(Kc^Z '3- ^6"^

Bulletin 456

March, 1942

Causes, Effects and Control
of Defoliation on Tomatoes

James G. Horsfall. and John W. Heubeeger

^gnculiwral ?£xpjriin«ttt Station

CONTENTS

PAG|;

Materials and Methods • • . . 184

Measuring Intensity of Infection , 184

Causes of Defoliation on Tomatoes 185

Fungi and Insects 185

Abnormal Physiology 186

Shading 186

Age of Tissue 186

Fruit Load 188

Nutrition 190

Effects of Defoliation 191

Total Yield 191

Marketable Yield or Quality 192

Relation to Type of Market 193

Poor Quality Types 193

Diseases on Fruit 194

Ripening 195

Rate of Ripening 197

Rate of Fruit Production 198

Harvest Peaks 198

Discussion 199

Combating Defoliation Diseases With Fungicides 202

Injuriousness of Fungicides 202

Performance of Fungicides 205

Fungicidal Value 205

Tenacity 206

Protective Value 207

Derivation of Protective Coefficient 208

Timing 212

Coverage 214

Practical Aspects 217

To Spray or Not to Spray 217

Materials to Use 217

When to Spray 218

How to Apply Materials 219

Varieties 219

I'Crtilizcrs 219

Air Drainage 220

Miscellaneous Suggestions 220

S I • M .\I A R\' 220

J.ITERATLRK ClTEl) 0 ) >

\ i

Causes, Effects and Control of Defoliation
on Tomatoes^

James G. Horsfall and John W. Heubergek

MORE than thirty years of research have been directed at ex-
plaining the paradox first perceived by Lloyd and Brooks (24)
that bordeaux mixture reduces the yield of tomatoes despite its
obvious value in controlling the defoliation diseases. This paradox
has been particularly baffling in the light of the fact that bordeaux
mixture is widely used to improve the yield of potatoes, a plant in a
related genus. Farmers would like to reduce the ravages of the de-
foliation diseases, but they have generally not dared for fear of yield
reduction from the fungicides. Lloyd^ and Brooks (24) initiated a
persistent fallacy in stating that bordeaux seems to cause the tomato
plants to "continue growth rather than ripen early fruit."

Boyle (1) in 1913 advanced the explanation that the defoliation
of non-sprayed vines caused them to "ripen their crop quicker than
bordeaux-sprayed vines, so that a larger portion was picked ahead
of the killing frost." Edgerton (6) applied the term "delayed ripen-
ing" to the phenomenon that underlies the paradox. This label has
crept into practically all subsequent papers on the subject of tomato
spraying.

The fatalism induced in the subject by the concept that spraying-
delays the ripening of tomatoes appears to have slowed progress on
the problem, because very few papers except those by Wilson (41,
42, 43, 44) have appeared on tomato spraying after completion of the
work that happened to be under way when the theory was advanced.

In 1929 the problem of defoliation diseases of tomatoes was
taken up at the suggestion of Dr. Charles Ohupp, Extension Plant
Pathologist at Cornell University, who pointed out the need for a
practical control of the diseases. During the intervening 13 years,
research has been conducted on various aspects of the problem.
Several portions of the results have been published (7 to 22), but
an effort will be made in this paper to summarize the important in-
formation extant on the causes, effects and control of defoliation
diseases of tomatoes as it applies in the Northeast. Particular em-
phasis will be placed on the problems of ripening of diseased and
sprayed tomatoes, and on the problems that are involved in the testing
and development of new fungicides for tomato spraying.

^ The results reported herein are based on research that began in 1929. The
earlier phases were conducted while the writers were associated with the New York
State Agricultural Experiment Station at Geneva, N. Y. The facilities provided
by the Director of that Station and by the Department of Plant Pathology are
gratefully acknowledged.

The research was done in cooperation with the Crop Protection Institute.

184 C onnectieut Experiment Station Bulletin 456

MATERIALS AND METHODS^

The essentials of the field technique have already been published
(20). Only a summaiy is needed here. Begun in 1929, the work has
continued in the laboratory, g-reenhouse and field every year except
1933. Disease data were scarce in the early j'ears of the work because
disease itself ■ was scarce in the test plants until September during
most 3'ears up to 1935 when a mild outbreak occurred. It was neg-
ligible again in 1936, but it was serious in 1937 and epiphytotic in
1938. The incidence was low in 1939, but epiphytotic again in 1940
and 1941 in Connecticut.

