|
February, 1940
UNIVERSITY OF FLORIDA
AGRICULTURAL EXPERIMENT STATION
GAINESVILLE, FLORIDA
WILMON NEWELL, Director
FLOWERING, FRUITING, YIELD
AND GROWTH HABITS
OF TUNG TREES
By R. D. DICKEY and WALTER REUTHER
Fig. 1.-The difference in manner of fruit production which has led to the tentative
distinction of "cluster" (right) and "single" (left) types. Foliage partially removed to
show fruit. (From Bul. 280.)
Single copies free to Florida residents upon request to
AGRICULTURAL EXPERIMENT STATION
GAINESVILLE, FLORIDA
Bulletin 343
EXECUTIVE STAFF
John J. Tigert, M.A., LL.D., President of
the University3
Wilmon Newell, D.Sc., Directors
Harold Mowry, M.S.A., Asst. Dir., Research
V. V. Bowman, M.S.A., Asst. to the Director
J. Francis Cooper, M.S.A., Elitor3
Jefferson Thomas, Assistant Editors
Clyde Beale, A.B.J., Assistant Editor3
Ida Keeling Cresap, Librarian
Ruby Newhall, Administrative Manager6
K. H. Graham, Business Manager3
Rachel McQuarrie, Accountants
MAIN STATION, GAINESVILLE
AGRONOMY
W. E. Stokes, M.S., Agronomist'
W. A. Leukel, Ph.D., Agronomists
G. E. Ritchey, M.S., Associate2
Fred H. Hull, Ph.D., Associate
W. A. Carver, Ph.D., Associate
John P. Camp, M.S., Assistant
Roy E. Blaser, M.S., Assistant
ANIMAL HUSBANDRY
A. L. Shealy, D.V.M., Animal Husbandman' 3
R. B. Becker, Ph.D., Dairy Husbandman3
L. M. Thurston, Ph.D., Dairy Technologist3
W. M. Neal, Ph.D., Asso. in An. Nutrition
D. A. Sanders, D.V.M., Veterinarian
M. W. Emmel, D.V.M., Veterinarian3
N. R. Mehrhof, M.Agr., Poultry Husbandmans
W. G. Kirk, Ph.D., Asso. An. Husbandman3
R. M. Crown, B.S.A., Asst. in An. Husb.3
P.. T. Dix Arnold, M.S.A., Assistant Dairy
Husbandman3
L L Rusoff, M.S., Asst. in An. Nutrition3
CHEMISTRY AND SOILS
R. V. Allison, Ph.D., Chemist' 3
Gaylord M. Volk, M.S., Chemist
F. B. Smith, Ph.D., Microbiologists
C. E. Bell, Ph.D., Associate Chemist
H. W. Winsor, B.S.A., Assistant Chemist
J. Russell Henderson, M.S.A., Associate3
L. H. Rogers, M.S., Asso. Biochemist
Richard A. Carrigan, B.S., Asst. Chemist
ECONOMICS, AGRICULTURAL
C. V. Noble, Ph.D., Agricultural Economist' 3
Bruce McKinley, A.B., B.S.A., Associate
Zach Savage, M.S.A., Associate
A. H. Spurlock, M.S.A., Assistant
ECONOMICS, HOME
Ouida Davis Abbott, Ph.D., Specialist1
Ruth Overstreet, R.N., Assistant
R. B. French, Ph.D., Associate Chemist
ENTOMOLOGY
J. R. Watson, A.M., Entomologist'
A. N. Tissot, Ph.D., Associate
H. E. Bratley, M.S.A., Assistant
HORTICULTURE
G. H. Blackmon, M.S.A., Horticulturist'
A. L. Stahl, Ph.D., Associate
F. S. Jamison, Ph.D., Truck Horticulturist3
R. J. Wilmot, M.S.A., Specialist, Fumigation
Research
R. D. Dickey, M.S.A., Assistant Horticulturist
J. Carlton Cain, B.S.A., Asst. Horticulturist
Victor F. Nettles, M.S.A., Asst. Hort.
PLANT PATHOLOGY
W. B. Tisdale, Ph.D., Plant Pathologist'
George F. Weber, Ph.D., Plant Pathologists
L. O. Gratz, Ph.D., Plant Pathologist
Erdman West, M.S., Mycologist
Lillian E. Arnold, M.S., Assistant Botanist
BOARD OF CONTROL
H. P. Adair, Chairman, Jacksonville
W. M. Palmer, Ocala
Chas. P. Helfenstein, Live Oak
R. H. Gore, Fort Lauderdale
N. B. Jordan, Quincy
J. T. Diamond, Secretary, Tallahassee
BRANCH STATIONS
NORTH FLORIDA STATION, QUINCY
J. D. Warner. M.S.. Agronomist Acting in
Charge
R. R. Kincaid, Ph.D., Asso. Plant Pathologist
Elliott Whitehurst, B.S.A., Asst. An. Husb.
Jesse Reeves, Farm Superintendent
CITRUS STATION, LAKE ALFRED
A. F. Camp, Ph.D., Horticulturist in Charge
John H. Jefferies, Superintendent
Michael Peech, Ph.D., Soils Chemist
L. H. Greathouse, Ph.D., Chemist
B. R. Fudge, Ph.D., Associate Chemist
W. L. Thompson, B.S., Asso. Entomologist
W. W. Lawless, B.S., Asst. Horticulturist
R. K. Voorhees, M.S., Asst. Plant Path.
EVERGLADES STATION, BELLE GLADE
J. R. Neller, Ph.D., Biochemist in Charge
J. W. Wilson, Sc.D., Entomologist
F. D. Stevens, B.S., Sugarcane Agronomist
Thomas Bregger, Ph.D., Sugarcane
Physiologist
Frederick Boyd, Ph.D., Asst. Agronomist
G. R. Townsend, Ph.D., Plant Pathologist
R. W. Kidder, B.S., Asst. An. Husbandman
W. T. Forsee, Ph.D., Asso. Chemist
B. S. Clayton, B.S.C.E., Drainage Engineer'
SUB-TROPICAL STATION, HOMESTEAD
W. M. Fifield, M.S., Horticulturist Acting in
Charge
S. J. Lynch, B.S.A., Asst. Horticulturist
Geo. D. Ruehle, Ph.D., Asso. Plant Pathologist
W. CENTRAL FLA. STA., BROOKSVILLE
W. F. Ward, M.S., Asst. An. Husbandman
in Charge-
FIELD STATIONS
Leesburg
M. N. Walker, Ph.D., Plant Pathologist in
Charge
K. W. Loucks, M.S., Asst. Plant Pathologist
Plant City
A. N. Brooks, Ph.D., Plant Pathologist
Cocoa
A. S. Rhoads, Ph.D., Plant Pathologist
Hastings
A. H. Eddins, Ph.D., Plant Pathologist
Monticello
Samuel O. Hill, B.S., Asst. Entomologist2
Bradenton
Jos. R. Beckenbach, Ph.D., Truck Horticul-
turist in Charge
David G. Kelbert, Asst. Plant Pathologist
Sanford
R. W. Ruprecht, Ph.D., Chemist in Charge,
Celery Investigations
W. B. Shippy, Ph.D., Asso. Plant Pathologist
Lakeland
E. S. Ellison, Meteorologist2
B. H. Moore, A.B., Asst. Meteorologists
'Head of Department.
