J. Agronomy & Crop Science, 156, 110—116 (1986)© 1986 Paul Parey Scientific Publishers, Berlin and HamburgISSN 0044-2151

Contribution from the Plant Sciences Department of the Universityof Arizona/Arizona Expt. Stn. Journal Article No. 4139

University of Arizona, Tucson, AZ 85721

Drought Tolerance Aspects in Pearl Millet

Y. M. IBRAHIM, V. MARCARIAN and A. K. DOBRENZ

Authors' addresses: Dr. Y. M. IBRAHIM, College of Agriculture Studies, P.O. Box 71 Shambat,Khartoum, North, Sudan, Dr. V. MARCARIAN, Praia (ID), Department of State, Washington,D.C. 20520 and Prof. Dr. A. K. DOBRENZ, Plant Sciences Department, University of Arizona,

Tucson, Arizona 85721.

With 2 tables

Received July 27, 1985; accepted September 1, 1985

Abstract

Different indices for selection of lines grown under low and high water levels were examinedfrom both theoretical and practical standpoints. Pearl millet {Pennisetum americanum (L.) K.Schum) was planted under a sprinkler irrigation gradient in 1983 and 1984. Equations weredeveloped for drought tolerance for both yield and forage production. Correlation coefficientswere developed for yield and yield components.

The equations showed that drought will generally result in decreased productivity of bothgrain and forage. Grain yield was found to be positively correlated with seed weight and withthe number of seeds in the main head, and negatively correlated with plant height under lowwater level. Dry matter was positively correlated with tiller number.

Key Words: Pennisetum americanum (L.) K. Schum, Gradient line. Water stress.

Introduction

Pearl millet {Pennisetum americanum (L.) K. Schum) is grown extensively in the aridand semi-arid tropics as a dry land crop. Once established, millet is very drought tolerant,with the ability to survive in dry ecological zones.

Drought resistance is the overall process by which a plant is able to grow, survive, andproduce a harvestable yield when subjected to severe moisture stress. There is a lot ofconfusion in the literature in drought resistance partitioning. PANDEY et al. (1984A)separated drought resistance into two components, drought avoidance and droughttolerance. RosiELLiand HAMBLIN(1981) defined tolerance to stress as the difference in yieldbetween stress and non-stress environments, while the mean productivity is the averageyield in stress and non-stress environments.

Selection for drought tolerance is one of the main problems that has faced researchers.Several techniques have been used to obtain positive results. The most widely used criteriafor high and stable performance are mean yield, regression response and deviation fromregression. FiNLAYand WILKINSON (1963) proposed that regression coefficients approaching

U.S. Copyright Clearance Center Code Statement: 0044-2151/86 /5602-0110$02 .50 /0

Drought Tolerance Aspects in Pearl Millet 111

zero indicated stable performance. LANGERet al. (1979) proposed the use of variety ranges(highest mean yield minus lowest mean yield) as a crude measure of stability across variableenvironments. According to RosiELLiand HAMBLIN(1981) lines with high tolerance to stresswould be expected to have low regression coefficient stability parameters.

Various other indices have been used to differentiate genotypic response to water stress.Yield and yield components are useful indicators for drought resistance (ROSIELLI andHAMBLIN 1981; VIDAL and ARNOUX 1981; STEWART et al. 1983; GARRITY et al. 1984; and

PANDEY et al. 1984A). O'TOOLE and MOYA (1978) reported that two visual scoring tech-niques, one based on leaf firing and the other on leaf drying, were highly correlated withleaf water potential. SHIOWLONG (1981) added that screening by use of leaf firing ratiowould be simple, reliable, and desirable. Drought tolerance index was considered by VIDAL

and ARNOUX (1981) as a good basis for which to define new selection criteria for droughttolerance.

The line source sprinkler procedure described by HANKS et al. (1976, 1978, 1980) hasbeen very useful for applying various levels of irrigation water to experiments. It hasrecently been used successfully to screen a large number of genotypes of millet andsorghum (INTSORMIL 1982, 1983, 1984). Many techniques have been used to indicate waterstress intensity in crops, but few have been evaluated in pearl millet. The objectives of thispaper are to examine different aspects of drought tolerance and to determine possibleparameters that can be used for selection.

Materials and Methods

A sprinkler gradient line was set up at Tucson, Arizona in 1983 and 1984 to evaluate theresponses of pearl millet at high and low water levels. The experimental procedure was describedpreviously (IBRAHIM et al. 1985). Data were collected for yield and yield components andcorrelation coefficients were determined.

Theoretical

Grain yield: Considering the weight of grain at low water level as Yi, weight of grain at highwater level as Y2, and drought tolerance index as Y3, then:

Y3 = Y1/Y2 X 100 [1]If weight of grain per plant is Y4 and total shoot dry weight of plant is Y5, then:

Ye = Y4/Y5 X 100 [2]where Y^ is the harvest index.