In all years at least four replicate plots of ten or more plants
each were used for each treatment. A wheelbarrow hand sprayer
was used prior to 1934. but from 1934 on, excepting where otherwise
stated, the spraj^s were applied with a power outfit with three nozzles
per row, 300 pounds pressure, 300 gallons per acre. Spra5^s were
standardized at one pound of copper per 50 gallons making six
pounds of copper per acre per application. The standard of ref-
erence was 4-4-50 bordeaux mixture.

On some plots in 1938, 1939 and 1941 a knapsack sprayer (Cali-
spray) developing 150 pounds pressure was used.

In taking yield records the apparent!}^ ripe fruits were picked
once a week, counted and weighed. The picking posed a technique
problem not yet completely solved. An attempt was made to pick
only i-ipe red fruit, but this was not easy on defoliated plants, where
the fruits invariabl}^ developed an orange cast. As a result the
criterion of ripening was not always the same for all plots. The
picked fruits were frequently sorted for cracks, or fruit diseases or
spray injury. ' At the end of the season the green fruits also were
picked, counted and weighed. In some years the green weight of
vines was also recorded at the end of the season. In sonie seasons
the fruit was graded according to U. S. standards.

MEASURING INTENSITY OF INFECTION

In studying the defoliation disease of tomatoes, it became neces-
sary to measure the intensity of disease attack. In assessing the
value of any treatment, it was necessary to know how many fungous
penetrations had been prevented and how mucli tlie intensity of in-
fection had been reduced by the treatment.

'i'he defoliation disease of tomatoes is such an interestin.g dis-
ease in this connection that a separate study of this aspect of the
problem has been made (18). The chief problem involved was to
procure adcijuate data quickl3^ Counting actual penetrations (leaf
spots) was f(Mind to be accurate, but entirely too slow to be really

^ The writers are grateful for extensive assistance rendered by Messrs. A. E.
Dimond, R. O. Magie, A. D. McDonnell, G. E. Nutile, and R. F. Suit.

Messrs. G. and J. Nutile of Montowcse, Conn., have provided excellent colla-
boration in the form of a growing crop and facilities for fickl work in 1940 and
1941. This help is gratefully acknowledged.

Defoliation on Tomatoes 185

useful. In studies already made on clover leaf spots (12), it was
observed that the leaf dies when about 20 percent of the area has
been hit. It was assumed that in the case of tomatoes also the pro-
portion of dead leaves reflects directly the number of successful in-
fections produced by the fun^-us.

McKinney's method (27) for measuring disease attack was adopt-
ed, and a study was made of its precision. To avoid bias each plant
in the experimental area was- examined separately by walking cross-
wise of the treatments. Each plant was classified into one of five
categories of infection based on the leaf area killed by disease attack :
0 = disease-free or nearly so, l = one to 25 percent of leaf- area killed,
2 = 26 to 50 percent of leaf area killed, 3 = 51 to 75 percent of leaf
area killed, and 4=76 to 100 percent of leaf area killed. An infection-
index in percentage for any treatment is calculated hj the following
formula :

^ ^ summation category numbers ,^^

Index= ^ — ^-^^ X 100

no. plants x 4

The 4 in the denominator represents maximum disease and 100 is
used to convert to percentage. In dealiuig with fungicides the in-
fection index is subtracted from 100 to give percentage control Avhich
brings the data into line with other toxicological data.

It was found that this method gives precise results, especially
for a group of plots, and that its precision was satisfactory even
for different times.

CAUSES OF DEFOLIATION ON TOMATOES

The defoliation disease of tomatoes is easy to diagnose. The
leaves die and drop, opening up the center of the plant and exposing
the fruit to the sun. A study has been made of the various factors
that are involved in the causation of the disease, such as fungi and
insects, abnormal physiology and weather.