'In cooperation with U.S.D.A.
3Cooperative, other divisions, U. of F.
FLOWERING, FRUITING, YIELD AND GROWTH
HABITS OF TUNG TREES
By R. D. DICKEY and WALTER REUTHER1
CONTENTS
PAGE
IN TRODU CTION ....................... ... .............................................................. .................. -3
Y IELD ................................................. ........................................................................... 4
Original 10 Trees on the Experiment Station Grounds ........................ 4
N ew Tung Block ....................................... ....... ..................................... 7
FACTORS CORRELATED W ITH YIELD ............ .............................................................. 13
Fruiting Habit ..................................................................................... ....... 14
T ree Size .............................. ..................... ................... ........................ 14
Tree Conform ation ................. .................................... .......... ........... .... 14
F low ering H abit ............................ .................................... ....... .. ............... 15
HEIGHT OF BRANCHING ........... .................................................................. ... .. 18
O BSERVATIONS ......................... .......... ............................. ......... ...... ........ ..- ... ..... 24
DISCUSSION AND SUM MARY .................................................. ..................... ..... .. 26
L ITERATURE CITED ....................... .... .................... ..... ...... ........ .......... 28
INTRODUCTION
Knowledge relative to the continued progress in tung grow-
ing and of the large and constant demand for the oil by several
American industries has maintained and heightened the already
growing interest manifested in the commercial possibilities of
this crop. This interest has been further increased because of
the threat to the source of supply of this essential raw product
by the present disturbed political situation in China.
Due to the inconclusive results thus far obtained with asexual
methods of propagation under field conditions, it is safe to say
that, for some time to come, most of the commercial plantings
made in Florida will be from seed. This point is of considerable
importance because the work of Mowry (4)2 has shown that
wide variations exist in seedling trees and that these variations
have an important effect upon yield and oil production.
Fruiting habit as distinguished by the "single" and "cluster"
type (Fig. 1) was shown to have a decided effect upon yield
(6 and 7). Other variations which influence oil production are
size of fruit and seed, relative weights of seed and hull, ratios
of weight of kernel to testa and percentage of oil in the seed.
However, the most important individual factor that influences
oil production is total yield.
As Mowry (5) has pointed out, the success or failure of the
tung industry obviously depends on the ability of the grower
to produce at a fair margin of profit. Yield is the most im-
1Reuther, formerly Assistant Horticulturist at the Citrus Experiment
Station.
'Italic figures in parentheses refer to "Literature Cited" in the back
of this bulletin.
Florida Agricultural Experiment Station
portant single problem confronting the grower; this emphasizes
the need for a vigorous high-yielding variety of trees.
Since the productivity of the individual tree is of such great
importance, information relative to factors affecting it is of
paramount interest to the prospective planter of a tung grove.
Records covering a period of years have demonstrated that wide
variations exist among the yields of miscellaneous (unselected)
seedling trees growing under comparable conditions and that
these variations are fairly constant from year to year.
Whether or not this very important variation in yield is passed
on to the progeny is of major concern because of the influence
it may have on the future production of the tung grove, for
seed taken from unselected trees will undoubtedly include many
trees which are low yielding. The data presented here give
considerable information relative to this point as well as to tree
size, conformation, flowering and fruiting habit and height of
branching.
YIELD
ORIGINAL 10 TREES ON THE EXPERIMENT STATION GROUNDS
The original 10 trees on the Florida Experiment Station
grounds at Gainesville were planted in 1912 and 1914 with seed-
lings coming from the U.S.D.A.
TABLE 1.-AVERAGE ANNUAL Plant Introduction Gardens at
YIELD, AIR-DRIED HULLED SEED, Chico, California. These trees are
OF ORIGINAL 10 TREES ON HORTI-
CULTURAL GROUNDS, 1922-1938. spaced 10 feet apart in the row,
which has resulted in consider-
17-Year Average able crowding. Had they been
Tree I Seed Oil given sufficient space, it is thought
(Pounds) (Pounds) that the yields given in Table 1
1 18.6 6.4 would be greater, though the rela-
2 38.5 13.3 tive differences would remain
3 4.8 1.6
4 15.3 5.3 approximately the same. During
5 4.2 1.4 the early days of the tung in-
6 32.7 11.3
7 10.7 3.7 dustry seed from these trees was
8 24.0 8.3 widely distributed so that, di-
9 64.9 22.4
10 6.7 2.3 rectly or indirectly, they have
formed the basis for the planting
Avg. 22.0 7.6 stock of a considerable acreage.
Evaluation of the characteristics
of these trees should prove of worth, since similar variations
are to be expected in seedling trees from other sources.
Yield and Growth Habits of Tung Trees
An examination of Table 1 will show that the trees vary
greatly in that all-important characteristic of "ability to yield".
Over a period of 17 years the average annual production per tree
varied from 4.2 pounds to 64.9 pounds of hulled seed. Though
some of the trees have yielded well, others have yielded so low
that they must be classed as almost non-producers.
In Table 1 it will be noted that Tree 9 has outborne all others;
also it has grown to be the largest tree in the group. Adjacent
citrus trees, until they were removed in 1936, crowded Trees 1
to 7, but Trees 8, 9 and 10 have not had this competition. This
may account, in part, for the differences in yield between Trees
2 and 6 and Tree 9.
An examination of the data discloses the fact that the yield-
ing ability of the different trees has been a quite constant
factor. Data for the year 1922 are omitted because the seeds
of all of the trees were not accurately separated for that year,
and those for 1926, 1930, 1932 and 1935 are omitted because
of partial or total crop reduction due to frosts or freezing
temperatures. The 10 trees are divided into three groups for
each year by arranging the trees numerically from low to high
on the basis of yield. The three lowest yielding trees compose
Group 1, the three highest yielding trees make up Group 3,
with the four trees of medium yields comprising Group 2.
