallel cross. Heredity 27:247-264.

Rawal, K. M., and J. R. Harlan. 1971. The evolution of growth habit in Cynodon L. C. Rich. (Gramineae). Trans. 111. S. Acad. Sci. 64:110-118.

Robert, E. S., H. A. Joseph, and L. H. Stolzy. 1972. Wheat response to short-term heat stress and to soil oxygen stress at three stages of growth. Aaron. J. 64:450-452.

Salisbury, F. B., and C. Ross. 1969, Plant physiology. Wadsworth Pub, Co., Inc., Belmont, California.

53

Schroder, U. N. 1956. Photosynthesis and respiration of four pastures grasses as affected by moisture conditions and salinity. Proc. 10th Int. Grassland Congr. 181-187.

Shaffer, H. E., and R. A. Usanis. 1969. General least square analysis of diallel experiments. A computer program. Diallel program for ANOVA. Genetics Dept. Res. Rpt. #1. N.C.S.U. Raleigh. 61 p.

Sleper, D. A., and P. N. Drolsom. 1974. Analysis of several morphological traits and their associa­tions with digestibility in Bromus inermis Leyss. Crop Sci. 14:34-36.

Sleper, D. A., P. N. Drolsom, and N. A. Jorgensen. 1973. Breeding for improved dry matter digestibility in smooth bromegrass (Bromus inermis Leyss.). Crop Sci. 13:556-559.

Smith, J. D., and M. J. Kinman. 1965. The use of parent-offspring regression as an estimate of heritability. Crop Sci. 5:595-596.

Sprague, G. F. 1967. Quantitative genetics in plant improvement. Plant Breeding. The Iowa State Univ. Press, Ames. 315-344.

Sprague, G. F., and L. A. Tatum. 1942. General versus specific combining ability in single crosses in corn, J. Amer. Soc. Agron. 34:923-932.

Strickberger, M. W. 1971. Genetics. The Macmillan Co., New York. 868 p.

Teare, I. D., E. T. Kanemasu, W. L. Powers, and H. S. Jacobs. 1973. Water-use efficiency and its relation to crop canopy area, stomatal regulation, and root distribution. Agron. J. 65:207-211.

Theurer, J. C., and L. J. Elling. 1963. Comparative per­formance of diallel crosses and related second generation synthetics. I. Bacterial wilt resistance. Crop Sci. 3:50-53.

Viets, F. G., Jr. 1962. Fertilizers and the efficient use of water. Adv. in Agron. Edited by A. G. Norman. Acad. Press 14:223-264,

54

Weng, V. K., and F. A. Quinones. 1969. Water requirement in strains of vine-mesquite (Panicuia obtusum H.B.K.). Crop Sci. 9:412-415.

Wilson, D., and J. P. Cooper. 1969. Diallel analysis of photosynthetic rate and related leaf characters among contrasting genotypes of Lolium perenne. Heredity 24:633-649.

Wright, L. N., and A. K. Dobrenz. 1970. Efficiency of water use and seedling growth tolerance of boer lovegrass, Eragrostis curvula (Schrad.) Nees. Crop Sci. 10:1-2.

Yates, F. 1947. Analysis of data from all possible reciprocal crosses between a set of parental lines. Heredity 1:287-301.