/ A METHOD FOR SALVAGING BIRD DAMAGED PEARL MILLET RESEARCH
J.A. Pudelko, D.L. Wright, I.D. Teare / ,
ABSTRACT 4/ .
Pearl Millet [Pennisetum qlaucum (L.)] is a potentially-
productive, high-quality grain crop that is highly susceptible to
bird damage in small plots and in areas around the outside of large
fields. Even when research plots are in the center of a large
field of pearl millet, treatments that change the physiological
maturity (specifically the soft dough stage which red winged
blackbirds seem to prefer and seek out) of certain plots will
result in their destruction. Our objective was to relate pearl
millet grain yields with head length and seed size measurements of
undamaged panicles selected for specific head lengths (15, 12, and
9 inches) with linear regression for predicting grain head yields
per unit area and estimation of grain yield per acre of bird
damaged pearl millet research plots. This research was conducted
on a Norfolk sandy loam located on the North Florida Res. and Educ.
Ctr., Quincy FL with HGM-100 (W.W. Hanna, Tifton, GA) pearl millet
hybrid. Three hundred and sixty pearl millet (HGM-100) panicles
that were not damaged by birds were selected at random for three
different lengths of panicle (15, 12, and 9 inches in length) for
J.A. Pudelko; Agric. Univ. Inst.of Soil Cult. and Plant
Prod.,Mazowiecka 45/46, 60-623 Poznan', Poland:
D.L. Wright, I.D. Teare; North Florida Res. and Educ. Ctr. Quincy,
FL 32351 (Dept of Agronomy, Inst. of Food and Agric. Sci., Univ. of
Florida, FL 32611) Florida Agric Exp. Stn. Res. Rep. No. NF 93-12.
grain yield and linear regression analysis. A simple linear
equation is presented for predicting grain head yields: Y = 6.98 +
191.22 X, where Y = pearl millet head yield and X = head length.
Grain head yields can be predicted if all the head lengths are
measured in a unit area and then converted to yield per acre with
a P < 0.0001.
Pearl millet is a potentially-productive high-quality grain or
silage crop (Burton et al.,1986 and Kumar et al., 1983). It is
grown under low-input management conditions (noncrusting sandy
soils with little fertilizer and limited water; Payne et al., 1990)
and fits the summer growing season presently occupied by crops such
as soybean, peanuts, sorghum, tropical corn, bahiagrass, and
bermudagrass in a year-round multiple cropping system of the
southeastern United States.
Two major problems have been demonstrated by Wright et al.
(1993). First, HGM-100 is a small seeded crop with the need for
uniform depth of planting which can be remedied by improved planter
engineering and careful planter adjustment. Second, is the problem
of the crops susceptibility to extensive bird damage to maturing
panicles (the milk stage is the most susceptible stage) (Wright et
The objectives of this study was to find a parameter
persistent after bird damage for accurately predicting pearl millet
grain yields for salvaging small plot research that had been
successfully conducted up to the milk stage and bird invasion.
MATERIALS AND METHODS
These studies were conducted in 1993 on a Norfolk sandy loam
(fine, loamy siliceous, thermic Typic Kandiudult) located on the
North Florida Research and Education Center, Quincy, Florida. The
soil has a compacted layer located 8 to 14 inches below the
The pearl millet hybrid used in this study was HGM-100,
developed as a grain pearl millet by W.W. Hanna (1991), Tifton,
Georgia. Pearl millet seed was no-till planted in a weed fallow
field with a Brown Ro-Til implement with KMC planters in a
completely randomized block design with six replications on 29 May
1993. Before the millet was planted, the weeds were burned down
with applications of Round-up (7 May) at 2 pt/A and Gramoxone (21
S May) at 3 pt/A. Seed of pearl millet were planted 3/4" deep at 4
lbs/A (322 000 seeds/A planted) with an emergence of approximately
177 000 plants/A (55% emergence). Plots were 24' X 30' with eight
rows 36" apart.
Five hundred lb of 5-10-15 fertilizer/A was applied on 15 May
before planting. Ammonium nitrate was sidedressed to the side of
the row at 120 lb N/A on 16 July. Prowl @ 1 qt/A + Atrazine @ 2
qt/A was used for weed control (Wright et al., 1993). Herbicides
were applied between stage 1 and 2, about 12 days after planting
when millet was between 3 and 5 inches tall.
Pearl millet heads were measured from top to bottom of panicle
as illustrated in Fig. 1. Twenty pearl millet heads were carefully
selected for each of three specific head lengths (9, 12, and 15
inch) and replicated six times with concomitant measurements of
head grain yields and counts of heads per unit area which were used
for regression analysis. Pearl millet heads were harvested on 28
Sept, dried in a greenhouse, and threshed with a clover threshing
machine that required 20 pearl millet heads per sample for the
Little rain occurred throughout the growing season for this
rainfed experiment. A total of 19.0 inches of rainfall was
recieved during the pearl millet growing season from 29 May to 28
Aug, 1993. Rainfall events and amounts are shown in Fig. 2.