Results for the several years are given in Table 2. It is evident
that, for each year, the group in which any tree may fall is
determined entirely by its yield. If it were merely a matter
of chance as to its yield, and therefore the group into which it
fell each year, it would be expected that over a period of 12
TABLE 2.-YIELD GROUP DISTRIBUTION FOR ORIGINAL 10 TREES FOR THE
12 YEARS UNDER CONSIDERATION.1
Tree 1923 1924 1925 1927 1928 1929 1931 1933 1934 1936 1937 1938
1 1 1 22 2 2 2 2 2 3 2 2 3
2 3 3 3 3 3 3 3 3 3 3 3 3
3 1 2 1 1 1 1 1| 1 1 1 2 1
4 1 1 1 2 2 2 2 2 2 2 2 2
5 2 1 2 1 1 1 1 1 1 1 1 1
6 3 3 3 i2 3 3 3 2 3 2 2
7 2 2 2 2 1 2 1 2 1 2 2 1 2
8 3 2 2 2 8 2 2 i 2 2 2 1 3 2
9 2 3 3 3 | 3 3 3 3 3 3
10 2 2 1 1 2 1 2 1 1 2 1
'Yield groups-1, low group; 2, medium group; 3, high group.
Florida Agricultural Experiment. Station
LEGEND:
YIELD GROUPS
HIGH -0 MEDIUM- .
GooD -+ Low -0
17 34
17 0
16 33
16 0 l
15 32
15 -+-
14 31
14 a 0
13 30
13 0 D
12 29
12 O D
It 28
11 O O
10 27
10 0 0
9 26
.9 0 0
8 25
8 0 0
7 24
7 0 +
6 23
6 0 0
5 22
5 0 0
4 21
4 0 0
3 20
3 0 a
2 19
2" + 1n
Tree
NumbP A 18
I DI
SA
3Rows A B
10f2 19 136
00 +
Fig. 2.-Map showing the
101 118 135 yield groups and the fertil-
S + e izer treatments given the
trees in New Tung Block.
Rows numbered 1 to 17 and
100 117 134 lettered from A to H, trees
0 Q numbered 1 to 136.
The fertilizers used were
99 16 133 as follows:
0 + ]
Plot 1, Rows 1, 2, 3-
Steamel bonemeal and cot-
98 115 132 tonseed meal, equal amounts.
il + +
Plot 2, Rows 4 and 5-
97 1f 131 5-8-4 mixture (same as Plot
+ 5) with addition of four
pounds of air-slaked lime
four times during 16 years.
96 113 130 No fertilizer applied until
* + D 1930. Lime applied 1923,
1926, 1930 and 1934.
95 112 f29
t O Plot 3, Rows 6 and 7-
5-8-4 mixture (same as Plot
5) with one application of
94 iti 128 lime as above at planting.
n +
Plot 4, Row 8-Check, no
93 110 127 fertilizer.
Plot 5, 'Rows 9 and 10-
92 109 126 5-8-4 mixture of 2% nitrate
I of soda; 1% sulfate of am-
monia; 2% cottonseed meal;
8% superphosphate, and 4%
91 108 125 muriate of potash.
+ [ M
Plot 6, Rows 11 and 12-
90 107 124 Stable manure.
Plot 7, Rows 13 and 14-
S23 5-8 4 mixture of 3% nitrate
89 106 123 of soda ; 2% cottonseed meal;
D 0 0 8% superphosphate, and 4%
muriate of potash.
88 105 122
* 0 5 Plot 8, Rows 15, 16 and 17
-Steamed bonemeal.
87 104 121
+ *
86 103 120
+ + *
) E F G H
Yield and Growth Habits of Tung Trees
years each tree under consideration would appear in all three
groups at one time or another during this period.
This is not the case, as Tree 2 appears in the group showing
the highest production (Group 3) in 12 out of 12 years, ,Tree 9
in 11 out of 12 years, and Tree 6 in eight out of 12 years.
Conversely, taking into account the three trees producing the
smallest crop in each year (Group 1), Trees 3 and 5 will be
found in this group 10 years and Tree 10 eight out of the 12
years. It will be noted that none of the trees mentioned above
appear in both groups at any time during the 12 years under
consideration. Furthermore, Trees 4, 7 and 8 are just as con-
sistent in their yielding performance as the trees mentioned
above, as they fall in the medium (Group 2) 9 out of the 12
years under consideration and at no time are they to be found
throughout the range of possibilities. There is only one ex-
ception to this, that being Tree 1, which is to be found in all
three groups. While its general performance is Group 2, it is
found in Group 1 two years and in Group 3 two years.
The mathematical chance that any one tree, such as No. 2,
would fall in Group 3 12 times in succession is one to more
than 500,000. The probability that Tree 5 would fall in Group 1
by chance nine successive times is 1 in 19,683. With nine of
the 10 trees under consideration showing a striking consistency
in their "yield performance" the indication is clear that it is
not a chance happening. In explanation of this fact it is be-
lieved that the trees have an inherited fruiting capacity. Trees
2, 6 and 9 may be spoken of as inherently high-yielding trees,
and Trees 3, 5 and 10 as inherently low-yielding trees.
A study of a planting (the New Tung Block) coming from
a mixture of seed from these trees furnishes some additional
information of interest in this regard and confirms the trends
suggested by the data of Tables 1 and 2.
NEW TUNG BLOCK
Trees used to plant this block came from a mixture of seed
from the original 10 trees and possibly from some other trees
growing upon the Experiment Station grounds.
This block was planted in 1923 and consisted of 136 trees in
eight rows of 17 trees each (Fig. 2). It is located on a coarse,
deep sand (Norfolk) that had been growing field crops for sev-
eral years prior to planting to tung trees and should not be
considered as capable of producing maximum yields. Weights
Florida:.Agricultural Experiment Station
of air-dried hulled seed were taken over the duration of the
experiment. Because of partial or total crop reduction due to
frosts or freezing temperatures, the years 1930, 1932, 1935 and
1938 are eliminated from consideration. Only those years are
included-1927, 1928, 1929, 1931, 1933, 1934, 1936 and 1937-
which represent more nearly the true ability of the trees to
bear under the existing environmental conditions when un-
affected by frost hazards.
