t, n. .. - ..
‘ Hm; sqnqfibhu
. ‘

in.
:43:
A: m

I.<I.{.

 

:w”

u ‘ 1 .
I.
5...... .99.“. :l
3. t {3.1...
. .3 ‘
J1: u.

r,
.
9
‘
v
.
-
V
..
a

“uuauuun

5...): 2.1
71rd,... 3.13 .339...)

 

 

'7‘ iEStS

MIC HIG IGAN STATE UNIVER

IIUIUHIHHIIII HI I ”MINI”HIJHHIIHHM

31293 01018 5803

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

AGRONOMIC METHODS FOR GROWING BLACK LOCUST

(ROBINIA PSEUDOACACIA L.)
AS A PERENNIAL FORAGE CROP
presented by

Robert P. Barrett

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in ForEStry

(Md W

Major professoU

 

 

[hue November 17. 1993

 

MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove We checked from your record.
TO AVOID FINES return on or More date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mmF—W

 

 

 

MSU is An Affirmative Action/Equal Opportunity lnetitmon

”3-9.1

AGRONOMIC METHODS FOR GROWING BLACK LOCUST (ROBINIA
PSEUDOACACIA L.) AS A PERENNIAL FORAGE CROP

Robert P. Barrett

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

Doctor of Philosophy

Department of Forestry

1993

ABSTRACT

AGRONOMIC METHODS FOR GROWING BLACK LOCUST (ROBINIA
PSEUDOACACIA L.) AS A PERENNIAL FORAGE CROP

By

Robert P. Barrett

The nitrogen-fixing black locust tree (139m pseudoacacia L.) is extremely high yielding,
rapid growing, and adaptable, widely naturalized, and has potential for use in agroforestry systems,
including forage production, short rotation biomass, fuelwood, beekeeping, soil improvement, and
alley cropping. Wildlife and livestock will consume it, although it has tannins which reduce protein
digestibility, and stipular spines. Despite these faults, genetic diversity is great and the
reproductive age is low, allowing rapid selective brwding. The concept of ideotype breeding is
examined, and black locust is evaluated in terms of an ideal model forage tree.

Field trial treatments involving five spacings, two harvest heights, two years of first
harvest, and eight harvest dates during the growing season were applied to plots of black locust
planted in 1988, during the years 1988 to 1991. Effects on yields per plot and per plant, the dry
weight percentage, and survival are reported.

Harvesting two times a year was found unsatisfactory. Harvest dates early in the season
were unfavorable because of low regrowth productivity, while those in July or August were
rejected because of low survival. A single harvest date as late in the growing season as possible

was advised.

Harvesting the first season reduced future productivity and survival, often severely. For
every aspect of growth, yield, and survival, the harvest height of 30 cm was superior compared to
5 cm. Continued improvement with establishment periods longer than one season and heights
above 30 cm was projected.

The narrowest spacing tested (10 cm) gave superior yields the first season only. The
widest (50 cm) gave yields which were lower the first and second seasons, equal the third, and
superior the fourth. Four year total yields did not vary with density. Survival was highest at 50

cm.

 

 

ACKNOWLEDGEMENTS

To my original Committee: Dr. James W. Hanover, Dr. Michael A. Gold, Dr. Donald 1.
Dickmann, all of the Department of Forestry, and Dr. Oran Hesterman of the Department of Crop
and Soil Sciences, who assisted in the design of the experiments and conducted my comprehensive
examinations. Especially to my Advisor, Dr. Hanover, for arranging funding, a place to conduct
the experiments, additional labor for planting and harvesting, and for inspiring me to investigate
Rpm pseudoacacia in the first place. His death in 1992 was a great loss to all who knew him,
and to the science of Forestry.

To the final members of my Committee: Dr. Michael A. Gold, Dr. Donald 1. Dickmann,
and Dr. Daniel Keathley, all of the Department of Forestry, and Dr. Richard Harwood of the
Department of Crop and Soil Sciences, who conducted the final examination on this dissertation.

To the Forestry Department staff members I have worked with over the years who have
shown me the ropes, shown me where to find things at the Tree Research Center, and given me
good advice: Tesfai Mebrahtu, Ray Miller, Paul Bloese, Roy Prentice, Randy Klevikas, Phu
Nguyen, and Sue Plesko.

To my family, who have always encouraged scholarship. I have wanted to be a scientist of
some sort for as long as I can remember. My grandfather, Theodore H. Elferdink, Sr., introduced
me to the wonders of nature and taught me how to identify common trees when I was two. He had
the foresight to buy advanced science books for me before I could even read, confident that I would
enjoy them some day. My grandmother, Pearl Barrett, gave me a set of model dinosaurs when I
was a child, long before dinosaurs became popular. My parents, Pat and Mary Barrett, deserve
much of the credit for what I have accomplished, and deserve thanks for more than I can possibly
list here. My wife, Linda, has taught me everything I know about word processing, and a great
deal more that I keep forgetting. Her patience through nine drafts of this dissertation has been

remarkable.

iii

TABLE OF CONTENTS

List of tables ...................................................................................................................... viii
List of figures ....................................................................................................................... xi
INTRODUCTION TO BLACK LOCUST ............................................................................ 1
BLACK LOCUST IN AGROFORESTRY ................................................... l
INTERDISCIPLINARY NATURE OF THE RESEARCH ........................... 2
FEATURES OF BLACK LOCUST ............................................................. 3
Adaptations and range ...................................................................... 3
Insect Pests ...................................................................................... 3
Genetics and cultivars ...................................................................... 4
Flowering ......................................................................................... 5
Use as bee forage ............................................................................. 5
Uses of wood ................................................................................... 5
Uses for soil conservation ................................................................. 6
Uses for animal feed ......................................................................... 7
EXPERIMENTAL OBJECTIVES ............................................................... 8
ORGANIZATION OF THIS DISSERTATION ........................................... 9
REFERENCES CITED .............................................................................. 12
EFFECTS OF HARVEST DATE ON SURVIVAL AND FORAGE YIELD OF BLACK
LOCUST ................................................................................................................ 16
ABSTRACT .............................................................................................. 16
INTRODUCTION ..................................................................................... l6
PREVIOUS RESEARCH ........................................................................... 17
Black locust ................................................................................... l7
Leucaena ....................................................................................... 17
Other tropical species ..................................................................... 18
Difficulty of applying tropical results in Michigan .......................... 19
EXPERIMENTAL OBJECTIVES ............................................................. 19
MATERIALS AND METHODS ................................................................ 20
General information for all trials ..................................................... 20
Row plots at eight harvest dates ...................................................... 21
Dry weight/fresh weight comparisons ............................................. 24
RESULTS .................................................................................................. 25
Height growth ................................................................................ 25
Dry weight yields ........................................................................... 25

iv

Dry weight productivity per day ..................................................... 28

Dry weight productivity per plant and per plant per day .................. 32
Changes in dry weight percentage ................................................... 35
Survival afier harvesting ................................................................ 40

Canopy coverage ............................................................................ 40
DISCUSSION ............................................................................................ 43
Growth, yield, and survival ............................................................ 43
Defoliation and refoliation .............................................................. 48
Applications for forage management ............................................... 53
REFERENCES .......................................................................................... 55

EFFECTS OF HARVEST HEIGHT ON SURVIVAL AND F ORAGE YIELD OF

BLACK LOCUST .................................................................................................. 59
ABSTRACT .............................................................................................. 59
INTRODUCTION ..................................................................................... 60
Literature review ............................................................................ 60
Objectives ...................................................................................... 62
MATERIALS AND METHODS ................................................................ 64
General information for all trials ..................................................... 64
Harvest height trials ....................................................................... 65
RESULTS AND DISCUSSION ................................................................. 66
Relation of optimal density to harvest height ................................... 66

Dry weight yields ........................................................................... 66

Dry weight percentage .................................................................... 73
Survival after harvesting ................................................................ 75
CONCLUSIONS ........................................................................................ 78
REFERENCES CITED .............................................................................. 79

EFFECTS OF SPACING ON SURVIVAL AND FORAGE YIELD OF BLACK

LOCUST ................................................................................................................ 82
ABSTRACT .............................................................................................. 82
INTRODUCTION ..................................................................................... 82
Literature review ............................................................................ 83
Objectives ...................................................................................... 85
MATERIALS AND METHODS ................................................................ 85
General information for all trials ..................................................... 85
Directly sown trial .......................................................................... 87
Transplanted trial ........................................................................... 87
RESULTS AND DISCUSSION ................................................................. 88
Dry weight yields per plot .............................................................. 88

Dry weight yields per plant ............................................................. 93
Composition of harvested biomass .................................................. 95
Survival after harvesting ................................................................ 96
Optimal density ............................................................................ 102
Defoliation severity and solar interception ..................................... 102
Spacing recommendations for stand establishment ........................ 103
REFERENCES CITED ............................................................................ 104

EFFECTS OF YEAR OF FIRST HARVEST ON SURVIVAL AND FORAGE YIELD

OF BLACK LOCUST .......................................................................................... 107
ABSTRACT ............................................................................................ 107
INTRODUCTION ................................................................................... 108

Literature review .......................................................................... 108
Objectives .................................................................................... 109
MATERIALS AND METHODS .............................................................. 110
General information for all trials ................................................... 110
Grid plots at two harvest heights ................................................... 1 12
Grid plots at five spacings ............................................................ 112
Row plots with five seed sources .................................................. 113
RESULTS AND DISCUSSION ............................................................... 114
Dry weight yields per plot ............................................................ 114
Dry weight yields per plant ........................................................... 1 19
Changes in dry weight / fresh weight ratio .................................... 123
Number of branches harvested in row plots ................................... 123
Survival afier harvesting .............................................................. 125
Interactions of year treatments with other factors .......................... 129
Delaying the initial harvest beyond the second growing season ...... 129
REFERENCES CITED ............................................................................ 131
AN IDEOTYPE FOR TREE SPECIES GROWN AS PERENNIAL FORAGE CROPS,

WITH SPECIAL REFERENCE TO BLACK LOCUST ....................................... 133
INTRODUCTION ................................................................................... 133
DEFINITION OF IDEOTYPE ................................................................. 134
THE IDEOTYPES OF CULTIVATED PLANTS .................................... 134
GENERAL OBSTACLES TO IDEOTYPE BREEDING .......................... 135
IDEOTYPES OF CULTIVATED TREES ................................................ 136
PERENNIAL F ORAGE CROP TREE IDEOTYPE ................................. 138
REFERENCES CITED ............................................................................ 142

CONCLUSIONS ............................................................................................................... 144
CONCLUSIONS FOR EACH TREATMENT AND ASPECT ................. 144
Ideotype breeding goals ................................................................ 144
Harvest date ................................................................................. 145
Harvest height .............................................................................. 145
Spacing ........................................................................................ 145

Year of first harvest ..................................................................... 145

Long term trends .......................................................................... 146

ITEMS FOR FURTHER STUDY ............................................................ 146
Harvest date ................................................................................. 146
Harvest height .............................................................................. 147
Spacing ........................................................................................ 147

Year of first harvest ..................................................................... 148
Spacing by harvest height interaction ............................................ 148
Spacing by harvest date interaction ............................................... 149
Harvest date by harvest height interaction ..................................... 149

Year of first harvest by seed source interaction ............................. 149

Long term trends .......................................................................... 149

vi

Height growth .............................................................................. 150

Yields .......................................................................................... 150
Yields per plant ............................................................................ 151
Yields per day .............................................................................. 151
Dry / fresh weight proportion ....................................................... 151
Survival ....................................................................................... 152
Canopy coverage .......................................................................... 152
Juvenile stage ............................................................................... 153
RECOMMENDATIONS ......................................................................... 154
Crop establishment ....................................................................... 154
Experimental logistics .................................................................. 154

vii

LIST OF TABLES

Table 2.1. Effect of harvest data 1989 - 1990 on May 31, 1991 canopy height. .................... 26

Table 2.2. Dry weight yields (g/mz) for those plots harvested the same date in 1989 and
1990, and comparisons of yields, numbers of survivors, and weights per plant
expressed as ratios. ................................................................................................. 27

Table 2.3. Effect of first harvest date in 1989 on yields (g/mz) at later harvests in 1989,
and cumulative yields. ............................................................................................. 29

Table 2.4. Daily dry weight productivity rates (g/mzlday) in 1989 and 1990 in relation to
the number of days from bud break, for plots harvested on the same date both
years. ...................................................................................................................... 30

Table 2.5. Daily dry weight productivity rates (g/mzlday) in 1989 and 1990 in relation to
days from the theoretical zero date on May 29, 1989 or July 7, 1990 ........................ 31

Table 2.6. Daily dry weight productivity rates (g/mzlday) after the first harvest on June 15,
1989, and July 12, 1990. ......................................................................................... 33

Table 2.7. Efl‘ect of first harvest date on dry weight yields (g/mz) and daily dry weight
productivity rates (g/mz/day) for second harvests in 1989 and 1990. ........................ 34

Table 2.8. Dry weight/fresh weight percentages for entire plots in 1987, 1989 and 1990,
and for small samples in 1991, according to date and days from bud break ............... 36

Table 2.9. Dry weight / fresh weight percentages for sets of individual plants in 1989 and
1990, compared to days from bud break. ................................................................. 37

Table 2.10. Dry weight / fresh weight percentages of leafy and woody portions of material
harvested in September, and the percentage of the total sample represented by the
dry weight of the woody portion. ............................................................................. 39

Table 2.11. Numbers of plants harvested per m2 and percentage of original plants

surviving for those plots harvested the same date 1989-1990, for first and second
harvests. ................................................................................................................. 41

viii

 

Table 2.12. Effect of harvest date on average canopy coverage percentages from visual
estimates on September 28, 1989, May 30 and October 3, 1990, and May 31,
1991, related to number of days from first harvest to first frost in 1989 or to
recording in 199 ...................................................................................................... 42

Table 3.1. Dry weight yields (g/mz) for square grid plots at five planting densities
harvested at 5 cm and 20 cm in late August 1988, 65 days from sowing. .................. 67

Table 3.2. Dry weight yields (g/mz) for square gn'd plots harvested at 5 cm and 25 cm in
early July and at 5 cm and 30 cm in late August 1989, and at 5 cm and 30 cm in
late June and late August 1990. ............................................................................... 69

Table 3.3. Dry weight yields (g/plant) from square grid plots at five planting densities
harvested at 5 cm and 30 cm in late August 1989, late June 1990, and late August
1990. ...................................................................................................................... 71

Table 3.4. Annual total dry weight yields (g/plant) from square grid plots at five planting
densities harvested at 5 cm and 30 cm in 1989 and 1990. ......................................... 72

Table 3.5. Dry weight percentage of fresh weight for square grid plots harvested at 5 cm
and 20-30 cm in 1988-1990 ..................................................................................... 74

Table 3.6. Average number of plants harvested from square grid plots at five planting
densities harvested at 5 cm and 20 cm in late August 1988, 65 days from sowing,
and at 5 cm and 25 cm in early July 1989. ............................................................... 76

Table 3.7. Average number of plants harvested from square grid plots at 5 cm and 25 or
30 cm starting a year after sowing (three replications) .............................................. 77

Table 4.1. Dry weight yields (g/m’) for square grid plots at five planting densities
harvested at 25 cm in early July and at 30 cm in late August 1989, late June and
late August 1990. .................................................................................................... 89

Table 4.2. Annual and cumulative dry weight yields (g/mz) for square grid plots at five
planting densities harvested in 1989, 1990, and 1991. .............................................. 91

Table 4.3. Single harvest and cumulative dry weight yields (g/mz) from grid plots of 4 and
16/m2 ..................................................................................................................... 92

Table 4.4. Single harvest and annual dry weight yields per plant (g) from grid plots at five
planting densities harvested at 25 cm in early July and at 30 cm in late August,
1989, late June, and late August, 1990. .................................................................. 94

Table 4.5. Number harvested (plants/m2) from grid plots at five planting densities in
subsequent harvests relative to the number harvested in July, 1989 ........................... 97

Table 4.6. Number harvested (plants/m2) each time 1988 - 1991 from grid plots at five
planting densities ..................................................................................................... 98

Table 4.7. Efl‘ect of 1990 plot population (plants/m2) on 1991 plot population, and ratio
of survival from 1990 to 1991, for square grid plots originally planted at five
spacings, and transplanted plots originally planted at three spacings. ........................ 99

Table 4.8. Numbers of survivors (plants/m2) and ratio of survival from the previous date
for transplanted seedlings at two population densities. ............................................ 101

Table 5.1. Dry weight yields per plot (g/mz) from grid plots with first harvests in the
seedling year (group 1) or a year after sowing (group 2) ......................................... 115

Table 5.2. Single harvest and cumulative dry weight yields (g/m2) from grid plots of 4
and l6/m2 first harvested the seedling year (group 1, 1988 - 1991) or a year after

sowing (group 2, 1989 - 1991). ............................................................................. 116
Table 5.3. Dry weight yields per plot (g/mz) from row plots with first harvests in the

seedling year (group 3) or a year after sowing (group 4) ......................................... 118
Table 5.4. Dry weight yields per plant (g) from grid plots with first harvests in the

seedling year (group 1) or a year after sowing (group 2) ......................................... 120
Table 5.5. Dry weight yields per plant (g) from row plots with first harvests in the

seedling year (group 3) or a year after sowing (group 4) ......................................... 122
Table 5.6. Numbers of upright branches cut per plant and per plot from row plots with

first harvests in the seedling year (group 3) or a year after sowing (group 4) ........... 124
Table 5.7. Number of plants harvested from grid plots with first harvests in the seedling

year (group 1) or a year after sowing (group 2) ...................................................... 127
Table 5.8. Number harvested fi'om row plots with first harvests in the seedling year

(group 3) or a year after sowing (group 4). ............................................................ 128
Table 6.1. Root, stem, and foliage features of a perennial forage crop tree ideotype ............ 139

Table 6.2. Propagation, reproduction, and stand establishment features of a perennial
forage crop tree ideotype. ...................................................................................... 140

Table 6.3. Environmental adaptation features of a perennial forage crop tree ideotype ........ 141

 

LIST OF FIGURES

Figure 2.1. Dry weight percentage of fresh weight for samples harvested in four seasons,
related to days from bud break. ............................................................................... 38

xi

INTRODUCTION TO BLACK LOCUST

BLACK LOCUST IN AGROFORESTRY

Agroforestry is the science of including trees in farming systems to obtain more favorable
results than are possible with trees, crops, or livestock alone. The main techniques include alley
cropping, with annual crops grown between hedges of trees, taungya, with crops grown in forest
plantations while the trees are young, silvopasture, with timber trees grown on grazing land,
tropical gardens, with a wide variety of shade tolerant food, medicinal, and ornamental crops
grown under large fruit or timber trees, and fodder trees, which are cultivated and periodically
trimmed to feed livestock. Scientific research on alley cropping has concentrated on tropical
nitrogen fixing leguminous trees, primarily leucaena (Qucaena leucocephala (Lam) de Wit), but

other species in the genera Acacia, Calliandra, Ervthrin_a_, Gliricidia, Leucaena, and Sesbania have

 

 

also been investigated. Because of their high-protein foliage, these same genera have been
researched as fodder trees. These and other multi-purpose trees, which are useful for fuelwood,
soil enrichment or conservation, fodder, shading coffee or cacao plantations, and sometimes for
food, have been promoted throughout the tropics for use in agroforestry systems that are intended
to be more sustainable and are sometimes more profitable than growing annual crops or timber
trees alone.

Agroforestry systems have received less attention in the temperate zone than in the tropics
because there is little room for trees in mechanized cash crop monocultures. Industrialized
countries also have readily available chemical fertilizers and pesticides, no major crops that require
shade, and few subsistence farmers. In addition to these economic factors, there are few nitrogen

fixing plants that grow to tree size that can survive freezing temperatures. The black locust

2
(Mug pseudoacacia L.) is a frost tolerant, productive, and adaptable nitrogen fixing tree. Thus,
it is the most likely species to play the multi-purpose role in temperate climates that leucaena has in
the tropics. Both leucaena and black locust have rapid growth from seed or coppice, rhizobial
nitrogen fixation, dense wood, high protein forage, and a reputation for adaptability (Barrett and
Hanover 1991).

INTERDISCIPLINARY NATURE OF THE RESEARCH

It should be recognized that the treatments applied to black locust in these experiments
were more typical of those employed in the science of agronomy than in forestry. The product
under consideration was not wood but livestock feed, and it was harvested annually. The only
element of conventional forestry involved was the species, which grows as a tree in nature but can
also be cultivated as field crop. The emerging scientific discipline of agroforestry involves growing
trees as part of a diversified farming system, in conjunction with livestock and field crops. The
cultivation of woody species as crops to produce commodities other than wood could be considered
a branch of agroforestry, or of agronomy. It could also be considered an extension of the forestry
technique known as short rotation intensive culture, which has been developed to produce high
yields of woody biomass for fuel and fiber. Then again, perennial leaf crops such as tea, qat, betel,
coca, and yerba mate are claimed by the discipline of horticulture. The easiest solution would be
to designate black locust as a "new crop" (Barrett e_t a_1. 1990) and leave it at that. Because my
chosen area of study lies at the intersection of several established disciplines, I propose the name

"dendro-agronomy" to specify the practice of growing woody species under agronomic conditions.

FEATURES OF BLACK LOCUST

Adaptations and range

Black locust is a multi-purpose leguminous tree noted for rapid juvenile growth. It is
adaptable to a wide range of growing conditions, and has a tendency to escape from cultivation.
Although it requires full sunlight and grows best on limestone soils, deep taproots and symbiotic
nitrogen fixation enable it to thrive on difficult sites, including acidic mine spoil banks and sand
dunes (Fowells 1965; Miller et al. 1987; Barrett and Hanover 1991). Black locust tolerates more
acidic soil than most other North American deciduous tree species, and is recommended for
planting at pH levels as low as 4.0 (Sutton and Dick 1987).

Native to southeastern North America, black locust has become naturalized in temperate
regions around the world. In the early 1600's it was imported to France and England, and rapidly
spread through Europe (Peabody 1982; Barrett and Hanover 1991). By 1880 it had become
naturalized in New Zealand, and it has been grown as an ornamental in the Australian states of
Tasmania and Victoria, and used for soil conservation and ornamental purposes in both Israel and
Zimbabwe. It was introduced to Kenya in 1913, and in 1962 specimens were reported as
growing well in the Nairobi Arboretum (Streets 1962).

Black locust appears to reach the limit of its productive range where the frost-free
growing season is under 100 days (Miller et a1. 1987). In India black locust grows best from
1500 - 2000 m altitude with 1150 - 1750 mm annual rainfall (Gupta 1986). The National
Research Council (1983) asserts that it has been successfully grown with 300 - 400 mm rainfall.
It avoids drought stress by deep taproots, leaflet folding, and leaf abscission. It will root sprout

or coppice if cut down or damaged by fire (Gustafson 1935; Fowells 1965).

Insect Pests
In North America, the major pest attacking black locust is the locust stem borer beetle

Megacvllene robiniae Forster. The larvae tunnel through the heartwood. ofien causing the trunks

4
to break in high winds and generally making the lumber unattractive (Hoffard and Anderson 1982;
Galford 1984; Barrett and Hanover 1991). In Hungary, where the beetle is not present, black
locust is a major timber tree (Keresztesi 1988). The locust twig borer Ecdflolopha insiticiana
Zeller, a moth larva, forms galls in the new growth and can cause undesirable crooked trunks. In
addition to the locust stem and twig borers, black locust is sometimes damaged by cowpea aphids
Apfls craccivora Koch, two leaf beetles, the leaf mining larvae of the moth Parectopa Lgbinella,
and the leaf-miner beetle Mtg dorsalis Thunberg (Meng 1934; Hoffard and Anderson

 

1982; Genys and Harman 1990; Hofi‘ard 1992).

Genetics and cultivars

Hopp (1941) classified black locust trees into three major groups: spreading, palmate, and
pinnate growth forms, with the crooked spreading form by far the most common. Bongarten
(1992) reported a system for uniquely describing 40 black locust clones growing on the same site,
by classifying four characteristics of juvenile growth: branch sinuosity, branch angle, frequency of
lateral branching, and degree of apical dominance. The shipmast locust, Robinia pseudoaeaeia
var. rectissima (L.) Raber, is a well-known example of the pinnate form. Huang e_t a_l. (1975)
called it an ecological variant, not a genetic variety, based on soluble protein gel-electrophoresis
data. Several researchers have noted that the characteristics of shipmast locust change when
clones are planted in other regions (Bongarten 1992). Many special-purpose black locust cultivars
have been bred in Hungary from the diverse naturalized gene pool (Keresztesi 1977, 1983, 1988).
Rehder (1940) described 24 ornamental varieties of black locust, mostly from Europe and
cultivated since the 19th century, with unusual flowering habits or morphology. Four varieties
were listed as spineless. Bean (1980) noted that although "an extraordinary number" of varieties
have been grown in Europe, only three to four dozen of them were being cultivated at the Kew
Gardens near London.

Because black locust is cross-pollinated by bees, desired clones must be reproduced

vegetatively. Major gennplasm collections include over 900 accessions at Michigan State

5
University, over 100 at the University of Georgia, and hundreds in Hungary (Barrett and Hanover
1991). Range-wide provenance/progeny tests of over 400 accessions in Michigan showed highly
significant variation in five traits (height, diameter, number of leaders, spine length, and winter
dieback), but no differences due to region of origin (Mebrahtu and Hanover 1989; Barrett e_t a_l.
1990). Two prostrate mutants have been described which sometimes retained their horizontal habit
when vegetatively propagated (Kriebel 1960; Davis and Keathley 1988). Spineless seedlings have
been observed, but they are shrubby and grow slowly (Barrett and Hanover 1991). A mutant
spineless shoot described by Kim and Lee (1974) grew slowly when grafted. In plantation 92.13 at
Michigan State University's Sandhill Research Area, 24 spineless clones have been planted by the
author in a permanent archive (19 survived to 1993), together with other unusual phenotypes and

hybrids that vary in leaf form, leaf color, flower color, and flowering time.

flew—Grins

Depending on growing conditions, black locust will bloom within two to five years from
sowing. Flowering time in Michigan is in late May and early June. The fragrant white bee-
pollinated pea-like flowers grow in clusters 10 - 20 cm long and supply much nectar which makes
excellent honey. Bencat (1986) found the average dimensions of black locust inflorescences to be
10.7 cm long by 5.5 cm wide, with a peduncle length of 2.4 cm and a fresh weight of 3.0 g. Total

floral biomass fresh and dry weights were calculated at up to 12.7 and 1.5 tons/ha, respectively.

Use as bee forage

Black locust has many positive features as a bee forage, but the time of bloom is too early
for domestic honey bees to take full advantage of (Ayers 1992). Keresztesi (1977) noted that in

good flowering years, black locust honey accounts for about 75% of the market in Hungary.

Uses of wood
The heavy, hard wood is useful for fuel, charcoal, parquet flooring, posts, poles, and pulp.

Because of black locust's adaptability, inexpensive propagation by seed, rapid juvenile growth,

6
high heat content of the wood, and prolific regrowth after cutting, it is ideally suited for woody
biomass plantings. Commercial energy production may eventually become one of its primary uses
in the US. (Miller 91 n 1987).

Black locust wood has a specific gravity of 0.66 to 0.71, a caloric content of 4.7 kcal/g, an
extractives content of 6-9% of dry weight, and a very low volumetric shrinkage value of 10-1 1%.
It is far easier to dry without warping than other commercial high density woods such as oak
(Stringer 1992). The durability of black locust wood makes it suitable for railroad ties, mine
timbers, and structures in contact with the soil. Heartwood extracts have conferred decay
resistance on other woods equivalent to the effect of commercial wood preservatives (Smith et al.

1989; Smith 1992).

Uses for soil conservation

At present, surface mine reclamation is probably the main use for which black locust is
planted (Bridgen 1992). In the past it has been widely employed for erosion control, windbreaks
and shelterbelts, and to increase soil fertility for other species. Black locust will thrive in sand and
has been used extensively for stabilizing coastal dunes in the US. (Gustafson 1935; Kroodsma
1937), southern France, Romania, Cyprus, and Japan (Streets 1962). Estimates of black locust's
annual nitrogen input per hectare range from 30 kg (Boring and Swank 1984) to 109 kg (Dawson
et al. 1982), with 56 kg (Ike and Stone 195 8) and 70 kg (I-Iaines 1978) also reported. Bridgen
(1992) cited five authors who reported average nitrogen fixation rates ranging from 26 to 274
kg/ha/year. Soil in a 36-year-old Indiana plantation had an average annual nitrogen input of 19 kg
per hectare. From peak acetylene reduction rates of individual nodules, it was estimated that with
the most efficient Rhizobium strains the nitrogen fixation rate could be increased over five times to
above 100 kg/ha/year (Reinsvold and Pope 1985).

Black locust has been interplanted to improve growing conditions for high value timber
trees such as Sitka spruce (Picea sitchensis (Bong) Carr.) (Streets 1962) and black walnut

(Juglans nigra L.) (Fowells 1965). Schlesinger and Williams (1984) concluded that the rapid

7
growth of black locust becomes detrimental to walnuts, unless competition is controlled by locust
borers, poisoning, or other silvicultural interventions. Timely girdling of interplanted black locust
improved walnut grthh by up to 135%, but only on the poorest sites.

Black locust is being tested as the hedge component in an alley cropping system with
maize, in Ohio. Year-old seedlings were transplanted 150 cm apart in hedges 540 cm apart, with
five rows of maize between. The alley cropping arrangement had 28% fewer maize plants than a
solid stand, but yielded 18% less. This was attributed to better access to light for the rows
adjacent to the young trees. Soil moisture was adequate throughout the growing season, but
competition for light and water was anticipated as the trees grew larger (Ssekabembe and

Henderlong 1991).

Uses for animal feed

A wide variety of wildlife and livestock consume the high-protein foliage, although it is not
suitable as a sole feed. In a listing of 19 Indian tree fodders containing over 15% crude protein

(CP), black locust rated third at 20.45% average CP (Singh and Negi 1987). CP content and

1 protein digestion inhibitors decline with the age of the leaves, so the content of digestible protein
remains relatively constant at around 10% (Singh and Negi 1987; Barrett and Hanover 1991).

Black locust has tannins which reduce protein digestibility, and stipular spines on the
stems which are flexible when green but rigid when mature. Processing can separate the leaflets
for a higher protein feed component, or grind the spines and woody stems into digestible fragments
(Barrett and Hanover 1991). Black locust leaves are used for livestock feed in the Republic of
Korea and in Bulgaria (Keresztesi 1983, 1988) and for commercial angora rabbit production in
China (Cheeke 1992). In the highlands of northern India it is a common fodder tree. Branches
above the reach of livestock are cut when other green forages are scarce, and the wood is used later
for fuel (Barrett e_t a1. 1990). Black locust leaf meal is exported from China for use in Japan as a
xanthophyll pigment source for poultry feed. It is used in small amounts (3 - 5%) for coloring

only, and not for its nutrient content (Cheeke 1992).

8

In one laboratory test, ground black locust foliage and woody stems were found chemically
comparable to alfalfa (Mediggo gtjv_a L.), with 22 - 24% CP, 7% lignin, and 4.2 kcal/g
(Baertsche 91 al_. 1986). In published feeding trials, black locust has always been found inferior to
alflafa. Horton and Christensen (1981) compared black locust leaf meal with alfalfa meal in a
feeding trial with lambs. Both contained 20% CP and 4.5 kcal/g; CP digestibility was 27.0% in
black locust and 69.6% in alfalfa, and caloric utilization was 42.5% and 61.7%, respectively.

They deduced that in black locust leaf meal only the hemicellulose component of the cell walls was
digested.