Fungi and Insects

Three fungi have been found attacking tomato foliage in the
experimental fields. In the ascending order of importance these are
Cladosporiwn fulvum, Septoria lycopersici and Alternaria solani.
GladosfoHum fulvwn is rare. Septoria lycopersici has occurred spo-
radically, but it cannot be considered to have been a major factor
in defoliation during the period 1929 to 1941. A survey of the litera-
ture indicates that Septoria played a more important role in the cau-
sation of the disease prior to 1929 than it a^Dpears to have played
since 1929.

Alternaria solani has probably been responsible for 90 percent
of the defoliation during the same period. Apparently, this organ-
ism has captured the major role from Septoria during the last decade
and it api^ears to be still on the increase as a disease producer in
tomatoes.

186 Connecticut Ed:periment Statimi Bulletin 456

Flea beetles and apliids also cause some defoliation in Connecti-
cut although neither Avas much of a factor in the plots in western
New york. Flea beetles were found by Heuberger and Dimond
(9) in 1941 to be serioush^ involved in the defoliation problem be-
cause they punctured the leaves and opened the road to infection.
"W. H. ]\Iartin (25) has indited them also for transporting spores.

Abnormal Physiology

It is becoming increasingly clear that abnormal physiology of
the plant is associated with the causation of defoliation of tomatoes.
It is not yet clear, however, whether the abnormal physiology is a
jDrimary cause of leaf abscission or whether it contributes to suscep-
tibility to fungous invasion. Some people seem to feel that the
problem is primarily one of abnormal physiology. One of the im-
portant reasons for thinking so is that the defoliation disease in bad
years is seldom held more than 50 percent in check by the best fun-
gicides. If fungi were the prime movers in the etiology of the trouble,
better control should be obtained.

Shading. Simple shading of the foliage has been offered as the
cause of the disease, because shading favors the abscission of leaves.
The importance of shading per se is certainly minor, as shown by
the thousands of acres of unstakecl tomatoes where the foliage is
exceedingly dense but where no defoliation occurs. Moreover, in
years when the disease becomes serious the plants are opened to the
sun, but this does not stop the process of defoliation. This is not
to say, of course, that shading may not overbalance a situation that
borders on susceptibilit3^

Age of Tissue is certainly one of the most important variables
in the susceptibilitv of tomatoes to defoliation h\ fungi, as suggested
by Moore (28). '

Tomato seedlings in the cold frame are sometimes attacked by
Alternaria, but this seldom or never occurs until toward the end of
their seedbed life j^vhen the tissues are becoming old and hardened.
Tlie disease on such lieavily attacked plants has been observed to dis-
apjjear as if by magic as soon as the plants are moved into the iiehl
and they begin to grow again with much vigorous young tissue. The
disease reappears, however, as soon as growth begins to slow down
and f)ld tissue begins to predominate. Volunteer plants that start
late are not aifected as seriously as the ohk'r plants tliat make up
the crop. Age of tissue a])pears to be concerned in (he case of the
early susceptibility of staked ])l;iiits. (Jood air drainage normally
i-ediices attacks by k'af diseases. Slakci] plants have bettei- air ih-ain-
age than ground plants, and this shows up as favorable to them b)'
the end of the season. In 1939 the iicrceiilage of disease reached 83
for the (ground jjlanls but only (Vl lor ihc staked plauls. Tjikewise in
1940 it rear.hed OS and (iO, re'speclively.

Despite the a<lvan(age of air drainage, however, staked plants are
attackcil ciirlicr in the season than <:i(iiiiid iiliinl^ in tlic >;niii' licid.

Defoliation on Tomatoes

187

Q.

loocororoiovoo

Q.

CM CM c^i "J^ f~^ c^i i^i 00

^or^coONOOoor^Lo

<

O lO CO "^^ "^ O ro O

lo" O m5 ^ CM CO CA O
cocorOfO<^CM<N|ro

Z5

O r^ CO P CO O ^ 00
u-j rrj \d ^ iJ^ Cvi '-< CO

^^

( 1

11)

0)

n

H=

o

1-

->.