When the data for the eight-year average of yield for in-
dividual trees are subjected to statistical treatment by the
analysis of variance method (9), the values for mean squares
obtained are presented in
TABLE 3.-ANALYSIS OF VARIANCE. Table 3. Since the fertil-
Source of Degrees of Mean square izer plots are not ade-
variance. freedom of variance quately replicated or ran-
7 domized, it is not possible
Rows ............. 7 193.331
Treatments-. 16 97.086 to evaluate the variance
Remainder, j.| ., 112 225.020 contributed by the treat-
-ments alone-no doubt soil
differences contribute to
the value calculated to some extent. The variance due to rows
(positional factors such as soil differences and competition from
other trees) may be in part influenced by the fact that most
of Row.A (see Fig., 2) is adjacent to a horse lot.
The calculated mean squares show that when variance in yield
due to treatment or rows is removed, there still remains a major
source of yield variance which is believed to be caused by the
inherent genetic constitution of the tree. Moreover, this dif-
ference in yield capacity is probably not due to a difference in
tree size, because it is evident from Figure 3 that there is but
little relation between the cross-sectional area of trunk and the
average yield of trees. Therefore, it seems probable that the
yield capacity of a given seedling tree is related to such factors
as the type and amount of inflorescence produced. It is not
believed that any of the low-bearing trees fail to set fruit be-
cause of lack of adequate pollination, since very few functional
female flowers have ever been observed to absciss or fail to
develop into fruit.
The yield data of the 136 trees in this block were divided
arbitrarily into four groups for each year under consideration
-1927, 1928, 1929, 1931, 1933, 1934, 1936 and 1937. This was
done by arranging them numerically from low to high on the
Yield and Growth Habits of Tung Trees
* *
*..
C .
C *
* C
t C f
p
5.
0 0* 0 0
P..q V .. ^- "P -
iyaJl ,Id paaC' P'rl"H p'i-'P --** 'H 4 PUnOr "' pl^a 's^'o^^^^b
a 0)
a-
U) w
E
o r
a-I-
r,
o
t '
0
0
L EI
a o
0 S
*. S
'J l
0
? *S
a:
CCO
? -^
10 Florida Agricultural Experiment Station
basis of yield. The 34 lowest yielding trees compose Group 1
(low group), the 34 trees next in order compose Group 2 (med-
ium group), Group 3 (good group) consists of the 34 trees next
in order above Group 2, and Group 4 is composed of the 34
trees producing the greatest yields (high group). It is evident
that, for each year under consideration, the group in which
any tree may fall is determined entirely by its yield.
Much the same trend is followed in regard to consistency of
"yield performance" as was shown for the original 10 trees
(Table 2). Data for 20 representative trees are presented in
Table 4, which show the consistency with which they fall in a
certain "yield group". The odds have been calculated to show
the mathematical chance of their having fallen by accident
successively in the same group.
TABLE 4.-YIELD GROUP DISTRIBUTION FOR 20 TREES OF NEW TUNG BLOCK
FOR THE EIGHT YEARS UNDER CONSIDERATION.1
Number Tree Year I Calculated
I No.2 1927 19281 19291 19311 19331 19341 19361 1937 odds
High Group
1 12 4 4 4 4 4 4 4 4 1 to 65,536
2 37 4 4 4 3 4 4 4 4 1 to 310
3 22 3 4 4 4 4 4 3 4 1 to 1,024
4 17 2 4 4 4 4 4 4 4 1 to 16,384
5 27 4 4 4 4 4 3 4 1 1to 256
Good Group
6 52 3 4 3 3 3 3 3 4 1 to 1,024
7 87 3 3 3 2 3 3 3 2 1 to 128
8 49 3 3 2 3 4 3 3 3 1 to 80
9 36 3 3 3 3 3 2 3 4 1to 1,024
10 95 3 2 3 3 3 3 2 2 1 to 256
Medium Grou
11 94 2 1 2 2 2 2 2 2 1 to 4,096
12 72 2 3 2 2 2 2 2 3 1 to 1,024
13 75 3 2 2 2 2 3 2 2 1 to 272
14 66 2 2 2 1 1 2 2 2 1 to 128
15 105 2 2 2 1 3 1 2 2 1 to 80
Low Group
16 7 1 1 1 1 1 1 1 1 1 to 65,536
17 32 1 1 1 1 1 1 1 2 1 to 16,384
18 65 1 2 2 1 1 1 1 1 1to 1,024
19 121 1 1 1 1 2 1 1 2 1 to 272
20 123 1 1 1 1 2 1 1 3 1 to 272
WYield groups-4, high group; 3, good group; 2, medium group; 1, low group.
2Refers to numbers given trees in Fig. 2.
The odds that any one tree would by chance fall in the same
group eight times in succession are 1 to 65,536 (Table 4). Since
'Yield and Growth Habits of Tung Trees
this happens with nine of the 136 trees in this block, it is cer-
tainly indicated that there is some reason other than just a
chance happening. If the yields of the trees in this block are
entirely a matter of chance, it would be expected that, during
the eight years under consideration, each tree would appear in
each of the groups at least once, yet only 22 of the 136 trees
do this. It is of further significance that, in many cases, when,
for any year, a tree drops out of the group average, it is into
the group immediately above or below the group average, e. g.,
Numbers 2, 6 and 15 (Table 4).
The seed which produced the trees used to plant this block
came from about the same genetical stock as many of the older
plantings made in Florida. Therefore, they should serve as an
index to what would be expected to be found in other groves
in which the seed came from the same source, whether it be
five acres or 1,000 acres.
40
34
32
30
26
S24
r22
I20
i (8 CLASSES
S2. 0.5-- .5 LBS.
2.. 56- .05
3. 0.6-15.5 -
14 4. 15.6-20.5
5 20.6-25.5 -
S6 25 -30.5
7 30.6-35.5
10 1 35.6-40.5 -
2
0 1 2 3 4 5 6 7
CLA55ES
Fig. 4.-Frequency distribution curve of yield of air-dried hulled seed of trees in
New Tung Block. Average annual yield for 8 years under consideration.
In this connection it is of interest to note that the data pre-
sented in Figure 4 and Table 5 show that practically 90 percent
of the trees fall in the lowest four classes as regards yield-
that is, have an average annual yield for the eight years under
consideration of less than 20.6 pounds of air-dried hulled seed.
Only 10 percent fall in the highest four classes, that is, from
20.6 to 40.5 pounds. An. examination of the low-yielding trees
Florida Agricultural Experiment Station
TABLE $.--THE FRUITING HABIT AND THE AVERAGE ANNUAL YIELD FOR
THE 8 YEARS CONSIDERED OF THE TREES IN THE NEW TUNG BLOCK
(WEIGHT IN POUNDS OF HULLED, AIR-DRIED SEED).