Forage: Considering shoot dry weight at low water level as Y7 and shoot dry weight at highwater level as Yg, then:

Y9 = Y7/Y8X1OO [3]where Y9 is the relative yield of dry matter. When the dry weight of dead and burned leaves isYio and the dry weight of all plant leaves is Y^, then:

Y,2 = YIO/YHX1OO [4]

where Y12 is the leaf firing ratio.Water use efficiency = (WUE): If the weight of seed per plant or per hectare is Y,3 and total

amount of applied water (cm) is Y^, then:

Y,5 = Y,3/Y,4 [5]where Y,5 is WUE for grains. When shoot dry weight per plant or per hectare is Y^, then:

Y,7 = Y,6/YH [6]

where Y,7 is the WUE for dry matter.

112 IBRAHIM, MARCARIAN and DOBRENZ

Results

To select for drought tolerance for yield, a high drought tolerance index (Y3) is desirablebecause it indicates that weight of grain at low water level (Yi) will be high (equation 1).Harvest index (Y ) should also be high at low water level. This is important as it willincrease the proportion of the harvestable grain (equation 2).

High relative yield of dry matter (Y9) is the main concern in forage production due tothe increased dry weight at low water level (Y7) compared to the dry weight at high waterlevel (equation 3). On the other hand leaf fire ratio (Y12) should be low under stress. This isimportant for both grain and forage production as it will minimize the number of deadleaves and increase the number of functioning ones (equation 4).

Water use efficiency (WUE) is a useful relative term in drought selection. Under stressconditions the main concern is the production per unit of applied water rather thanabsolute production. WUE for grain (Y15) and for dry matter (Y17) could be found for bothhigh and low water levels. Selection for drought tolerance is for genotypes with high Yjsand Yi7 under low water level (equations 5 and 6) and for genotypes with a decreasingamount of water use under stress.

Yield as measured by the number of seeds per plant was positively correlated (0.001)with number of seeds per head and grain size (Table 1) in 1983. Number of leaves in themain stem was also correlated with yield at the 0.01 level. There was a negative correlation(0.05) between yield and length of head and height of the main stem (Table 1). Thecorrelation of other parameters was not significant. At high water level yield was positivelycorrelated with length of head, number of seeds in the main head and number of leaves inthe main stem (Table 1).

Data of 1984 showed a positive correlation (0.001) of yield to grain size (Table 2).Number of seeds per head and number of the leaves in the main stem were positivelycorrelated with yield at 0.05 level of significance. There was a negative correlation (0.01)between yield and length of head. Other parameters were not correlated with yield(Table 2). At high water level yield was positively correlated with head exsertion andnumber of seeds in the main head (Table 2).

Dry matter was positively correlated with number of tillers at the 0.01 level and withnumber of productive tillers at the 0.05 level in the 1983 season (Table 1). Otherparameters were not significantly correlated. The 1984 season showed a positive correla-tion (0.001) of dry matter with number of seeds per head and with number of tillers at the0.01 level of significance (Table 2). Other parameters were not significantly correlated withdry matter. Head exsertion was positively correlated (0.001) with length of head andnumber of productive tillers in both seasons (Tables 1 and 2).

Discussion

The practicality of the equations defined in this paper depends on the responses of thecrop to water stress and particularly when used under the sprinkler gradient. Lines withhigh tolerance to stress would be expected to have a high drought tolerance index, highharvest index, high relative yield, and high efficiency in water use. On the other hand,firing ratio should be low under stress to reduce the number of non-functioning leaves,however, as stated by PANDEY et al. (1984B) increased leaf senescence will result in lesswater use.

Yield was positively correlated with number of seeds per head, grain size, and numberof leaves on the main stem. In contrast, VIDAL and ARNOUX (1981) found a negativecorrelation between seed weight and seed number in soybeans. SINGH and KANEMASU (1980)

Drought Tolerance Aspects in Pearl Millet 113

R

O* «

R

R

2

R

RO

8

"S

8RO

(3

<uoo

T:<UCO

v_ _ Ti y j ^ M

J— PC

c4

:

co

gu

o P

I I I I I I I I I I I I I I I I I O O

C/3 0/5 t/5

O 1-1 O OT-< c n • ^ CM

o o o o

I I I I I I I tON

d

NO O O - ^CM CM

o o o

I I I I I I I I I I I

•V

ON ooin (Nd d

ONn

OOON 000 C O O C N l ' ^

d d d d d d d d

V.- t-7 I-T- -.V iV I-T- -.V

00 00 00 I—I fv. cO CM NOm t N O C M i n r O 000

d d d d d d d d

ON CO 0 0 CO ON 0 0 •<*•Tl- T-1 CM ^ - CO CN O

O O O O O O <^ C3 Q O O

'•• * CO ^ CO , CO CO *•J.- VSr h^. •A- 1-7 -;- 1-7- 1-7 *-,'.- i; ^ •>.; it 4^ * /-< i; -fr• ^ OOON COON S O O N S O O O C O O OTt- r O O r O O c O O r v | O r M ' ^