b

1

■^

u

<— )ir)000"^00
'^ LO CM 00 ^_ t> CO -^

w CO cm' vd o "-o 00 cj

,-1 CM <>.l Cvl '-I

LOiOioOOoO'O

lopc'jcfNp'O'-joo

lOli-jUOOOl-OU^lO

irjONio'OO-^t^co
CM CM cm' MD !>; t< OS Cvi

.— I CM ON Ol CO i— I -^ ■

OOOi-OOu->iou-j

vq lo T-H 00 CM Tf CO lo

O CM ^ lO ■*' T-H .-H O

t-icoCM'^CvlT-irtO

r^ -* c^i p u-) 00 CM T-H

CM CO CO "^ O O O O

lOirjiOOOOCSiO

T^o"^oqco<— iCMt-<

T-it^V],— IT— (OOO'O

" ^ ' ■ ri O VD 00 ^
'Sb'&'bi „ "O -c -o -o

ffi

Ph

H

Q

w

Q-H-

co^-^0;CO OnOOOO

^ \0 U-) -^ T^ VO LO '^ CO

'S

"to
'o

H

'^ O

lO (^1 O O CNl 1-M rj- (M ^H
u^OnCOvOmD OsONt^r^

t< o 00 ^ CO o o CN] rv]'

•rt Cvl ^H CM CM C^l C^) CM CM

0)

E
2

CM CO <0 O CO CO 00 CO ,—1

c^i u-i o "^ cm" <>i t< ^ r<

OnOOOnOn OOCDcO

^ „ 00 ^ ^ rt

'5
c

6

C^ vo ^O VO CM ON o t^ vo
p O u^ CO CM_ lO p CM_ to

o o J>! t>l t>I o\ CO w in

XI

E

2

p-^incocN] co^CMt^
co-^r^ooN c^joo^'

VO^iO^co \0^i0co

'5

£-0

00 c

T^_ 00 r^ CM ^_ CO ON -^ --H

XI

E
2

CM o ^ "j^ ^ P '^ ""; ■*.

o\ ^ c^i ON CO o r^ ■^' cm'

CM ^ ^ ■* to ^ -^ ii-> lO

O

^§

Z

CNJ CO ^ —1 C<1 ^

is

Or-HCMCO^ Ot-iCMco^

c

0)

E

Plants

not
Sprayed

Plants
Sprayed
Once
Early
With
Bordeaux

188

Comucticut Experhnent Station

BuJJetm 456

In 1940, for example, staked plants reached 50 percent defoliation
by mid-July, but this level of defoliation was not reached hj the
ground plants until mid-August. This may be related to age of
tissue, because much of the 3'oung tissue is pruned out of staked
plants, and the basal leaves soon become tough and hard, resembling
old leaves.

Fruit Load. Fruit load is concerjied with susceptibility to de-
foliation. It is a common observation first published by Samson
(33) for Septoria lycoperslc/ that defoliation strikes most heavily
on the plants that bear the largest number of fruits. The same re-
lation seems to hold for Alternmna solani. What the farmer calls
"bull"' plants, i. e. those that refuse to set fruit, seldom are aU'licteci
seriously with defoliation. Similarly, the disease seldom or never
attacks a crop in the field until after the plants begin to set fruit.

The effect of fruit load was first tested experimentally in 1040.
The fruits were removed as they appeared from two groups of ten
plants each in a large field. When the defoliation disease struck in
late August, it practically passed over the defruited plants, so that
they were strikingl}' green in September when the checks were almost
dead. At first it might seem as if the result was an artifact due to
the increased foliage that the plant pushes out in response to de-
fruiting. Such was not the case, however, because the stems were also
freer of spots on the deflorated than on the non-deflorated plants.

10 16 22 28 34 40 46
NO RIPE FRUITS PER PLANT

'ir.iKi-; 1. Kelation of l<>t;il fniit load
to susceptibility (if tomatoes
to flefoiiatioii by AUernaria
solani.

2 6 10 14 18 22 26
NO RIPE FRUITS PER PLANT TOSEPT 8

tRK 2. Relation nf fruit 1 o a <1
ciimulativi' to .September 8 to
susceptibility f)f tomatoes to
defoliation by . / //(';■ >/ <7 r i a
so Inn i.