Num-I Tree Fruiting Num-
ber Yield No.1 habit2 b er Yield
Tree
No.
Fruiting
habit
1 40.5 12 C 51 14.2 132 C
2 35.2 17 C 52 14.0 113 C
3 33.0: 27 C 53 13.7 49 C
t4 33.0 68 C 54 13.5 80 C
5 31.4 119 C 55 13.0 70 C
6' 25.9 9 C 56 12.9 79 C
7 25.2 85 S 57 12.6 97 S
8 24.9 31 S 58 12.4 87 S
9' 24.4 6 C 59 12.4 118 C
10 22.3 3 C 60 12.3 53 S
11 22.3 10 C 61 12.3 71 C
12 22.3 44" C 62 12.2 42' C
13 22.2 58 C 63 11.9 136 S
14 21.1 102 S 64 11.8 56 C
15 19.9 5 'C 65. 11.8 86 C
16 19.7 64 C 66 11.7 81 C
17 19.7 37 C 67 11.7 15 S
18 19.7 39 C 68 11.6 116 C
19 19.6 22 C 69 11.3 130 S
20 19.3 48 S 70 11.3 117 C
21 19.2 13' S 71 11.3 60 S
22 19.0 8 S 72 11.2 89 C
23 19.0 28 C 73 11.2 73 C
24 18.9 43 C 74 11.2 83 C
25 18.8 35 C 75 10.9 108 S
26 18.7 100 C 76 10.7 55 S
27 18.6 99. C 77 10.6 20 S
28 18.3 11 S 78 10.5 67 C
29 18.3 16 C 79 10.5 90 C
30 18.3 78 S 80 10.4 95 S
31 18.1 112 C 81 10.2 98 S
32 17.9 101 S 82 10.1 77 S
33 17.7 51 C 83 10.0 33 S
34 16.6 4 S 84 10.0 54 C
35 16.1 111 C 85 9.9 129 C
36 16.1 114 C 86 9.8 29 C
37 16.1 115 C 87 9.7 1 C
38 15.8 91 C 88 9.4 18 S
39 15.7 57 S 89 9.3 72 S
40 15.6 59 C 90 9.1 14 S
4 41 15.4 24 S 91 8.9 109 C
42 15.2 473 C : 92 8.8 76 C
43 15.2 36 S 93 8.8 94 S
44 -15.1 2 C 94 8.7 19 S
45 15.1 74 S 95 8.6 75 S
46 14.8 52 S 96 8.5 82 S
47 14.8 40 S 97 8.5 128 C
48 14.7 61 C 98 8.4 125 S
49 14.4 103 C 99 8.4 62 C
50 14.3 41 C 100 8.2 131 C
'Refer3 to numbers given trees in Fig. 2.
2Fruiting habit: C-Cluster type. S-Single type.
aYield records lost for Trees 30 and 47 in 1933, and Trees 13 and 99 in 1934. Tree 92
killed in 1934, and Trees 42 and 44 killed in 1937 by the Clitocybe mushroom root rot fungus.
Yield and Growth Habits of Tung Trees.
TABLE 5.-THE FRUITING HABIT AND THE AVERAGE ANNUAL YIELD FOR
THE 8 YEARS CONSIDERED OF THE TREES IN THE NEW TUNG BLOCK
(WEIGHT IN POUNDS OF HULLED, AIR-DRIED SEED)'-Concluded.
Num-I | Tree Fruiting ] Num-' Tree I Fruiting
ber Yield I No. habit her Yield No. I habit
101 8.0 30" S 119 5.3 126 S
102 7.8 133 S 120 5.1 26 S "
103 7.7 21 C 121 4.9 110 C
104 7.6 921 S 122 4.8 123 C
105 7.3 96 S 123 4.8 127 S
106 7.3 105 C 124 4.5 65 S
107 7.2 122 C 125 3.9 121 S
108 7.1 84 S 126 3.8 7 .S
109 7.0 66 S 127 3.6 104' S
110 6.7 134 C 128 3.2 .69 S,
111 6.5 63 C 129 3.2 106 C
112 6.3 45 S 130 3.1 93 S
113 6.2 135 S 131 3.0 120 S
114 5.9 38 S 132 2.8 50. S
115 5.8 88 S 133 -1.9 '46' C
116 5.7 107 C 134 1.8 32 S.
117 5.6 124 C 135 0.6 25 S
118 5.4 34 S 136 0.6 23 S
1Refers to numbers given trees in Fig. 2.
2Fruiting habit: C-Cluster type. S-Single type.
"Yield records lost for Trees 30 and 47 in 1933, and Trees 13 and'99 in 1934.. Tree 92
killed in 1934, and Trees 42 and 44 killed in 1937 by the Clitocybe mushroom root rot fungus.
reveals that, though they are in a vigorous growing condition,
many are practically devoid of fruit. From the discussion given
in connection with Tables 2 and 4, it has been shown that yield-
ing ability of a given tree is quite consistent from year to year.
It is recognized that the cultural care given a tung tree is re-
flected in its yield, to a considerable extent, but the, range of
variation in yield due to this cause is undoubtedly much lower
in the low-bearing trees. It is obvious that any practice which
will increase the percentage of high-bearing trees in a planting
is highly desirable.
The data presented here show very strongly that the inherit-
ance of a tree determines to a considerable extent its ability to
yield and emphasizes the great importance in using care in
selecting seed for planting purposes.
FACTORS CORRELATED WITH YIELD
In many tung plantings, particularly those from unselected
seed, there has been found a high percentage of low-producing
trees. It would seem desirable to identify and remove these
"boarder trees" from the planting, replacing them with prolific
trees. Also, in the selection of trees for future propagation, a
Florida Agricultural Experiment Station
certain characteristic of growth may be desirable, while another
may be undesirable. Therefore, it would be advantageous to
be able to distinguish the one from the other and to know what
effect, if any, they may have upon the yield of the tree. With
this in mind, the trees in the New Tung Block have been care-
fully studied in an attempt to determine whether or not yields
may be correlated with some characteristics of growth or con-
formation.
FRUITING HABIT
It has been, shown that fruiting habit as distinguished by the
"single" and "cluster" type (Fig. 1) has a decided effect upon
yield. However, not all trees which would be classified as cluster
type are desirable, nor can it be said that all single type trees
are undesirable. In general, though, it is the best character
found, so far, upon which to base the selection of a superior
strain. The data given by Mowry (4), Newell, et. al. (6 and 7)
and that contained in Table 5 definitely show this.