No adverse effects of fwding black locust to rabbits were observed by Harris e_t a_l. (1984),
but the apparent CP digestibility of 54.6% was significantly lower than the 78% measured for
alfalfa. Black locust leaf meal fed to growmg chicks reduced growth, but not as much after
autoclaving. Leetins were suspected of inhibiting digestion (Cheeke e_t a_1. 1983). Singh and Negi
(1987) stressed the role of condensed tannins in binding protein, and suspected the alkaloid robin of
additional inhibitory effects. Cheeke (1992) blamed low protein digestibility on tannins and other
phenolic compounds, but noted that the low dry matter digestibility of black locust compared to
alfalfa was not directly due to tannins. Tannins could act indirectly, however, by inhibiting

digestive enzymes of herbivores or their rumen microflora.

EXPERIMENTAL OBJECTIVES

The nitrogen-fixing black locust tree is extremely high yielding, rapid growing, and
adaptable, being widely naturalized in temperate regions. It has potential for use in agroforestry
systems, including forage production, short rotation biomass, fuelwood, beekeeping, soil
improvement, and alley cropping. Cultivars for specific purposes have long been used in Hungary,
and hold great promise for the future, considering the existing genetic diversity and the relatively
short generation time of the species. Current problems involving spines, crooked trunks, locust

borer damage, and protein digestion inhibitors can be addressed by selective breeding and improved

9
at a much faster rate than is possible with other temperate zone tree species. Together, these
features make black locust an attractive subject for scientific research in general and for
agricultural evaluation in particular.

A set of experiments was designed to obtain basic agronomic data on black locust as a
crop plant. Field trials were conducted on monoculture plantings to identify the optimal agronomic
management practices for producing green leafy material. Treatments involving five spacings, two
harvest heights, two years of first harvest, and eight harvest dates during the growmg season were
applied to plots of black locust planted in 1988. Two trials to compare seed sources were also
established, but they will be discussed here only in regard to spacing and year effects. The effects
of seed sources were rarely statistically significant, and have been reported elsewhere (Barrett
1992).

Harvest operations were done by hand on small plots to simulate mechanical harvesting at
a constant height. Fresh weight, dry weight, and number of plants cut were recorded at each
harvest, from 1988 to 1991. Records were also kept for height, number of stems, diameter, flesh
and dry weight, and survival for a set of individual seedlings harvested at different dates in 1989
and 1990. Numbers of harvested branches were counted in 1988-1990 in one of the seed source
trials.

The results of these experiments have direct applicability for the production of livestock
feed and short rotation woody biomass from black locust. The findings also have implications for
other species, and other uses of black locust, including direct grazing, alley cropping, and soil

conservation.

ORGANIZATION OF THIS DISSERTATION

This dissertation is organized into seven chapters. This introductory chapter described the

unusual features and multiple uses of black locust, explains what experiments were done, and why.

10
Details of where, when, and how field trials were conducted are contained in a separate materials
and methods section for each experiment.

The next four chapters cover agronomic treatments involving variations in time of harvest,
harvest height, spacing, and year of first harvest. The efl‘ects of these treatments over three or four
years upon yields per plot and per plant, the dry weight percentage, and survival are reported. The
year treatments interacted with spacing and harvest height, and spacing interacted with harvest
height. These interactions are mentioned where statistically significant. Chapter two reports the
efl‘ects of eight harvest dates during the growing season. Six of the harvest dates were compared
over two years (1989 - 1990) for yields and over three years (1989 - 1991) for survival. Chapter
three reports the effects of the height of harvest above the ground. Two heights were compared
over three growing seasons (1988 - 1990). Chapter four reports the effects of spacing or initial
population density. A square grid planting system was employed to give a single distance between
plants at each spacing. Five spacings were tested over three years (1988 -l990), and two over four
years (1988 - 1991). Chapter five reports the effects of the year of first harvest upon sets of plots
that were also used to investigate seed sources, harvest heights, and the spacing between plants.
First harvests were taken in 1988 or 1989, and the effects were recorded through 1990 or 1991.

Chapter six contains a literature review on ideotype breeding, and explains how it relates
to selecting and improving trees for use as crop plants. Desirable features for for a woody species
grown as a perennial forage crop are combined from published lists into a detailed ideotype, or
ideal model. The characteristics of black locust are then compared with this ideotype and
evaluated.

Chapter seven is a forward-looking conclusion. It briefly summarizes the main findings of
the experiments, and then examines a host of unanswered questions. As is typical in scientific
investigation, many more questions have been raised than answered. These include some of the
original hypotheses where the data are inconclusive, new mysteries that appeared during the field

trials, and ideas investigated by others using other species. Because the ideas and results of

11
previous researchers were cited in previous chapters, and the conclusion is, a synthesis, it does not
contain citations. Next, I present a short list of priorities for future research on black locust as a
perennial forage crop. Finally, I offer a few insights for firture researchers on working with black
locust. I ofl‘er both my conclusive results and my subjective opinions to those who will continue
research on multi-purpose nitrogen fixing tree species, in the hope that such plants can make a

greater contribution to human needs on this increasingly crowded planet.

12
REFERENCES CITED

Ayers, G. S. 1992. Characteristics of black locust as bee forage. pp. 243-251 in: Hanover, J.
W., K. Miller, and S. Plesko, eds. Proceedings: International conference on black locust:
biology, culture, & utilization. Department of Forestry, Michigan State University.

Baertsche, S. R., M. T. Yokoyama, and J. W. Hanover. 1986. Short rotation, hardwood tree
biomass as potential ruminant feed - chemical composition, nylon bag ruminal
degradation and ensilement of selected species. J. Animal Sci. 63:2028-2043.

Barrett, R. P. 1992. Forage plantation management for black locust. pp. 259-272 in: Hanover, J.
W., K. Miller, and S. Plesko, eds. Proceedings: International conference on black locust:
biology, culture, & utilization. Department of Forestry, Michigan State University.

Barrett, R P., and J. W. Hanover. 1991. Robinia pseudoacacia: a possible temperate zone
counterpart to Leucaena? pp. 27-41 in: Garrett, H. E., ed. Proceedings of the 2nd
Conference on Agroforestry in North America, Springfield, MO, August 18-21, 1991.

Barrett, R P., T. Mebrahtu, and J. W. Hanover. 1990. Black locust: a multi-purpose tree species
for temperate climates. in: Janick, J ., and J. E. Simon, eds. Advances in new crops. Timber
Press, Portland, OR.

Bean, W. J. 1980. Trees and shrubs hardy in the British Isles, 8th edition. John Murray, London.

Bencat, T. 1986. Biomasa a variabilita kvetu agata bieleho (Robini_a pseudoacacia L.) na Juznom
Slovensku. Biologia (Bratislava) 41:33-46. (in Slovakian with English abstract)

 

Bongarten, B. C. 1992. Genetic variation in black locust within its native range. pp. 78-97 in:
Hanover, J. W., K. Miller, and S. Plesko, eds. Proceedings: International conference on
black locust: biology, culture, & utilization. Department of Forestry, Michigan State
University.

Boring, L. R., and W. T. Swank. 1984. Symbiotic nitrogen fixation in regenerating black locust
(Robinia pseudoacacia L.) stands. Forest Science 30:528-537.

Bridgen, M. R. 1992. Plantation silviculture of black locust. pp. 21-32 in: Hanover, J. W., K.
Miller, and S. Plesko, eds. Proceedings: International conference on black locust: biology,
culture, & utilization. Department of Forestry, Michigan State University.

Cheeke, P. R., M. P. Goeger, and G. H. Arscott. 1983. Utilization of black locust (Robinia
pseudoacacia) leaf meal by chicks. Nit. Fix. Tree Res. Reports 1:41.

Cheeke, P. R. 1992. Black locust forage as an animal feedstuff. pp. 252-258 in: Hanover, J. W.,
K. Miller, and S. Plesko, eds. Proceedings: International conference on black locust:
biology, culture, & utilization. Department of Forestry, Michigan State University.

Davis, J. M., and D. E. Keathley. 1988. In vitro propagation of a black locust tree with an unusual
phenotype. Nitrogen Fixing Tree Research Reports 6:65-67.

l3
Dawson, J. 0., T. A. White, P. J. Dzialowy, and R. A. Herendeen. 1982. Nitrogen accretion in

surficial soil of a young plantation of Robinia pseudoacacia. Forestry Research Report No.
82-7, Agr. Exp. Sta., Univ. of Illinois.

' Fowells, H. A., ed. 1965. Silvics of forest trees of the United States. USFS, USDA, Ag. Handbook
No. 271.

Galford, J. R. 1984. The locust borer. USFS, USDA, Forest Insect and Disease Leaflet 71.

Genys, J. B., and D. M. Harman. 1990. Racial diversity of black locust in growth rate and in
susceptibility to the locust twig borer. N. J. Appl. For. 7:3 8-43.

Gupta, R. K. 1986. Establishment and management of fodder trees in temperate and sub-temperate
regions for agro/social forestry. pp. 163-180 in: Khosla, P. K., S. Puri, and D. K.
Khurana, eds. Agroforestry systems - a new challenge. Indian Society of Tree Scientists,
New Delhi.

Gustafson, A. F. 1935. Composition of black locust leaf mold and leaves and some observations on
the effects of black locust. Agronomy Journal 27:237-239.

Haines, S. G. 1978. Potential uses ofnitrogen-fixing plants in forestry. pp. 1-19 in: S. G. Haines
(ed.). Nitrogen fixation in southern forestry. International Paper Co., Bainbridge, GA.

Harris, D. J ., P. R. Cheeke, and N. M. Patton. 1984. Evaluation of black locust leaves for growing
rabbits. Journal of Applied Rabbit Research 7 :7-9.

Hofl‘ard, W. H., and R. L. Anderson. 1982. A guide to common insects, diseases, and other
problems of black locust. USF S, USDA, Forestry Report SA-F R 19.

Hofi‘ard, W. H. 1992. Insect pests of black locust. pp. 44-49 in: Hanover, J. W., K. Miller, and S.
Plesko, eds. Procwdings: International conference on black locust: biology, culture, &
utilization. Department of Forestry, Michigan State University.

Hopp, H. 1941. Growth-fem variation in black locust and its importance in farm planting. Journal
of Forestry 39:40-46.

Horton, G. M. J., and D. A. Christensen. 1981. Nutritional value of black locust tree meal
(Robimjafipseudoacacia) and alfalfa meal. Canadian Journal of Animal Science 61 :503-
506.

Huang, F.-H., F. C. Cech, and R. B. Clarkson. 1975. A comparative investigation of soluble
protein polymorphism in Robinia pseudoacacia roots by polyacrylamide gel
electrophoresis. Biochemical Systematics and Ecology 3: 143-147.

 

Ike, A. F., and E. L. Stone. 1958. Soil nitrogen accumulation under black locust. Soil Science
Society of America Proceedings 22:346-349.

Keresztesi, B. 1977. Robinia pseudoacacia: the basis of commercial honey production in Hungary.
Bee World 58:144-150.

 

14
Keresztesi, B. 1983. Breeding and cultivation of black locust, Robinia pseudoacacia, in Hungary.
For. Ecol. and Mgmt. 6:217-244.

Keresztesi, B. 1988. Black locust: the tree of agriculture. Outlook on Agriculture 17:77-85.

Kim, C. S., and S. K. Lee. 1974. Studies on characteristics of selected thornless black locust
(Robinia pseudoacacia L.). Research report of the Institute of Forest Genetics 11:1-12. (in
Korean)

Kriebel, H. B. 1960. A prostrate mutant of black locust. Journal of Forestry 58:222.

Kroodsma, R. F. 1937. The permanent fixation of sand dunes in Michigan. Journal of Forestry
35:365-371.

Lee, H. S., and H. S. Kang. 1980. A study on the effect of supplemental plant nitrogen sources on
the digestibility of rice straw by sheep. Korean Joumal of Animal Sciences 22(2): 109-1 17

(in Korean, English summary).

Mebrahtu, T., and J. W. Hanover. 1989. Heritability and expected gain estimates for traits of
black locust in Michigan. Silvae Genetica 38: 125-130.

Meginnis, H. G. 1934. The effect of cultivating young black locust. Journal of Forestry
32:569-571.

Miller, R. 0., P. D. Bloese, and J. W. Hanover. 1987. Black locust: a superior short-
rotation intensive culture species for biomass production in the Lake States. Inst. of
Gas Technology, 11th Annual Meeting on Energy From Biomass and Wastes, March
16, 1987, Orlando, FL.

National Research Council. 1983. Firewood crops: shrubs and trees for energy production,
Volume 2. National Academy Press, Washington, D. C.

Peabody, F. J. 1982. A 350-year-old American legume in Paris. Castanea 47299-104.

Rehder, A. 1940. Manual of cultivated trees and shrubs hardy in North America. Macmillan, New
York.

Reinsvold, R. J., and P. E. Pope. 1985. Association of Robinia pseudoacacia and Rhizobium:
potential nitrogen accretion. pp. 245-250 in: Proc. 5th Central Hardwood Forest Conf.

Schlesinger, R. C., and R. D. Williams. 1984. Growth response of black walnut to interplanted
trees. Forest ecology and management 9:235-243.

Singh, 8., and S. S. Negi. 1987. Tree forages in animal rations. pp. 178-190 in: Khosla, P. K., and
D. K. Khurana, eds. Agroforestry for rural needs. Indian Society of Tree Scientists, New
Delhi.

Smith, A. L. 1992. Chemistry of the extractives of the black locust tree. pp. 208-217 in: Hanover,
J. W., K. Miller, and S. Plesko, eds. Proceedings: lntemational conference on black locust:
biology, culture, & utilization. Department of Forestry, Michigan State University.

15
Smith, A. L., C. L. Campbell, M. P. Diwakar, J. W. Hanover, and R. 0. Miller. 1989.
Extracts from black locust as wood preservatives: A comparison of the methanol
extract with pentachlorophenol and chromated copper arsenate. Holzforschung 43:293-

296.

Ssekabembe, C. K., and P. R. Henderlong. 1991. Below-ground interactions in alley cropping;
appraisal of first-year observations on maize grown in black locust (Robinia
pseudoacacia L.) alleys. pp. 58-73 in: Garrett, H. E., ed. Proceedings of the 2nd
Conference on Agroforestry in North America, Springfield, MO, August 18-21, 1991.

Streets, R. J. 1962. Exotic forest trees in the British Commonwealth. Clarendon Press, Oxford.

Stringer, J. W. 1992. Wood properties of black locust (Rebinia pseudoacacia): physical,
mechanical, and quantitative chemical variability. pp. 197-207 in: Hanover, J. W., K.
Miller, and S. Plesko, eds. Proceedings: lntemational conference on black locust: biology,
culture, & utilization. Department of Forestry, Michigan State University.

Sutton, P., and W. A. Dick. 1987. Reclamation of acidic mined lands in humid areas. Advances in
Agronomy 41:377-405.

EFFECTS OF HARVEST DATE ON SURVIVAL AND FORAGE YIELD OF
BLACK LOCUST

ABSTRACT

A harvest date trial was conducted over two years on plots of black locust trees planted at
25 - 35/m2, cut at 25 - 30 cm height on eight dates 46 - 156 days after bud break. Green biomass
yield increased at about 12 g dry weight/mZ/day from late June to late September, even after height
growth ceased. Plots harvested early and cut a second time had low regrowth productivity, under 3
g dry weight/mzlday. Only September harvest dates were comparable between years, because of
reduced vigor and survival the second year. Dry weight increased from about 20% of fresh weight
in June to about 50% in September. Survival steadily decreased over three years, but was highest
with September harvests. To obtain the highest green biomass yields each year and maintain
adequate survival, a stand of black locust seedlings should be harvested as late in the growing
season as possible, but before the leaves are killed by frost. Physiological reasons for this

recommendation are discussed.

INTRODUCTION

Black locust is a very rapid growing nitrogen fixing tree which is useful for a variety of
products. Because of its high yields of high protein foliage it is being investigated as a perennial
forage crop. To grow black locust for green biomass or forage, optimal harvest times must be
determined. The relation of yield to height growth and to survival rates needs to be clarified. Field

trials must be repeated on the same dates for more than one season, to determine how weather

17
conditions and any variation in survival affect yields. The dry weight percentage, important for

utilization as silage, is known to increase with age during the season, but may vary between years.

PREVIOUS RESEARCH
Seasonal effects have been observed for many species, where survival, coppice shoot
numbers, or growth rates were highest when cut while dormant and lowest when cut in midsummer

(Blake 1983).

Black locust

Meginnis (1940) transplanted year-old black locust swdlings to the field and decapitated
them at four dates, three during dormancy. Cutting back the tops in May, one month after
planting, gave the lowest survival and heights at two years after transplanting. Bencat (1992)
measured the change in mineral content of black locust leaves from late spring to late autumn. Ca
content increased 139%, K content decreased 41%, and little variation was found in the contents of
Fe, Mg, Na, Pb, or Zn. The leaves were collected from trees 10 - 15 years old.

No other information on the effect of harvest date on black locust is available. Therefore,
reports on other nitrogen-fixing leguminous trees must be consulted. All such research has been

done with tropical species.

Leucaena

Yields of leucaena (Leucaena leucocephala (Lam) deWit) usually increase with harvest
interval length (Guevarra e_t 11. 1978; Krishnarnurthy and Mune Gowda 1982, 1983; Jeyaraman e_t
a_l. 1989; Field and OeMatan 1990; de Lucena Costa e_t a_l. 1991), though not always significantly
(Topark-Ngarm 1983; Evensen 1985). In all trials, 20 - 50 day intervals gave the lowest yields.
Lower yields with frequent cutting may follow high mortality from stress (Osman 1986), and have
also been attributed to relatively more time spent in slow growth while recovering from cutting

(Guevarra e_t_ _a_l. 1978; Home 91 a1 . 1985). Some results showed a peak around 60 days. while

1 8
other trials gave a peak at 90 days. All intervals over 90 days gave declining yields (T akahashi
and Ripperton 1949; Osman 1981a, 1986), except for one with a unique bimodal pattern.
Mendoza e_t a. (1983) found a decline at about 84 days, compared with 56 and 112 days.

Horne e_t a. (1985) noted a strong interaction between harvest frequency and planting
density in leucaena. To maximize leaf yields, harvesting should be done on closed canopies with
the highest possible leaf area index, just before shading causes lower leaves to abscise. At higher
densities where the canopy closes earlier, shorter harvest intervals should raise leaf yields.

When expressed by weight, the yield of dry forage gives no indication of the proportion of
woody material present. Lignification afi‘ects the protein and fiber contents, and thus the nutritive
value of the forage. By defining the age of the material, harvest interval exerts a major effect on
wood content. Factors such as cultivars, fertilization, and season can also affect wood proportion
(Osman 198 lb). Leucaena yield data should be assumed to include woody stems unless specified
otherwise.

From a review of early leucaena data, Hill (1971) concluded that the optimum harvest
frequency for maximum yield was about four times per year, or every 90 days. Brewbaker (1987)

concluded that yields are maximized at 70 - 90 day intervals, depending on temperature.

Other tropical species
Galang et a1. (1990) harvested stands of Sesbania sesban (L.) Merrill in subtropical

Queensland at intervals of 28, 42, and 56 days. Total biomass increased significantly with interval
length, but leaf dry weight remained about the same, giving a declining percentage of leaf in the
biomass.

Seedlings of Dalbergia sis__s___oo Roxb. were defoliated three times at monthly intervals,
removing 0%, 25%, 50%, or 75% of leaves without cutting the stems. The 25% defoliation gave a
large increase in nitrogenase activity, while greater defoliation percentages reduced it below the

control level. Increased stem and root weights were recorded with 25% and 50% treatments, but

l9
treatments progressively reduced root collar diameter and leaf weight. Stem height was not
affected (Pokhriyal e_t a_l. 1992).

Year-old seedlings of the leguminous tree Paraserianthes falcataria (L.) Nielsen produced
significantly more biomass when coppiced at intervals of 84 days compared to 42 days. Yields of
the 42-day group declined severely after the fifth of eight harvests, while the 84-day group
maintained its yield level (Rogers and Rosecrance 1992).

Hardesty gt a1. (1988) cut coppice sprouts of five tree species (including some legumes) at
difi‘erent times during the year in northeast Brazil, where drought limits the growing season to the
period from January to June. Cutting late in the growing season (May) gave the lowest stem, leaf,
and total yields the next year in all cases. Yields were maximized by cutting in the early dry
season (July) or mid dry season (November), depending on species. Leafweight was affected less

than stem weight by the time of coppicing.

Difficulg of applying tropical results in Michigan
Because tropical crops are usually limited by rainfall rather than temperature, and may be

grown all year with irrigation, results from tropical leguminous trees can only give rough
guidelines for temperate zone fodder trees. A harvest interval of 90 days is not practical when the
growing season for black locust in Michigan is about 160 days. The season extends fi'om bud

break in late April until the first killing frost in late September or early October.

EXPERIMENTAL OBJECTIVES

To determine which harvest schedule would give the highest yields of green biomass over
two years, an experiment was conducted with eight harvest dates. It was designed to determine the
optimal time for harvesting black locust new growth, in order to maximize cumulative yields and
long-term survival. The patterns of daily productivity, dry weight percentage increase, and canopy

coverage were also investigated, to test the following hypotheses:

20

1. Biomass yield increases with height, and levels off when height growth ceases.

2. Daily productivity rates (g dry weighUm2/day) are directly related to solar radiation,
and vary significantly between harvest dates due to changing weather conditions and day lengths.

3. Daily productivity rates for the same harvest date vary significantly between years due
to weather differences.

4. Daily productivity rates per plant (g dry weight/plant/day) increase each year.

5. Dry weight percentage (dry weight/fresh weight) increases in direct proportion to time
from bud break.

6. Survival varies significantly with harvest date.

7. Autumn and spring canopy coverage (area covered by leaves/total area) vary

significantly with harvest date.

MATERIALS AND METHODS

General information for all trials

A bulk mixture of equal numbers of swds from four Michigan sources was planted. The
sources were selected on the basis of availability, and not for any judgement of quality. The seed
accessions were 442 fi'om Russ Forest in Cass County, 445 from the comer of Sandhill and
Hagadorn roads, and 446 from the Dansville State Game Area, both in Ingham County, and 450
from Kellogg Forest in Kalamazoo County. Prior to sowing, the seeds were scarified in
concentrated sulfuric acid for 50 minutes. They were directly sown in late June 1988 at the
Michigan State University Tree Research Center in East Lansing (420 41' N, 840 28' W). The
nursery beds were surrounded by evergreen windbreaks and had been firmigated with methyl
bromide. The soil was mapped as a complex of soil series including Riddles, Hillsdale, Owosso,
and Marlette, all of which were well drained with 2% to 6% slopes and sandy loam topsoil.

Subsoils were sandy loam, loam, or clay loam (Soil Conservation Service 1979).

2 1

Prolonged heat and drought in 1988 impaired growth and sprinkler irrigation was
necessary. After normal weather returned in mid August, growth was rapid until early September,
with swdlings reaching heights of 60-120 cm. In 1989, 1990, and 1991 growing conditions were
favorable.

Because the soil had been fumigated, there were few weed problems in 1988. In the spring
of 1989 and 1990, the plots were sprayed with glyphosate (trade name Roundup) at 0.7 kg active
ingredient/ha and DCPA (trade name Dacthal) at 8.4 kg active ingredient/ha. Potassium sulfate (0-
0-50) fertilizer was added at the rate of 225 kg/ha (112 kg/ha K20) on May 7, 1990, to bring soil
potassium levels up from around 90 kg/ha. No other fertilizer was applied.

In each plot, all new growth above the designated harvest height was cut with hand tools
and bagged to determine the fresh weight. The paper bags of harvested material were dried in an
oven at 600C for at least three days before measuring the dry weight. Weights per plant were
calculated fi'om the number of plants harvested. Multiple sprouts from the same root were counted
as one plant. Shoots from different roots were counted as separate plants. Due to the presence of
suppressed seedlings, the total survival in 1989 was often higher than the number harvested, but
few suppressed seedlings survived to 1990. Survival to 1991 was often higher than the number
harvested that year because of rabbit damage on the smallest plants.

All trials were planted using a randomized complete block design. Analysis of variance,
Duncan's multiple range tests, and other statistical operations were conducted on NCSS computer
programs (I-Iintze 1987). The probability level reported for each factor below a table is the
probability of the difference between treatments occurring by chance, according to F-test

comparisons.

Rowplots at eight harvest dates
Early in the 1989 season, 48 plots 1.00 m long by 2.00 m wide were selected within two
parallel sown areas each 45 m long by 2.5 m wide. Utilizing a randomized complete block design,

the 48 plots were divided into six replications, each containing eight plots within 15 m, which were

22

randomly assigned to the eight harvest dates. Sections with uneven plant distribution were avoided
and plot borders adjusted so that each plot contained either 55 (:1: 4, with two plots under 50) or 70
(:i: 4, with two plots at 64) evenly distributed surviving plants. Three plots from the low-
population group and three fiom the high-population group were cut on each of the harvest dates.
Plots were spaced 10 cm to 150 cm apart, and the plants between the plots were cut when the
second adjacent plot was harvested. This caused many of the earliest harvested plots to be quickly
overshadowed by taller neigthrs, creating ideal conditions for rabbits to stay hidden and consume
the new growth. Consequently, survival to 1990 was poor in the earliest plots and five had no
survivors at all (four from June 29 and one from July 13).

In 1989 new growth from the 20 cm stumps was harvested at 25 cm the first time, and at
30 cm ifharvested a second time. All harvests in 1990 and 1991 were made at 30 cm. In 1990 all
plots had closed canopies before the first harvest and had very little leaf area remaining afterwards,
while for the second harvest in early October, canopy coverage depended on the time since the first
harvest and the vigor of the plants. In 1991 the plots were allowed to grow undisturbed until
harvested in September.

Harvest dates in 1989 were usually at 14-day intervals: June 15, June 29, July 13, July 27,
August 10, August 24, and September 7. The eighth and final harvest was delayed until
September 28 to be nearer to the anticipated time of the first frost. Plots first cut in June were cut
again in August, and plots cut before mid August were harvested in early October. Most plots
were out twice and some were out three times, in an attempt to obtain higher yields. After the
unexpectedly early first frost on September 24 some leaflets had already dropped, so total dry
weight of the last harvest on September 28 would have been greater if it had been taken five days
earlier.

The plots out June 15 and June 29 were abandoned in 1990 because of poor survival, and
the practice of harvesting some plots three times a year was discontinued. For each of the

remaining plots, harvest dates in 1990 were one day earlier than in 1989 according to the calendar,

23
but four days later when counted from the respective days of bud break in the spring. In 1990 the
harvests were conducted on July 12, July 26, August 9, August 23, September 6, and September
27.

The 1991 bud break was 25 days earlier than in 1989, but growth was very slow until mid
May. In 1991 the plots were harvested by replications between September 11 and 30, so that all
harvest date treatments were treated alike.

The plots of the late September harvest were often adjacent to areas that were not cut in
1989. The uncut trees began growth at 1.5 to 2.5 m height, and grew laterally as well as upward
in 1990. Four of the six plots harvested on September 27, 1990, were discovered to be shaded by
these more vigorous neighbors when harvested (the degree of shading was not quantified), although
they had not been shaded at the start of the 1990 growing season. The offending adjacent trees
were cut back to 50 cm before the 1991 season, but they still shaded the late September plots on
20% to 70% of the plot area at harvest in 1991. Only the four plots shaded on 30% of their area
or less were included in the 1991 average for comparison with previous years. For comparisons
between treatments within the years 1990 and 1991, five replications were used.

Daily weight increments were calculated from both the number of days since growth began
in spring (April 30, 1989, April 25, 1990, and April 5, 1991), and the number of days since the
first harvest was taken, subtracting the yield and number of days of the earliest harvest from later
measurements on other plots.

Because growth was very slow early in the season and more rapid at harvest time, daily
productivity rates (g dry weight/mz/day) were also calculated from the "theoretical zero date".
This date was found by regressing the yields at all first harvests on the number of days since bud
break to locate the date where the regression line crossed the X-axis.

Plots harvested before September had produced new foliage on regrowth shoots, and these
were harvested at the end of the season in early October in order to maximize annual yields. To

obtain daily productivity rates up to this last harvest in 1989, the dry weights were divided by the

 

24

number of days between the previous harvest and the first frost on September 24, which halted
growth. In 1990 days were counted from the previous harvest to the final harvest on October 4,
which was before the first frost. Rabbit herbivory reduced the number of plots available for
measuring, so statistical operations were .

Percentage canopy coverage was estimated visually, at 10% intervals, on September 28,
1989, May 30, 1990, October 3, 1990 and May 31, 1991. Although no measurements were made
of the leaf area index (LAI), the percentage canopy coverage can be assumed to be much lower
than the LAI, because the overlapping of leaves was not taken into account as it is when

calculating the LAI. Average canopy height was measured May 31, 1991.

Du weightlfresh weight comparisons
Fresh and dry weights were recorded from black locust plots harvested at 30 cm on July 9,

l4, and 27, 1987. The seeds had been sown into fumigated soil in July, 1986, on the same site
used for the later trials. The plots were 1.0 m2, with an average of 86 surviving and 71 harvested
seedlings in 1987, replicated five times in a randomized complete block design.

Fresh and dry weights were recorded from all row plots harvested in 1989, 1990, and
1991. Additional data, for purposes of illustrating seasonal trends, were obtained from a set of
107 seedlings within and between the row plots, selected before the 1989 growing season. All
seedlings harvested on a particular date were used to calculate dry/fresh weight ratios. Five to 17
individuals were cut and weighed on each of the harvest dates in 1989, and three to 16 on each date
in 1990. Harvest dates in 1989 were June 29, July 5, July 13, July 28, August 10, August 17,
August 24, September 7, September 28, and October 5. Harvest dates in 1990 were June 22, July
19, July 26, August 10, August 24, September 7, September 27, and October 5. The first harvest
in 1989 on each plot or individual was made at 25 cm, and all later harvests were made at 30 cm.

In 1991, fresh and dry weights were recorded for samples of new growth on 18 dates from
June 14 to October 23. Harvests on five dates (September 11, 17, 23, 27, and 30) were made at 30

cm on sets of six plots in a single replication of the row plots. Harvests on the other 13 dates were

25
made on three to eight scattered individual seedlings, and took all new growth regardless of the
previous harvest height. All samples had fresh weights of over 1000 g, except for those taken July
19 (237 g) and October 17 (816 g).

RESULTS

Height growth

Growth in 1988 was slow until mid August when adequate rain and seasonable
temperatures returned, then rapid until early September. Seedling heights in all replications
averaged 5 - 10 cm in mid July, 20 - 30 cm in mid August, and 80 - 90 cm in mid September.
Growing conditions were ideal in 1989, 1990, and 1991, with rainfall well distributed.
Undisturbed coppice sprouts reached 250-300 cm each year.

After harvesting new growth from replicated plots on the same dates for two years, the
average canopy height of new growth the following spring was lower with August harvests and

significantly higher using latest harvest date (Table 2.1).

On weight yields
In 1989, dry weight yields at first harvests increased at a steady rate of 150-200 g each 14

days, with a slight slowing in the 21 days before the September 28 harvest (Table 2.2). In 1990
the average yields were much lower until September, when they reached about 90% of the 1989
levels. This was mainly due to poor survival, but weight per plant showed an additional loss of
vigor in the survivors harvested before September. Averages for the two dates in September 1990
show the proportion of survivors from 1989 to 1990 was much lower than the ratio of dry weights,
meaning that the average plant had more space in 1990 and had the vigor to grow and occupy it.
Although the September 28 harvest date plots always had higher yields per plot and per plant than

the September 7 plots. the differences were never statistically significant.