Defoliation on Tomatoes

189

The experiment was elaborated in 1941, using Scarlet Dawn
tomatoes. The fertilizer was 3-12-6 applied as side dressing in bands,
500 pounds per acre two weeks after transplanting, and 500 pounds
three weeks later. Blossoms began to form in the field about July 1.
Beginning on July 10, when each plant had set two or three small
fruits, four replicate five-plant plots, randomized in blocks, were
laid out for each of the treatments (Table 1). All fruits were re-
moved from some plants and these were kept essentially free of fruit
until late August when defloration was discontinued. Other groups
of plants were deflorated beginning and ending progressively later.
As a result there were groups of plants carrying a few fruits all
season, a few at the beginning of the season, a few at the end of the
season, and all intergrading conditions.

Data indicate a fairly general relation between fruit load and
magnitude of infection (Figures 1 and 2). This finding agrees with
the generalization already noted that susceptibility is associated with
fruit load. It is surprising, however, that the agreement is as good
as it is, because of the variable introduced as to when the fruits
were set.

10 20 30 40
NO. OF DAYS

Figure 3. Relation of number of days
that fruit was removed to sus-
ceptibility of tomatoes to de-
foliation by Alternaria solani.

It is to be noted that the checks reached the peak of harvest on
September 8. If the total yields up to September 8 are plotted
against magnitude of infection (Figure 2) two distinct curves appear,
one for plants carrying fruits at the first of the season (labelled e
for "end defloration") and one for plants carrying fruits at the end

190 Connectieut Ex peri merit Station Bulletin 456

of the season (labelled b for '"beoin defloration""). These curves
show clearly that, for eciual numbers of fruit, the plants carrying
fruit early showed more disease than those carrying them late.

This fact suggested that the critical element here is actually the
number of daj's that fruits were picked off. The longer the j^lants
were defiorated, the less disease they developed. Since blooming
began about July 1, this can be calculated (Table 1). When these
data, are plotted a.gainst magnitude of infection (Figure 3) the points
come ver}' close indeed to a fit on the curve, showing that the num-
ber of daj^s that fruit were picked off is actuallj^ more critical in
predisposition to infection than the number of fruit finally set. This
fact confirms the observation noted above that disease seldom attacks
until after the onset of fruiting.

Observations indicate that disease usually begins to be somewhat
aj^parent about mid-July to August 1 in western New York and in
southern Connecticut. This shows several interesting correlations.
Since blossoms are set toward the end of June, this gives the disease
two or three weeks to develop after fruits begin to appear on the
vines. Steier (36) in Maryland has approaclied it differently. He
says that disease begins to remove the leaves in about 65 to 80 days
from planting. Transplanting usually begins about ^lay 20. Sixty-
five days from May 20 is July 23. Timing exiDeriments of sprays
indicates that July 10 is early enough in most 3'ears. Allowing
two weeks for incubation, this means that defoliation could be ex-
pected to begin about July 24.

NutHtion. A significant correlation of infection and nutrition
is worth noting here. Practical men believe as noted above that Al-
ternaria on tomatoes is increasing in importance, at least in the North-
east. Publications from Experiment Stations tend to confirm this. It
is suggested that this increase may be due in part to a strong trend
in the farmers' practice toward reducing the nitrogen, and increasing
the phosphorus in the fertilizer, in an effort to bring about higher
fruit loads. Although this practice may increase yield of fruit per
acre it may also increase susceptibility to Alteniana solanl at the
same time. Possibly the nutrition balance lias been disturbed. Sev-
eral fields were noted in Connecticut in 1911 where the production
of the plants was enormous but the picked jdeld was low because
the disease was so bad.

This problem has been tentatively explored experimentally. In
1040 two groups of ten Scarlet Dawn plants each in a field were
heavily fertilized witli sodium nitrate (one-half i)()und per jdant)
on July 2 and again on Jul}' 25. The base fertilizer at planliu.g tiuie
was 1000 pounds of 3-12-6 a[)plied in bands at planting time. The
nitrated plants grew luxuiiantly and fruited poorly as was to be
expected on the basis of current fertilizer recounnondations. Alter-
nai'ia attackcfl the field strongly in August and del'oliated the checks
eai'ly in S(!l)teMiber, leii\ing the niti'ated plants as green ishnuls in
the field. The gi'eeii island ell'ect, of course, was due in part to the
excessive vegetation hut, sinec the stems on the treateil plants were

Defoliation on Tomatoes 191

freer of disease than those on the checks, it follows that the treatment
had imparted resistance. Additional work should be done on the
effect of nitroigen and phosphorus nutrition on susceptibility.