Of the 34 highest yielding trees in the block, 24 are of the
cluster type while 10 are of the single type. Twenty-three of
the 34 lowest yielding trees in the block are of the single type
and 11 are of the cluster type (Table 5).
TREE SIZE
Can size be used as an index to the ability of a tung tree to
yield? To determine this, the low and high-yielding trees were
carefully compared. It was found that there is no correlation
between size as evidenced by cross-sectional area of the trunk
(Fig. 3) and the ability of a tung tree to yield. Similar scatter
diagrams have been made for height and spread which show
approximately the same lack of correlation. It is evident that
size alone cannot be used as an index for ascertaining the high
or low-producing trees in a planting or in the selection of de-
sirable strains for future propagation.
TREE CONFORMATION
All trees in this block were carefully studied to determine if
yield could be correlated with any particular tree conformation.
Some of the factors considered were type of branching and shape
of head (whether round or pyramidal, compact or loose, upright
or drooping). No particular conformation was found to be corre-
lated with yields.
Yield and Growth Habits of Tung Trees
FLOWERING
HABIT
Tree No. 10 of
the original 10
trees undoubtedly
has sufficient ter-
minals, were it a
prolific tree, to
produce a heavy
yield. However,
an examination
of the inflores-
cences of this tree
at time of flower-
ing reveals that,
though blossom
buds have been
differentiated on
the apices of the
central parts of
the cymes, which
is the position of
the pistillate (fe-
male) flower if
present, a consid-
erable percent are
n o n functional.
This condition is
manifested in
several ways; for
example, though
the bud occupy-
ing the central or
apical part of the
inflorescence i n
the early stages
of its develop-
ment appears to
be normal, it is
soon evident that
no further growth
of the bud is be-
Fig. 5.-Inflorescence from Tree 10 of the original 10 trees
in which an apparently normal bud occupies the apex of the
central stalk. This bud will soon dry up and absciss.
Fig. 6.-Inflorescence from Tree 10 of the original 10 trees
showing well developed central stalk but with apical bud dead.
Florida Agricultural Experiment Station
ing made (Fig. 5). In a comparatively short time this bud
and its accompanying joint of the inflorescence dries up and
abscisses, without opening. In another case the central stalk
is well developed, but the apical bud soon dies, shrivels and
abscisses (Fig. 6). A third variation is where the entire cen-
tral stalk of the
inflorescence and
its accompanying
flower buds make
no growth, soon
die, shrivel and
absciss (Fig. 7).
Though it is
not known that
Sthe terminal buds
That function in
this way are
actually differen-
tiated into pistil-
late- buds (that
is, whether they
are staminate or
pistillate), it is
known that a
Fig.-7.-Inflorescence from Tree 10 of the original 10 trees great majority of
in which the central stalk and its accompanying flower buds
have made no growth, and is shriveled and dead. the buds appear-
ing on the apices
of the central parts of the inflorescences of this tree are non-
functional.- Using the last five years that the tree has fruited
as ~a basis of computation, it is estimated that this tree has
had fr6m 1,000 to 3,000 flowering terminals each year, yet- dur-
ing.this time an average of approximately 60 terminals have
pr6diiced fruit for each year. A similar condition has been
observed on several low-yielding trees in the New Tung Block.
It has been reported (2 and 3) that certain other species of
Aleurites, though essentially monoecious, show a decided ten-
dency toward dioeciousness, which would indicate that this
genus tends to be heterozygous as far as this character is con-
cerned. Thus it seems possible that the low yield of a tree such
as No. 10 is due to the inheritance of the tendency toward dioeci-
ousness, making it in reality practically a male tree.
Yield and Growth Habits of Tung Trees
Another condition has been observed on several trees in this
block which, it is believed, is a somewhat different morphological
manifestation of the tendency toward dioeciousness. In this
type the central stalk of the inflorescence is well developed but,
in the position where the pistillate flower normally should be,
a staminate (male) flower has been developed instead (Fig. 8).
On certain trees the number of inflorescences in which this has
taken place is high, which would account for the comparatively
low yield of trees of this type.- It has been observed that this
is a very consistent character which is further substantiated by
the yield record of these trees.
Fig. 8.-Inflorescence from Tree 23 of the New Tung Block showing well developed
central stalk terminated by xa staminate flower which has eight petals. Non-terminal
staminate flowers of this inflorescence have 5 petals.
Abbott (1) reports that "the number of petals varies from
five to nine, but eight is by far the most common number in the
female flower; while the number of petals varies from five to
eight, but five is the most common number in the male flower".
In this regard it might,.be mentioned that counts show that
Florida Agricultural Experiment Station
staminate flowers occupying the position on the central stalk
of the inflorescence normally occupied by pistillate flowers have
over 90 percent of the flowers with six or more petals. Other
counts show that over 99 percent of the staminate flowers which
are in a non-terminal position have five petals to a flower.
It was observed that there is a large and consistent difference
in the volume of bloom produced by many trees in this block.
So profuse is the flowering of some trees that they have the
appearance of being a solid mass of bloom, while other trees
produce relatively little bloom. Every gradation between these
two types is to be found. A careful check showed that there
is no consistent correlation between volume of flower produc-
tion and yield that could be used as an index for the selection
of high-yielding trees. The number of normal pistillate flowers
produced is the critical factor affecting yield. It seems prob-
able in view of the data herein considered that this is a factor
inherent in the genetic constitution of the tree rather than the
result of condition of vigor of the tree as has been suggested by
Abbott (1). It is not intended to imply, however, that the dif-
ferentiation of the amount of flower primordia is not influenced
by the condition of tree vigor. As yet no direct evidence is at
hand indicating that the condition of vigor of tung trees can
markedly influence the relative proportions of staminate and
pistillate flowers differentiated.
HEIGHT OF BRANCHING
Observations made and records kept by Rolf Buckley of the
Alachua Tung Oil Company show that a large percent of the
trees that branch low
TABLE 6.-TYPE OF BRANCHING IN THE in the nursery, "low
FIELD OF HIGH AND Low HEADED SEEDLINGS. headed" trees, when
Planted in the field cut
Height of Number Branching in field back to from 6 to 8
branching trees No. low No. high
in nursery planted headed headed inches above the ground
I and subsequently
High headed 418 1 417 trained to a single
Low headed 714 713 1
Low headed 182 181 1 trunk, will again branch
in this manner (Table
6 and Fig. 9), while
"high headed" trees treated in the same manner will again
branch as high headed trees (Table 6 and Fig. 10). One block
of 418 trees was planted with high headed seedlings which had
Nit
^*
' W1
Fig. 9.-Low branching tung tree in the field
which was selected in the nursery as a "low headed"
seedling. Stick is 7 feet high. Photographed in the
fourth year after planting. (Courtesy Alachua Tung
Oil Co.)