d d d d d d d d d d d d d d

CO T-H

CO \D

C O C O C O C O C O C O C O C O C O COOO

< ^h ^3 * rv. \O CO ON CM ON

o o o o Q o d d d d d d d

Vf it A* it ICH 4 H

00 ro r\ CM m m

o d d d d d d do

* ^ iiJ. .•.. .«..

coco C/5CO coco coco coco cocoCOOtx CNj PO sD O <N

CO CO ro <N 0 00 0 0 0 0 0

bO bX)

CD

bX)

0 0 0 0 0 0cp o o cp

lS-2 15-2

H

zPCCOz CO COo

oo

H-lz

bis

oo

6<u

I•-•-I

o

CO

oo

bx>CO

odrt

CO

(U

mo

u

'ICO

c

CO

o

ooz

114 IBRAHIM, MARCARIAN and DOBRENZ

be

RO

R

RO

"•2

IoO

-Si

C <u

00

So -S <U

to

L

NO

co

2CL,

. J^ V)

o <u3

I I I I I I I I 1 1 I I I I I I I I o o

>*# .*«

o o o tOO

I I§ i § SO IN O ^

o o S*

in o(N O

C/5 C/D CO C/5

^ ^ A ^fN4 o m ON1—I ( N -< o

o o o o

I I I I I I I

CO ^ CO _, CO ;7; S <^

O N v o T - H \ o o o o i n r t - i nT-H f* o f* T-H m (N ' ^ o

o o o T ° T o o

o

CO

<N O

CO1

in O '—' - ^ O T-H T-lT-H T-H O ( N O • r - CSI

o o o 00 00 00

C/5CO

o r^ T—• i n CNI T-H T-H T-H T-H

OO OC) OO d p p o o p p0 po op pp1 1 I I T T

"• CO C/D CO CO C/5 * CO CO CO CO CO CO ^ CO"" *^r *~^ ' 7" '"T' ^^ ^~7 ^~7 ^~^ ^7 7 7 "^

* ' • *1» ^F . ^r . ^ ^ ^r ^ ^r , *^ ^ j ^ j ^ ^ ^ j ^^^» ^ * ^L^J TTf ^C^rf* • ffi ^ ™ ^ ^ ™ ^ " I ^ ^ " ^ r 1 ' * F^^ n^^ ^ ^ ^ ^ ^ ^ ^ ^ r • p ' • * • '

Ovrf- T-HOO sOfV, f O O fNJr'^ T - H O inON CSmr O < N T - H O T-HT-H T-HpV, T - H O < N f S T — I T - H f ^ ' ^

d> o Q d> <^ <^ o d PP PP So PP

CO CO CO "• CO * * CO CO C/1 CO CO V5

* -,V- -k A ^ A^ •>'; fi^ i: -k Vc (t-i * |t_i ^ ^ < H ^

dd dd dd dd <S d do do oo00 00

"wa ^ "w)

Q

COCO

QC/5 CO C/5 1:^ ^

Z Z O 1 '-'-'

4

Vi

c3

CUD

<u

II

oo

artU

"Sto

CO

*

qd

4

U

qd

c

CO

atO

eg"S

COz

Drought Tolerance Aspects in Pearl Millet 115

considered head number as a good index of drought tolerance. Increased head number willresult in smaller seeds and it is thought that by directing the photosynthate to the mainhead only, seed size and density can be improved.

The height of the main stem was negatively correlated with yield in 1983 (Table 1). Ingeneral tall cultivars are low in yield and better in forage production. This agreed with thefindings of OWONUBI and KANEMASU (1982) that tall sorghum plants had high evapotranspi-ration and high water use efficiency for dry matter while dwarf plants were water efficientfor grain production.

In conclusion, the indices defined in this paper could be very useful in defining differentselection criteria for drought tolerance in millet. These screening processes would be usedfor other crops if supported by different visual techniques and good understanding ofdrought tolerance in such crops. The use of a sprinkler gradient line is a valuable techniquein such studies.