26
Table 2.1. Effect of harvest date 1989 - 1990 on May 31, 1991 canopy height.

 

 

Harvest date Height (cm)
July 13 66b
July 27 56bc
Aug. 10 440d
Aug. 24 34d
Sep. 7 66b
Sep. 28 92a

 

Means within the same column followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

27
Table 2.2. Dry weight yields (g/mz) for those plots harvested the same date in 1989 and 1990, and
comparisons of yields, numbers of survivors, and weights per plant expressed as ratios.

 

 

 

 

Harvest Dry weiMmz) 1990/1989 Ratios

date 19891 19392 19902 Yield ’ Survival Wt./plant
June 14 203r

June 28 34lf

July 13 554e 554d 243b .44b .48b 8%
July 27 741d 722d 361b .SOb .63ab .83b
Aug. 10 905.: 938C 326b .35b .5 lb .7lb
Aug. 24 1116b 1110bc 43% .40b .59ab 6%
Sep. 7 1277a 1206ab 1095a .91a .71a 1.33a
Sep. 28 1433a 1363a 1225a .903 .70a 1.28a

 

1 Means of all seven replications harvested in 1989.

2 Means of all five replications harvested in 1990.

Means within the same column followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

 

28
Including additional harvests did not change the pattern of higher yields with later first
harvests (T able 2.3). Total dry weight accumulation was about 150 - 200 g each 14 days, with

slower rates early and late in the season.

Du weight productivity per day
When measured fi'om the date of bud break, daily dry weight productivity in 1989 showed

a plateau in June and a higher plateau from late July onward, although the means always increased
fi'om June to September (Table 2.4). Daily productivity data from those plots harvested on the
same dates both years show the plateau starting in mid July 1989. In 1990, July and August plots
gave a plateau of very low productivity. A second, higher plateau in the 1990 rates was observed
in September, which was equivalent to the 1989 rates.

The correlation of 1989 dry weight fi'om the eight first harvest dates with days from bud
break to harvest was very high (r = .9492), and the Y-intercept was negative (-721). The slope of
the regression line was 24.5 g/day. At this rate, average dry weight would have reached zero on
May 29, the "theoretical zero date". The correlation of 1990 dry weight from the first harvest with
days from bud break to harvest was not as high (r = .8501), and the Y-intercept was much more
negative (-2002). The slope of the regression line was 27.9 g/day. At this rate, average dry weight
would have reached zero on July 7.

With the first 29 days of the 1989 growing season removed, counting dry weight from
May 29 instead of May 1, the results were much different (Table 2.5). The daily increments were
not significantly different for any harvest dates. Growth was very linear, departing little from a
rate of 12 g/mZ/day. Growth in 1990 showed wide fluctuations which were not statistically
significant, except for the overly high rate on July 14, which was only five days after the
theoretical zero date.

Results from 1989 showed a pattern similar to the results of a harvest date trial in 1987.
First harvests were made on July 9, 14, and 27, 1987. With bud break presumed to have been at

May 1, first harvests were conducted 70. 75, and 88 days later. Regressions of the dry weight on

 

29
Table 2.3. Effect of first harvest date in 1989 on yields (g/mz) at later harvests in 1989, and
cumulative yields.

 

 

 

 

Dry weiglflg/mz)

First harvest date August October All 1989
June 15 121a 125a 432f
June 29 38b 49b1 400f2
July 13 7lab 625e
July 27 99ab 840d
Aug. 10 62b 967cd
Aug. 24 1 1 16bc
Sep. 7 1277ab
Sep. 28 1433a

 

1 The average total for the June 29 plots was 49, but three of seven plots had lower yields because
they did not survive to the end of the season. The average for the four surviviving plots was 86.

2 The average total for the June 29 plots was 400, but three of seven plots had lower yields

because they did not survive to the end of the season. The average for the four surviviving plots
was 894.

Means within the same column followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

30
Table 2.4. Daily dry weight productivity rates (g/m2/day) in 1989 and 1990 in relation to the
number of days from bud break, for plots harvested on the same date both years.1

 

 

 

Days from Rate (g/m2/day)

Harvest date bud break2 1989 1990
June 15 46 4.4c

June 29 60 5.7c

July 13 74 7.5b 3.1b
July 27 ' 88 8. lab 3%
Aug. 10 102 9.1ab 3.0b
Aug. 24 116 9.6a 3.6b
Sep. 7 130 9.3ab 8.1a
Sep. 28 151 9.0ab 7.9a

 

1 Means were averaged fi'om five replications only.

2 Elapsed days for 1989; add five days for 1990.

Means within the same column followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

 

3 1
Table 2.5. Daily dry weight productivity rates (g/m2/day) in 1989 and 1990 in relation to days
fi'om the theoretical zero date on May 29, 1989 or July 7, 1990.

 

 

 

1989 1990

Days from Rate Days from Rate
Harvest datel zero date (g/m2/day) zero date (g/m2/day)
June 15 17 12a
June 29 31 11a
July 13 45 12a 5 49a
July 27 59 12a 19 18b
Aug. 10 73 12a 33 10b
Aug. 24 87 13a 47 9b
Sep. 7 101 13a 61 18b
Sep. 28 122 12a 82 15b

 

1 Harvest dates in 1990 were one day earlier than in 1989.

Means within the same column followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

 

32
the number of days gave the equation y = -409+12.4x, with r = .61. This gave a theoretical
starting date 33 days past May 1, on June 3. There were significant differences between harvest
dates for yield, but not for yield per day. Daily increments in 1987 ranged from 5.4 - 7.0 g/mZ
starting May 1, and from 10.2 - 12.6 g/m2 starting June 3, which were close to those measured in
1989 (Table 2.5).

When the first harvest date was used as the baseline, the 1989 numbers fell within a
narrow range, with no significant differences (Table 2.6). This method highlighted the extreme
fluctuations in 1990, with the August dates depressed significantly below September dates, and
below the same dates in 1989.

In 1989, dry weights of the last harvests taken in October generally increased with time
since the first harvest, and daily productivity varied little (Table 2.7). In 1990, both parameters
increased with time since the first harvest. Daily productivity of regrowth was very low compared

to productivity before the plots were cut (Tables 2.4, 2.5, and 2.6).

Dy weight productivig per plant and per plant per day

While the average number of surviving plants was quite even for the first harvests of 1989,
with 48 - 53 per plot, this was not the case in 1990, after severe winter mortality had produced a
range of 22 - 36 per plot (11 - 18/m2). In order to fairly compare 1990 productivity rates with
each other and 1989 rates, the numbers of survivors must be taken into account.

The daily productivity rates per plant did not vary significantly by harvest date in 1989,
averaging 0.35 g. In 1990, plots harvested in July and August averaged 0.26 g, while those
harvested in September averaged significantly higher at 0.45 g. In 1991, when all plots were
harvested in September, daily rates averaged 0.53 g per plant, with no significant variation. Due to
the extreme loss of vigor in 1990 for plots at earlier dates, only the September harvest dates are
comparable between years. For September plots, daily productivity per plant rose 15 - 27% each

year from 1989 to 1991.

 

33
Table 2.6. Daily dry weight productivity rates (g/m2/day) after the first harvest on June 15, 1989,
and July 12, 1990.

 

 

 

1989 1990

Days from Rate Days from Rate
Harvest date1 first cut flm2/day) first cut (g/m2/day)
June 29 14 9.9a
July 13 28 12.5a
July 27 42 12.5a 14 8.4bc
Aug. 10 56 12.5a 28 3.0d
Aug. 24 70 13.0a 42 4.7cd
Sep. 7 84 12.8a 56 15.2a
Sep. 28 105 11.7a 77 12.8ab

 

1 Harvest dates in 1990 were one day earlier than in 1989.

Means within the same column followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

 

34
Table 2.7 . Effect of first harvest date on dry weight yields (g/m2) and daily dry weight productivity
rates (g/m2/day) for second harvests in 1989 and 1990.

 

 

 

1989 1990
Days of Yield Rate Yield Rate

Harvest datel growth2 (g/m2) @2/dayL (g/m2) (g/m2/day)
June 28 86 116 1.3

July 13 72 72 1.0 101 1.4
July 27 58 107 1.8 66 1.0
Aug. 3 51 66 1.3

Aug. 10 44 60 1.4 12 0.2
Aug. 17 37 53 1.4

 

1 First harvest dates in 1990 were one day earlier than in 1989.

2 Elapsed days since first harvest in 1989; add 10 days for 1990.

Because of variable numbers of plots, statistical analysis was not possible. Dates with only one
surviving plot were not included.

 

35
Changes in dgy weight percentage

The dry weight/fresh weight percentage (D/F) represents the dry matter proportion of the
green biomass. D/F is low early in the season when all new growth is succulent, but increases as
tissues become lignified.

The D/F rose in a step-wise manner (Table 2.8). In the 1989 season there were two major
increases, after 60 days and again after 130 days. However, the D/F was well correlated with days
from bud break (r = .9564) with an average rate of increase of 1% dry matter each five days. In
the 1990 season there was one much more extreme increase after 121 days. The final D/F in 1989
and 1990 was about the same, but the longer growmg season in 1991 gave a higher final D/F.
Data from individual trees showed a similar pattern of rapid increases and plateaus (T able 2.9).
While green leaves remained, D/F stayed below 50%, but increased to about 60% after leaves
abscised.

When matched against days from bud break rather than calendar dates, the differences
between years were reduced (Figure 2.1). Each year it took about 70 - 90 days to reach a dry
weight proportion of 30%, 130 - 140 days for 40%, and 160 - 170 days for 50%.

When separated, the leafy upper stems had the same D/F as the leafless lower stems until
late in the season, when the D/F of the more woody lower stems increased at a faster rate (Table
2.10). In 1989 the tops and lower stems had the same D/F on August 24, but after that the
percentages varied. The rapid increase in the proportion of leafless stem was due to the shedding
of shaded lower leaves. In 1990, leafy and lower stems were not separated at the September 6

harvest, but such data were taken at the last harvest on September 27 and from a separate sample
on September 25. The D/F percentages resembled those in late September, 1989, and at
September 11, 1991. In 1991 five replications were harvested on different dates, so the variation
also includes the effects of location within the plantation. A smaller sample harvested October 17,

with many rachi but few leaflets remaining, had the same D/F throughout. In 1991 the D/F of

 

36
Table 2.8. Dry weight/fresh weight percentages for entire plots in 1987, 1989 and 1990, and for
small samples in 1991, according to date and days from bud break.

 

 

 

 

Days from Percent dryweight
Harvest date bud break1 1987 1989 1990 19912
June 15 46 21c 28
June 29 60 22c 29
July 9 70 27c 29
July 13 74 30b 29d 26c 303
July 27 88 36a 31c 28b 333
Aug. 10 102 35b 28b 383
Aug. 24 116 35b 29b 41
Sep. 7 130 36b 43a 46
Sep. 28 151 45a 44a 49

 

1 Elapsed days for 1987 and 1989; add 5 days for 1990 and 25 days for 1991.
2 Statistical analysis was impossible in 1991 due to a lack of replication.

3 Data were interpolated because the actual harvest date was not within one day of the date
indicated.

Means within the same column followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

37
Table 2.9. Dry weight / fresh weight percentages for sets of individual plants in 1989 and 1990,
compared to days from bud break.

 

 

 

Days from Percent dry weight

Date bud break1 1989 1990
June 29 60 23

July 5 66 29

July 13 74 28

July 19 80 24
July 27 88 32 30
Aug. 10 102 36

Aug. 17 109 35

Aug. 24 116 34 36
Sep. 7 130 36 40
Sep. 25 148 41
Sep. 28 151 44 39
Oct. 6 158 47 47
Nov. 11 199 612

 

1 Elapsed days for 1989; add 5 days for 1990.

2 Dormant with no foliage remaining.

 

38

55 """""""""" 41997 +19'.' {-1 999 1991 """""""""

 

50°""'"""""""'°'"""°'°'"""°'°"°°°"°'°'""""‘-§§:23"%’
.. 41>
7’
545""y‘, """""
Q) “’5
a
E40 ------------------------------------------ :77} --------- iii. °°°°°°°°°°
E 23:.“ . “JCT . ..........................
9.3
d)
O.

 

60 80 1 00 1 20 140 160
Days from bud break

Figure 2.1. Dry weight percentage of flesh weight for samples harvested in four seasons, related to
days fi'om bud break.

 

39
Table 2.10. Dry weight / fresh weight percentages of leafy and woody portions of material
harvested in September, and the percentage of the total sample represented by the dry weight of the
woody portion. ‘

 

 

 

Percent d9! weight Woody
Harvest date Total Leafy Woody proportion
Sept. 7, 1989 36 35 40 06
Sept. 28, 19891 44 42 51 35
Sept. 25, 1990 41 38 51 31
Sept. 27, 1990 44 40 53 36
Sept. 11, 1991 44 41 54 31
Sept. 17, 1991 50 47 58 32
Sept. 23, 1991 51 46 59 46
Sept. 27, 1991 49 44 58 41
Sept. 30, 19911 57 53 60 51
Oct. 17, 19911 60 60 60 84

 

‘Harvestedafterthefirstfi’ost.

 

 

40
woody lower stems increased faster before frost, and the D/F of leafy upper stems increased faster

after frost, until the ratios equalized again by mid October.

§urvival after harvesting

Average plant numbers per m2 declined from 30 at the time the plots were selected, to
seven at the last harvest in September, 1991 (Table 2.11). Mortality in summer, between the first
harvests on various dates and the last harvest in early October, was higher than over the winters.
During the 1989 growing season mortality varied little by harvest date, but survival to 1990 was
highest with September harvests, and survival to 1991 was significantly poorer for plots harvested

in August.

Canopy coverage

Canopy coverage was assessed by visual inspection from a position directly above the plot,
and quantified at 10% intervals. Canopy coverage at the end of the season was usually incomplete
due to mortality after harvesting, or lack of time for regrowth to restore the canopy (Table 2.12).
In September, 1989, the plots harvested earlier generally had more coverage, with a steep decline
for harvest dates after mid August. Canopy coverage on May 30, 1990 reflected the number and
the vigor of the survivors, which were highest with September harvests. Canopy coverage on
October 3, 1990 varied directly with regrowth season length. Canopy coverage on May 31, 1991
varied significantly, with the highest values for harvests in September and the lowest for harvests
in August.

Canopy coverage on May 30, 1990 was poorly correlated to survival over the previous
winter (r = .3222). However, when only plots with over 33% coverage were considered, there was
a better correlation (r = .8581). In 1991 the correlation of spring canopy percentage with survival
was higher than in 1990 (r = .6201). When only the August and September plots were considered,
the 1991 correlation improved (r = .8965). For both years, more complete spring coverage was

closely related to better survival, while plots with poor coverage had quite variable survival.

4 1
Table 2.11. Numbers of plants harvested per m2 and percentage of original plants surviving for
those plots harvested the same date 1989-1990, for first and second harvests.

 

 

 

Harvest Original 1989 1990 1991
date density lst 2nd lst 2nd Sept.
July 13 29 24( 83ab) 17( 60a) 11( 40b) 9( 301) 8( 27a)
July 27 V 31 24( 77b) 19( 62a) 15( 48ab) 9( 281) 8( 253)
Aug. 10 31 27( 87a) 21( 69a) 13( 44ab) ‘ 5( 181) 4( 15b)
Aug. 24 32 26( 80ab) 15( 47ab) 2 ( 6b)
Sep. 7 32 26( 83ab) 18( 58a) 9( 28a)
Sep. 28 32 25( 78ab) 17( 55a) 11( 35a)
Mean 302 252 19 16 8 7

 

1 Statistical analysis was impossible because of difi‘ering numbers of plots per harvest date.
2 Means include two additional harvest dates in June.

Means within the same column followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

First harvests were taken on the date indicated, while second harvests were taken in early October
from plots with adequate regrowth. In 1991, all harvests were taken in September.

42
Table 2.12. Efl‘ect of harvest date on average canopy coverage percentages from visual estimates
on September 28, 1989, May 30 and October 3, 1990, and May 31, 1991, related to number of
days from first harvest to first frost in 1989 or to recording in 1990.

 

 

Harvest Days of 1989 1990 1990 1991
date regrowthl % Sep. % May % Oct. % May
July 13 72 66 18 66 78b
July 27 58 85 3o 43 60c
Aug. 3 5 l - 75 30

Aug. 10 44 75 23 33 36d
Aug. 17 37 7o 20

Aug. 24 3o 40 13 0 20d
Sept. 7 16 o 43 o 88ab
Sept. 28 -52 O 63 0 100a

 

1 Elapsed days for 1989; add 10 days for 1990.

2 The last harvest date was five days after the first frost in 1989, and five days before the canopy
coverage recording date in 1990.

Statistical analysis was impossible because of variable numbers of plots per harvest date in 1989
and 1990.

Means within the same column followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

43

First harvest yield varied directly with May canopy coverage the same year, and was well
correlated in 1990 (r = .8544) and 1991 (r = .7499). Canopy coverage in May, 1990 had some
predictive value for 1990 yields within each of the harvest dates. July and August plots tended to
have above average 1990 yields if coverage in May was above 15%, and yielded much above
average if coverage was above 25 %. Average May canopy coverage for each of the four July and
August harvest dates ranged fiom 14 to 20%. September plots had greater average coverage in
May, 1990, with all above 25%. Those with coverage above 65% or below 35% yielded above or

below average, respectively, in 1990.

DISCUSSION

GrowthI yieldI and survival
Height growth ended in late August or early September, 120 - 130 days from bud break,

followed by abscission of apical buds. Although no new apical growth was being produced, and
shaded lower leaves were abscising, above-ground dry weight continued to increase. Isarasenee _e_t
a1. (1983) observed that the increasing leaf weight on leucaena plants during autumn in subtropical
Queensland was due to the increasing weight of existing leaves, not increasing numbers of leaves.
In this study, where leaves and stems were not weighed separately, it is possible that either or both
the leaves and stems gained dry weight.

Although no distinct yield plateau appeared in September as had been hypothesized, yields
at the last two dates were never significantly different. These dates were three weeks apart, but
yields were significantly different for five of six earlier pairs of dates in 1989 that were only two
weeks apart.

Complications involving harvesting after frost in 1989, and shading from adjacent taller
plots every year, must have reduced the harvested weights of the late September plots, but the

amount is impossible to calculate. Without those reductions in potential yield, it is likely that

44
yields would have increased significantly from early to late September, but the results remain
inconclusive.

While June plots would have yielded more and survived better if not shaded by adjacent
uncut trees, the strategy of harvesting black locust three times a season appears too deleterious
(and tedious) to justify further investigation. Harvesting twice a season, in late June or early July
and again in late August, was also unfavorable, yielding (in adjacent trials) at most about half the
biomass of the plots harvested on September 28 (Barrett 1992). Cutting coppice sprouts back to
the stump once or twice during the growing season significantly reduced total biomass yields, but
not leaf yields, in three tree species in Brazil. The yield differences between one and two harvests
were not significant (Hardesty and Box 1988). It is not likely that the untested method of
harvesting in early July and again in late September would yield as much as a single harvest in late
September. Almost any harvest during the growing season will remove enough leaves to reduce
photosynthesis, and a severe defoliation can result in carbohydrates being consumed faster than
they are produced, as will be explained below. Therefore, to maximize yields and minimize the
labor required per unit of biomass produced, a single late harvest is recommended. But if the
harvest is too late, problems with freezing are encountered. Mebrahtu (1992) found that
photosynthesis rates in mature black locust leaves declined to almost zero the day after a freeze,
but then recovered over the next week to about 75% of the normal rate, and that photosynthesis can
continue to the end of October. Although undamaged leaves continue fiinctioning, exposed leaves
abscise when damaged by frost, so that the harvestable yield can be greatly reduced from one day
to the next. Therefore, the optimal time to harvest is just before the first killing frost.

Survival varied with harvest date, as had been hypothesized, being significantly reduced by
August harvests. Survival to 1991 was poor for all treatments, but improved greatly once densities
fell below 10/m2, which could be due to individual plants growing larger and stronger. Whether

high mortality continues or not, it is clear that populations above 10/m2 cannot be maintained with

45
annual harvesting. In black locust forage stands, it appears that harvesting more than once a year
is too frequent for optimal survival.

The hypothesis that significant differences in daily productivity rates would be found was
confirmed, but not for the reasons expected. Although changes in daily dry weight productivity in
1989 could be related to weather, a full comparison between years was impossible. Productivity
rates during 1989 varied significantly when calculated fi'om the date of bud break, and were
inversely related to day length. However, when rates were calculated from the theoretical zero date
or from the baseline established at the first harvest, they did not vary significantly from June to
September as day length decreased.

Significant variation in daily dry weight increments was predicted between years, and it
occurred, not from variable weather, but because most 1989 treatments greatly reduced 1990
survival and vigor. Comparisons between years were not valid except for harvests made in
September, where results were consistent. Mebrahtu (1992) recorded the photosynthesis rate on 60
black locust seedlings at the M.S.U. Tree Research Center in a nursery bed adjacent to this trial,
and found the average photosynthesis rate in 1990 to be higher than in 1989 during July, but lower
in August and September. From mid July to mid September, 1989, the average rate rose about
40%. Three factors could reconcile Mebrahtu's observations with the results of this study.
Mebrahtu's measurements were taken on plants from different seed sources, which were
transplanted much farther apart, frequently irrigated, and shaded longer each morning by a
windbreak to the east than were the plots in this study. The photosynthesis rate was measured on a
leaf area basis, and it is possible that leaf area in the more crowded plots declined between July and
September due to abscission of more shaded lower leaves than were replaced by new growth. It is
also possible that net carbon assimilation did increase in the plots of this study, but was diverted to
below-ground parts which were not harvested, as was found in Mrs by Hinckley e_t a_l. (1989).

Leucaena yields per day were highly correlated with temperature in Australia and Hawaii

(Brewbaker 1987), while the yield of Paraserianthes falcataria in Western Samoa varied directly

 

46
with rainfall (Rogers and Rosecrance 1992). In this study, black locust yields per day in 1989
were highly correlated with mean temperature (r = 0.75), but better correlated with pan
evaporation (r = 0.85). On average, each additional 1 mm/day of pan evaporation was associated
with about 2 g/mZ/day additional dry weight. This finding may indicate that the effect of higher
temperatures is not simply to allow more rapid carbon accumulation, but to promote more rapid
transpiration which was associated with more rapid growth. It should be noted that the black
locust plants were well established and did not appear to suffer from a shortage of moisture at any
time after 1988. A possible mechanism for increased transpiration to accelerate growth is through
the transport of cytokinins from the roots to the leaves (McNaughton 1983); mineral nutrition may
also contribute. The high correlations with temperature and pan evaporation support the
hypothesis that the daily growth rate varies directly with solar radiation. Linear relationships
between dry matter production and absorbed solar radiation have been shown for Eucalyptus
QM Labill. (Linder 1985) and many agricultural crops (Monteith 1977). Unfortunately, in
this study the available solar radiation data were unreliable due to inadequate equipment
maintenance.

The hypothesis that daily productivity rates per plant increase each year was confirmed,
although yields and productivity per area declined. With continued mortality of some of the
surrounding plants each winter, it is likely that growth rates of survivors would continue to
increase.

The hypothesis that the dry weight percentage (D/F) increases in direct proportion to time
from bud break was supported by the multi-year average, but not in any particular year. In a
climate with a longer growing season, the D/F would probably rise even higher before frost than
the 51% recorded in 1991. After frost, the D/F rises due to the loss of frozen leaflets and the
drying of those that are damaged. Kinch and Ripperton (1962) found the D/F of leucaena forage
varied between 20.4% and 31.4%, with no clear pattem, as harvest interval varied between 63 and

l 17 days. Jeyaraman e_t a_l. (1989) obtained dry weight percentages of 22.6%, 26.1%, and 28.9%

47
with harvest intervals of 40, 50, and 60 days, respectively, for irrigated leucaena. Black locust
harvested 60 days after bud break in 1989 had only 23% dry weight. The greater water content
would be expected fiom slower growth in spring for a plant in a temperate climate.
When harvested 16 times over five years, the dry weight of leucaena forage varied between
50% and 52% of flesh weight when averaged by year. The range was 55 - 58% for 51m

lebbeck (L.) Benth. and 56 - 61% for Delonix regia, which were cut 17 times. There was no clear

 

trend over the five years (Oakes and Skov 1962). Similarly, there was no clear trend over three
years from 1989 to 1991 in the black locust data from this study. It appears unlikely that the D/F
of the forage would vary with the age of the stumps.

Stem lignification proceeded through the entire season, but no difference in the dry weight
percentages of leafy upper and woody lower stem portions appeared until after height growth had
ceased. This finding could simply be due to the re-allocation of photosynthate to cambial growth
and/or storage in the stems after new growth ceased. However, as the D/F increased more rapidly
in leafless lower stems during September, the succulent stem tips became harder and less flexible,
and presumably less watery as well. To account for the observed low rates of D/F increase in the
leafy tops between the cessation of height growth and the first frost, the leaves must have
maintained a high water content while the attached stems became less succulent.

In the annual cycle of most deciduous woody plants, proteins and nucleic acids are broken
down into amino acids as the leaves senesce in autumn, translocated to the stem, re-assembled into
storage proteins in the bark, and re-mobilized as a nitrogen source for rapid leaf growth the next
spring (Bray 1983; Feller 1990). During autumn, the solute content in the phloem changes from
primarily carbohydrates to primarily amino acids, which are mainly asparagine and glutamine
(Bray 1983). The possibility that black locust does not follow this pattern is suggested by the fact
that the leaves remain green in autumn, and abscise green when killed by freezing. If chloroplasts
are maintained, then the leaves do not senesce as in other species. Mebrahtu (1992) recorded

photosynthetic rates about 25% lower after frost than before, but noted that the difference may

48
have been due more to low temperatures than to leaf damage. Photosynthesis was observed to
continue through the month of October. A speculative explanation for this behavior would lie in
the ability of black locust to fix nitrogen, which is obtained from the rhizobial nodules in exchange
for carbohydrates. If obtaining new nitrogen from well-fed nodules is more efficient than recycling
it from senescent leaves, then continuing photosynthesis as far into autumn as possible would be a
valuable adaptation. An alternate explanation is that Michigan is so far north of the native range
of black locust that the leaves are killed by freezing before the shorter days of autumn can induce a
signal to senesce. This hypothesis is countered by the observation that black locust grown in a

heated greenhouse without supplemental lighting will stay green all winter.

Defoliation and refoliation

Results confirmed the hypothesis that autumn and spring canopy coverage vary
significantly with harvest date. Variation in autumn was related to the time available for regrowth,
and in spring to the numbers of survivors. However, more variation in spring canopy coverage
was observed than could be explained by the survival rate; survivors from some harvest dates were
more vigorous than from others. A decline in vigor due to refoliation and regrowth after leaf
removal has been demonstrated in other temperate zone species. Measurable factors related to low
vigor include low levels of stored carbohydrates, increased winter dieback of new growth, reduced
leaf areas, reduced growth rates, and increased death rates.

Refoliation in late summer is hazardous, for often the time remaining before frost is too
brief to regain the carbohydrate reserves consumed to produce the new growth, much less the
reserves needed to survive the winter and grow rapidly the next spring. Wargo (1981b) found
higher mortality rates after defoliation in three tree species were associated with lower root starch
contents. In xylem tissues of sugar maple, the autumn starch content, as a percentage of the
control, in shoots and roots, respectively, was 98% and 97% after complete defoliation on June 6,
81% and 59% after defoliation July 15, and following defoliation July 29 only 27% and 1%. The

concentration in the shoots was 38% of that in the roots of the undefoliated control group (Gregory

49
and Wargo 1986). Complete defoliation reduced starch reserves in roots of sugar maple saplings
severely if done in mid June or mid July, but had little effect if done in late August (Parker 1981).

In deciduous trees in late smnmer and early autumn, older shoots that have ceased growth
sufier no drain of resources to produce new foliage, so carbohydrate reserves might be
accumulated at a more rapid rate than in actively growing plants. No matter whether the
composition of phloem sap in autumn is dominated by remobilized nutrients from the senescent
leaves, or by carbohydrates from continued photosynthesis, it is clear that the death of tender new
growth at the first fi'ost would terminate either process.

A major factor in variable survival in these trials was the indeterminate character of
refoliating shoots. While black locust is considered to have indeterminate growth, compared to
other tree species, all shoots that began growth before July generally ceased new growth in late
August or early September. However, refoliating shoots that began growth after June stayed
succulent right up to fi'ost, which wilted them immediately. The same process is observed with
black locust seedlings that germinate too late in the summer, indicating that a stem may require a
certain number of days of growth (over 50) before its apical meristem will go dormant.

Normally, growth of black locust is affected by day length. Kramer (1936) reported that
seedlings growing outdoors in North Carolina exposed to long days (duration not specified)
continued growing rapidly until killed by freezing temperatures in November, while those with the
normal day length ceased growth in September and were not injured. Wareing and Roberts (1956)
found that the duration of cambial activity in seedlings, after height growth has ceased, depends
upon day length. Their results indicated that photoperiod is sensed by mature leaves and not by the
stem or buds. It is possible that younger leaves on refoliated stems lack the ability to distinguish
short days, which would explain why they grow until frozen.

Gregory and Wargo (1986) found that the timing of defoliations of sugar maple (_A_c_g
saccharum Marsh.) between May 27 and August 5 affected the survival of buds over the next

winter. With earlier defoliation, terminal shoots refoliated and their buds survived the winter, but

50
with progressively later defoliations, more buds on refoliated shoots died, and the next years
growth depended more on axillary buds formed before defoliation. In sugar maple, however,
defoliations after early August did not stimulate regrowth the same year, which reduced winter
dieback of temrinal shoots. This dieback was probably related to low food reserves and the
inability of new buds to properly harden in the shorter growing season before winter (W argo
1981a, b). In black locust, survival of buds appears less important than the extent of winter
dieback of stems or their consumption by herbivores, but these processes were promoted by late
harvests.

The effects of low carbohydrate reserves and/or a shortage of live buds are often expressed
the next spring in a reduction of leaf area. In 12 to 15 year old (4 - 6 m tall) specimens of sugar
maple, white oak (013m ally; L.), and black oak (Quercus velutm Lamarck), refoliation was
induced by 100% defoliation in May, June, and July. Leafarea measured the following June was
most severely reduced by June and July defoliations in the oak species, and by June defoliations in
sugar maple. Refoliation did not usually follow defoliation in mid-August, and if so, the few leaves
produced were small (Wargo 1981a).

The average size of individual primary leaves in years following severe defoliations was
reduced in red oak (Quercus rubra_ L.) and red maple (m ELM L.). When defoliations were
repeated for three successive years, the total leaf area per tree was reduced only after the second
year with 50% and 75% defoliations, and after the third year with 100% defoliation (Heichel and
Turner 1976). In this study, canopy coverage at the end of the season in plots of the later harvest
dates was much lower in 1990 than in 1989, showing the effect of a second year of 100%
defoliation in late summer. It appears likely that fewer and/or smaller leaves were produced by
black locust stumps with lower vigor, although leaf numbers and areas were not measured.

Since spring canopy coverage is a measure of both survival and vigor, it is also an
indicator of yields later in the same season. The correlation was best for 1990 yields because the

September plots, which had significantly higher spring coverage, also had a longer period of

5 1
growth before harvest. Ifcanopy coverage is measured at the optimal time, about 30 - 40 days
after bud break, it could be an excellent indicator of biomass yields later in the season. It is
possible that the proportion of canopy remaining after an early summer harvest could predict the
yield at the second harvest two or three months later, but no data were taken to test this hypothesis.