RESULTS AND DISCUSSION
Wright et al. (1993) experienced extensive bird damage to
pearl millet in some small plot research in 1992 and used a grain
to silage-without grain ratio from an undamaged pearl millet
herbicide study to estimate grain yield from other bird damaged
research plots. This estimate was better than nothing, but a
better predictor of bird damaged pearl millet yield was needed. In
1993, they tested the null hypothesis that pearl millet grain yield
per head could be predicted from head length measurements.
Six replications of 20 non-bird damaged pearl millet grain
heads of specific lengths (15, 12, or 9 inches) were carefully
threshed and grain yield per head and grain weight per seed were
found to be significantly different for each head length (Table 1).
A simple linear regression equation was developed to predict
head yield from head length: Y = 6.98 + 191.22 X, where Y = pearl
millet grain yield (lb/head) and X = head length (inches) with a
* correlation coefficient (r) = 0.96 and P < 0.0001. A simple linear
equation was also developed to predict head yield from grain size
(seed/lb): Y = -1.782 + 1116.44 X, where Y = pearl millet grain
yield (Ib/head) and X = seed weight (lb)/1000 seeds with a
correlation coefficient (r) = 0.90 and P < 0.0001.
When grain size (seed weight (lb)/1000 seeds) and head length
were used in a multiple regression analysis, the equation developed
was: Y = -0.034 + 0.004 X, + 0.763 X2 where Y = pearl millet
grain yield, Xl = head length (inches) and X2 = grain seed size
(seed weight (lb)/1000 seed) with a correlation coefficient (r) =
0.92 and P < 0.0001.
We agree that the best measure of grain yield is from
undamaged pearl millet heads per unit area, but using predictions
S of head grain yield from head length measurements of a specified
unit area can salvage previously collected time consuming research
that is irretrievably damaged in the soft dough stage by birds
The most useful equation for predicting head grain yield--is the
simple linear regression where head length explains 92 % of the
variation in head grain yield.
Our thanks to E. Brown Senior Laboratory Technician; North
Fla. Res. and Educ. Ctr. Univ. of Fla., Quincy, FL; computer
processing, statistical analysis, and data illustration.
Burton, G.W., A.T.Primo, and R.S. Lowrey. 1986. Effect of
clipping frequency and maturity on the yield and quality of
four pearl millets. Crop Sci. 26:79-81.
Hanna, W.W. 1991. Pearl millet-a potentially new crop for the
U.S. In Abstracts of Technical Papers, No. 18, Southern
Branch ASA, 2-6 Feb 1991, Ft. Worth:, TX.
Kumar, K.A., S.C. Gupta, and D.J. Andrews. 1983. Relationship
between nutritional quality characters and grain yield in
pearl millet. Crop Sci. 23:232-234.
Payne, W.A., C.W. Wendt, and R. J. Lascano. 1990. Root zone water
balance of three low-input millet fields in Niger, West
Africa. Agron. J. 82:813-819.
Pudelko, J.A., I.D. Teare, and D.L. Wright. 1993. Induced stress
on pearl millet vs. head length. Fla. Agric. Exp. Stn. Rep.
No. NF 93-13:1-17.
Wright, D.L., I.D. Teare, F.M. Rhoads, and R.K. Sprenkel.- 1993.
Pearl millet production in a no-tillage system. p. 152-159.
In P. Bollich (Ed.) 1993 Southern Conservation Tillage
Conference for Sustainable Agriculture. June 15-17, Monroe,
LA. SB 93-1.
* Table 1. Pearl millet grain yield, and grain size for six
replications of each head length (9, 12, and 15 inch); Quincy, FL,
Head Grain Grain
length' size2,4 yield3'4
Rept. (inch) (lb/1000 seeds) (lb/head)
1 15 0.0150 0.0467
2 15 0.0139 0.0377
3 15 0.0139 0.0398
4 15 0.0142 0.0452
5 15 -0.0146 0.0382
6 15 0.0135 0.0410
x 0.0142 A 0.0414 A
1 12 0.0122 0.0225
2 12 0.0120 0.0220
3 12 0.0132 0.0221
4 12 0.0130 0.0218
5 12 0.0143 0.0299
6 12 0.0133 0.0314
x 0.0130 B 0.0250 B
1 9 0.0109 0.0130
2 9 0.0086 0.0144
3 9 0.0102 0.0154
4 9 0.0109 0.0111
5 9 0.0095 0.0112
6 9 0.0090 0.0096
X 0.0098 C 0.0124 C
SThree specific head lengths selected at random from non-bird-
damaged pearl millet. Each replication is the mean of 20 pearl
2 Grain size (seed weight (lb)/1000 pearl millet seed) of the three
specific head lengths.
SGrain yield was collected for each 20 heads per replication,
divided by twenty and expressed as yield/head.
4 Mean values in columns followed by the same letter are not
significantly different at the 5% level of significance.
Figure 1. Length of pearl millet head measured as illustrated.
I I %
136 151 166 1
81 196 211
Day of Year
226 241 256
Figure 2. Rainfall during the 1993 pearl millet growing season in
relation to rainfall amounts and dates of events.
.......... ..... ............ ...... .. ... ~..... .. ............... ..--....------
-- -- ------~ -. ..... .. .... ....
.. ..... .. ... ..