Zusammenfassung

Aspekte der Trockenheitstoleranz bei Kolbenhirse

Verschiedene Indices zur Selektion von Linien, die unter den Bedingungen einesniedrigen und hohen Wasserangebotes angezogen wurden, wurden unter Berlicksichtigungtheoretischer und praktischer Uberlegungen getestet. Kolbenhirse {Pennisetum ameri-canum (L.) K. Schum) wurde unter Regnerbewasserung in den Jahren 1983 und 1984angebaut. Es wurden Gleichungen fiir die Trockenheitstoleranz im Hinblick auf Ertragund Futterproduktion entwickelt. Ferner wurden Korrelationskoeffizienten fur Ertragund Ertragskomponenten bestimmt.

Die Gleichungen zeigten, dafi Trockenheit grundsatzlich zu einer Abnahme der Pro-duktivitat sowohl fiir den Korn- als auch den Futterertrag fiihrten. Der Kornertrag warpositiv korreliert mit dem Samengewicht und mit der Anzahl Samen in dem Hauptfrucht-stand; eine negative Korrelation mit der Pflanzenhohe war fiir niedriges Wasserangebotnachzuweisen. Die Trockenmasseproduktion war positiv korreliert mit der Anzahl Be-stockungstriebe.

References

FiNLAY, K. W., and G. N. WILKINSON, 1963: The analysis of adaptation in a plant breedingprogram. Aust. J. Agric. Res. 14, 742—754.

GARRITY, D . P., C . Y. SULLIVAN, and D. G. WATTS, 1984: Changes in grain sorghum stomataland photosynthetic response to moisture stress across growth stages. Crop Sci. 24,441—446.

HANKS, R. J., J. KELLER, V. P. RASMUSSEN, and G. D. WILSON, 1976: Line source sprinkler forcontinuous variable irrigation crop production studies. Soil Sci. Soc. Am. J. 40,426—429.

, G. L. ASHCROFT, V. P. RASMUSSEN, and G. D. WILSON, 1978: Corn production asinfluenced by irrigation and salinity — Utah studies. Irrig. Sci. 1, 47—59.

, D. V. SiSSON, R. L. HURST, and K. G. HUBBARD, 1980: Statistical analysis of results fromirrigation experiments using the line-source sprinkler system. Soil Sci. Soc. Am. J. 44,886—888.

IBRAHIM, Y. M. , V. MARCARIAN, and A. K. DOBRENZ, 1985: Evaluation of drought tolerance inpearl millet {Pennisetum americanum (L.) K. Schum) under sprinkler irrigation gradient.Field Crop Res. (in print).

INTSORMIL, US AID Title XII CRSP, 1982: Annual report, Yuma, Arizona. The Univ. ofArizona, Tucson, AZ 85721.

116 IBRAHIM, MARCARIAN and DOBRENZ, Drought Tolerance Aspects in Pearl Millet

, 1983: Annual report, Yuma, Arizona, The Univ. of Arizona, Plant Science Dept.,Tucson, AZ 85721.

, 1984. Annual report, Yuma, Arizona, The Univ. of Arizona, Plant Science Dept.,Tucson, AZ 85721.

LANGER, I., K. J. FREY, and T. BAILEY, 1979: Associations among productivity, productionresponse, and stability indexes in oat varieties. Euphytica 28, 17—24.

O'ToOLE, J. C , and T. B. MOYA, 1978: Genotypic variation in maintenance of leaf waterpotential in rice. Crop Sci. 18, 873—876.

OwONUBi, J. J., and E. T. KANEMASU, 1982: Water use efficiency of three height isolines ofsorghum. Can. J. Plant Sci. 62, 35—46.

PANDEY, R. K., W . A. T. HERRERA, and J. W. PENDLETON, 1984A: Drought response of grainlegumes under irrigation gradient: I. Yield and yield components. Agron. J. 76,549—553.

, , A. N. VILLEGAS, and J. W. PENDLETON, 1984B: Drought response of grainlegumes under irrigation gradient: III. Plant growth. Agron. J. 76, 557—560.

ROSIELLI, A. A., and J. HAMBLIN, 1981: Theoretical aspects of selection for yield in stress andnon-stress environments. Crop Sci. 21, 943—946.

SHIOW-LONG, T., 1981: Comparison on the physiological character of drought resistance amongsorghum varieties. J. Agric. Assoc. China. 116, 15—28.

SINGH, P., and E. T. KANEMASU, 1980: Soil water, plant water and temperature relation of millet{Pennisetum americanum L. Leeke) and their relationship to crop yields. Third annualreport. Improvement of pearl millet. Manhattan, KS 66506, pp. 75—-93.

STEWART, B . A., J. T. MUSICK, and D. A. DUSEK, 1983: Yield and water use efficiency of grainsorghum in a limited irrigation-dry land fanning. Agron. J. 75, 629—634.

VIDAL, A., and M. ARNOUX, 1981: Drought tolerance process in soybeans. Biologia Plantarum.23(6), 434—441.