Estimating canopy coverage visually, or measuring solar interception with a light meter,
would be much easier than measuring the leaf area index of a plot directly and destructively, as had
been originally planned but never done for lack of time. As shown by previous work with annual
crops (Monteith 1977) and Eucalyptus trees (Linder 1985, Linder g 11. 1985), the correlation of
solar interception and yield should be a fi'uitful area for future research, as it would give a common
reference point for the effects of harvest time, height, and plant spacing, and all their complex and
confusing interactions.

The growth rate, and thus the harvestable yield, declines with any defoliation, but its
timing strongly affects future conditions. Wargo (1981a,b) summarized results from artificial
defoliation studies on deciduous trees, and concluded that defoliation is most injurious when the
new leaves are just fully expanded and the tree is growing rapidly. At this time the demand for
energy is greatest and food reserves are lowest, prior to being restored by photosynthate from the
new leaves. The efl‘ect of the timing of defoliation depends on the species, its phenology, and the
length of the growing season remaining, which influence whether the stems refoliate.

Defoliation followed by refoliation in deciduous trees reduces growth through several
mechanisms. Productive leaves are absent for a time, and food reserves (primarily starch) are
metabolized to maintain living tissues. About three to four weeks after defoliation the new leaves
begin growth, but the total area of replacement leaves is significantly smaller. The refoliated tree
is out of phase with the seasons: weather is usually hotter and drier, day length is declining, and
fewer days remain before frost. Because buds intended for the next spring have opened, new buds
must be formed before autumn. Newly formed tissues may be physically and chemically

immature, and thus more likely to die during the winter. Food reserves may be too low to maintain

 

'7‘]

52
living tissues during dormancy. The following spring, fewer and smaller leaves are produced, food
reserves are lower, and mineral imbalances may also arise. The tree becomes more vulnerable to
pests and diseases that it would normally resist. These factors combine to reduce the vigor of
refoliated trees, and can be compounded by repeated defoliations in succeeding years (W argo
198 1a,b).

By Wargo's description, defoliation of black locust should be more hannfirl in June, when
leaves are young, rather than in mid summer. Rapid growth was recorded during three to four
months in this study, yet the results varied widely during this period. According to Parker (1981),
after defoliation, the recovery or death of an individual tree depends on several crucial factors:
health and vigor, how much foliage is removed, timing during the growing season, abundance of
pathogens and insects, and the number of successive years of defoliation. In this study, the timing
of the harvest determined the percentage of leaf area removed, which directly affected the health
and vigor of the tree. Refoliation in mid summer firrther reduced vigor, for even had the regrowth
not been harvested in October, rabbits and winter dieback would have claimed the immature stems.
When repeated a second year, defoliation resulted in higher winter mortality. Variation between
harvest dates can be explained by two major factors: whether defoliation was total (in June it was
not), and whether refoliation occurred (in September it did not).

Black locust plants cut in June suffered only partial defoliation (25 - 75 %), so had much
less intemrption of growth than plots cut later. Recovery was not delayed by removal of all active
apices, as might be expected, because the lower branches were not damaged. A few vertical stems
escaped harvest, and most horizontal branches assumed a vertical orientation to become new
leaders. In addition to this process, several modifications of plant metabolism and of the external
environment that can contribute to rapid recovery from partial defoliation were examined by
McNaughton (1983). These included more rapid photosynthesis in remaining leaves, mobilization
of stored carbohydrates and redirection of current photosynthate to produce new growth, and the

greater availability of cytokinin hormones, nutrients, water, and sunlight per unit of remaining leaf

53
area. Heichel and Turner (1983) documented increased C02 assimilation rates in residual
primary (30 - 60% higher) and regrowth (50% higher) foliage of defoliated red oak and red maple.
Partially defoliated plants of the legume Vign_a m_ungg (L.) Hepper compensated for foliage losses
by growing new leaves and increasing the areas and photosynthetic rates of remaining leaves
(Pandey and Singh 1981). Increased photosynthetic rates and leaf areas have also been reported in
remaining foliage of common bean (Phaseolus v_u_lga_ri§ L.) (Alderfer and Eagles 1976). It is
possible that black locust displays similar improvements in photosynthetic efficiency after partial
defoliation, in addition to the observed production of new leaves. Although the low growth rates of
refoliated shoots argue against such compensation mechanisms having a major effect on yields,
they may have been important for winter survival.

Another factor in the observed reduction of vigor following mid summer harvests, which
was not addressed by previous researchers, could be the effect of temperature on carbohydrate
consumption. Plots cut in June retained some foliage, but plots out later in the summer lost all
their leaves. Unshaded stumps and roots became much hotter, enhancing respiration in the
remaining stems, roots, and rhizobial nodules. Pawlick (1989) warned against coppicing at the
hottest time of the year, because plants already affected by drought will recover poorly from the
added stress. Although drought was not a factor in these trials after 1988, average daily

temperatures were lower in September than in July or August each year.

Applications for forage management

It can be concluded that in central Michigan, July and especially August harvests are more
stressful to black locust plants than June or September harvests, resulting in lower survival and
reduced yields the next year. Plants cut in July and August are less likely to have adequate
carbohydrate reserves for winter survival or vigorous growth the next spring. They expend much
of their reserves to refoliate, but do not have time to collect an adequate return on the investment.
They miss out on sunshine with their canopy destroyed twice, first by harvesting and again by the

first frost. They suffer greater losses from respiration in sunny weather, and can lose all their new

54
growth to herbivores or winter dieback. Though survival in plots harvested twice, in June and
September, may equal that in plots out only in September, the yield would be much lower.
Therefore, to obtain the highest yields each year and maintain adequate survival, a green biomass
crop of black locust coppice growth should be harvested as late in the growmg season as possible,
but before any leaves are killed by frost.

55

REFERENCES

Alderfer, R. G., and C. F. Eagles. 1976. The effect of partial defoliation on the growth and
photosynthetic efficiency of bean leaves. Bot. Gaz. 137 :35 1-355.

Barrett, R. P. 1992. Forage plantation management for black locust. pp. 259-272 in: Hanover, J.
W., K. Miller, and S. Plesko, eds. Proceedings: lntemational Conference on Black Locust:
Biology, Culture, & Utilization. Department of Forestry, Michigan State University.

Bencat, T. 1992. Nutrient content of Robinia pseudoacacia leaves in Slovakia. Nitrogen Fixing
Tree Res. Reports 10:190.

 

Blair, G., D. Catchpoole, and P. Home. 1990. Forage tree legumes: their management and
contribution to the nitrogen economy of wet and humid tropical environments. Advances in
Agronomy 44:27-54.

Blake, T. J. 1983. Cappice systems for short-rotation intensive forestry: the influence of cultural,
seasonal, and plant factors. Aust. For. Res. 13:279-291.

Bray, C. M. 1983. Nitrogen metabolism in plants. Longrnan, London.

Brewbaker, J. L. 1987. Leucaena: a multipurpose tree genus for tropical agroforestry. pp. 289-323
in: Steppler, H. A., and P. K. R. Nair, eds. Agroforestry: a decade of development.
ICRAF, Nairobi.

Evensen, C. I. 1985. Effects of season on forage yield variation in Leucaena. Leucaena research
reports 6:88-90.

Feller, U. 1990. Nitrogen remobilization and protein degradation during senescence. pp. 195-222
in: Abrol, Y. P., ed. Nitrogen in higher plants. John Wiley & Sons Inc., New York.

F erraris, R. 1979. Productivity of Leucaena leucocephala in the wet tropics of North Queensland.
Tropical Grasslands 13:20-27.

Galang, M. C., R. C. Gutteridge, and H. M. Shelton. 1990. The effect of cutting height and
frequency on the productivity of Sesbania sesban var. nubica in a subtropical environment.
Nitrogen Fixing Tree Res. Reports 82161-164.

Gregory, R. A., and P. M. Wargo. 1986. Timing of defoliation and its effect on bud development,
starch reserves, and sap sugar concentration in sugar maple. Can. J. For. Res. 16:10-17.

Guevarra, A. B., A. S. Whitney, and J. R. Thompson. 1978. Influence of intra-row spacing and
cutting regimes on the growth and yield of Leucaena. Agronomy Journal 70: 1033-1037.

Hardesty, L. H., and T. W. Box. 1988. Defoliation impacts on coppicing browse species in north-
east Brazil. Journal of Range Management 41:66—70.

56

Hardesty, L. H., T. W. Box, and J. C. Malechek. 1988. Season of cutting affects biomass
production by coppicing browse species of the Brazilian caatinga. Journal of Range
Management 41 :477-480.

Heichel, G. H., and N. C. Turner. 1976. Phenology and leaf growth of defoliated hardwood trees.
pp. 31-40 in: Anderson, J. F., and H. K. Kaya, eds. Perspectives in forest entomology.
Academic Press, New York.

Heichel, G. H., and N. C. Turner. 1983. C02 assimilation of primary and regrowth foliage of red
maple (Acer rubnrm L.) and red oak (Qgercus rubra L.): response to defoliation.
Oecologia 57:14-19.

Hill, G. D. 1971. Leucaena leucocephala for pastures in the tropics. Herbage Abstracts 41:111-
119.

Hinckley, T. M., R. Ceulemans, J. M. Dunlap, A. Figliola, P. E. Heilrnan, J. G. Isebrands, G.
Scarascia-Mugnozza, P. J. Schulte, B. Smit, R.F. Stettler, E. VanVolkenburgh and B. M.
Wiard. 1989. Physiological, morphological and anatomical components of hybrid vigor in
Populus. In: K.H. Kreeb, H. Richter and T. M. Hinckley (eds), Structural and functional
responses to environmental stresses, pp. 199-217, SPB Academic Publish, The Hague,
The Netherlands.

Hintze, J. L. 1987. Number cruncher statistical system user's manual, version 5.0. J. L. Hintze,
Kaysville, UT.

Horne, P. M., D. W. Catchpoole, and A. Ella. 1985. Cutting management of tree and shrub
legumes. in: Blair, G. J., D. A. Ivory, and T. R. Evans, eds. Forages in Southeast Asian
and South Pacific agriculture. Australian Center for Intl. Agr. Research, Proceedings No.
12.

Isarasenee, A., H. M. Shelton, R. M. Jones, and G. A. Bunch. 1983. Accumulation of edible
forage of Leucaena leucocephala cv. Peru over late summer and autumn for use as a dry
season feed. Leucaena research reports 4:3.

Jeyaraman, S., S. Purushothaman, and M. Govindaswamy. 1989. Optimum clipping interval for
Leucaena leucocephala var. K8 under sewage effluent irrigation. Leucaena research
reports 10:26.

Kinch, D. M., and J. C. Ripperton. 1962. Koa haole production and processing. Hawaii
Agricultural Experiment Station, University of Hawaii, Bulletin 129.

Kramer, P. J. 1937. Photoperiodic stimulation of growth by artificial light as a cause of winter
killing. Plant Physiology 12:881-883.

Krishnamurthy, K., and M. K. Mune Gowda. 1982. Effect of cutting and frequency regimes on the
herbage yield of Leucaena. Leucaena research reports 3:31-32.

Krishnamurthy, K., and M. K. Mune Gowda. 1983. Forage yield of Leucaena var. K8 under
rainfed conditions. Leucaena research reports 4:25.

57

Linder, S. 1985. Potential and actual production in Australian forest stands. pp. 11-35 in: Bowen,
GO, and E. K. S. Narnbiar, eds. Nutrition of plantation forests. Academic Press.

Linder, S., R. E. McMurtrie, and J. J. Landsberg. 1985. Growth of eucalypts: a mathematical
model applied to Eucalyptus globulus. pp. 117-126 in: Tigerstedt, P. M. A., P. Puttonen,
and V. Koski, eds. Crop physiology of forest trees. University of Helsinki, Helsinki,
Finland.

McNaughton, S. J. 1983. Compensatory plant growth as a response to herbivory. Oikos 40:329-
336.

Mebrahtu, T. 1992. Growth and photosynthesis of black locust. pp. 149-170 in: Hanover, J. W., K.
Miller, and S. Plesko, eds. Proceedings: lntemational Conference on Black Locust:
Biology, Culture, & Utilization. Department of Forestry, Michigan State University.

Mebrahtu, T., and J. W. Hanover. 1991. Variations in coppicing and root sucker formation in
black locust. Nitrogen Fixing Tree Res. Reports 9: 1 19-121.

Meginnis, H. G. 1940. Effect of top pruning on survival and early growth of black locust. Joumal
of forestry 38:30-36.

Mendoza, R. C., T. P. Altamarino, and E. O. Javier. 1983. Herbage, crude protein, and digestible
dry matter yield of ipil-ipil (I_. leucocephala cv. Peru) in hedgerows. Leucaena research
reports 4:69.

Monteith, J. L. 1977. Climate and efficiency of crop production in Britain. Phil. Trans. Royal
Society of London B. 281:277-294.

Oakes, A. J ., and O. Skov. 1962. Some woody legumes as forage crops for the dry tropics. Trop.
Ag. (Trinidad) 39:281-287.

Osman, A. M. 1981a. Effects of cutting interval on the relative dry matter production of four
cultivars of Leucaena. Leucaena research reports 2:33-34.

Osman, A. M. 1981b. Leafzstem ratio in Leucaena. Leucaena research reports 2:35-36.

Osman, A. M. 1986. Leafzstem ratio in Leucaena - a final report. Leucaena research reports 7 :91-
93.

Pandey, R. K., and V. B. Singh. 1981. Influence of partial defoliation on dry-matter production
and seed yield of urd bean (Vigpa mungo L. Hepper). J. Agr. Sci., Cambridge 97:437-443.

Parker, J. 1981. Effects of defoliation on oak chemistry. pp. 219-225 in: Doane, C. C., and M. L.
McManus, eds. The gypsy moth: research towards integrated pest management. Forest
Service Technical Bulletin 1584, USDA, Washington, DC.

Pawlick, T. 1989. Coppice with care. Agroforestry Today 1(3): 15-17.

58

Pokhriyal, T. C., S. P. Chaukiyal, U. Singh, and G. S. Bisht. 1992. Effect of defoliation on
acetylene reduction activity in Dalbergia sissoo. Nitrogen Fixing Tree Res. Reports
10:155-157.

 

Rogers, 8., and R. Rosecrance. 1992. Coppice management of Paraserianthes falcataria in Western
Samoa. Nitrogen Fixing Tree Res. Reports 10:178-179.

Soil Conservation Service, 1979. Soil survey of Ingham County, Michigan. USDA, Washington
DC.

Takahashi, M., and J. C. Ripperton. 1949. Koa haole (Eucaena glauca): its establishment,
culture, and utilization as a forage crop. Bulletin 100, Hawaii Agr. Exp. Station.

Topark-Ngamr, A. 1983. Testing shrub legumes for forage crop in Northeast Thailand. Leucaena
research reports 4:77-78.

Wareing, P. F., and D. L. Roberts. 1956. Photoperiodic control of cambial activity in Robinia
pseudoacacia L. New Phytologist 55:356-367.

Wargo, P. M. 1981a. Defoliation and tree growth. pp. 225-240 in: Doane, C. C., and M. L.
McManus, eds. The gypsy moth: research towards integrated pest management. Forest
Service Technical Bulletin 1584, USDA, Washington, DC.

Wargo, P. M. 1981b. Defoliation, dieback, and mortality. pp. 240-248 in: Doane, C. C., and M. L.
McManus, eds. The gypsy moth: research towards integrated pest management. Forest
Service Technical Bulletin 1584, USDA, Washington, DC.

EFFECTS OF HARVEST HEIGHT ON SURVIVAL AND FORAGE YIELD OF
BLACK LOCUST

ABSTRACT

Black locust seedlings planted at five spacings in 1988 were harvested for forage in two
trials. Plots were cut twice each year in late June (or early July), and in late August, from 1988 to
1990. Harvest height treatments were at 5 cm and 20 - 30 cm. In August, 1988, harvesting at 5
cm gave significantly higher forage yields per plot and per plant than at 20 cm. Winter mortality
was significantly greater at 5 cm (73%) than at 20 cm (under 10%). This trial was discontinued in
1989. Harvesting during the year of establishment should not be practiced.

In the second trial, dormant intact stems were cut at 5 cm or 20 cm in March, 1989.
Harvests at 5 cm gave equal or lower yields than at 25 - 30 cm, and cumulative yields for 1989 and
1990 were significantly reduced. Yields declined from 1989 to 1990, but more severely at 5 cm
(74%) than at 25 - 30 cm (20%). Yields per plant were significantly greater when out at 30 cm,
and increased from 1989 to 1990.

The effect of harvest height on dry weight percent of fresh weight was slight but sometimes
statistically significant. Mortality was low for the 20 - 30 cm plots before 1991, and for the 5 cm
plots between harvests in the summer. Mortality was high over the winters for 5 cm plots, and
from 1990 to 1991 at 5 cm (84%) and 30 cm (56%). Harvesting at 5 cm is not recommended, due

to reduced future survival and yields.

59

60
INTRODUCTION

Literature review

Black locust (Mini; pseudoacacia L.) is a nitrogen-fixing leguminous tree, suitable for
cultivation for biomass and high-protein forage. It grows rapidly from seed or coppice, and has
often been recommended for short-rotation intesive culture. When tree species are grown at close
spacings and cut at short intervals, the stumps tend to die quickly. Cultural procedures are a major
factor in the health and productivity of coppice stands (Blake 1983). Because little previous
research on black locust is applicable to forage production, this review includes much data from
leguminous trees in the tropics, especially concerning the cultivation of leucaena (Qucaena
leucocephala (Lam) de Wit) for forage.

Meginnis (1940) observed that cutting dormant black locust seedlings at ground level
significantly increased mortality compared to pruning at 20 cm. Survival of seedlings cut at 20 cm
was not significantly different from uncut controls. Mebrahtu and Hanover (1991) compared
regrowth of black locust seedlings cut at 1 cm or 25 cm. Height did not affect the timing of
coppice sprouts from the stumps in spring. None of the 25 cm stumps had sucker sprouts from the
roots, but all the 1 cm stumps did, averaging 7.9 suckers per stump. Most suckers survived to the
end of the season. Above ground dry weight was 14.3 tons/ha/year for both heights.

Many researchers have reported on yields of leucaena leaves in relation to harvest height,
and in six cases the highest yields were obtained at the highest height tested: 30 cm (Pathak L211.
1980), 50 cm (Petheram ital. 1982), 90 cm (Ezenwa and Cobbina 1991), 100 cm (Pathak and
Patil 1983; Sampet 1992), and 120 cm (Isarasenee L211. 1984). The highest biomass yields
(leaves and stems combined) occurred at the greatest cutting height tested in seven reports: 75 cm
(Relwani 3311. 1983), 90 cm (Ezenwa and Cobbina 1991), 100 cm (Field and OeMatan 1990),
120 cm (Mohatkar and Relwani 1985), and 150 cm (Krishnamurthy and Mune Gowda 1982,

1983). Mendoza et al. ( 1983) obtained fodder dry weight yields of 10.7 tons/ha/year at 15 cm,

6 1
15.8 at 150 cm, and 23.6 at 300 cm. This is the only available report of a harvest height above
150 cm.

F our leucaena researchers have reported maximum yields at less than the maximum
harvest height. Osman (1981) found an optimal harvest height below 150 cm. Harvest between 45
and 90 cm was recommended since yields at those heights did not significantly differ. Tukel and
Hatipoglu (1989) tested 20, 40, and 60 cm harvest heights, obtaining significantly higher leaf
yields at 40 cm the first year and at 60 cm the second year. The average yields of leaves and total
biomass did not vary significantly between 40 and 60 cm. Ezenwa and Cobbina (1991) found
higher yields of cv. Cunningham at 90 cm than at 45 or 15 cm (not significant) during the first year
of harvesting only. In the second year, the 45 cm height yielded most, and significantly more than
at 15 cm. Horne and Blair (1991) found no significant differences in yields of leaves or biomass,
or the leaf percentage, in cv. Cunningham planted at 20,000/ha in Indonesia when harvested at 30
or 100 cm at irregular intervals of 43 to 64 days. It was proposed that the long establishment
phase (four years) before the first cut allowed the accumulation of large carbohydrate reserves in
the roots, contributing to the rapid recovery fi'om the initial cut at 30 cm.

Several authors have summarized previous research and made harvest height
recommendations for leucaena as a forage crop. Blair _eLa_1. (1990) cited nine earlier reports and
put the optimal range at 90 - 120 cm, affinning that promoting branching for a solid canopy and
timely harvest before lower leaves are shaded out appear important for high yields. Hegde (1983)
recommended 90 - 100 cm for hand harvest, but 30 - 50 cm for mechanized harvest. Brewbaker
(1987) reconunended hedges be maintained at above 60 cm to retain some foliage and vigorous
side branches, shrubby Hawaiian types be cut as low as 10 em, but arboreal types be cut above 25
cm. NFTA (1985) advised starting harvesting hedges at 25 cm to stimulate branching, and slowly
raising the cutting level to the recommended 80 - 100 cm. Cutting the hedge lower again afler two

or three years was said to rejuvenate the plants.

62

In addition to leucaena, the effects of harvest heights have been investigated in other
tropical leguminous trees. Leaf, stem, and total dry weights varied significantly among three
species in Samoa, but not between harvest heights of 100 and 150 cm, except for the higher leaf
weight at 150 cm in Gliricidia m (Jacq.) Walp. (T ekle-Haimanot $11. 1991). Desmodium
wigs, a shrub from South Asia, was tested in Belize at harvest heights of 5, 25, and 50 em,
spaced 100 cm apart, with three harvest intervals. Leaf, stem, and total dry weights did not differ
between the 25 and 50 cm heights, which always yielded significantly more than the 5 cm height.
Yields at 5 cm did not recover greatly after the dry season, and were reduced by weeds. Survival
rate increased with harvest height from 25 to 50 em, but not significantly (Lazier 1981).

A_cac_ia MM Willd. trees in China, averaging 9 m tall and 8 cm dbh, were cut at O, 30,
6O , and 100 cm above ground level. Harvesting at ground level gave significantly lower results
for all features measured: percent of stumps sprouting, number of sprouts per stump, length and
diameter of coppice shoots, and 12-month stump survival. Significant differences among the 30,
60, and 100 cm heights were found for number of sprouts per stump and 12-month stump survival,
with the highest levels recorded at 100 cm (Huang 1989).

Eight species of Acacia native to Australia were coppiced after growing four years in

 

Malawi. Survival six months later was 7% if cut at 10 cm, 29% at 30 cm, and 40% at 50 cm
(Ngulube 1991). Paraserianthes falcataria (L.) Nielsen always produced more biomass when

harvested at 100 cm than at 30 cm in atrial in Western Samoa (Rogers and Rosecrance 1992).

Obiectives

If black locust is to be grown as a woody perennial intercrop or field crop for green
biomass or forage, the optimal mechanized harvest height must be determined. Field trials must be
repeated for more than one season, to determine how the harvest height affects survival and yields.
The dry weight percentage, important for preservation as silage, may vary with height. The inverse
relation of individual plant biomass to stand density has been well documented for some woody

species at short rotations of one to seven years (Cannell and Smith 1980), but harvesting leaves

63
twice a year has rarely been investigated by foresters. The optimal density, defined as the lowest
planting density to give the maximum yield, may vary with height and age. Experience with other
species, primarily arboreal cultivars of leucaena, indicates that the ideal harvest height for black
locust may be found in a similar range of 90 - 150 cm. This was not considered practical for
mechanized harvest with conventional equipment designed for hay or silage, so this experiment was
conducted with lower harvest heights (5 cm and 20 - 30 cm) that could be attained by hay
harvesting equipment commonly used on Michigan farms.

An experiment was designed with two harvest heights, at five population densities, to
address the following hypotheses.

For black locust plants of up to four years of age:

1. Optimal density decreases with harvest height.

2. Optimal density decreases with time, but the pattern differs according to harvest height.

3. Survival rate increases with harvest height.

4. Survival rate increases with time for all harvest heights.

5. Yield per plant increases with harvest height.

6. Yield per plant increases with time for all harvest heights.

The experiment was also used to test the response to different harvest heights in plants of
difl‘erent ages. It was thought that stems of swdling and coppice origin would differ in their yields,
numbers harvested, and dry weight percentages, and would respond differently to the harvest height
treatments. These responses were monitored in order to address the following hypotheses.

For intact first year seedlings:

7. Dry weight yields per plot decline with height.

8. Dry weight yields per plant decline with height.

9. Number of livivng plants harvested declines with height.

10. Dry weight percentage declines with height.

For regrowth from coppiced stumps:

64
11. Dry weight yields per plot increase with height.
12. Dry weight yields per plant increase with height.
13. Number of livivng plants harvested increases with height.

14. Dry weight percentage increases with height.

MATERIALS AND METHODS

general information for all trials
A bulk mixture of equal numbers of swds from four Michigan sources was planted. The

sources were selected on the basis of availability, and not for any judgement of quality. The seed
accessions were 442 fi'om Russ Forest in Cass County, 445 from the comer of Sandhill and
Hagadorn roads, and 446 fiom the Dansville State Game Area, both in Ingham County, and 450
fi'om Kellogg Forest in Kalamazoo County. Prior to sowing, the seeds were scarified in
concentrated sulfuric acid for 50 minutes. They were directly sown in late June 1988 at the
Michigan State University Tree Research Center in East Lansing (420 41' N, 840 28' W). The
nursery beds were surrounded by evergreen windbreaks and had been fumigated with methyl
bromide. The soil was mapped as a complex of soil series including Riddles, Hillsdale, Owosso,
and Marlette, all of which were well drained with 2% to 6% slopes and sandy loam topsoil.
Subsoils were sandy loam, loam, or clay loam (Soil Conservation Service 1979).

Prolonged heat and drought in 1988 impaired growth and sprinkler irrigation was
necessary. After normal weather returned in mid August, growth was rapid until early September,
with seedlings reaching heights of 60-120 cm. In 1989, 1990, and 1991 growing conditions were
favorable.

Because the soil had been fumigated, there were few weed problems in 1988. In the spring
of 1989 and 1990, the plots were sprayed with glyphosate (trade name Roundup) at 0.7 kg active

ingredient/ha and DCPA (trade name Dacthal) at 8.4 kg active ingredient/ha. Potassium sulfate (0-

65
0-50) fertilizer was added at the rate of 225 kg/ha (112 kg/ha K20) on May 7, 1990, to bring soil
potassium levels up from around 90 kg/ha. No other fertilizer was applied.

In each plot, all new growth above the designated harvest height was cut with hand tools
and bagged to detemiine the flesh weight. The paper bags of harvested material were dried in an
oven at 600C for at least three days before measuring the dry weight. Weights per plant were
calculated from the number of plants harvested. Multiple sprouts from the same root were counted
as one plant. Shoots fi'om different roots were counted as separate plants. Due to the presence of
suppressed seedlings, the total survival in 1989 was often higher than the number harvested, but
few suppressed seedlings survived to 1990. Survival to 1991 was often higher than the number
harvested that year because of rabbit damage on the smallest plants.

All trials were planted using a randomized complete block design. Analysis of variance,
Duncan's multiple range tests, and other statistical operations were conducted on NCSS computer
programs (Hintze 1987). The probability level reported for each factor below a table is the
probability of the difference between treatments occurring by chance, according to F-test

comparisons.

Harvest height trials

Four replications were planted in a randomized complete block design. Ten plots were
included in each replication, representing all combinations of the two harvest heights and the five
spacings. In June, 1988, seeds were planted by hand in a grid pattern, 10, 12.5, 16.7, 25, or 50 cm
apart, to give 1.0 m2 plots of 100, 64, 36, 16, or 4 seedlings, respectively. With one or two border
rows at the same spacings on all sides, the planted area was 1.5 m wide and covered 2.25 m2 per
plot. Multiple seedlings were common because two to four seeds were sown at each grid position.
Empty grid positions were replanted with seeds in early July, and with seedlings in early August.
The effect of spacing treatments was included in the analysis of variance, but it will not be reported

here unless the interaction with harvest height was statistically significant.

66

In late August 1988, 65 days from sowing, the plots were harvested at either 5 cm or 20
cm above the ground. At this time the seedlings were mostly 20 - 30 cm tall. Regrowth was weak,
and was consumed by rabbits during the winter. Most 5 cm stumps died and were frost-heaved out
of the ground before spring. Survival to 1989 was recorded, and the trial was terminated.

A duplicate set of plots was established as described for 198 8, but not harvested that year.
In March, 1989, the dormant plants were cut back to 5 or 20 cm, and the stems left to decay on the
ground. Coppice shoots fi'om three of the four replications were harvested at 5 and 25 cm between
June 28 and July 5, and again at 5 and 30 cm between August 23 and 31, 1989. The fourth
replication was discarded because of variations in treatment.

In 1990 first harvests occurred between June 18 and 22, and second harvests between
August 16 and 27, all at 5 and 30 cm. In 1991 the survivors were counted for both heights, and

the trial was terminated.

RESULTS AND DISCUSSION

Relation of ogtimal densig to harvest height

To investigate the relation of harvest height to optimal population density, five spacing
treatments were combined with the two height treatments. The hypothesis that optimal population
density declines with harvest height was not supported. Rather than decreasing, the optimal
density increased with harvest height in 1988, from 64/m2 at 5 cm to lOO/m2 at 20 cm (Table 3.1).
In the other trial, the interaction between height and density was never statistically significant, so
there was no variation in optimal density with height in 1989 and 1990. The pattern of change in

Optimal density did not vary with harvest height.

DE! weight yields
In 1988, harvesting at 5 cm about doubled the yields obtained at 20 cm (Table 3.1).

Yields generally increased with density, but when cut at 20 cm, yields showed a plateau from 16 to

67
Table 3.1. Dry weight yields (g/mz) for square grid plots at five planting densities harvested at 5
cm and 20 em in late August 1988, 65 days from sowing.

 

 

 

 

 

Planting density Harvest height
per square meter 5 cm 20 cm
4 22a 16a
16 125b 72b
36 147b 67b
64 261C 80b
100 245C 144C
Mean 160 76
Factor Probability
Harvest height .0000
Planting density .0000
Interaction .0002

 

Means in the same column followed by the same letter are not significantly different at the 0.05
level according to Duncan's Multiple Range test. Probability levels of the effects of factors and
their interaction were determined by F test.

68
64/m2. The treatment factors and their interaction gave highly significant results. These plots
could not be compared again because of poor survival at 5 cm.

In the other trial, the effects of harvest height and planting density treatments on yield were
highly significant in 1989 and 1990, but the treatment interaction was not statistically significant.
In 1989 the 5 cm harvests removed all regrowth, but the higher harvests did not. When cut at 25
cm in July and at 30 cm in August, 5 cm of new growth was retained above the stump each time.
However, the 5 cm harvests did not result in higher yields (Table 3.2). For the plots cut at 25 cm
in July, 1989, the greater number of retained leaves gave an advantage for regrth before the
August harvest, which was nullified by harvesting the new growth 5 cm higher than the July
stumps. If all regrowth had been harvested in August, the higher cutting height probably would
have yielded significantly more at every harvest. In 1990 and in total for two years, yields were
significantly higher at 30 cm. Two year totals were 589 g/m2 at 5 cm and 1112 g/m2 at 30 em.
After the 1990 season the experiment was terminated because of poor survival at 5 cm.

The hypotheses that dry weight yields per plot decline with height in intact seedlings, and
increase with height for coppice regrowth, were both confirmed. Yields declined from 1989 to
1990, although growing conditions were similar and fertilizer was added at the start of the 1990
season. Forage production was reduced more severely in the 5 cm (73%) than in the 30 cm plots
(20%).