Fig. 10.-High branching tung tree in the field which was
selected in the nursery as a "high headed" tree. Seedling of
Tree Ax in Table 7. Photographed in the second year after
planting.
1
Florida Agricultural Experiment Station
been selected from the nursery on this basis. These were the
progeny of a single tree (Tree Ax, Table 7). Two other blocks
of 714 and 182 trees were planted with seedlings which were
selected from the nursery as low headed trees.
Seed from a single parent tree produced both high and low
branching trees in the nursery (Figs. 11 and 12 and Table 7).
However, the ratio of
each type produced will
vary depending upon the
parent tree, e.g., one tree
which is high headed (it
is 7 feet 3 inches to the
first branch) has approx-
imately 98 percent of its
seedlings branching as
high headed trees in the
nursery (Tree Ax, Table
7), while another tree
which is low headed has
approximately 95 percent
of its seedlings branch-
ing as low headed trees
(Tree J2-18, Table 7).
The data presented
i strongly indicate that
the height of branching
and its subsequent effect
upon the height of head
is a definite genetical
character.
It has been observed
that, in the nursery, a
high headed tree is in-
dicated in two ways: (1)
those trees which branch
I high and (2) those trees
which fail to branch
(Figs. 11 and 13). In
taking the data pre-
sented in Table 7, all
Fig. 11.-"High headed" seedling of Tree Ax in trees that were less than
Table 7. Seed planted February 1939, photograph 22 feet in height and
taken October 25, 1939. 2/ feet in height nd
Yield'and Growth Habits of Tung Trees
unbranched were not counted, while all unbranched trees above
this height were counted as high headed trees. This arbitrary
standard of classification was adopted since experience has shown
that the number of low headed trees which reach this height and
have not given some evidence of branching is very small.
TABLE 7.-TYPE OF BRANCHING IN THE NURSERY OF SEEDLINGS OF
20 PARENT TREES.
Height Height of head of progeny
Number of
head High headed Low headed
_____Total Percent -Total Percent
Ax High headed 1455 98.0 23 2.0
F1-26 High headed 227 100.0 0 0.0
F1-11 High headed 656 99.2 5 0.8
Ay High headed 930 99.3 7 0.7
D1- 6 High headed 372 99.2 3 0.8
E,-26 High headed 172 98.9 2 1.1
F1-10 High headed 256 98.1 5 1.9
C1- 8 High headed 367 97.3 10 2.7
B,-16 High headed 114 97.4 3 2.6
C2-15 High headed 243 92.7 19 7.3
J2- 7 Low headed 215 72.9 80 27.1
B2-17 Low headed 38 67.9 18 32.1
BM- 5 Low headed 76 58.9 53 41.1
Bz-ll Low headed 99 43.6 128 56.4
F2-14 Low headed 56 39.2 87 60.8
F2-17 Low headed 37 18.7 161 81.3
F2-18 Low headed 32 .13.1 213 86.9
L2-11 Low headed 14 5.8 229 94.2
Ls-19 Low headed 29 6.5 419 93.5
J2-18 Low headed 11 4.9 213 95.1
All trees in the three blocks for which the height of head is
recorded in Table 6, that did not branch during their first grow-
ing year in the field, were cut back in the manner previously
suggested to prevent the formation of cartwheels. -The growth
made the following year enabled them to branch at their natural
height, that is, as high or low headed trees (Figs. 9 and 10)
as indicated by their previous branching in the nursery. If
this had not been done the data given in Table 6 would have
been materially affected, because these trees would have
branched at this lower level upon resumption of growth in the
spring, and therefore would have been classed as low headed
trees. It is not to be inferred that the central leader is sup-
pressed in all trees which fail to branch their first year in the
field, for such is not the case. A whorl of branches will be
Fig. 12.-"Low headed" seedling of Tree Ax in Table 7. Seed
planted February 1939, photograph taken October 25, 1939.
Fig. 13.-Unbranched seedling of Tree Ax in Table
7, indicating that it is a high headed tree. Seed planted
February 1939, photograph taken October 25, 1939.
Yield and Growth Habits of Tung Trees
formed from the terminal bud of the unbranched trunk which
resembles the cartwheel condition described in the following
paragraph except that the main axis of growth is continued
upward (Fig. 16). This type of tree usually has a much stronger
crotch union than the true cartwheel; however, it does not
make as desirable a type of tree as that shown in Fig. 9.
When a high or low headed type of tree does not make suffi-
cient growth to branch during its first year in the field often
an undesirable condition arises. As this type of tree begins
growth the following year, often the central leader is suppressed
and makes no further growth and several branches are put out
at approximately the same level on the trunk which produces
a tree which is commonly spoken of as a "cartwheel" (Fig. 14).
Fig. 14.-A two-year-old tung tree showing "cartwheel" branching which resulted from the
failure of a high headed tree to branch during first year in the field.
Florida Agricultural Experiment Station
This forms a very weak crotch, with the result that, when such
a tree begins to fruit, almost invariably it is split to the ground
(Fig. 15). To prevent such a condition those trees which have
failed to branch should be cut back to from six to eight inches
above the ground. The resulting growth the following season
should be strong enough so that branching takes place as de-
sired.
Fig. 15.-Tung tree with "cartwheel" branching split by weight of fruit.
(Courtesy Alachua Tung Oil Co.)
Some growers now train their trees to a single trunk and
with certain growers it is desirable that the scaffold of branches
be formed from approximately two and one-half to four and
one-half feet from the ground (low headed type) rather than
from four and one-half to eight feet or more (high headed
type). If a low or high headed tree is desired, it will be neces-
sary to select from the nursery those trees which have branched
in the manner desired. The seed from some parent trees will
produce a much higher percent of low branching trees than
will others.
OBSERVATIONS
Two experimental plantings have been made with seed coining
from Trees 2 and 9 of the original 10 trees; both are excep-
tionally high yielding trees. An examination of these trees
when in fruit revealed a much higher percent of good bearing
Yield and Growth Habits of Tung Trees
trees in these plantings than is present in a planting such as
the New Tung Block in which the seed came from an unselected
source. There are very few trees in these plantings which are
practically without fruit while, in the New Tung Block and other
plantings originating from similar unselected seed sources, there
is a high percent of low-yielding trees and many which are prac-
tically devoid of fruit (Table 5 and Fig. 4).
Fig. 16.-"Modified cartwheel" type of branching in which the central axis is well
developed but a whorl of branches resulting from failure of the tree to branch during its
first growing year in the field was produced. Photographed during the third year after
planting.
Similar observations of two large commercial plantings made
in recent years in Florida in which considerable care was exer-
cised in the selection of seed for planting from trees of known
high yield performance furnish further evidence of the desira-
bility of such care in seed selection. The percent of high-yielding
Florida Agricultural Experiment Station
trees present in these groves and the yields thus far obtained
indicate this.
DISCUSSION AND SUMMARY
The exact flower behavior in the genus Aleurites needs fur-
ther study and elucidation, as accurate information in regard
to this point for all species, with the possible exception of fordi,
is incomplete and often lacking. For example, A. cordata has
been reported as essentially monoecious with a decided tendency
toward dioeciousness. It has been.observed that certain indi-
viduals of A. fordi are essentially staminate trees. In this con-
nection it is thought possible that trees are in existence which
are essentially female or at least will have a much larger number
of pistillate flowers per inflorescence than any yet recorded.
Evidence for the verification of such a condition within the
species A. fordi is lacking; however, it is possible that it does
exist and merely awaits discovery.
The terms "high headed" and "low headed" types of tree
denote a distinct spacing of branching in which the branches
are formed on the trunk at the nodes during the current grow-
ing season and are subsequently spaced along the trunk (Figs.
11 and 12). The two types differ primarily in the height at
which the head is formed and in the length of the internodes.
As a rule the high headed type has shorter internodes which
makes a more compact head.
None of the characters studied here can be used arbitrarily
as an index to the measurement of the desirability of a tree for
further propagation and breeding use. However, it is apparent
that of these total yield offers perhaps the best index, especially
when coupled with the cluster fruiting habit.
It should be pointed out that not all cluster type trees are
desirable, nor are all single type trees undesirable. Tree 9 of
the original-10 trees produces its fruit singly, yet it has far
outborne any other trees of which authentic yield records are
available. It is obvious that this tree should not be excluded
from the breeding program simply because it produces its fruit
singly. Actually, it does not matter how the fruits are borne
(whether single or cluster) if the tree in question produces a
high total yield.
In the selection of trees for propagation and breeding pur-
poses, in addition to total yield, some other characters of
importance that should be considered are: tree vigor, confor-
Yield and Growth Habits of Tung Trees
nation, fruiting habit, fruit size and oil content. Pickett and
Brown (8) have shown that there is wide variation in quality
of oil from individual trees as measured by the iodine number
of the oil. This point is of considerable importance and should
be given consideration.
The merits of Tree 2 of the original 10 trees are well estab-
lished. Data presented by Mowry (4) and subsequent observa-
tions by the writers have established that a high percentage of
the seedlings of this tree are similar to the parent tree in many
characters, e. g., shape and size of fruit, fruiting habit and
general habit of growth. If it were to be assumed that this
tree or its progeny did not meet a certain arbitrary standard
of growth habit desired and it therefore were to be discarded
for propagation purposes, it seems obvious that, in this case,
the desirability of this character should be discarded for the
much more important one of total yield. Furthermore, if any
given tree should have the conformation which is considered
desirable but should not have in its genetic constitution the
ability to bear heavily, it is not as desirable for propagation
as a tree which has proven its prolificacy regardless of its
conformation.
It is not meant to imply that yield should be the sole basis
of future selection. This character combined with the other
characters that will make for a greater percent of oil in the
fruit will eventually produce the highest returns. The ideal tree
would combine these factors with desirable characters of size,
conformation and type of branching.
From the data herein presented it can be seen that there is
a consistent difference in bearing ability of seedling tung trees
which is due to the difference in the genetical constitution of
the trees. The high percentage of low yielding trees in plant-
ings from which the seed came from an unselected source, as
opposed to the increased number of good bearing trees in plant-
ings in which the seed were selected from high bearing strains,
has been pointed out. The necessity for exercising care in the
selection of seed for propagation purposes from trees of known
high yield performance cannot be too strongly emphasized.
No particular correlation was found between the ability of a
tree to yield heavily and tree size, conformation and volume of
bloom produced. Fruiting habit as distinguished by the "single"
and "cluster" type is the best index yet available as to the ability
of a tree to fruit heavily.
Florida Agricultural Experiment Station
Certain flowering habits have been correlated with unfruit
fulness. It has been suggested that these may be manifestation
of a tendency toward dioeciousness.
Data are presented which indicate that the height of branch
ing of nursery seedlings and its subsequent effect upon the high
of head is a definite genetical character.
LITERATURE CITED
1. ABBOTT, C. E. Fruit-bud development in the tung-oil tree. Jour. Agi
Res. 38: 679-695. 1929.
2. HOH, H. C. Genus Aleurites in Kwangtung and Kwangsi. Lingnai
Sci. Jour. 18: 3: 303-327. 1939.
3. LEGROS, J. Cutivation of Aleurites, wood-oil trees. Int. Rev. Agr.
Mon. Bul. Agr. Sci. and Practice 26: 129-160. Rome. 1935.
4. MOWRY, HAROLD. Variation in the tung-oil tree. Univ. of Fla. Agr
Exp. Sta. Bul. 247. 1932.
5. MOWRY, HAROLD. The role of Experiment Stations in the development
of a tung-oil industry. Farm Chemurgic Council-Gulf Coast Chem
urgic Conference and Tung Oil Association of America. File No
69: 164-167. Oct. 20, 21. 1936.
6. NEWELL, WILMON, HAROLD MOWRY and R. M. BARNETTE. The tung-oi
tree. Univ. of Fla. Agr. Exp. Sta. Bul. 221. 1930.
7. NEWELL, WILMON, HAROLD MOWRY and R. M. BARNETTE: revised b3
A. F. CAMP and R. D. DICKEY. The tung-oil tree. Univ. of Fla
Agr. Exp. Sta. Bul. 280. 1935.
8. PICKETT, T. A., and W. L. BROWN. Oil variations of tung tree. Univ.
of Georgia Agr. Exp. Sta. Circ. 115. 1938.
9. SNEDECOR, GEORGE W. Statistical methods. The Collegiate Press, Inc.,
Ames, Iowa. 1937.
|