Possible factors giving an advantage to greater harvest heights are the vigor of regrowth
due to greater carbohydrate reserves in remaining stems (Meginnis 1940; Hardesty and Box 1988),
and more branching (Brewbaker 1987; Blair e_t ad. 1990) to maintain at least a partial canopy after
harvest which raises photosynthesis and suppresses weeds. Total defoliation, as usually occurred
at the 5 cm height, has been shown to reduce future growth in temperate species, including English
oak (mercus robur L.) (Hilton e_t a_l. 1987), red oak (Quereus rubra; L.), and red maple (_Ac_er_
_ru_b_ru_m L.) (Heichel and Turner 1976, 1984). Defoliation by leaf miner insects can be fatal to

mature black locust if the first flush of leaves and the regrowth are destroyed for two consecutive

69
Table 3.2. Dry weight yields (g/mz) for square grid plots harvested at 5 cm and 25 cm in early
July and at 5 cm and 30 cm in late August 1989, and at 5 cm and 30 cm in late June and late
August 1990.

 

 

 

June-July August Cumulative
Harvest height Harvest height Harvest height
Year 5 cm 30 cm 5 cm 30 cm 5 cm 30 em
1989 354 504 112 115 466 619
Probability level .0002 .7955 .0002
1990 43 172 80 321 123 493
Probability level .0000 .0000 .0000
2 year total 397 676 192 436 5 89 1 112
Probability level .0000 .0000 .0000

 

The probability level for the effect of harvest height was determined by F test.

70
years (Hoffard 1992). The death of black locust seedlings, which have smaller reserves, was more
rapid in this study.

The primary factor working to reduce yields at greater heights is leaf abscission below the
canopy before harvest (Blair g a_l. 1990). This occurs when the harvest interval is too long for the
combination of height and spacing, and lower leaves become heavily shaded and senesce. In black
locust, senescent leaves turn bright yellow before dropping.

Because plot yields are influenced by the number of survivors, comparison of yields per
plant is necessary to determine whether a harvest height treatment had an effect on the vigor of the
survivors. Because each remaining plant has more space after some die, a slight increase in yield
per plant would be expected each year.

Dry weight yields per plant of intact seedlings declined with height, as predicted. In 1988,
the average dry weight yield per plant at 5 cm was 4.1 g, significantly higher than the average of
2.6 g harvested at 20 cm. Yields were greater at 5 cm because all plots were the same height
before treatment. The interaction with density treatments was not significant because the seedlings
had little time to crowd each other.

The hypothesis that regrowth yield per plant increases with harvest height was clearly
supported (Tables 3.3 and 3.4), but yield per plant did not increase with time regardless of height.
At 5 cm the average yield per plant declined from 1989 to 1990, and was too low to justify
harvesting at all in 1991. The average July 1989 yield for the 5 cm group was 12.8 g per plant,
significantly lower than the average of 20.3 g per plant harvested at 25 cm. At the second harvest
in August 1989, the effect of height was not significant, but the interaction of height and density
treatments was (Table 3.3). This was because yields per plant were nearly equal for all but the
widest spacing (50 cm apart), which gave half the yield per plant at 5 cm as at 30 cm. The
significant interaction continued in 1990. Yields per plant at the June 1990 harvest for the 5 cm

group were about half those for the 30 cm group, and continued lower at the second harvest.

71
Table 3.3. Dry weight yields (g/plant) from square grid plots at five planting densities harvested

at 5 cm and 30 cm in late August 1989, late June 1990, and late August 1990.

 

 

 

 

 

Planting density August 1989 June 1990 August 1990
per square meter 5 cm 30 cm 5 cm 30 cm 5 cm 30 cm
4 10.23 21.03 4.1ab 33.4a 7.6ab 75.7a
16 7413 7.8b 5.0a 10.3b 11.9a 18.4b
36 3.5c 3.9b 3.5ab 6.8b 8.8ab 10.5b
(,4 2.10 2.0,, 1.9b 4.5b 3.7a 7.2b
100 2.30 1.9!, 1.8b 3.3b 2.9b 6.6b
Mean 51 7.3 3.2 11.7 7.0 23.7
Factor Probability Probability Probability
Harvest height .0801 0000 .0152
Planting density .0000 0000 .0081
interaction .0375 9000 .0133

 

Means in the same column followed by the same letter are not significantly different at the 0.05
level according to Duncan's Multiple Range test. Probability levels of the effects of factors and
their interaction were determined by F test.

72

Table 3.4. Annual total dry weight yields (g/plant) from square grid plots at five planting

densities harvested at 5 cm and 30 cm in 1989 and 1990.

 

 

 

 

Planting density 1989 1990
per square meter 5 cm 30 cm 5 cm 30 cm
4 18.03 30.0a 11.5ab 109.9a
16 12.0b 15.1b 16.8a 28.8b
36 8.4bc 11.9b 12.4ab 17 .9b
64 4.50 6.40 5.8b 12.1b
100 3.80 6.00 4.8b 10.1b
Mean 9.3 13.9 10.1 35.8
Factor Probability Probability
Harvest height .0001 .0021
Planting density .0000 .0009
Interaction .0125 .0018

 

Means in the same column followed by the same letter are not significantly different at the 0.05

level according to Duncan's Multiple Range test. Probability levels of the effects of factors and
their interaction were determined by F test.

73

For the cumulative yields per plant in 1989 and 1990, both height and spacing treatments
produced highly significant differences, and the interaction was also significant (Table 3.4). At
most spacings, total yields per plant in 1990 increased slightly over 1989 in the 5 cm plots, but
about doubled for the 30 cm plots from 1989 to 1990. The widest spaced plants at 4/m2 departed
from these patterns. In June, 1990, the 5 cm group averaged one eighth of the yield per plant in the
30 cm group (Table 3.3). In August, 1990, yields of the 30 cm plots increased to 10 times the
average of the 5 cm plots. When compared with the 1989 totals, or with the other spacings at the
same harvest height, it is clear that growth of the 4/m2 plants was retarded by harvesting at 5 cm

and promoted by harvesting at 30 cm.

Du weight gercentage
The dry weight/fresh weight percentage (D/F) represents the dry matter proportion of the

green biomass. The D/F was low early in the season when new growth was succulent, but
increased as plant tissues became lignified. Differences between height treatments were
statistically significant at both harvests in 1989, but not in 1988 or 1990 (Table 3.5), so the
hypothesis that the D/F of intact swdlings declines with height was rejected. The results from
coppice regrowth were inconclusive, since the D/F increased significantly with height in 1989, but
not in 1990. This is similar to the experience of Holden _e_t a_l. (1989), who reported the dry matter
percentage in M macrantha was 7% at 25 cm and 12% at 50 cm at one harvest, but equal at
two other harvests.

The slight difference in the WP appears to depend on the volume harvested, since the drier
group always had the greater dry weight as well (T ables 3.1 and 3.2). Greater dry weight from an
equal area requires larger stems with a higher proportion of wood, and thus a higher D/F.
Although the D/F was sometimes significantly different between treatments, this represented at
most an extra 2% moisture by weight. Under field conditions, such a difference could be erased by

a short exposure to the sun after cutting, and would be of no concern.

74
Table 3.5 . Dry weight percentage of fresh weight for square grid plots harvested at 5 cm and 20-
30 cm in 1988-1990.

 

 

 

Harvest height Probability
Date 5 cm 30 cm level
August 1988 23.8 23.2 .1838
July 1989 23.4 25.5 .0025
August 1989 23.5 24.7 .0035
Agust 1990 25.9 26.3 .5030

 

Data not suitable for comparison in June 1990 because of rain which biased fresh weight
measurements.

75
Survival after harvesting

The number of plants surviving was always equal to the number harvested at 5 em, but
this was not true for the 20 - 30 cm plots. Uncut live plants in the 20 - 30 cm plots were not
counted because they made no contribution to yield, did not shade their taller neighbors, and
probably used a negligible fi'action of the water and nutrients consumed by the entire plot.
Although the number harvested sometimes increased from one harvest to the next, it was a close
estimate of survival in the 20 - 30 cm plots fi'om 1989 onward.

At the first harvest in August, 1988, all seedlings exceeded 5 cm but about one quarter had
not yet reached 20 cm (1' able 3.6). Those not cut at 20 cm continued to grow for the rest of the
season and survived well over the winter, giving an increase of 18% over the 1988 level when
harvested in July 1989. Winter mortality was 73% for seedlings cut at 5 cm, and under 10% for
those in the 20 cm plots.

The estimate of under 10% mortality was calculated as follows. The 5 cm plots averaged
51 seedlings in August, 1988. The plots harvested at 20 cm in 1988 averaged 46/m2 harvested in
July, 1989, representing 90% of the 51/m2 presumed to have been present in 1988. An unknown
number of additional plants survived but were not tall enough to harvest.

In the other trial, mortality was low for the 30 cm plots before 1991, and for the 5 cm plots
between harvests in the summers, but high over the winters for 5 cm plots (Table 3.7). The non-
significant increase between June and August in 1990 resulted when the new growth of some weak
seedlings did not reach 30 cm by the first harvest, but passed 30 cm before the second. After four
harvests, survival was still adequate to maintain a closed canopy in the 30 cm plots, but was so
low in the 5 cm plots that the comparison trial was ended.

Survival rates increased with harvest height, as predicted, but not with plant age. Two
more hypotheses, that the number of intact seedlings harvested declines with height, and that the

number of coppiced plants cut at any single harvest increases with height, were confirmed.

76
Table 3.6. Average number of plants harvested from square grid plots at five planting densities
harvested at 5 cm and 20 cm in late August 1988, 65 days from sowing, and at 5 cm and 25 cm
in early July 1989.

 

 

 

 

 

Planting density August 1988 July 1989

per square meter 5 cm 20 em 5 cm 25 cm
4 4.0 3.3 1.3a 4.0a
16 21.0 21.5 6.3ab 27.3b
36 40.3 32.3 13.5bc 42.3bc
64 75.0 47.0 21.0cd 58.00
100 112.5 88.3 _ 26.0d 96.5d

Mean 50.6 38.5 13.6 45.6

Factor Probability Probability

Harvest height .0002 .0000

Planting density .0000 .0000

Interaction .0070 .0000

 

Means in the same column followed by the same letter are not significantly different at the 0.05
level according to Duncan's Multiple Range test. Probability levels of the effects of factors and
their interaction were determined by F test.

77
Table 3.7. Average number of plants harvested from square grid plots at 5 cm and 25 or 30 cm
starting a year afier sowing (three replications).

 

 

Date 5 cm 25-30 cm Probability
July 1989 40.9 38.0 .0592.
Aug. 1989 32.8 32.6 .8501
June 1990 14.7 27.3 .0458
Aug. 1990 13.8 ' 30.3 .0406
July 19911 2.2 13.5

 

1 Counted but not harvested.

78
At both heights, mortality up to July 1989 was three to four times higher if the first cut
was made in August 1988 compared with March 1989. It is likely that delaying the first cut even
longer would result in larger stronger plants and improved survival (Pawlick 1989). The same
outcome would probably result fi'om harvesting at heights above 30 cm. Holden £31. (1989)

reported 100% survival in Sesbania macrantha coppiced at 50 em, but not at 25 cm or 10 cm.

 

Large 5933 mangium trees cut at 30 cm survived significantly better than those cut at ground
level, but not as well as those cut at 60 or 100 cm (Huang 1989). Survival of eight species of
m after six months was 7% if cut at 10 cm, 29% at 30 cm, and 40% at 50 cm (Ngulube
1991). Although survival of black locust seedlings cut at 20 cm by Meginnis (1940) was not
significantly difi‘erent fiom uncut controls, those seedlings were widely spaced and were out only
once. Under the strain of crowding and repeated harvesting, it is likely that black locust survival
would be improved by harvest heights above 30 cm.

CONCLUSIONS

Because of severe mortality the first winter, a harvest height of 5 cm or below is
recommended only if the plantation of black locust is treated as an annual crop. For every aspect
of growth, yield, and survival that was measured, the harvest height of 30 cm is superior to 5 cm in

the long term. It is reasonable to project improved performance with heights above 30 cm.

79
REFERENCES CITED

Blair, G., D. Catchpoole, and P. Home. 1990. Forage tree legumes: their management and
contribution to the nitrogen economy of wet and humid tropical environments. Advances in
Agronomy 44227-54.

Blake, T. J. 1983. Coppice systems for short-rotation intensive forestry: the influence of cultural,
seasonal, and plant factors. Aust. For. Res. 13:279-291.

Brewbaker, J. L. 1987. Leucaena: a multipurpose tree genus for tropical agroforestry. pp. 289-3 23
in: Steppler, H. A., and P. K. R Nair, eds. Agroforestry: a decade of development.
ICRAF, Nairobi.

Cannell, M. G. R., and R. 1. Smith. 1980. Yields of minirotation closely spaced hardwoods in
temperate regions: review and appraisal. Forest Science 26:415-428.

Ezenwa, I. V., and J. Cobbina. 1991. Strategic pruning of leucaena for optimum herbage yield.
Leucaena research reports 12:75-77.

Field, S. P., and S. S. OeMatan. 1990. The effect of cutting height and pruning fi'equency of
Leucaena leucocephala hedgerows on maize production. Leucaena research reports 11:68-
69.

Gregory, R. A., and P. M. Wargo. 1986. Timing of defoliation and its effect on bud development,
starch reserves, and sap sugar concentration in sugar maple. Can. J. For. Res. 16:10-17.

Hardesty, L. H., and T. W. Box. 1988. Defoliation impacts on coppicing browse species in north-
east Brazil. Journal of Range Management 41:66-70.

Hegde, N. 1983. Leucaena forage management in India. pp. 73 -78 in: Leucaena research in the
Asian - Pacific region. IDRC 211e, Ottawa.

Heichel, G. H., and N. C. Turner. 1976. Phenology and leaf growth of defoliated hardwood trees.
pp. 31-40 in: Anderson, J. F ., and H. K. Kaya, eds. Perspectives in forest entomology.
Academic Press, New York.

Heichel, G. H., and N. C. Turner. 1984. Branch growth and leaf numbers of red maple (Acer
rubrum L.) and red oak (Quereus rubra, L.): response to defoliation. Oecologia 62: 1-6.

Hilton, G. M., J. R. Packham, and A. J. Willis. 1987. Effects of experimental defoliation on a
population of pedunculate oak (Quereus robur L.). New Phytol. 107:603-612.

Hintze, J. L. 1987. Number cruncher statistical system user's manual, version 5.0. J. L. Hintze,
Kaysville, UT.

Hoffard, W. 1992. Insect pests of black locust. pp. 44-49 in: Hanover, J. W., K. Miller, and S.
Plesko, eds. Proceedings: lntemational conference on Black Locust: Biology, Culture, &
Utilization. Department of Forestry, Michigan State University.

80
Holden, S. R, S. Lungu, and J. Volk. 1989. Growth rate and coppicing ability of Sesbania

macrantha after repeated prunings at various heights. Nitrogen Fixing Tree Res. Reports
7: 132-134.

Home, P. M., and G. J. Blair. 1991. Forage tree legumes IV: productivity of Leucaena/grass
mixtures. Aust. J. Agric. Res. 42:1231-1250.

Hun-1g, S. 1989. Effect of coppicing at different heights on regeneration of Acacia mangium Willd.
Nitrogen Fixing Tree Res. Reports 7:52-54.

Isarasenee, A., H. M. Shelton, R. M. Jones, and G. A. Bunch. 1984. Accumulation of edible

forage of Leucaena leumhala cv. Peru over late summer and autumn for use as a dry
season feed. Leucaena research reports 5:3.

Kl‘islinamurthy, K., and M. K. Mune Gowda. 1982. Effect of cutting and frequency regimes on the
herbage yield of Leucaena. Leucaena research reports 3:31-32.

I(I‘iShnamurthy, K., and M. K. Mune Gowda. 1983. Forage yield of Leucaena var. K8 under
rainfed conditions. Leucaena research reports 4:25.

L«azier, J. R. 1981. Efi‘ect of cutting height and frequency on dry matter production of
Codariocaly; moides (syn. Desmodium gyloides) in Belize, Central America. Tropical
Grasslands 15:10-16.

Mebrahtu, T., and J. W. Hanover. 1991. Variations in coppicing and root sucker formation in
black locust. Nitrogen Fixing Tree Res. Reports 9: 1 19-121.

lNileginnis, H. G. 1940. Effect of top pruning on survival and early growth of black locust. Journal
of Forestry 38:30-36.

Mendoza, R. C., T. P. Altamarino, and E. O. Javier. 1983. Herbage, crude protein, and digestible

dry matter yield of ipil-ipil (I_. leucocephala cv. Peru) in hedgerows. Leucaena research
reports 4:69.

Mohatkar, L. C., and L. L. Relwani. 1985. Effect of plant population, stubble height, and number

of cuttings on the growth, seed, forage, and firewood production of Leucaena K8.
leucaena research reports 6:40-41.

NFTA. 1985. Leucaena forage production and use. Nitrogen Fixing Tree Association, Waimanalo,
HI.

Ngulube, M. 1991. Evaluation of coppicing ability of promising Austalian acacias for dryland
agroforestry in Malawi. Nitrogen Fixing Tree Res. Reports 9:81-82.

Osman, A. M. 1981. The effect of cutting height on the dry matter production of Leucaena.
Leucaena research reports 2:37-38.

Pathak, P. S., and B. D. Patil. 1983. Leucaena research at the Indian Grassland and Fodder

Research Institute (IGFRI). pp. 83-88 in: Leucaena research in the Asian - Pacific region.
IDRC 2116, Ottawa.

81

Pathak, P. s., P. Rai, and R. Deb Roy. 1980. Forage production from koo-babool (L leucocephala

(Lam) De Wit.). 1. Effect of plant density, cutting intensity and interval. Forage Research
6:83-90.

Pawlick, T. 1989. Coppice with care. Agroforestry Today 1(3): 15-17.

Petheram, R J ., M. Pandjaitan, and J. Liano. 1982. Leucaena forage production trials based at
Ciawi, Indonesia. Leucaena research reports 3:45-46.

Relwani, L. L., L. C. Mohatkar, G. K. Bhavsar, and R. V. Nakat. 1983. Effect of spacings and
stubble heights on forage yields of K-8 and Cunningham. Leucaena research reports 4:41.

Rogers, S., and R. Rosecrance. 1992. Coppice management of Paraserianthes falcataria in Western
Samoa. Nitrogen Fixing Tree Res. Reports 10:178-179.

Sm‘pet, C. 1991. Productivity of Lanigm maximum - Leucaena leucocephala mixture under cut
and carry system. Leucaena research reports 12:93-94.

Soil Conservation Service. 1979. Soil survey of Ingham County, Michigan. USDA, Washington
DC.

Tekle-Haimanot, A., C. W. Weeraratna, and E. V. Doku. 1991. Evaluation of three tree legume
species in Western Samoa. NFI'RR 9:71-74.

Time], T., and R Hatipoglu. 1989. The effects of intra-row spacings and cutting heights on the
yield of Leucaena leumphala in Adana, Turkey. Journal of Range Management 42:502-
503.

 

EFFECTS OF SPACING ON SURVIVAL AND FORAGE YIELD OF BLACK
LOCUST

ABSTRACT
Black locust seeds were planted in 1988 to give plots of 4, 16, 36, 64, and 100/m2 (40,000

- 1,000,000/ha) in a square grid pattern. Plots were cut back to 20 cm while dormant before the
1989 growing season, and were harvested in early and late summer at 30 cm the next two years.
Plots of 4 and 16/m2 were harvested at 50 cm in 1991. Effects of spacing on plot yields were
highly significant at the early summer harvest in 1989, and significant at the late summer harvest in
1989, but were not significant in 1990 and 1991. Effects of spacing on the proportion of dry
matter in fresh material were minimal, although sometimes statistically significant. Numbers of
plants harvested changed little fi'om early to late summer, but the decline was greater over winter.
Survival rates were significantly higher at 4/m2. The different spacings became less distinct with
time, and in 1991 the numbers harvested per plot at all spacings narrower than 4/m2 were not
significantly different. To establish black locust as a perennial forage crop, a wide spacing of

transplanted seedlings or a narrow spacing of seeds would be advised.

INTRODUCTION

The black locust tree (Robinit gseudoacacia L.) is noted for its rapid juvenile growth,

 

adaptability to difficult sites, and high protein content in new growth. Efforts to cultivate the
species as a perennial crop for fuel, fiber, or forage require more information on optimal

agronomic practices. These trials were conducted in order to determine the proper spacing between

82

83
Plants in the field. Trials were monitored for four seasons, to detemrine how the treatments

affected survival and yields.

Literature review
Many researchers have noted how the size of black locust seedlings depends on the spacing
between them (Wilson and Bailey 1957; Geyer _e_t Q. 1987; Genys and Harman 1990) or the
volume of the pot in which single plants are grown (Mebrahtu and Hanover 1991), but the species
has not previously been tested as a forage crop.

The multi-purpose nitrogen-fixing leucaena tree (Eucaena leucocephala (Lam) de Wit)

has been extensively researched in the tropics and behaves much like black locust under
cultivation. Much attention has been given to the optimal plant spacing for leucaena forage. Over
a range of planting populations, yields have typically been highest at the greatest density. Such
results at maximum populations of 20,000/ha have been reported by Visuttipitakul gt fl. (1983),
Mohatkar and Relwani (1985), and Yantasath e_t 11. (1985); at 40,000/ha by Pathak e_t a_l. (1980),
Van Den Beldt (1982), Relwani gt a_l. (1983b), Ella e_t a_l. (1989), and Tukel and Hatipoglu (1989);
at 62,500/ha by Cooksley and Goward (1988); at about 66,000/ha by Ferraris (1979); at
73,000/ha by Lahiri (1983); and at 100,000/ha by Relwani e_t a_l. (1982).

Trials with leucaena planted at over 100,000/ha often gave the highest green biomass
yields at populations other than the highest tested. Jayararnan e_t a_I. (1988) obtained significantly
higher yields at 66,666/ha, while yields at 33,333 and 133,333/ha were not significantly different
from each other. Desai e_t a_l. (1988) grew three strains of leucaena at five densities from 33,333 to
200,000/ha. The Cunningham cultivar yielded best at 100,000/ha, but K8 and Peru yielded best at
50,000/ha. Yields from rows 50 cm apart were 43 - 57% of rows 100 cm apart, confounding the
population effects. Shih and Hu (1981) found green biomass yields of four strains at four densities
from 50,000 - 200,000/ha rose with population in the first two years. Yields averaged 24% higher
at 200,000/ha than at 50,000/ha, although the differences were not significant. In the third through

fifth years there were no significant yield differences between spacing treatments (Shih e_t a_l. 1989).

84
Jiang and Liu (1991) obtained significantly higher yields at 83,000/ha than at 42,000/ha or
I 6 7, GOO/ha using cv. Salvador.

At higher populations of over 200,000/ha, yields usually increased with plant numbers.
Guewarra g g. (1978) found forage and stem yields increased significantly with population in the
range fi'om about 110,000 - 330,000/ha when averaged over all harvest intervals and cultivars.
Relwani g g1. (1983a) compared two cultivars at two sites, at 111,111, 166,666, and 333,333/ha,
obtaining the highest forage yields at the highest density. Ghatnekar 91 gl. (1983) attempted

gIOVving leucaena at 16,000,000/ha (2.5 by 2.5 em), but survival was about 134,000/ha.

Savory e_t _a__l. (1980) concluded from experimental results in Malawi that the optimal
forage population varied by region. In the Shire Highlands about 200,000/ha served best, but in
the hotter and drier Lower Shire Valley about 100,000/ha was optimal. Hegde (1983) considered a
leucaena population of 100,000 - 150,000/ha ideal for forage production, with only marginal yield
increases at higher densities. The National Research Council (1984) judged forage production best
at planting densities of 75,000 - 140,000 plants/ha. Horne e_t a_l. (1985) concluded that most
researchers reported better yields from leucaena planted at higher densities, but gave no upper
limit. In populations above 100,000/ha the stems are probably crowded enough that new growth
after each harvest closes the canopy so quickly that additional plants would not improve the speed
of recovery.

Spacing trials have been reported with tropical leguminous trees other than leucaena.
Highest yields were obtained at the highest density tested (40,000/ha) for the tropical legume
fodder trees Calliandra calothyzsus Meissn., Gliricidia; sggiium (Jacq.) Walp., and leucaena, but not
for Se;lmn_ia grandiflora (L.) Poir. Sesbw yields were low and did not increase much at higher
densities because of poor survival after cutting, although the yield at the first harvest had been

much greater at 40,000/ha (Ella e_t a_l. 1989). Gliriciclij sggium yielded the most biomass and

 

crude protein at the maximum density (20,000) (Karim and Savill 1991). Although individual tree

growth rates were inversely related to population density. the highest yields of leaves, stems, roots,

85
and total biomass were obtained at the highest density tested (40,000/ha) for Albizig Lebbeck (L.)
Benth. at 6, 12, 18, and 24 months of growth (Parrotta 1988).

Q1) iectives
The optimal density, defined as the lowest population per square meter to give the highest
yield, was expected to be affected by the age of the plantation, due to rates of canopy closure and
Wrences in mortality. An experiment was designed to test the following hypotheses:
1. Optimal density is higher at earlier ages, when more plants are nwded to form a closed
canopy.
2. Mortality over time increases with density, due to competition stress.
3. Survival is greater with older plants.
4. There is some final optimal density, which will be reached when mortality of surplus
plants ceases.
5. Dry weight percentage decreases with density.

6. Stem percentage decreases with density.

MATERIALS AND METHODS

General information for all trials

A bulk mixture of equal numbers of seeds from four Michigan sources was planted. The
sources were selected on the basis of availability, and not for any judgement of quality. The swd
accessions were 442 from Russ Forest in Cass County, 445 from the comer of Sandhill and
Hagadom roads, and 446 from the Dansville State Game Area, both in Ingham County, and 450
from Kellogg Forest in Kalamazoo County. Prior to sowing, the seeds were scarified in
concentrated sulfuric acid for 50 minutes. They were directly sown in late June 1988 at the
Michigan State University Tree Research Center in East Lansing (420 41' N, 840 28' W). The

nursery beds were surrounded by evergreen windbreaks and had been fumigated with methyl

86
bromide. The soil was mapped as a complex of soil series including Riddles, Hillsdale, Owosso,
and Marlette, all of which were well drained with 2% to 6% slopes and sandy loam topsoil.
Subsoils were sandy loam, loam, or clay loam (Soil Conservation Service 1979).

Prolonged heat and drought in 1988 impaired growth and sprinkler irrigation was
necessary. After normal weather returned in mid August, growth was rapid until early September,
wit}:- seedlings reaching heights of 60-120 cm. In 1989, 1990, and 1991 growing conditions were
favorable.

Because the soil had been fumigated, there were few weed problems in 1988. In the spring

of 1989 and 1990, the plots were sprayed with glyphosate (trade name Roundup) at 0.7 kg active
ingredient/ha and DCPA (trade name Dacthal) at 8.4 kg active ingredient/ha. Potassium sulfate (0-
0-50) fertilizer was added at the rate of 225 kg/ha (112 kg/ha K20) on May 7, 1990, to bring soil
potassium levels up from around 90 kg/ha. No other fertilizer was applied.

In each plot, all new growth above the designated harvest height was cut with hand tools
and bagged to determine the fresh weight. The paper bags of harvested material were dried in an
oven at 600C for at least three days before measuring the dry weight. Weights per plant were
calculated fi'om the number of plants harvested. Multiple sprouts from the same root were counted
as one plant. Shoots from different roots were counted as separate plants. Due to the presence of
suppressed seedlings, the total survival in 1989 was often higher than the number harvested, but
few suppressed seedlings survived to 1990. Survival to 1991 was often higher than the number
harvested that year because of rabbit damage on the smallest plants.

All trials were planted using a randomized complete block design. Analysis of variance,
Duncan's multiple range tests, and other statistical operations were conducted on NCSS computer
programs (Hintze 1987). The probability level reported for each factor below a table is the
probability of the difference between treatments occurring by chance, according to F-test

comparisons.

87
Directly sown trial
Three replications with five treatments were planted in a randomized complete block
design. In June 1988 seeds were planted by hand in a grid pattern, 10, 12.5, 16.7, 25, or 50 cm
apart, to give 100, 64, 36, 16, or 4 seedling positions per 1.0 m2 plot, respectively, with one or two
border rows on each side. The size of the planted area was about 4.0 m2 for each plot. Multiple
seedlings were fiequent because two to four seeds were sown at each grid position. Gaps were
replanted with seeds in early July, and with seedlings (transplanted out of designated pathways) in
early August.
In March 1989, the previously uncut plots were cut back to 5 or 20 cm, and the stems left
to decay on the ground. Plots were harvested at 25 cm between June 28 and July 5, and again at
30 cm between August 23 and 31, 1989. In 1990, first harvests were taken between June 18 and
22, and second harvests between August 16 and 27, all at 30 cm. In 1991 first harvests were taken
from plots of 4 and 16/m2 only, between July 10 and 17, at a height of 50 cm. Because of shading
from nearby uncut plots and feeding by rabbits, no second harvest was possible, and the

experiment was terminated.

Transglanted trial
Seedlings sown in 1987 were transplanted in June and July, 1988 into 1.0 m2 plots with

either four plants 50 cm apart or 25 plants 20 cm apart. With one border row at the 50 cm
spacing, or two at the 20 cm spacing, the total space required for each plot was 4.0 m2. There
were two plots of each of six seed sources in each of three replications, giving 36 plots. Data from
three more plots planted at 64/m2 were included in cases where a statistically balanced design was
not required. Seed source differences were tested in the analysis of variance, but are not reported
as they had no statistical significance.

No fertilizer was applied. Chemical weed control was attempted but was ineffective, and
the 4/m2 plots had vigorous grass growing between and often above the seedlings. These seedlings

were harvested at 30 cm in July and September 1988, 1989, and 1990. They were counted but not

88
cut in 1991. No more data could be obtained because the research plots were destroyed without
notice in June 1992.
Although the yield differences between spacings were usually statistically significant, it
was judged that weed competition varied between treatments and caused much of the difference.
Therefore, only data concerning survival rates and the composition of harvested material are
presented fi'om this trial. Mortality soon after planting was attributed to severe damage by

herbivores, so survival was calculated from the number living in September, 1988.

RESULTS AND DISCUSSION

Dgy weight yields ger glot
Effects of spacing treatments were highly significant in July and non-significant in August,

1989 (1‘ able 4.1). In July, yields were very low at 4/m2, significantly higher at 16/m2, again
significantly higher at 36/m2, and plateaued at higher densities. In 1990, the yields did not vary
significantly with spacing at either harvest.

Harvesting in July 1989, long after the canopy closed, resulted in a poor second harvest.
Most lower leaves had already abscised, and a closed canopy was restored slowly. In 1990,
harvesting in late June retained more lower leaves, restoring a complete canopy sooner and giving a
much larger second harvest. In leucaena, a strong interaction between time of harvest and planting
density has been shown. To maximize leaf yields, the crop must be harvested when the leafy
biomass is greatest, just before shading causes lower leaves to abscise. Leaf yields will be lower if
harvest is delayed, but more frequent harvesting will increase the proportion of time spent growing
with less than a complete canopy, and also reduce cumulative yields. At higher densities where the
canopy closes earlier, earlier harvests would yield more (Home e_t a_l. 1985; Blair e_t a_l. 1990).
Castillo gt g1 . (1979) reported higher leucaena herbage yields at higher populations, especially with

shorter harvest intervals.

89 ‘
Table 4.1. Dry weight yields (g/mz) for square grid plots at five planting densities harvested at
25 cm in early July and at 30 cm in late August 1989, late June and late August 1990.

 

 

 

Planting density 1989 1990

per square meter July August June Augqst
4 233a 117ab 163a 330a
16 412b 125ab 178a 305a
36 6950 134b 181a 293a
64 5580 85a 174a 314a
100 623C 114ab 164a 363a

Mean 504 115 172 321

Probability .0004 .2221 .9960 .9305

 

Means within the same column followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

90

Sumberg (1986) found that forage yield differences between six spacing treatments 4 - 50
cm apart (4 - 625/m2) in Nigeria with Gliricfl sepium (Jacq.) Walp. were significant at the first
three harvests but not at the fourth. Regrowth after each harvest was during the rainy season, so
drought was not involved. It is likely that the canopy had closed at all spacings by the fourth
harvest, as occurred in these trials by the third harvest.

Cooksley and Goward (1988) found leucaena yield increased three-fold when density
increased 20-fold from 3100 to 62,500 plants/ha. With a 25-fold increase in density in this study,
the yield of black locust forage doubled in 1989, and remained level in 1990 (1‘ able 4.2), when the
difi‘erences in canopy coverage between treatments were slighter. Although growing conditions
were excellent both years, and fertilizer was added at the start of the 1990 season, yields declined
fiom 1989 to 1990 for all spacings but 4/m2, with an average reduction of 20%.

For the plots of 4 and 16/m2 harvested in 1991, yield differences due to spacing were only
significant in July, 1989 (T able 4.3). Although cumulative yields of these plots did not vary
significantly by density after 1989, the much higher yields at 4/m2 in 1991 indicated that this
spacing would likely yield significantly higher over a longer period.

Savory e1 a_1. (1980) reported that leucaena forage yields tend to rise over the first four
years, presumably due to greater root penetration. Yields have risen over the first three years in
some black locust trials, but in this study average yields declined the third year (1990). In the
fourth year, only the two widest spacings were harvested, and only once, in July. Still, it is clear
that the plots of 4/m2 gave higher yields each year, and the l6/m2 plots did not.

It is unclear whether the lower yields in 1990 were due to changed harvest dates, the age of
the plants, fewer plants surviving, or weather conditions. The pattern of declining yields in
measured plots may also have been due to shading by the more vigorous plants in the outside
border rows, so that the forage yield of the entire plantation may have continued rising. Border

rows also tended to have better survival than the measured plots. Future research should

91
Table 4.2. Annual and cumulative dry weight yields (g/mz) for square grid plots at five planting
densities harvested in 1989, 1990, and 1991.

 

 

Planting density Two-year Three-year
per square meter 1989 1990 Total 1991 Total
4 350a 493a 843 579a 1422a
16 536b 482a 1019 3163 1335a
36 829d 474a 1303
64 643bc 489a 1 132
100 737cd 527a 1264
Mean 619 493 1 1 12
Probability .0006 .9882 .0265 .4937 .8616

 

Means within the same column followed by the same letter are not significantly different
at the P=0.05 level according to Duncan's Multiple Range Test.

92
Table 4.3. Single harvest and cumulative dry weight yields (g/mz) from grid plots of 4 and
16/m2 .

 

 

Harvest 47m2 16/m2 Probability
7/1989 233 412 .0002
8/1989 1 17 125 .4305
6/1990 163 168 .1957
8/1990 330 305 .9761
7/1991 579 316 .2445

Total 1422 1336 .4780

93
investigate patterns of plant growth and competition, both above and below the ground, in relation

to yield levels and survival.

Dgy weight yields ger glant

Yields of plots were influenced by the number and vigor of survivors. Comparison dry
weight of yields per plant was used to determine at what population the plants start to compete.
Because each remaining plant has more space after some die, a slight. increase in yield per plant
would be expected over the years.

At both harvests in 1989, the effect of spacing treatments was highly significant (Table
4.4). August 1989 yields were 20 - 35% of July yields, except at 4/m2 where it was over 50%.
Yields at the early summer harvest in 1990 were about half those at the second harvest that year.
Population density effects were highly significant in June and significant in August. Only the yield
per plant at the 4/m2 spacing was significantly different from an adjacent spacing at either 1990
harvest. Total yields per plant in 1990 averaged higher than in 1989, but the average was biased
by the 4/m2 plots, while plots at higher densities yielded about the same both years. Spacing
treatments produced highly significant differences in 1989 and 1990 totals. Again, only the yield
per plant at the 4/m2 spacing was significantly different from an adjacent spacing. While yields
per plant at all other spacings had level or declining trends from 1989 to 1990, the 4/m2 plots
increased sharply. The 4/m2 plants were still growing to fill the available space between them
above ground, and presumably below ground also. This growth continued into 1991.

August 1989 yields were 20 - 40% of July yields, except at the lowest population, where
more leaves were retained below the harvest height. At the higher populations, almost all lower
leaves had already dropped from shading, and recovery was slower. While yields at the June
harvest in 1990 were about half those in July 1989, which were cut two weeks later, the yields of
plants at 4/m2 were almost as high in 1990. At the second harvest of 1990, yields per plant were
about double those at the first harvest or at the second harvest of 1989, reflecting fewer plants at

the three highest densities in 1990 than in 1989, plus a longer period of regrowth.

94
Table 4.4. Single harvest and annual dry weight yields per plant (g) from grid plots at five
planting densities harvested at 25 cm in early July and at 30 cm in late August, 1989, late June,
and late August, 1990.

 

 

 

Planting density 1989 1990
per square meter July August Mean June August Mean
4 40.6a 21.0a 60.0a 33.4a 75.7a 109.9a
16 22.0b 7.8b 30.2b 10.3b 18.4b 28.8b
36 19.3b 3.9b 23.7b 6.8b 10.5b 17.9b
64 9.80 2.0b . 12.80 4.5b 7.2b 12.1b
100 9.50 1.9b 11.90 3.3b 6.6b 10.1b
Mean 20.2 7.3 27.7 11.7 23.7 35.7
Probability .0001 .0022 .0000 .0009 .0341 .0106

 

Means within the same column followed by the same letter are not significantly different at the

P=0.05 level according to Duncan's Multiple Range Test.

 

95
Comgosition of harvested biomass

The dry weight/fresh weight ratio (D/F) represents the dry matter proportion of the green
biomass. Spacing treatments had no significant effect on D/F in grid plots. In the transplanted
plots, a significant difference due to spacing was found in the dry weight percentage at the late
summer (September) harvest in 1990. The D/F averaged 36.6% at 4/m2 and 34.4% at 25/m2.
Difl‘erences were not significant at three earlier harvests. The D/F was predicted to decrease with
density, but the difference was rarely statistically significant. The slight differences observed
appeared to depend on the size of harvested stems. Spacing at 4/m2 produced larger stems with a
higher proportion of wood, and usually a higher D/F percentage.

In the grid plots, the proportion of woody stem in the harvested material in 1989, did not
vary significantly by spacing. Woody stems comprised 20 - 30% of harvested material in early
July, 5 - 15% of the regrowth harvested in late August, and 15 - 25% of the total biomass
harvested in 1989. In the transplanted trial, the proportion of woody stems in mid July 1990
averaged 27% at 4/m2 and 23% at 25/m2, a statistically significant difference.

The proportions of dry weight and of stems in the harvested black locust forage had been
predicted to decline as populations increased, but this did not always occur. It is possible that
abscission of shaded lower leaves in the denser plots, shortly before harvest in June 1989, changed
the stem proportions of the samples. A similar result with leucaena was reported by Guevarra e_t
11. (197 8). Under conditions of a closed canopy and a long harvest interval, leafy forage and stem
yields both increased significantly with population, but forage yield rose more slowly. Two other
leucaena reports, though not mentioning increased leaf loss before harvest at higher populations,
show results consistent with such a process. Jayararnan e_t ll (1988), using leucaena cv. K8,
found no variation in the protein content of dry matter from densities of 33,000 to 133,000/ha.
This would indicate that the proportion of woody stems changed little with density. J iang and Liu
(1991) found the stem proportion of leucaena cv. Salvador was significantly lower (under 40%) in

the range of 21,000 - 83,000/ha than at 167,000/ha (41%) or 17,000/ha (45%).

96
Severe defoliations have been observed to reduce the regrowth of stems more than leaves in
red oak (Ms {11113 L.) and red maple (A_cer rub—rum L.) (Heichel and Turner 1984), and the
Brazilian species Mm caesalpinia, Caesalginia gymgdalis, and Auxemma oncocalfl
(Hardesty and Box 1988), although the opposite was found in English oak (Quereus robur L.)
(Hilton 1987). In this study, plots with narrow spacings tended to suffer greater defoliation at each

harvest than plots with wide spacings.

§urvival after harvesting

In the grid plots, the number harvested was not exactly equal to survival, as numbers
increased slightly in August 1990 (Table 4.5). The greatest reductions occurred over the winters.
While mortality was low (6 - 21%) between harvests in the summer of 1989 at the narrower
spacings, plots at 4/m2 had a 13% increase in numbers harvested. The weakest seedlings, which
tendw to die under competition at narrower spacings, lived to be harvested in the plots of 4/m2.
The increase between June and August in 1990 resulted when the new growth of some weak
seedlings did not reach 30 cm by the first harvest, but passed 30 cm before the second. This
increase was not significantly affected by spacing treatments.

Survival to 1991 was generally poor, but significantly higher for plots at 4/m2. The
narrower spacing treatments lost about half of their remaining plants between August 1990 and
July 1991, and the different spacings became less distinct (Table 4.6). In 1989 and 111 1990 the
numbers per plot typically were not significantly different from consecutive spacing treatments, but
differed significantly from a second higher or lower spacing. In 1991 the original spacings had less
influence, as the narrowest spacing (100/m2) differed sigrificantly only from the widest (4/m2).

Therefore, plots were grouped by the number of survivors in August 1990, and mortality
over the following winter was calculated (Table 4.7). Winter mortality after the 1990 season was
less severe in the plots where individual plants were the largest and strongest from being spaced
farthest apart. Plots with over 15 plants in 1990 all suffered 54 - 59% mortality. Correlation of

numbers present in August 1990 and July 1991 gave a survival rate of 38% (r = .7749).

97

Table 4.5. Number harvested (plants/m2) from grid plots at five planting densities in subsequent
harvests relative to the number harvested in July, 1989.

 

 

 

Date Number planted Jger m2

4 16 36 64 100 Mean Probability
July 1989 1.00 1.00 1.00 1.00 1.00 1.00 ----
Aug. 1989 1.13a .91b .94b .79 .91b .90 .3766
June 1990 .91a .91a .68 .71 .75 .70 .5575
Aug. 1990 1.06a .89 .81 .84 .81 .75 .5348
July 1991 .77a .43b .39b .38b .37b .47 .0850
Aug/June 1.17a .99a 1.20a 1.20a 1.053 1.12 .3462
19901

 

1 The number harvested in August, 1990, relative to the number harvested in June, 1990.

Means witlnin the same row followed by the same letter are not significantly different at the
P=0.05 level according to Duncan's Multiple Range Test.

Table 4.6. Number harvested (plants/m2) each time 1988 - 1991 from grid plots at five planting

 

 

 

densities.

Bate Number planted per m2

Mu}; 4 16 36 64 100 Mean Probability‘
M289 6a 19ab 36bc 570d 72d 38 .0037
3.3124282 7a 17ab 34bc 43c 62d 33 .0006
13.1mm 5a 17ab 25b 40c 500 27 .0007
All-18.1299 6a 17ab 29be 460d 54d 30 .0016
W 4a 83b 14ab 20b 21b 13 .0877

 

Means witlnin the same row followed by the same letter are not significantly different at the

P=0.05 level according to Duncan's Multiple Range Test.

99

Table 4.7. Effect of 1990 plot population (plants/m2) on 1991 plot population, and ratio of
survival from 1990 to 1991, for square grid plots originally planted at five spacings, and
transplanted plots originally planted at three spacings.

 

 

 

Number 1990 1991 Survival
of plots Mean Rye Mean Range ratio
3 5.66 4-7 4.0 4-4 I .71
3 17.5 16-20 7.25 5-11 .41
3 34.5 32-38 16.0 6-24 .46
3 57.75 50-73 24.25 1740 .42
191 3.4 1-8 3.3 0—8 .97
151 15.0 11-19 13.4 9-17 .89
51 26.4 20-36 21.0 12-27 .80

 

Statistical analysis was impossible because of differing numbers of plots.

1 Transplanted plots.

100
Regression of the survival rate on the number present in August 1990 gave a decline of 6% for
eaCh 10 additional plants (r = -.4534).

In the other trial, the seedlings transplanted at 4/m2 had a significantly higher survival rate
over three years (78%) than those at 25/m2 (61%) (T able 4.8). Compared to two year survival in
tile trial with sown plots, three year survival in this trial was nearly the same at 4/m2, but at 25/m2
was about double the rate for any of the spacings narrower than 4/m2. Survival rates from 1990 to
1 991 generally declined with density, but were much higher than in the directly sown trial (Table
4- - 7). Correlation of numbers of transplanted seedlings present in August 1990 and July 1991 gave

a survival rate of 79% (r = .9753). Regession of the survival rate on the number present in
August 1990 gave a decline of 5% for each 10 additional plants (r = -.2425).

Because crown closure and competition for water and nutrients occur more rapidly at narrow
Spacing, the added strain of cutting in coppice stands typically increases mortality, as shown in
several studies with Eucalyptus and MS species (Blake 1983). The mortality rate increased
vvith density in these black locust trials also, as had been predicted. This was true whether

considering planting density (Tables 4.5, 4.6, and 4.8) or the number actually present (Table 4.7).

 

In Glirigg sepium. survival through four forage harvests in the first 15 months was near total
vvith under 12 seedlings per meter of row. At the highest density (62,500/ha) with 25 per meter of
row, only 60% survived (Sumberg 1986). Survival of leucaena below 62,500/ha (6.25/m2) was
not affected by density nor by planting arrangement (Cooksley and Goward 1988). At such low
densities, it is unlikely that black locust mortality would vary, either.
The hypothesis that survival is greater with older plants was not supported. The lowest
Survival rates were found in the last interval for the sown plots (Table 4.5) and in the first, fourth,
and last of five intervals for the transplanted plots (Table 4.8). Lower mortality in some plots after

1990 can be attributed to wider spacing of the survivors, but not directly to plant age.

101

Table 4.8. Numbers of survivors (plants/m2) and ratio of survival from the previous date for
transplanted seedlings at two population densities.

 

Mean survivors

Survival ratio

 

 

Date 4/m2 25/m2 4/m2 25/m2 Probability
Sept. 1988 3.89 22.44
July 1989 3.44 21.28 .88 .95 .1915
Sept. 1989 3.39 19.33 .99 .92 .0580
Jtlly 1990 3.39 18.11 1.00 .94 .1028
Sept. 1990 3.17 15.17 .94 .85 .0408
Sept. 1991 3.06 13.22 .94 .88 .3744
m year

survival rate .78 .61 .0247

 

102

Ogtimal densig

The hypothesis that optimal density is higher at earlier ages was confirmed, as it declined
fiom 36/m2 in 1989 to 4/m2 in 1990. No significant yield difl‘erences between spacings remained
in 1990, the third growing season. This experiment was designed to find the final optimal density,
where the yield per area is equivalent to that in more crowded plots and mortality of weak plants
ceases, but the search was inconclusive. In these trials, survival at the widest spacing was
S ignificantly higher, but still below 80% over three or four growing seasons. Though it is possible
that mortality would cease at some time after four years, the experiment gave no evidence that a
final optimal density actually exists. Ifit does exist at a harvest height of 30 cm, mortality would

be expected to cease somewhere below five plants per square meter.

Defoliation sever-1g and solar intercegtion

The distance between plants not only influenced competition between them, but affected
tlne percentage of foliage removed by harvesting. Even when black locust was cut before the
shaded lower leaves abscised in late June, the denser spacings (l6/m2 and above) lost more of their
leaves than the 4/m2 plots did. The 4/m2 plots had enough space between the plants so that lower
leaves could be illuminated adequately, and they remained green until harvest. The denser plots
lost their lowest leaves before the first harvest in 1989, and had very few remaining after harvest,
generally 0 - 10%, compared to 25 - 75% for the 4/m2 plots. Later harvests were taken at an

earlier stage of gowth so as to retain a partial canopy (about 10 - 25% of leaf area) in the denser

plots.
The canopy in 4/m2 plots did not close until shortly before harvest in 1990 and 1991, but

little sunlight was observed reaching the ground, due to the greater height of the stems and the

oblique angle of the sun at 42° north latitude. The absence of a closed canopy may account for
continued vigorous growth and better survival in 1991, since leaves around the sides of the plants
did not abscise from shading, providing more photosynthate per plant and possibly per plot.

M eb rahtu (1989) reported that the light saturation level of black locust was above the highest level

103
tested (1900 umol/m2/second of photosyntheticallly active radiation), and that similar levels had

been observed only for lodgepole pine (Pinus contorta Dougl.) among woody plants, which usually

 

reach saturation at 400 to 1100 umol/mZ/second. To expose more leaves to the sun and increase
photosynthesis, Mebrahtu recommended wider spacing in plantations. Since the plants at the
widest spacing showed no evidence of exhaustion or crowding in this study, it remains unknown
how many years they would remain productive. Plants at all narrower spacings were clearly

exhausted by the fourth growing season, after two years of harvesting.

§gacing recommendations for stand establishment

A perennial forage crop requires enough plants per area to maintain a closed canopy at the
harvest height. With the 30 cm height used in this experiment, the widest spacing tested (50 cm
apart) was adequate the third gowing season, when yield differences were not significant.
Survival was also best at this spacing. With higher cutting, even wider spacings could be used, so
transplanting seedlings may become an economical means of establishing a crop. But considering
the low cost of the seed and the loss of potential yields before canopy closure, narrow spacing of
seeds is to be recommended over wide spacing. A thick sowing would probably also be more
profitable than transplanting, because of the much higher expenses for seedlings and the labor to
plant them. A more thorough comparison of sowing or transplanting black locust has been
published (Barrett 1992). The grower should plant enough seed to give a density of over 35/m2
when ready to begin harvesting at 30 cm in the second growing season, or should transplant

seedlings at l - 4/m2 depending on the intended harvest height.

104
REFERENCES CITED

Barrett, R. P. 1992. Forage plantation management for black locust. pp. 259-272 in: Hanover, J.
W., K. Miller, and S. Plesko, eds. Procwdirngs: lntemational Conference on Black Locust:
Biology, Culture, & Utilization. Department of Forestry, Michigan State University.

Blair, 0., D. Catchpoole, and P. Home. 1990. Forage tree legumes: their management and
contribution to the nitrogen economy of wet and humid tropical environments. Advances in
Agronomy 44:27-54.

Blake, T. J. 1983. Coppice systems for short-rotation intensive forestry: the influence of cultural,
seasonal, and plant factors. Aust. For. Res. 13:279-291.

Castillo, A. C., F. A. Moog, and D. C. Avante. 1979. Efl‘ects of row arrangement and plant density
on herbage production and gowtln of ipil-ipil. Philippine Jourrnal of Animal Industry
34:75 -8 1 .

Cooksley, D. G., and E. A. Goward. 1988. Efl’ect of plant density and spatial arrangement on the
yield of Leucaena leucocephala cv. Peru in subcoastal soutln-eastem Queensland. Austr. J.
of Exp. Ag. 28:577-585.

Desai, S. N., J. S. Desale, R B. Khot, and S. K. Patil. 1988. Effect of spacing on Leucaena forage
yields. Leucaena research reports 9:47.

Ella, A., C. Jacobsen, W. W. Stur, and G. Blair. 1989. Effect of plant density and cutting
fi'equency on the productivity of four tree legumes. Tropical Grasslands 23:28-34.

Ferraris, R. 1979. Productivity of Leucaena leucocephala in the wet tropics of North Queensland.
' Tropical Grasslands 13:20-27.

Genys, J. B., and D. M. Harman. 1990. Racial diversity of black locust in growth rate and in
susceptibility to the locust twig borer. N. J. Appl. For. 7 :38-43.

Geyer, W. A., L. Melichar, and C. E. Long. 1987. Preemergent herbicide trials with direct-seeded
black locust gown in different soils. Journal of Arboriculture 13:105-107.

Ghatnekar, S. D., D. G. Auti, and V. S. Kamat. 1983. Biomanagement of leucaena plantations by
Ion Exchange (India) Ltd. in: Leucaena research in the Asian - Pacific region. IDRC 211e,
Ottawa.

Guevarra, A. B., A. S. Whitrney, and J. R. Thompson. 1978. Influence of intra-row spacing and
cutting regimes on the growth and yield of Leucaena. Agronomy Journal 70: 1033-1037.

Hardesty, L. H., and T. W. Box. 1988. Defoliation impacts on coppicing browse species in north-
east Brazil. Journal of Range Management 41:66-70.

Hegde, N. 1983. Leucaena forage management in India. in: Leucaena research in the Asian -
Pacific region. IDRC 211e, Ottawa.

 

105
Heichel, G. H., and N. C. Turner. 1984. Branch gowth and leaf numbers of red maple (Acer
rubrum L.) and red oak (Qgercus rubra L.): response to defoliation. Oecologia 6221-6.

Hilton, G. M., J. R. Packham, and A. J. Willis. 1987. Effects of experimental defoliation on a
population of pedunculate oak (Quereus robur L.). New Phytol. 107:603-612.

Hintze, J. L. 1987. Number cruncher statistical system user's manual, version 5.0. J. L. Hintze,
Kaysville, UT.

Horne, P. M., D. W. Catchpoole, and A. Ella. 1985. Cutting management of tree and shrub
legumes. in: Blair, G. J ., D. A. Ivory, and T. R. Evans, eds. Forages in Southeast Asian
and South Pacific agriculture. Austr. Center for Intl. Ag. Res. Proceedings No. 12.

Jayaraman, S., S. Purushotaman, and M. Govindaswamy. 1988. Optimum planting geometry for
production of Leucaena fodder in South India. Leucaena research reports 9:50.

J iang, H., and G. Liu. 1991. Fodder yield experiments with Leucaena at various densities.
Leucaena research reports 12:92.

Karim, A. B., and P. S. Savill. 1991. Effect of spacing on gowth and biomass production of
Gliricidia segium (Jacq) Walp in an alley cropping system in Sierra Leone. Agroforestry
Systems 16:213-222.

Lahiri, A. K. 1983. Results of trial of some leguminous plants in the Himalayan foothill region of
West Bengal. Leucaena research reports 4:29-30.

Mebrahtu, T. 1989. Gas exchange characteristics of black locust (Robinia pseudoacacia L.) in
relation to gowth, development, and environmental factors. Unpublished Ph.D.
dissertation, Michigan State University, East Lansing, MI, USA.

Mebrahtu, T., and J. W. Hanover. 1991. Effects of nitrogen level on photosynthesis, growth, and
leaf traits of black locust. NFTRR 9: 130-132.

Mohatkar, L. C., and L. L. Relwani. 1985. Effect of plant population, stubble height, and number
of cuttings on the gowth, seed, forage, and firewood production of Leucaena K8.
Leucaena research reports 6240-41.

National Research Council. 1984. Leucaena: promising forage and tree crop for the tropics.
Second edition. National Academy Press, Washington, DC.

Parrotta, J. A. 1988. Early growth and yield of Albizi_a_ lebbecjk at a coastal site in Puerto Rico.
Nitrogen Fixing Tree Res. Reports 6:47-49.

Pathak, P. S., P. Rai, and R. Deb Roy. 1980. Forage production from koo-babool (L leucocephala
(Lam) De Wit). 1. Effect of plant density, cutting intensity and interval. Forage Research
6:83-90.

Relwani, L. L., S. S. Deshmukh, D. Y. Khandale, and R. V. Nakat. 1982. Effect of spacing
management on the yield and quality of forage of K-8 cultivar. Leucaena research reports
3:40.

106
Relwani, L. L., L. C. Mohatkar, G. K. Bhavsar, and R. V. Nakat. 1983a. Effect of spacings and
stubble heights on forage yields of K-8 and Cunningham. Leucaena research reports 4:41.

Relwani, L. L., L. C. Mohatkar, and R. V. Nakat. 1983b. Effect of plant populations and
frequency of watering on the gowth and yield of trees (Hawaiian Giant K8). Leucaena
research reports 4:38-40.

Savory, R., J. A. Breen, and C. I. A. Beale. 1980. Leucaena as a forage crop on small farms in
Malawi. pp. 411-414 in: Le Houerou, H. N., ed. Browse in Africa: the current state of
knowledge. ILCA, Addis Ababa, Ethiopia.

Shih, W.-C., and T.-W. Hu. 1981. The yields of forage of Leucaena leucocephala in Taiwan.
Leucaena research reports 2:55-56.

Shih, W.-C., T.-W. Hu, and H.-M. Wu. 1989. The influence of planting density on different
strains of Leucaena leumghala for forage production. Leucaena research reports 10:61-
62. .

Soil Conservation Service, 1979. Soil survey of Inglnam County, Michigan. USDA, Washington
DC.

Sumberg, J. E. 1986. Alley farming with Gliricidig sepium: gennplasm evaluation and planting
density trial. Trop. Agric. (Trinidad) 63:170-172.

Tukel, T., and R. Hatipoglu. 1989. The effects of intra-row spacings and cutting heights on the
yield of Leucaena leucocephala in Adana, Turkey. Journal of Range Management 42:502-
503.

Van Den Beldt, R. J. 1982. Litterfall as a function of population in a l-year old Leucaena (K8)
planting. Leucaena research reports 3:95.

Visuttipitakul, S., S. Chairnongkol, and K. Yantasath. 1983. Field trial of Leucaena in Chiang Mai
province. Leucaena research reports 4:81.

Wilson, J. D., and C. Bailey. 1957. Soil treatments may increase stand of locust seedlings. Ohio
Farm and Home Research 42:44—46.

Yantasath, K, W. Supatanakul, I. Ungvivhian, S. Charnsawad, S. Chantrasiri, S. Patanavibul, C.
Hyakitkosol, S. Prompetchara, N. Pithakamop, and P. Chalemnklin. 1985. Species trials
of NFT. NFTRR 3:48-56.

EFFECTS OF YEAR OF FIRST HARVEST ON SURVIVAL AND FORAGE
YIELD OF BLACK LOCUST

ABSTRACT

Black locust seeds were sown at five spacings (40,000 - 1,000,000/ha) in a square grid
pattern. Plots were harvested at 20 cm height two months after sowing (group 1), or were cut back
to 20 cm while dormant before the second growing season (group 2). All were harvested in early
and late summer at 30 cm the next two years, and plots of 4 and 16/m2 were harvested at 50 cm
the fourth year. Black locust seeds from five sources were sown in rows 20 cm apart to give
25/m2. Plots were harvested as above, two months after sowing (group 3) or while dormant
(group 4).

Harvesting the first year gave slight yields and damaged future productivity. At the July
harvest the second year, the average yield of goup l was 75% of group 2, and only 90% when
combined with the first year's total, while the average yield of group 3 was 70% of group 4, and
93% when combined with the first year's total. Two-year totals were significantly higher for group
2, but not significantly different for groups 3 and 4. Yields declined the third year, by 43% for
group 1 and 19% for goup 2, but increased by 35% for group 3 and 52% for group 4. Third-year
yields were significantly higher in groups 2 and 4.

Effects of year treatments on the proportion of dry matter in fresh material were minimal,
although sometimes significant. Effects on the branching pattern were significant, but changed
over time. The number of plants harvested per plot increased 26% from the first to the second
year. Numbers harvested changed little from early to late summer. Survival from the second to

third year declined highly significantly with harvesting in the seedling year.

107

 

108

INTRODUCTION

The black locust tree M113 gseudoacacia L.) is noted for its rapid juvenile growth,
adaptability to diflicult sites, and high protein content in new gowth. Efforts to cultivate the
species as a perennial crop for fuel, fiber, or forage require more information on optimal
agronomic practices. These trials were conducted to determine the amount of gowth required
before initiating harvesting, and the interactions of year-of-first-harvest treatments with seed
sources, spacings, and harvest heights. Trials were monitored for three or four seasons, to

determine how the treatments affected survival and yields.

Literature review

Although older and larger diameter (greater than 20 - 40 cm) trees of many species
typically show reduced coppicing ability, regrowth is usually optimal from sapling-sized trees
(Blake 1983). However, Pawlick (1989) warned not to coppice tropical trees until they are well
established, and suggested an acceptable height of 300 - 400 cm, nine months to two years after
planting.

In black locust, reported maximum heights of 2.5 m for first year seedlings and 4.5 m for
the new gowtln from one year old stumps indicate that harvestable biomass increases with the age
of the plant. However, the same report documents wide variation among seed sources in branching
patterns, which affect the relation of weight to height, and root to shoot ratios averaging 0.41, with
a range of 0.27 to 0.55, which affect the vigor of coppice growth (Mebrahtu 1992). In a study of
25 black locust seed sources, first year heights averaged 84 cm and ranged from 64 to 98 cm. The
heights of second year and third year coppice sprouts (above the 30 cm stumps) ranged from 84 to
200 cm, and 120 to 270 cm, averaging 140 cm and 220 cm, respectively (Genys and Harman
1990).

After one year in the field in Uttar Pradesh, India, at a spacing of 50 cm by 50 cm, black

locust seedlings were found to consist of 31% stem, 24% branches, and 45% roots (collected down

109

to 45 cm). Black locust had the highest proportion of branches of six species tested (Bhatt and
Todaria, 1991).

Annual forage yields of Leucaena leumhala (Lam) de Wit. and Gliricidia m
(Jacq.) Walp. increased with age of first harvest up to 21 months, the longest period tested. The
yield of regrowth varied directly with stump size (Blair g a_l. 1990). Increasing stump diameters

and biomass production with age at first harvest have also been reported in Sesbania macrantlna

 

 

(Holden e_t g. 1989). A comparison of newly sown and one year old stands of M m
(L.) Merr. var. gngbjg revealed no significant yield differences at three harvest heights and three
interval lengths (Galang e_t 31. 1990).

In summary, variations among black locust seed sources in gowth rate, and the proportion
of first year gowth allocated to roots, stems, and branches will cause variations in the proportion
of biomass removed by a harvest treatment at a constant height. This could lead to significant
interactions between seed sources and years of first harvest treatments. Spacing treatments, which
affect the allocation of gowth to stems and branches (and possibly to roots as well) could interact
with the year treatments. Experience with other species indicates that the size of the first harvest
varies directly with the size and age of the plants, but the pattern of cumulative yields has not been

reported and requires more investigation.

Obiectives

It was expected that delaying the year of first harvest would improve yields in the second
year, but the pattern of cumulative yields fi'om all harvests was not known. It was also desired to
determine whether the age of the plants at first harvest leads to differences in mortality, the
percentage of dry matter, and the branching pattern. An experiment was designed to test the
following hypotheses:

l. Yields the second year are higher for plots not harvested the first year.

2. Cumulative yields for first and second years are similar to second year yields for plots

not harvested the first year.

110

3. Cumulative yields for first, second, and third years are lower than cumulative second
and third year yields for plots not harvested the first year.

4. Dry weight percentage the second year is lower for plots not harvested the first year.

5. Number of leaders harvested the second year is higher for plots harvested the first year.

6. Survival is not affected by harvesting the first year.

7. Survival rate increases with plant age.

8. Interactions of year-of-first-harvcst treatments and other treatments are not statistically

significant for any measured parameters.

MATERIALS AND METHODS

General information for all trials
A bulk mixture of equal numbers of seeds from four Michigan sources was planted. The
sources were selected on the basis of availability, and not for any judgement of quality. The swd

accessions were 442 fi'om Russ Forest in Cass County, 445 from the comer of Sandlnill and

 

Hagadom roads, and 446 fi'om the Dansville State Game Area, both in Ingham County, and 450
from Kellogg Forest in Kalamazoo County. Prior to sowing, the seeds were scarified in
concentrated sulfuric acid for 50 nninutes. They were directly sown in late June 1988 at the
Michigan State University Tree Research Center in East Lansing (420 41' N, 840 28' W). The
nursery beds were surrounded by evergeen windbreaks and had been fumigated with methyl
bromide. The soil was mapped as a complex of soil series including Riddles, Hillsdale, Owosso,
and Marlette, all of which were well drained with 2% to 6% slopes and sandy loam topsoil.
Subsoils were sandy loam, loam, or clay loam (Soil Conservation Service 1979).

Prolonged heat and drought in 1988 impaired growth and sprinkler irrigation was
necessary. After normal weather returned in mid August, growth was rapid until early September,
with seedlings reaching heights of 60-120 cm. In 1989, 1990, and 1991 growing conditions were

favorable.

111

Because the soil had been fumigated, there were few weed problems in 1988. In the spring
of 1989 and 1990, the plots were sprayed with glyphosate (trade name Roundup) at 0.7 kg active
ingredient/ha and DCPA (trade name Dacthal) at 8.4 kg active ingedient/ha. Potassium sulfate (0-
0-50) fertilizer was added at the rate of 225 kg/ha (112 kg/ha K20) on May 7, 1990, to bring soil
potassium levels up fi'om around 90 kg/ha. No other fertilizer was applied.

In each plot, all new gowth above the designated harvest height was cut with hand tools
and bagged to determine the fresh weight. The paper bags of harvested material were dried in an
oven at 600C for at least three days before measuring the dry weight. Weights per plant were
calculated fi’om the number of plants harvested. Multiple sprouts from the same root were counted
as one plant. Shoots from different roots were counted as separate plants. Due to the presence of
suppressed seedlings, the total survival in 1989 was often higher than the number harvested, but
few suppressed swdlings survived to 1990. Survival to 1991 was often higher than the number
harvested that year because of rabbit damage on the smallest plants.

All trials were planted using a randomized complete block design. Analysis of variance,
Duncan's multiple range tests, and other statistical operations were conducted on NCSS computer
programs (Hintze 1987). The probability level reported for each factor below a table is the
probability of the difference between treatments occurring by chance, according to F-test
conmafimxw.

Identity of groups referrred to in the text:

Trial Cut: 1988 Cut 1989
30 cm Group 1 Group 2
S cm died Data omitted
Spacings Group 1 Group 2

Seed sources Group 3 Group 4

 

112

Grid plots at two harvest heights

The field was divided equally for the two year-of-first-harvest treatments. Each treatment
was on one side of the field, so the effect of soil variation was uncontrolled. Within each side, a
randomized complete block design was used with 10 plots in each of four replications. Each
replication contained all combinations of two height treatments and five spacing treatments.
Spacing treatments were included in the analysis of variance, but will not be reported here.

In late August 1988, 65 days from sowing when most seedlings were 20 - 30 cm tall, plots
in one halfof the field were harvested at either 5 cm or 20 cm above the ground. The weak
regrowth was consumed by rabbits during the winter. Most 5 cm stumps died and were frost-
heaved out of the ground before spring, so those plots were discontinued. Survivors at 5 cm were
counted in July, 1989, but not harvested. Survivors at 20 cm were harvested in July, 1989 and at
later dates; results are given in the report of the spacing trial (Chapter 4).

In March 1989, the previously uncut plots in the other half of the field were cut back to 5
or 20 em (but not harvested and measured) while dormant. These plots were harvested at 5 and 25
cm between June 28 and July 5, and the numbers harvested were counted for comparison with the

plots first cut in 1988.

Grid plots at five spacings
This trial used a subset of the plots planted for the harvest height trial described above.

The field was divided for the two year treatnnents, with each side containing four replications in a
randomized complete block design. Each replication consisted of five plots representing all spacing
treatments. At the July harvest in 1989, yield data were recorded from three replications only, but
plant numbers were recorded from all replications. Cumulative yields which included July 1989
data are also reported for three replications, resulting in inconsistent tables.

Seeds were planted by hand in June, 1988, in a grid pattern, 10, 12.5, 16.7, 25, or 50 cm
apart, to give 100, 64, 36, 16, or 4 seedling positions per 1.0 m2 plot, respectively, with one or two

border rows on each side. The size of the planted area was about 4.0 m2 for each plot. Multiple

 

 

113

seedlings were frequent because two to four seeds were sown at each grid position. Gaps were
replanted with swds in early July, and with seedlings (transplanted out of designated pathways) in
early August.

In late August 1988, 65 days from sowing when most seedlings were 20 - 30 cm tall, plots
in one halfof the field were harvested at 20 cm above the ground. This set of plots harvested at 20
cm in 1988 was designated Group 1. In March 1989, plots on the uncut side of the field were cut
back to 20 cm while dormant, and the stems left to decay on the ground. This set of plots first cut
at 20 cm in 1989 was designated Group 2.

Plots in both groups were harvested at 25 cm between June 28 and July 5, and again at 30
cm between August 23 and 31, 1989. In 1990, first harvests were taken between June 18 and 22,
and second harvests between August 16 and 27, all at 30 cm. In 1991, because of the reduced
availability of labor, first harvests were taken fi'om plots of 4 and 16/m2 only, between July 10 and
17, at a height of 50 cm. Because of shading fiom nearby uncut plots and feeding by rabbits, no
second harvest was possible in 1991, and after counting survivors at all spacings, the trial was

terminated.

Row glots with five seed sources

The trial included as treatments five seed sources (four acessions and a mixture for the
control) and two years of first harvest. The field was divided in half for the year treatments, so the
effect of soil variation was uncontrolled. Each half contained four replications of five plots in a
randomized complete block design. In July 1989 yields were recorded from tlnree replications only.

Plots consisted of five rows 20 cm apart, with an area of 1.00 m2. Two outside border
rows on the east and west sides were planted with the bulk mixture, giving nine rows with a total
width of 1.8 m. Interior border plants on the north and south sides were from the same seed source
as the plot. Plots were sown by hand with 50 seeds to give stands of 25/m2, and small seedlings

were transplanted where necessary. Spacing of survivors within the five rows was random.

 

 

114

Half the field was harvested at 20 cm in August 1988. The other half was cut back to 20
cm while dormant in March, 1989. All plots were hand harvested at 25 cm in July 1989 and at 30
cm in August 1989, June 1990, and August 1990. The trial was ended at the August 1990 harvest.
During the August 1988 harvest, plots were thinned to at most 30 seedlings/m2. The unharvested
plots were not thinned, but mortality over the winter resulted in nearly equal populations on both
halves of the field.

RESULTS AND DISCUSSION

Dgy weight yields ger glot

In the grid plots, effects of year treatments in 1989 were highly significant in July and non-
significant in August, while the interaction with spacing treatments was not significant at either
time (T able 5.1). The July average yield of 380 g/m2 for goup l was 75% of the yield for group
2, and even combined with the 1988 average of 76 g/m2, it reached only 90% of the yield of goup
2. In August there was no significant variation in regowth yields; group 1 averaged 95% of group
2. The yields at both harvests in 1990 varied highly significantly with year of first harvest. The
goup 1 plots averaged 57% and 55% of the yields of group 2 in June and August, respectively.

Although gowing conditions were excellent botln years, and fertilizer was added at the
start of the 1990 season, yields declined from 1989 to 1990; the total 1990 yield was 43% below
1989 for group 1, compared with 19% below for group 2, so that group 1 yields declined from
79% to 55% of the yields of group 2. '

Cumulative yields from 1988 to 1990 of group 1 compared with group 2 ranged from 56%
at 4/m2 to 99% at 16/m2 (Table 5.2), averaging 77% (Table 5.1). The interaction of spacing and
year treatments was not statistically significant for any single harvest, but was significant for the

1989 total yield. For the subset of plots of harvested in 1991 (4 and l6/m2 only), yield differences

 

 

 

 

115

Table 5.1. Dry weight yields per plot (g/mz) from grid plots with first harvests in the seedling
year (group 1) or a year after sowing (group 2).

 

 

 

 

Year treatment group Probability of F-value

Harvest l 2 Year Spam Interactionr
8/1988 76 - - .0000 -
7/19892 380 504 .0015 .0004 .0641
8/1989 109 115 .6202 .2221 .5867

All 19892 489 619 .0005 .0000 .0196
6/1990 99 174 .0003 .8015 .5965
8/1990 179 327 .0002 .7083 .7739

All 1990 278 501 .0000 .7902 .6106
Tota12 857 1112 .0000 .0000 .1596

 

1 Interaction of first harvest year and spacing factors.
2 Three of four replications used.

 

116

Table 5.2. Single harvest and cumulative dry weight yields (g/rn2) from grid plots of4 and Wm2
first harvested the seedling year (group 1, 1988 - 1991) or a year after sowing (group 2, 1989 -

 

 

 

1991)

Probabilityof F-value
Harvest 1 2 l 2 Year Spacing Interalctio

n

8/1988 8 -- 60 -- ---- .0216 ----
7/1989 103 233 490 412 .5084 .0002 .0284
8/1989 112 117 131 125 .9760 .4305 .7275
6/1990 98 163 147 168 .0713 . 1957 .4623
8/1990 153 330 175 305 .0186 .9761 .6326
7/1991 252 579 126 316 .1366 .2445 .6640
Total 726 1422 1129 1336 .0754 .4780 .2853

 

T Interaction of first harvest year and spacing factors.

 

117

due to year of first harvest became significant only in August, 1990 (Table 5.2). Although
cumulative four-year yields of group 1 were 51% and 85% of the tlnree-year yields of goup 2 at
4/m2 and l6/m2, respectively, the year effect was not significant. The year by spacing interaction
was significant in July, 1989 only.

In the row plots, efl‘ects of year treatments were highly significant in July and non-
significant in August of 1989, while the interaction with swd sources was not significant at either
time (Table 5.3). For both harvests combined, year treatments were not statistically significant,
but the year by seed source interaction was. Group 3 yielded 70% as much as group 4 in July,
114% in August, 87% for the entire year, and 102% for 1988 and 1989 together. In 1990, effects
of the year treatment were significant at both harvests, and highly significant for both harvests
combined, while the interaction was not significant. Group 3 yielded 80% as much as goup 4 in
June, 76% in August, and 78% for the entire year. Average annual yield levels rose 35% for group
3 and 52% for goup 4 in 1990 compared to 1989. The tlnree-year total for goup 3 was 89% of
the two-year total for group 4, and not significantly different.

Savory e_t g. (1980) reported that leucaena forage yields tended to rise over the first four
years, presumably due to greater root penetration. Yields have risen over the first three years in
some of these black locust trials, including transplanted seedling plots which were not fertilized
(Barrett 1992; Chapter 4). In the directly sown plots at five densities, yields declined the third year
(Table 5.1), even with applied fertilizer. In the fourth year, 1991, only the two widest spacings
were harvested, in July only. Still, it is clear that the plots of 4/m2 gave higher yields each year, as
the canopy became more complete, and would have produced much more the fourth year if there
had been a second harvest (1‘ able 5.2). The yields of the l6/m2 plots did not continue rising, and
may have given their peak yields in the second year when the canopy first closed. Declining yields
in measured plots may have been due to shading by the more vigorous plants in the outside border

rows, so that the forage yield of the entire plantation may have continued rising.

 

118

Table 5.3. Dry weight yields per plot (g/mz) from row plots with first harvests in the seedling year
(group 3) or a year after sowing (group 4).

 

 

 

Year treatment group Probability of F-value

Harvest 3 4 Year Interaction1
8/1988 75 - - -
7/19892 226 322 .0089 .0760
8/1989 249 218 .1317 .6677
All 19892 438 501 .0968 .0422
6/1990 235 294 .0180 .6837
8/1990 357 469 .0150 .8465
All 1990 592 763 .0045 .8595
Tota12 1105 1240 .1032 .3676

 

1 Interaction of first harvest year and seed source factors.
2 Three of four replications used.

 

 

119

Genys and Harman (1990) recorded dry weights of third-year black locust coppice sprouts
of 1.3 to 7 .0 kg/m2, averaging 3.6 kg/m2 for 25 seed sources. Stump density averaged l7/m2, and
the weights included wood and bark but not foliage. Third-year coppice sprouts in the present
experiment were harvested in June and August, 1990. Cumulative 1990 yields, including foliage,
ranged from .278 kg/m2 in goup l to .763 kg/mZ in group 4. Factors that would account for the
much higher yields reported by Genys and Harman (1990) include the use of different seed sources,
a longer gowing season and different soils at their experimental site in Maryland, and harvesting E—
of coppice shoots only once a year, while dorrmnt, rather than twice a year during the gowing

season. Yield reductions of over 45% have been reported for harvesting twice compared to once l '

 

during the gowing season (Barrett 1992).

Two hypotheses, that not harvesting the first year gives higher yields the second year, and
that cumulative yields over three years would be lower in groups 1 and 3, were confirmed.
Another hypothesis, that cumulative yields for first and second years are not affected by year
treatment, was refitted by results from goups 1 and 2, but supported by results from groups 3 and
4. Yields were usually significantly higher for goups 2 and 4, and when all harvests from 1988
through 1990 were added together, goup 2 yielded highly significantly more than goup l, but

there was no significant difference between groups 3 and 4. But with the advantage of groups 2

 

and 4 increasing with time, it is likely that both groups would have yielded significantly more than

the other treatment over a longer period than the duration of these experiments.

Dgy weight yields ger glant

Yields of plots were irnfluenced by the number and vigor of survivors. Comparisons of dry
weight yields per plant were used to determine whether harvesting the first year had an effect on
vigor separate from the effect on survival.

In grid plots, year treatment effects were highly significant at the first harvest of 1989
(Table 5.4), but at the second harvest, the effect was not significant. Group 1 yields were 64% and

105% of Group 2 in July and August, respectively. In 1990, year treatment effects were

120

Table 5.4. Dry weight yields per plant (g) from grid plots with first harvests in the seedling year
(group 1) or a year after sowing (group 2).

 

 

 

Year treatment group Probability of F-value
Harvest 1 2 Year Interaction1
8/1988 2.6 - - -
7/19892 12.9 20.2 .0000 .0641
8/1989 7.7 7.3 .7095 .4851
6/1990 8.4 12.7 .0159 .0857
8/1990 14.6 22.9 .1097 .6239

 

1 Interaction of first harvest year and spacing factors.
2 Three of four replications used.

 

 

121

significant at the first harvest only. Group 1 yields were 66% and 64% of Group 2 in June and
August, respectively. The interaction with spacing was not significant at any time. In the subset
of plots of 4/m2 and l6/m2 harvested in 1991, the effects of the year treatment and the year by
spacing interaction were never statistically significant.

In row plots in July 1989, the effect of harvest year on weight per plant was highly
significant, with goup 3 yielding only 59% of goup 4 (Table 5.5). The effect was not significant
in August, 1989, when yields per plant were identical. The year effect was significant in both July
and August, 1990, when goup 3 yielded 80% and 77% of group 4, respectively. The interaction
with swd sources was never significant.

Yields per plant increased with each harvest in goups l and 3 from 1988 through 1990,
but the levels declined in goups 2 and 4 from July to August 1989, before increasing at the next
two harvests. Because each remaining plant has more space after some die, a slight increase in
yield per plant was expected over the years.

Genys and Harman (1990) stressed the importance of considering both weight per plot and
per plant when judging productivity of black locust. Their tallest seed source was the most
productive per plant at 256% of the mearn, but because tlnere were only eight such trees/m2
compared to an average of l7/m2, the productivity per plot was only 118% of the mean. Pecson
and Brewbaker (1991) noted that larger stumps in a leucaena spacing trial had more coppice
shoots, and that the correlation of number of shoots and stump basal area was highly significant.
When the first harvest of M macrantha was delayed three or four months (date of first
harvest not reported) during the growing season, average stump diameter increased from 24 to 42
mm (Holden gt g1. 1989).

Mebrahtu and Hanover (1989) found that black locust root cuttings of larger diameter
sprouted sooner and maintained a height growth advantage over cuttings of lesser diameter. Since
the results of this study indicated no sprouting time difference for coppiced stumps related to
diameter in 1989 or 1990, it is uncertain whether root biomass size, which is related to age at first

harvest, would have had an effect on growth in the field trials.

 

 

122

Table 5 .5 . Dry weight yields per plant (g) from row plots with first harvests in the seedling year
(group 3) or a year after sowing (group 4).

 

 

 

Year treatment group Probability of F—value
Harvest 3 4 Year InteractionI
8/1988 3.2 - - -
7/19892 8.9 15.1 .0000 .0713
8/1989 10.1 10.1 .9893 .4382
6/1990 10.2 12.8 .0126 .3337
8/1990 16.1 20.9 .0262 .9067

 

1 Interaction of first harvest year and seed source factors.
2 Three of four replications used.

 

 

123

In this study, the difference in weights per plant between year treatments, when significant,
always favored goups 2 and 4 which were not cut until the second year. Except for the interval

between July and August, 1989, the weight per plant increased with time, as expected.

Changes in dgy weight I fresh weight ratio

The dry weight/fresh weight ratio (D/F) represents the dry matter percentage of the geen
biomass. Significant differences in the D/F were found between the grid plot groups in 1989 only.
The D/F ratios for goups 1 and 2, respectively, were 23.4% and 25.5% in July, and 23.9% and
24.7% in August, 1989. For the row plot groups, the year effect was significant in July 1989, and
highly significant at the three following harvests, while the interaction with seed sources was never
significant. The D/F ratios for goups 3 and 4, respectively, were 24.6% and 28.3% in July, 1989,
24.6% and 25.0% in August, 1989, 24.5% and 25.1% in July, 1990, and 25.2% and 27.1% in
August, 1990.

The greatest difference between year treatments found in four seasons was 3.7% moisture
by weight, and it was usually under 2.0%. This slight difference in D/F appeared to depend on the
size of harvested stems. In every case of statistical significance, the goup first cut in 1988 had
less dry matter content than the goup cut in 1989. Delaying harvesting until the second season
produced larger stems with a higher proportion of wood, and a thus a higher D/F percentage. The

hypothesis that D/F the second year is lower for plots not harvested the first year was refuted.

Number of branches harvested in row plots

In 1989 and 1990 when the row plots were harvested at 30 cm, the number of upright
branches cut was recorded (Table 5.6), and the average number of branches per plant harvested
was calculated.

The number of branches per plot was significantly higher in group 3 than in group 4 in
July (17% higher) and August 1989 (31% higher) (Table 5.6). In June, 1990, the number per plot
was slightly higher for group 4, and in August group 3 had only 80% as many cut branches as

group 4. a highly significant difference. Numbers harvested increased from the first harvest to

 

 

124

Table 5 .6. Numbers of upright branches cut per plant and per plot from row plots with first

harvests in the seedling year (group 3) or a year after sowing (group 4).

 

 

 

Year treatment gLoup Probability of F-value

Harvest 3 4 Year Interaction1
8/1988

Plant 1.1 - - -

Plot 27 - - -
7/19892

Plant 2.2 2.3 .5492 .0230

Plot 55 47 .0500 .0071
8/1989

Plant 3.4 3.1 .0487 .0324

Plot 84 68 .0025 .0097
6/1990

Plant 4.4 4.8 .0665 .8843

Plot 98 110 .1233 .9129
8/1990

Plant 2.8 3.6 .0003 .7901

Plot 66 83 .0016 .9921

 

T Interaction of first harvest year and seed source factors.

 

 

125

June, 1990, but declined at the last harvest. The interaction with seed sources was highly
significant at both harvests in 1989, but not significant at either harvest in 1990.

The average number of branches harvested per plant varied significantly due to year
treatments at the August harvests both years, but not in the early summer harvests (Table 5.6).
Group 3 averaged fewer branches than goup 4, except in August, 1989. The widest variation was
in August, 1990, when goup 3 averaged 7 8% of goup 4. Average numbers increased from
August, 1988 to June, 1990, but declined in August, 1990. The interaction with seed sources was
significant at both harvests in 1989, but not significant at eitlner harvest in 1990.

Average stem numbers per plant and diameters declined as density increased in leucaena
plots (Cooksley and Goward 1988). The same effect on branch numbers was observed in black
locust plots (Barrett, unpublished data). The efl‘ect of delaying the first harvest was to produce
significantly fewer branches per plant and per plot in 1989 and significantly more in 1990. The
hypothesis was supported that the number of leaders harvested the second year is higher for plots
harvested the first year, but the pattern reversed in the tlnird year, making any conclusions
uncertain. The important factor related to yields is not the number of branches, but how rapidly
the plants can form a closed canopy, and it was clear that this could have been accomplished with
fewer than 10 plants/m2. When there were over 20 plants/m2, it mattered little how many

branches each plant had.

Survival gftgr harvesting
If all grid spaces had been filled with one seedling, the mean of all spacings would be

44 plants per plot. All grid plots were sown alike with extra seeds, giving an average of 51
for the 5 cm plots in 1988. It should be assumed that all treatments had similar surplus
germination, but that some seedlings died over the first winter. The group 1 plots harvested at
20 cm in 1988 averaged 46 plants/m2 harvested in July, 1989, representing 90% of the 51/m2
presumed to have been present in 1988. Therefore, winter mortality was about 10% for those

out at 20 cm (group 1), while it was 73% for seedlings cut at 5 cm in 1988.

 

 

 

126

Table 5.7 gives a false impression of survival, because at the first harvest in August,
1988, not all seedlings exceeded the 20 cm harvest height. Plants of group 1 that escaped
cutting continued to grow for the rest of the season and survived well over the winter, giving
an increase of 21% over the 1988 level when harvested in July 1989.

Year effects on the numbers harvested in grid plots were significant in July but not in
August, 1989, when goup 1 outnumbered group 2. In 1990, group 2 had significantly more plants
harvested at both dates. Group 2 had higher yields per plot and per plant botln years (Tables 5.1
and 5.4). The survival difl‘erence was not significant in 1991, and the interaction with spacing was

never significant. The most deaths occurred over the winters and in goup 1.

 

In row plots, the average numbers harvested in goup 3 rose from 1988 to July 1989, and
then declined at the next two harvests before rising again at the last harvest (Table 5.8). In group
4, numbers harvested increased at each harvest. These increases were caused by the height growth
of plants which were below the harvest height at one harvest and above it at the next, and not by
delayed germination of surplus seeds. By harvesting in early and late summer, the amount of
sunlight reaching the weak seedlings was increased, and the relative advantage of stronger
seedlings was reduced. Although numbers harvested were significantly lower for group 4
compared to group 3 at both 1989 harvests, there were no significant differences in 1990. The
interaction with seed sources was significant in August 1989.

The two hypotheses, that survival is not affected by harvesting the first year, and that
survival is higher with older plants, were refuted by data from grid plots, while the pattern in row

plots was inconclusive, except that numbers harvested could not continue increasing indefinitely.

 

 

127

Table 5.7. Number of plants harvested from grid plots with first harvests in the seedling year
(group 1) or a year after sowing (group 2).

 

 

 

 

Year treatment group Probability of F-value

Harvest 1 2 Year Interactionl
8/1988 38 - - -
7/1989 46 36 .0142 .1438
8/1989 35 31 .1045 .1953
6/1990 20 26 .0193 .1443
8/1990 23 29 .0373 .2083
7/1991 11 13 .4431 .2133

 

1 Interaction of first harvest year and spacing factors.

 

 

 

128

Table 5 .8. Number harvested from row plots with first harvests in the seedling year (group 3) or
a year after sowing (group 4).

 

 

 

Year treatment goup Probability of F-value
Harvest 3 4 Year InteractionI
8/1988 23.5 - - -
7/1989 25.1 19.8 .0006 .2368
8/1989 24.9 22.5 .0416 .0359
6/1990 22.6 23.0 .7179 .4366
8/1990 23.5 23.3 .8322 .3148

 

1 Interaction of first harvest year and seed source factors.

 

 

 

129

Interactions of year treatments with other factors

Many responses to year treatments varied according to seed source or spacing, so that the
last hypothesis, that no statistically significant interactions would be found, was refuted.
Examination of the eight tables in this chapter shows that out of 22 and 23 possible interactions
with spacing and seed sources, respectively, zero and two were significant at the 1% level, with
two and four more interactions significant at the 5% level. The interactions with spacing were
significant at the 5% level twice out of 22 times, which can easily be attributed to chance, while
only 17 of 23 interactions with seed sources were non-significant. Genotype by treatment
interactions were important, especially concerning branching behavior, and future researchers

should be alert for them.

Delaying the initial harvest beyond the second growing season

Pawlick (1989) warned not to coppice trees until they are well established, and suggested
an acceptable height of 300 - 400 cnn, nine months to two years after planting. The time of first
harvest treatments on black locust were on plants two months old and about 30 cm tall, compared
to plants about 100 cm tall with four months of growth. Under cultivation, black locust trees
would usually reach a height of 300 - 400 cm in the second growing season, 15 - 18 montlns after
sowing. Waiting to harvest black locust seedlings until the tlnird growing season would certainly
increase the annual yield, but it is uncertain when or if cumulative yields would surpass those of
plantings harvested the second year, and whether such a delay could be economically justified in
practice. Commercial production of leucaena for leaf meal in Malawi starts when the plants are
about 100 cm tall, one year after sowing; yields of leafy material are usually less than 0.3 tons dry
weight per ha (Savory e_t a_l. 1980). This would be under 30 g/m2, compared to 75 g/m7— of leafy
material harvested two months after sowing in this study, and far lower than the yields one year
after sowing. Thus, yields of black locust would appear to be adequate for commercial production.

Considering the high costs of either removing the large stumps from a depleted coppice

stand and replanting mechanically, or transplanting large numbers of replacement seedlings by

 

 

 

130

hand, delaying the first harvest until the tlnird year appears advisable if it substantially reduces
mortality. Another reason to consider a furtlncr delay is when direct sowing is not practical and the
trees must be established by transplanting. To reduce planting costs, fewer seedlings would be
planted at a wider spacing, necessitating a longer establishment period to close the canopy. Future
research should address how long to delay the first harvest, in combination with wider spacing

treatments than those tested here, and harvest heights above 30 cm.

 

 

 

131

REFERENCES CITED

Barrett, R. P. 1992. Forage plantation management for black locust. pp. 259-272 in: Hanover, J.
W., K. Miller, and S. Plesko, eds. Proceedings: Intennational conference on black locust:
biology, culture, & utilization. Department of Forestry, Michigan State University.

Bhatt, B. P., and N. P. Todaria. 1991. Biomass production in some leguminous taxa under a short
rotation cycle. Nitrogen Fixing Tree Res. Reports 9:4-5.

Blair, G., D. Catchpoole, and P. Home. 1990. Forage tree legumes: their management and
contribution to the nitrogen economy of wet and humid tropical environments. Advances in
Agronomy 44:27-54.

Blake, T. J. 1983. Coppice systems for short-rotation intensive forestry: the influence of cultural,
seasonal, and plant factors. Austr. For. Res. 13:279-291.

Cooksley, D. G., and E. A. Goward. 1988. Effect of plant density and spatial arrangement on the
yield of Leucaena leucocephala cv. Peru in subcoastal soutln-eastem Queensland. Austr. J.
of Exp. Agr. 28:577-585.

Galang, M. C., R. C. Gutteridge, and H. M. Shelton. 1990. The effect of cutting height and

frequency on the productivity of Sesbania sesban var. nubica in a subtropical environment.
Nitrogen Fixing Tree Res. Reports 8: 161-164.

Genys, J. B., and D. M. Harman. 1990. Racial diversity of black locust in gowth rate and in
susceptibility to the locust twig borer. N. J. Appl. For. 7:38-43.

Hintze, J. L. 1987. Number cruncher statistical system user's manual, version 5.0. J. L. Hintze,
Kaysville, UT.

Holden, S. R., S. Lungu, and J. Volk. 1989. Growth rate and coppicing ability of Sesbania
macrantha after repeated prunings at various heights. Nitrogen Fixing Tree Res. Reports
7: 132-134.

Mebrahtu, T. 1992. Growth and photosynthesis of black locust. in: Proceedings of lst lntemational
Conference on Black Locust, Michigan State University, East Lansing, MI, USA; June 17-
20, 1991.

Mebrahtu, T., and J. W. Hanover. 1990. The effect of root cutting size on time of sprouting of
black locust. Nitrogen Fixing Tree Res. Reports 8: 156-158.

Pawlick, T. 1989. Coppice with care. Agroforestry Today 1(3):15-17.

Pecson, R. D., and J. L. Brewbaker. 1991. Coppice shoot regrowth of leucaena. Leucaena
Research Reprots 12: l 19- 120.

Savory, R., J. A. Breen, and C. l. A. Beale. 1980. Leucaena as a forage crop on small farms in
Malawi. pp. 411-414 in: Le Houerou, H. N., ed. Browse in Africa: the current state of
knowledge. ILCA, Addis Ababa, Ethiopia.

 

H

 

 

132

Soil Conservation Service, 1979. Soil survey of Inglnam County, Michigan. USDA, Washington
DC.

 

 

 

AN IDEOTYPE FOR TREE SPECIES GROWN AS PERENNIAL FORAGE
CROPS, WITH SPECIAL REFERENCE TO BLACK LOCUST

INTRODUCTION
The black locust tree (Robinia pseudoacacia L.) has rapid gowtln, nitrogen fixation

capability, high protein content, a deep taproot, and many other features that would make it a good

forage crop plant. By gowing black locust seedlings at close spacings with annual coppicing at 30

cnn, annual dry weight yields of up to 14 tons/ha were obtained (Barrett 1992). Yield
improvements are possible from refinements in cultivation practices, and also from selective
breeding.

Selecting for high yielding ability from a diverse wild population can take two approaches,
conventional breeding or ideotype breeding. Under the conventional approach, large numbers of
genotypes gown under standard conditions are rated for productivity, and the best retained for the
breeding population. By selective pollination, a large number of progeny are produced, and the
cycle is repeated. Testing the progeny to select the parents is a variation on the conventional
approach. The ideotype approach is to define characteristics known or expected to improve
productivity, and then select individuals for the breeding population according to how closely they
conform to the list of desired features, rather than at random. By not measuring yield in the first
generation, a great deal of time and effort can be saved, and a larger population can be screened.
This is especially important when breeding trees, which can take many years to reach reproductive
age. Selecting without measuring yield is also an advantage for forage crops, because the seed-

bearing portion of the plant is generally removed by harvesting. When breeding black locust or

133

 

h

 

 

 

134

any other woody species for use as a forage crop, the ideotype approach should be seriously

considered.

DEFINITION OF IDEOTYPE

Ideotype refers to a biological model expected to perform well in a given environment. It
is an idea, and an ideal, which plant breeders strive for in cultivar development. Attaining this goal
is expected to result in higher yields of useful products, or at least higher profits to the gower.
The word ideotype should not be confused with idiotype, properly used to describe all the
hereditary influences on an organism, and more inclusive than the word genotype. Phonetically,
ideotype is accented on the second syllable, and idiotype on the first (Zevin 1975). In practice, an
ideotype must involve heredity, but it describes an unreached ideal, while an idiotype already
exists.

As Rasmusson (1987) wrote, "An ideotype is a hypothetical plant described in terms of
traits that are thought to enhance genetic yield potential. Ideotype bmding is defined as a method
of breeding to enhance genetic yield potential based on modifying individual traits where the
breeding goal (phenotype) for each trait is specified."

THE IDEOTYPES OF CULTIVATED PLANTS

Donald and Harnblin (1976) defined three types of ecosystems for cultivating plants, as
widely spaced plants without competition, as a mixed community in competition, and as a dense
monoculture. The criteria for high yields are different in each situation, requiring "isolation",
"competition", and "crop" ideotypes, respectively. The isolation ideotype may be useful for
orchard or landscape trees, but it is not pertinent to this discussion. The competition ideotype must
be able to gain a larger share of nutrients, water, and sunlight at the expense of its neighbors, so it

is rapid growing, tall, spreading, and leafy.

 

 

135

In contrast, the crop ideotype is erect and narrow, with small erect leaves, which do not
shade its neighbors. Leaves, roots, and branches of individuals are at the minimum level required
to fully exploit the resource by the entire community rather than by single plants, which must be
weak competitors. Plant parts not included in the harvest are a small portion of the total biomass.
In a dense monoculture, yield per area becomes more important than yield per individual plant.
Donald and Harnblin (1976), writing about cereal grains, emphasized that a successful cultivar will
have a geater biological yield at high density or fertility, and will maintain a high harvest index
(the harvested part being a high percentage of total biomass). These principles also apply to tree

plantations and forage crops.

GENERAL OBSTACLES TO IDEOTYPE BREEDING

Usually, ideotype breeding proceeds in three steps. First, the important traits are selected
and a phenotypic goal is set for each. Next, adequate genetic diversity for a breeding program
must be verified. Finally, each trait must be tested, in different genetic backgrounds and under a
variety of conditions (Rasmusson 1987). Then the traits that are known to improve yield can be
combined into cultivars. However, they do not always behave as expected.

Four types of trait interrelationships that often slow breeding programs are symmetry or
harmony in size of plant parts, compensation among plant parts, plciotropy, and genetic
background. These were noted as obstacles to ideotype breeding of cereal grains (Rasmussen
1987), but apply as well to any type of breeding program for any plant. Symmetry requires a plant
to maintain a balance among its components, so that it is impossible to enlarge the harvested
portion at the expense of the other parts. Compensation occurs in the yield of a product that
depends on more than one yield component. For example, an effort to raise the yield of seeds by
increasing the number per fruit may fail because fewer fruits or smaller seeds are produced. With
plciotropy, a gene that was presumed to raise yield actually has negative effects that reduce it.

When desired genes are available only in low-yielding parents, many breeding cycles are required

 

 

 

 

136

to incorporate the positive genes and remove the negative ones. It is possible that some genes will
be rejected as negative, when they would have raised yield if placed in a different genetic
backgound.

A general obstacle to tree breeding is the length of time required, both to select for quality
of the harvested portion, and to obtain the next generation. Because bf this, and the fact that
effective selection is limited to two or at most three traits at a time, most ideotype criteria must be

met by choosing a suitable multi-purpose tree at the species level.

IDEOTYPES OF CULTIVATED TREES

The two factors with the most influence on an ideotype are the agoecosystem where it is
to be gown, and the commodity or service to be produced. Many trees, especially legumes, can be
managed for multiple products, often simultaneously. Black locust has been grown for honey,
forage, fuel, poles, posts, pulpwood, and lumber.

The first tree crop ideotype published was for producing wood fiber under short rotation
intensive culture. Dickrnann (1975) described an ideotype with eight features:

. rapid juvenile gowth

. indeternninate shoot gowth

. upright single stem habit

. narrow crown with steep branch angle

. high shoot/root ratio (aided by good nutrition)

. ease of establishment and regeneration

. acceptable product quality

. freedom from major insect and microbial pests.

”Nam-BOON—

Later, Dicknnann (1985) expanded the ideotype by adding other qualities:

9. low competition between neighbors, resulting in:
A. stands with a narrow range of stem diameters
B. low mortality before harvest
C. high productivity per unit of area
10. high response to inputs, such as:
A. weed control before crown closure
B. adequate nutrient levels
1 l. nitrogen fixation or ability to dominate n-fixing intercrop

 

137

12. phenotypic plasticity allowing rapid crown closure but little competition later with
neighbors, OR
13. tolerance to common herbicides
14. optimal phenology for rapid gowth in environment:
A. rapid gowth early in season
B. leaves retained until late autumn
C. cambium active until late in season
15. delayed reproduction and/or minimal fi'uiting
16. short intemodes to maximize leaf number
17. strong taproot.

With the exception of conifer breeders in Finland, the ideotype concept has not been
popular among forest geneticists, and few tree ideotypes have been published (D. I. Dickmann,
personal communication). Whether the word ideotype was used or not, a number of authors have
provided lists of desirable features for cultivated trees. While the designated agoecosystems
ranged from short rotation intensive culture systenns for the temperate zone to subsistence farms in
the tropics, all the following publications were useful in preparing an ideotype for a tree grown as a
perennial forage crop.

F elker and Bandurski (1979) proposed a tree crop ideotype to nninirnize expenditures of
capital, fossil fuels, and machinery. Anderson and Zsuffa (1984) proposed 11 ideotype features
for biomass plantations of 1’ng and SE. Kang _e_t al. (1984) gave nine desirable criteria for
woody species suitable for alley cropping. Ranney e_t a_l. (1985) stressed several parameters to
evaluate in selecting species and cultivars for short rotation energy and fiber use, and Koski and
Vihera-Aamio (1986) mentioned 11 items. Von Maydell (1989) discussed eight features required
by food producing trees and shrubs in semi-arid regions, but the qualifications apply to multi-
purpose trees generally. Chuntanaparb and Ranganatlnan (1990) reported on the ideotypes of
multi-purpose trees preferred by farmers in six countries of south and southeast Asia. Ponce e_t 11.
(1991) investigated the multi-purpose tree species ideotypes preferred by farmers in two villages on
the island of Leyte, in the Philippines. Anderson 531 a_l. (1991) considered eight desirable attributes
for tree genera suitable for short rotation forestry in southern Ontario, and selected P_oanl_l_u_s, Ell,

Robinia, and Alnus. The six items that appeared most frequently in the above lists and ideotypes

 

138

were easy establishment, rapid growtln, deep roots, nitrogen fixation, rapid regeneration after

pruning or coppicing, and resistance to insects and disease.

PERENNIAL FORAGE CROP TREE IDEOTYPE

This list is a summary compilation of a number of previous lists. It must be kept in mind
that not all features are necessary for improving forage yields, that no single species or cultivar can
be expected to posess all of them, and that some of the features may be mutually exclusive. Two
debatable points fi'om Table 6.3 require clarification. If optimal phenology for rapid gowth is
taken to mean early bud break to nnaxirnize the length of the gowing season, tlnen the relatively late
bud break of black locust is not optimal. If it is defined as late bud break to avoid spring frost
damage, then black locust fails again, for in central Michigan in 1989 and 1992 the flower buds
were killed by frosts in May. Although the phenology of black locust may indeed be judged
inferior to most other tree species, its growth rate is more rapid despite tlnese failings. On another
point, most authors emphasize the vulnerability of black locust to borer insects. While insect
damage typically renders the wood unfit for commercial lumber, it would not impair the production

of forage from small or coppiced trees.

 

139

Table 6.1. Root, stem, and foliage features of a perennial forage crop tree ideotype.

 

Plant Part
Roots

Stems and
branches

Foliage

Feature

1. nitrogen fixation by symbiotic microbesl

2. deep, extensive root system to survive droughtl
3. adaptable surface root system to nourish plantl

. rapid gowth early in season4

. rapid regeneration after pruning stem tips2
indeterminate shoot gowthl

. short intemodes to maximize leaf number4

. leaves spirally arranged on stem2

. no thorrns or spines4

. cambium active until late in season3

. coarse bark to protect fiom wildlife damage4
9. good forage quality of new growth5

10. phenotypic plasticity, allowing rapid crown closure but little competition
later with neighbors2

11. rapid healing of wounds4

1. rapid leaf area gowth early in season4
2. rapid regeneration after pruning2
3. narrow or compound leaves to nninirrnize self-shading!
4. high light saturation levell
5. high net photosynthesis rate!
6. low light transmission through heavy foliageZ
7. long leaf retention time (until late autumn)4
8. rapid leaf position adjustmentl
9. good fodder qualities:
A. no hazardous toxins or allergins of livestock
B. low tannin content5
C. low lignin content
D. high caloric value5
E. high proteirn contentl
F. good protein quality
G. high digestibilityS

lTypical of black locust according to Hanover (1990).

2Typical of black locust based on observations by the author.
3Typical of black locust according to other published references.
4Not typical of black locust based on observations by the author.
5Not typical of black locust according to other published references.

 

 

140

Table 6.2. Propagation, reproduction, and stand establishment features of a perennial forage crop
tree ideotype.

 

Operations Features

Propagation and 1. early fi'uiting to allow rapid selective breedingl

reproduction 2. abundant seed crops with minimal effect on growthl
3. seeds easy to remove from fruitl
4. inexpensive seedl

5. readily available swdl

6. easily stored seed (not perishable, long viability)!

7. easily sown seedl

8. rapid gerrninationl

9. seedlings resistant to soil fungi

10. ease of vegetative propagationl

11. vigorous sprouting of root cuttingsl

12. vigorous rooting of stem cuttings4

13. case of surface-sterilization and nnicropropagationl

l4. vigorous cells in culture?!

15. vigorous regenerated plantlets from tissue culture2

16. strong inheritance of desired qualities (dominant genes)1

17. much genetic variation within speciesl

18. much genetic variation in hybridizable species3

 

Establishment 1. case of establishment and regeneration2

2. high response to inputs such as weed control, fertilizer2
3. efficiency of nutrient utilization

4. low danger of escape from cultivation4

5. ability to prevent soil erosion2

6. suitability for monoculture plantations2

7.

case of eradication when no longer usefn112

 

lTypical of black locust according to Hanover (1990).

2Typical of black locust based on observations by the author.
3Typical of black locust according to other published references.
4Not typical of black locust based on observations by the author.
5Not typical of black locust according to other published references.

141

Table 6.3. Environmental adaptation features of a perennial forage crop tree ideotype.

 

Adaptations Features

Growth rate 1. optimal phenology for rapid gowth in environment3.4(debmblc)
. rapid juvenile gowthl

. rapid crown closure3

. ability to outcompete wwdsl

#0319

Tolerances . saline goundwater

. saline soils

. low fertility soill

. drought stressl

. saturated soil6

. extreme low and high temperaturesl

frost2

. air pollutantsl

. coppicingZ

10. common herbicides used for weed control4

11. herbivore browsing2

12. intentional human damage or mis-management3
l3, insect pest52.5(dcbetablc)

14. diseases2

15. does not harbor agricultural pests or diseases2

\ooorn'axutewrot—

 

 

lTypical of black locust according to Hanover (1990).

2Typical of black locust based on observations by the author.
3Typical of black locust according to other published references.
4Not typical of black locust based on observations by the author.
5Not typical of black locust according to other published references.
6Not typical of black locust according to Hanover(l990).

 

142

REFERENCES CITED

Anderson, H. W., J. A. Rice, C. P. Chen, and J. A. Winters. 1991. Higlnspeed forestry as an
agroforestry option. pp. 109-115 in: Williams, P., ed. Agoforestny in North America/
Proceedings of the first conference on agroforestry in North America, August 13-16, 1989,
Guelph, Ontario, Canada.

Anderson, H. W., and L. Zsuffa. 1984. Genetically improved hardwoods as a potential energy
source in Ontario. pp. 211-214 in: Proceedings of Energex 84, the Global Energy Forum,
Regina, Saskatchewan, Canada; May 14-19, 1984.

Barrett, R. P. 1992. Forage plantation management for black locust. pp. 259-272 in: Hanover,
J. W., K. Miller, and S. Plesko, eds. Proceedings: International Conference on Black
Locust: Biology, Culture, & Utilization. Department of Forestry, Michigan State
University, East Lansirng, MI, USA; June 17-20, 1991.

Carlowitz, P. G. von. 1984. Multipurpose trees and shrubs, opportunities and limitations - the
establishment of a multipurpose tree data base. ICRAF working paper no. 17, Nairobi.

Chuntanaparb, L. and R. Ranganathan. 1990. Defining tree brwding objectives for multipurpose
tree species in Asia, Report # 10 in the Multipurpose tree species network research series.
Winrock lntemational - F/F RED, Bangkok.

Dickmann, D. I. 1975. Plant materials appropriate for intensive culture of wood-fiber in the North
Central region. Iowa State Journal of Research 49:281-286.

Dickmann, D. I. 1985. The ideotype concept applied to forest trees. in: Cannell, M. G. R., and J.
E. Jackson, eds. Attributes of trees as crop plants. Inst. Terrestrial Ecology, Huntingdon,
England.

Dicknnann, D. 1., K. Steinbeck, and T. Skinner. 1985. Leafarea and biomass in mixed and pure
plantations of sycamore and black locust in the Georgia Piedmont. Forest Science 31:509-
517.

Donald, C. M., and J. Hamblin. 1976. The biological yield and harvest index of cereals as
agronomic and plant breeding criteria. Advances in Agronomy 28:361-405.

Felker, P., and R. S. Bandurski. 1979. Uses and potential uses of leguminous trees for minimal
energy input agriculture. Economic Botany 33: 172-184.

Hanover, J. W. 1990. Physiological genetics of black locust (Robinia; pseudoacacia L.): a model
multipurpose tree species. pp. 175-183 in: Werner, D., and P. Muller, eds. Fast growing
trees and nitrogen fixing trees. Gustav Fischer Verlag, Stuttgart and New York.

 

 

Kang, B. T., G. F. Wilson, and T. L. Lawson. 1984. Alley cropping: a stable alternative to shifting
cultivation. lntemational Institute of Tropical Agriculture, Ibadan, Nigeria. Printed by
Balding + Mansell Ltd, England.

 

 

143

Koski, V. and A. Vihera-Aannio. 1986. Genn-plasms and conventional genetic programmes. Paper
presented at Biotechnology Activity Workshop. IEA, Uppsala.

Maydell, H. J. von. 1989 Criteria for the selection of food producing trees and shrubs in senni-arid
regions. pp. 66-75 in: Wickens, G. E., N. Haq, and P. Day, eds. New crops for food and
industry. Chapman and Hall, New York.

Ponce, E. R., L. B. Ponce, and L. A. Maurillo. 1991. Preferred characteristics of multipurpose tree
species: a case study with lowland and upland farmers in Leyte, Philippines. Report # 17
in the Multipurpose tree species network research series. Winrock lntemational -

F/F RED, Bangkok.

Rasmusson, D. C. 1987. An evaluation of ideotype breeding. Crop Science 27:1140-1146.

Ranney, J. W., R. D. Perlack, J. L. Trimble, and L. L. Wright. 1985. Specialized hardwood crops
for energy and fiber: status, impact, and need. Tappi Journal 68:36-41.

Zevirn, A. C. 1975. Editorial: idiotype or ideotype. Euphytica 24:565.

 

 

CONCLUSIONS

CONCLUSIONS FOR EACH TREATMENT AND ASPECT

Ideogyge breeding goals
The black locust has many of the qualities that are most useful for cultivated multipurpose

trees. Improvement of certain traits by selective breeding would produce cultivars or clones more
suitable for specific agricultural purposes, including alley cropping, short rotation biomass

production, and forage.

ALLEY QRQPPINQ

The most important traits to be added are:
1. Reduced propagation from cut roots.

2. Reduced extent of surface roots.

3. Spineless stems.

4. Increased nitrogen fixation.

SHORT ROTATIS 2N BIg QMASS

The most important traits to be added are:
1. Earlier bud break and/or later leaf drop.
2. Easier reproduction from stem cuttings.
3. Tolerance of saturated soil.

4. Resistance to insect pests.

FORAGE
The most important traits to be added are:

1. Higher digestibility of protein.

144

 

 

145

2. Higher digestibility of calories.
3. Spineless stems.
4. Longer retention of shaded lower leaves.

5. Improved horizontal branching for canopy closure.

Harvest date

Harvesting two times a year was found unsatisfactory. Harvest dates early in the season
were unfavorable because of low regrowth productivity, while tlnose in July or August were
rejected because of low survival. Since yields continued rising during September, a harvest date as

late as possible was indicated. Harvesting was necessary before frost caused the leaves to drop.

 

Harvest height
Harvesting at 5 cm damaged future survival and yields, except when done at the very end

of the gowing season. For every aspect of gowth, yield, and survival, the harvest height of 30 cm

was superior compared to 5 cm. Continued improvement with heights above 30 cm was projected.

Sgacing

With the 20 - 30 cm cutting height used in these trials, the narrowest spacing tested (10
cm) gave superior yields the first season only. The widest spacing tested (50 cm) gave yields
which were lower the first and second seasons, equal the third, and superior the fourth. Four
year total yields did not vary with density. The optimal spacing for survival was always the

widest.

Year of first harvest
In all trials, harvesting the first season reduced fiiture productivity and survival, often
severely. The small amount of forage available the first year was judged not worth the penalty

paid for harvesting it.

146

Long term trends

Survival rates did not increase over the years, although it is possible that beyond four
years they would improve as more space per plant becomes available. At a 30 cm harvest height,
yield per plant increased with time as survivors expanded into the spaces left by dying adjacent
plants, but this was not observed at a 5 cm harvest height. Yields per plot increased each year in

some trials and declined in others, but the pattern was not related to fertilizer application.

ITEMS FOR FURTHER STUDY

 

Harvest date

Several reasons have been proposed to explain why July and August harvests are more
stressful to black locust plants than June or September harvests, resulting in lower survival and
reduced yields the next year. Are these results due more to reduced carbohydrate storage or to
winter dieback? Which factors afi‘ecting net photosynthesis and carbohydrate storage are most
important, and which can be addressed by agricultural practices or breeding? Does disease play a
role in winter mortality?

Shoots that begin gowth before July generally respond to autumn signals and cease new
gowtln in late August or early September, but those that begin growth after June stay succulent
right up to frost, which wilts them immediately. In central Michigan, what is the latest date that
regrowth can begin and still enter dormancy normally? Does this date depend on day length (and
thus latitude if grown elsewhere), temperature, other weather factors, days from bud break,
genetics, harvest height, or other factors? It is possible that only the day length and leaf age
factors are involved; if the leaves are too young (under about 50 days of age) they may be unable
to detect the decline in day length and then signal the apex to cease growth.

Black locust leaves do not change color in autumn, indicating that they continue
functioning instead of senescing as in most other deciduous trees. Is this odd habit related to

nitrogen fixation? Can nodulated black locust plants obtain more nitrogen from the nodules by

147

continuing photosynthesis than from their own leaves by gradual senescence? How would the
leaves behave in autumn in a specimen that was not nodulated? Is the lower photosynthetic rate
recorded after fi'ost due to injury to the leaves, partial senescence, lower temperatures, or a
combination of these and other factors?

Complications involving harvesting after frost in 1989, and shading fi'om adjacent taller
plots every year, must have reduced the harvested weights of the late September plots, but the
amount is impossible to calculate. Witlnout those reductions in potential yield, would yields have

increased significantly from early to late September?

Harvest height
At what higher harvest height will yields begin to decline? Is this decline attributable to

 

competition for water, soil nutrients, light? Were diseases involved in the high deatln rate observed
at the 5 cm cutting height, and did the proximity of wounds to the soil surface favor their spread?
Is stem gowth fi'om coppiced stumps rapid only until a certain root:shoot ratio is restored, and

then much slower as a normal share of photosynthate is devoted to root gowth? In a tall crowded

 

stand with limited foliage on each plant, does maintenance respiration in the extra length of stems
reduce the amount of photosynthate available for gowth? Is the stern a major site for storage of
carbohydrates and/or nutrients over the winter? Is slower growth in taller plants related to the
passing of the juvenile stage? Questions of pathology and physiology arise, which can not be

answered by any of the data that were collected.

Spacing
Plants at spacings of 16 to 100/m2 were clearly exhausted by the fourth year, but how
many more years would plants at 4/m2 remain productive? Is it primarily competition for water,

soil nutrients, or light that makes the mortality rate increase with density in black locust?

148

Year of first harvest

The annual yield should increase when delaying the first harvest until the third growing
season, rather than the second. Would the cumulative yields ever exceed those of plantings
harvested the second year? Which treatnnent would be more profitable for the farmer? Future
research should address how long to delay the first harvest, in combination with wider spacing -
treatments than those tested here and harvest heights above 30 cm, in order to minimize the costs of

stand establishment.

Spacing by harvest height interaction
Because populations declined under the stress of frequent harvesting, especially over

winters, and never stabilized, it is uncertain whether an optimal density actually exists. Would
optimal density vary with harvest height as hypothesized? Would larger plants gown at wider
spacings and harvested at geater heights than tested in these trials survive longer and/or yield
more? Is the photosynthetic rate per unit of gound area maxirrnized by allowing space between
plants, so as to illuminate a geater number of leaves, or by maintaining a closed canopy?
Concerning yields, the main reason for significant interactions of spacing by harvest height
was that the plots of 4/m2 grew well at 30 cm and very poorly at 5 cm. Why did this occur?
Weeds and rabbits may have played a part, but the main reason for the extremely low yields of the
plants at 4/m2 cut at 5 cm was probably soil compaction from people stepping in the plots. The
1.5 m width of the plots was too far to step over, and the designated transverse pathways were
spaced 15 m apart. When measuring or harvesting the harvest height trial or the adjacent harvest
date trial, workers desired to switch pathways as quickly as possible to avoid being scratched by
spines of the tall growth on one side. Because the 4/m2 plots cut at 30 cm developed side branches
which impeded passage, the 5 cm plots were uniquely suited for use as shortcuts. Whether or not
this bias is severe enough to warrant discarding the data is academic; the conclusion that a 30 cm

harvest height is superior would not change. This problem could be avoided in the future by

 

 

149

providing more transverse pathways, or by making the paths between trials 2.5 - 3 .0 m wide
instead of 0.75 m.

Spacing by harvest date interaction

How do planting density and time of harvest interact? Is the optimal time for forage
harvest early in the season when shaded lower leaves are about to abscise, or is it at the end of the
season when the stems are woody and the leaves are about to abscise fi'om frost damage? The
answer is clear if the goal is to produce the most biomass, but if the goal is to produce the most

digestible nutrients for animal feed, the answer is not clear, and it may vary for different livestock

species.

Harvest date by harvest height interaction
In an unreplicated comparison, in September the plots cut at 5 cm yielded more than tlnose

cut at 25 cm. Was this a fluke or a typical result? How would yields and survival compare in the

long term?

Year of first harvest by seed source interaction
Seed sources often produced significant interactions with the year of first harvest

treatments, especially for branching behavior. Is genetic diversity adequate for breeding cultivars
that will tolerate harvesting the first season? Can heavily branching cultivars be planted at wider

spacings, allowing lower expenditures for plantation establishment?

Long term trends

Did random fluctuations give the illusion of rising or falling yields in different trials? How
many years can yields continue to rise before the limit is reached? Is mineral nutrient depletion a
concern in the long term? Would disease problems or innsect pest populations increase over the
years and eventually reduce yields? Can a closed canopy be maintained in the long term, or will
competition between plants result in gaps due to mortality of weaker specimens? Would

competition be reduced by planting a single clone, rather than a diverse mixture of seedlings?

 

 

150

Height growth
The hypothesis that the height of new growth is greater at lower cutting heights, but the

total height is less, was generally supported by the data, but no clear conclusion can be made
because the sample size was too small. Does this always occur?

The most rapid growth (averaged from ten plants) was in 1991, with a gain of 40 cm
between 47 and 56 days from bud break, averaging 4.4 cm/day. The most rapid height growth in
1989 and 1990 was recorded after 47 days from bud break. Is the peak growth rate at this time
due to hormonal factors, the leaf area, the rootzshoot ratio, or environmental conditions? Is it
because early gowtln requires less woody stems than later in the season, when more support is

needed to keep a taller and heavier stem upright? The speed record was 8.9 crn/day. How much

 

could agricultural treatments or breeding improve the gowth rate?

The slopes of the average height gowtln curves were very similar in 1989 and 1990, but
gowth was more rapid in 1991. Was this due to weather, plant age and size, or more space for
survivors following the severe mortality after the 1990 season?

What signals growth to stop in early September? Is it day length, the number of days since

bud break, soil moisture depletion, temperature, or a combination of factors?

ME».

After apical gowth ceased, dry weight yields continued increasing. Did stems, leaves, or
botln account for this increase? Were roots also gaining weight at this time? Although no distinct
plateau appeared as had been hypothesized, yields at the last two harvest dates were never
significantly different. Without the observed reductions in potential yield from frost and shading,
would normal yields increase significantly from early to late September?

Harvesting in late June or early July, and again in late August, yielded about half the
biomass of the September 28 plots. How would yields compare for leafy material only?

Declining yields in measured plots may have been due to shading by the more vigorous

plants in the outside border rows. so the forage yield of the entire plantation may have continued

 

151

rising. Future research should investigate patterns of plant gowth and competition, both above
and below the gound, in relation to yield levels. It is unclear whether the lower yields and survival
in 1990 compared to 1989 were due to harvest date effects, the age of the plants, or weather
conditions. What is the long term trend in yield levels? Can declining yields be raised by
fertilization? What are the optimal soil pH and fertility levels? Drought slowed growth of the new

seedlings in 198 8, but would a similar drought affect established plants as severely? How much

 

fartlner do the roots extend each year, vertically and laterally, and how is the pattern influenced by
the soil texture and drainage class? ;

Yields per plant H
Yields per plant generally increased with time. Was this due primarily to the geater -'

 

surface area occupied, or the presumed geater rooting depth? More space was available as

weaker neighboring plants died, but which was the limiting factor, light, nutrients, or water?

Yields per day
In 1989 the daily dry weight increment showed a plateau from late July onward.

Measurements of height gowth in 1989 were too infrequent to tell when it was fastest, but the
most rapid height gowth was recorded in June in 1990 and May in 1991. Was this apparent
inverse relation of height and dry weight growth rates due to weather, leaf area, or late season

secondary branching and stem lignification?

By I fresh weight proportion
What caused the dry weight percentage (D/F) to increase unsteadily each year in step-wise

fashion? In climates with a longer growing season, would D/F continue increasing beyond 50%?
Do larger stems with a higher proportion of wood always have a higher D/F?
Stem lignification proceeded through the entire season, but no difference in the D/F ratios

of leafy upper and woody lower stem portions appeared until after height growth had ceased.

 

152

During September, the succulent stem tips became less flexible and probably less succulent. Did

the leaves then became more succulent, and what physiological process would cause that?

Survival

Was it primarily competition for water, soil nutrients, or light that caused the mortality
rate to increase with density? Was the death rate inversely related to the content of stored
carbohydrates, as has been shown in other species? What is the normal seasonal pattern of
carbohydrate storage and depletion in black locust?

Was the severe mortality at 5 cm influenced by inadequate rooting deptln or volume,
possibly aggavated by evaporation from bare soil, finngal attack, herbivore attack, weed
competition, inadequate carbohydrate storage, hornnonal imbalance from total defoliation, frost
heaving, or a combination of these factors? By severely reducing the top in relation to the roots,
was root growth restricted to the extent that malnutrition or drought susceptibility resulted?

The planting density experiment was designed to find the optimal final density, where
yields were equivalent to those at higher densities and mortality of crowded plants ceased. In the
field, mortality at the widest spacing was lower but still substantial. Would a planting of identical
clones survive better tlnan the genetically diverse seedlings that were used? It is clear that
populations above 10/m2 cannot be maintained with annual harvesting at 30 cm, and the optimal
final density is probably below five plants/m2, if such a plateau even exists. What population
density can be maintained, and how many years are required to reach a stable level? How would
irnitial spacing, fertilization, climate, genetics, and different harvest times or heights affect the

pattern?

Canopy coverage
The canopy at 4/m2 only closed shortly before harvest in 1990 and 1991, but the taller

plants probably intercepted almost as much sunlight as those in more crowded plots. Did the
absence of a closed canopy account for the continued vigorous growth and better survival in 1991

observed at the 4/mz plots? If leaves around the sides of the plants do not abscise from shading,

 

 

 

153

will they provide more photosynthate per plant and per plot than in a closed canopy? At what
point would such an advantage disappear at wider spacings, lower harvest heights, or lower
latitudes, as more sunlight strikes the ground instead of the foliage? Plants at all narrower
spacings were clearly exhausted by the fourth year, but how many more years would plants at 4/m2
remain productive?

What density of planting would produce a canopy tlnick enough that chennical weed control
could be discontinued? The required density should decrease as the time span is extended or the
harvest height is raised, but how do these factors interact?

How does defoliation or shoot decapitation increase the rate of net photosynthesis in the
remaining leaves, and how can this phenomenon be utilized to best advantage? What proportion of
leaf area or canopy coverage is best to retain, and how is the optimal amount influenced by harvest
height, timing, or spacing? Is retaining some leaves at the last harvest beneficial or harmful in the
long temn, and how does this relate to harvest date or day length?

If measured at the optimal time, about 30 - 40 days after bud break, canopy coverage
could be an excellent indicator of biomass yields later in the season. This should be investigated
further. It is possible that the proportion of canopy remaining after an early summer harvest could
predict the yield at the second harvest two or three months later, but no data were taken to test this
hypothesis. The correlation of solar interception and yield should be a fruitful area for future
research, as it would give a common reference point for the effects of harvest time, harvest height,

plant spacing, and all their complex and confusing interactions.

Juvenile stage

Can black locust plants be kept in the juvenile stage indefinitely by annual coppicing? If
the harvest height is high enough to retain some branches, will these branches ever flower? Are the
rapid growth and profuse spines of the juvenile stage inseparable? Observations of trees growing

older, and of spineless mutants, indicate that they are. Could coppice growth possibly be rendered

 

 

154

spineless? Is slower gowtln in taller plants related to the passing of the juvenile stage? Can

gowth rates in mature trees be improved by preventing flowering?

RECOMMENDATIONS

Crop establishment

Recommended agronomic practices depend on whether black locust is gown as an annual
or a perennial crop, but the goal in both cases is to maintain a closed canopy for as much of the
gowing season as possible. An annual crop should be sown thickly (125 - 250 seeds/m2) to give
50 - 100 seedlings/m2, and harvested as near to the gound as possible to maximize yield.
Herbicides should be used before cultivating and planting the field the next year, so that the few
survivors do not become established.

A perennial crop requires enough plants to maintain a closed canopy at the harvest height.
With the 30 cm height used in this trial, the widest spacing tested (50 cm) was adequate the tlnird
gowing season and superior the fourth. With more time or higher cutting, even wider spacings
could be used, and transplanting may become economical. But considering the low cost of the seed
compared to the loss of potential yields the first and second seasons before the canopy closes, a

tlnicker sowing would be preferred if weed control permits it.

Experimental logistics
Future field trials with black locust should be designed so the plots can be planted,

sprayed, and harvested mechanically, or they should be conducted in a temperate region outside of
North America where labor costs are lower. The use of square plots with border rows made the
planted areas too wide to drive tractors or self-propelled harvesters over them, forcing us to use
hand tools to simulate mecharnical harvest. This caused much aggravation for the field workers,

who suffered from the heat if they wore thick clothing, and from frequent wounds if they did not.

 

 

 

155

A gasoline fueled hedge trimmer was tried once; it cut succulent stems with case, but had
difficulty getting tlnrough woody stems and the wire flags that marked plot boundaries. Many
stems and leaves were cut repeatedly, so that the small fragnents were diflicult to collect. All who
used the motorized trimmer found its weight, noise, and fumes so unpleasant that it was not tried
again. An electric model would be more suitable, if enough extension cords could be obtained.

If randomized plots are to be cut mechanically at different times, care must be taken to
allow room for the tractor to turn, so as not to damage the uncut plots. Also, uncut plots must be
kept fi'om shading the plots out earlier. Ifa small number of plots are cut early, rabbits and
rodents may destroy the regrowth, so fencing is advisable.

Finally, a major complication in most trials was that plants in border rows next to paths
grew larger and shaded the measured portions between them. This could be avoided by trimming

them frequently, as we did, or by using wider plots with narrow paths for the tractor tires.

.. 5.,
.9155.

 

 

 

nrcrtran smr: UNIV. Lrennnrss
lllllllllllllllllllllllllllllllllllllllll
31293010185803