It may be that the results on fruit load and nitrogen nutrition
are primarily to be explained on the basis of the physiologic age of
the tissues, because both are known to delay senescence of plant tis-
sue. Flower gardeners often pick blooms frequently to keep the
plants vegetating and producing more blooms.

EFFECTS OF DEFOLIATION

Having examined the causes of defoliation it is pertinent to
examine the effects. A clear understanding of the effects of defolia-
tion should be valuable in clearing up the mystery of "delayed ripen-
ing" said to be caused by spraying.

Loss of leaves by defoliation, of course, reduces green weight of
the plant. The leaves that are left have to assume the load of carry-
ing along the plant including the growing fruits. This is equivalent
to increasing the fruit load, and this probably ages the remaining
leaves, so that they become more susceptible than otherwise to Alter-
naria. The process then snowballs, resulting in complete defoliation
and, finally, death.

Total Yield

It is probable that defoliation can have little effect on total
fruit production in the Northeast, because most of the crop that can
be picked ahead of frost is already hanging on the vines before the
disease attack can become serious.

At first glance this statement seems at variance with the preced-
ing discussion that disease attack is associated with fruiting. The
inevitable lag, however, is tlie responsible agent. Fruits begin to
appear by late June, but their effect on the plant appears to go
through a lag period, so that initial infections do not begin until well
along in July. Several days are required for each spore generation
and as a result disease seldom attains sufficient momentum to induce
much defoliation until mid-August or later. By that time it is too
late to obtain fruits from the blossoms that set. Tagging experi-
ments with blossoms described in more detail below have shown that
nearly 60 days are required to ripen the crop. That means that fruits
set after August 1 have small chance of being picked ahead of frost
in the Northeast.

Three divergent approaches are possible in measuring the effect
of disease on yield: (a) comparison of diseased with healthy fields,
(b) comparison of diseased with healthy plants, and (c) comparison
of diseased non-sprayed plants with plants kept in various stages
of defoliation by different sprays.

Practical farmers use the method of comparing healthy fields
with diseased fields and comparing yields in disease years with yields
in disease-free years. They are all in agreement that the disease re-

192 Connecticut Experiment Station Bulletin 456

duces the yield of the fridt that they can pick, but this is a problem
in marketable fruit ^yhicll will be discussed below. No data is aA'ail-
able for making such comparisons of total yields of fields.

In 1929 the problem was investigated using individual plants.
In that year all the ripe fruits on each of the 552 plants in a spray
exjDeriment were picked each week. The intensity of infection on
September 18 was obtained for each plant as described. Consequently
data are available on the j'ielding performance of individual plants,
both sprayed and non-sprayed, that carried dili'erent amounts of
disease at the end of the season.

The total yields were assembled by disease categories for 112 non-
sprayed plants and for 88 plants sprayed only once early with bor-
deaux (Table 2). Results were clear cut and identical for the two
groups of plants, but the implication differs depending upon how
they are stated. The results may be stated in the form that the
yield of ripe fruits increases as the disease increases, or thej'^ niay be
said in the form that the most prolific plants developed the most
disease. This latter method of arranging the statement is probably
the more accurate, because other data just discussed show that the
prolific plants do develop the most disease.

Upon pursuing the matter further it appears that the green fruit
acts contrariwise. As disease increases gi'een fruit decreases. A
further step in the analysis shows that the proportion of ripe fruit
increases as disease increases. In practice this means that more fruits
on defoliated than on non-defoliated plants are picked ahead of frost.
Sometimes this statement is put in the form that they ripen ahead
of frost. This brings up the fallacy that disease accelerates ripening
because that explanation can be offered to account for the fact that
most of the crop on defoliated plants is picked ahead of frost.

This matter will be considered in more detail below.

Marketable Yield or Quality

If it seems difficult to measure the effect of disease on total
yield because of the complication of high yield, high disease, it is
even more difficult to measure the effects of disease on quality yield.
Quality in tomatoes is an ill-defined concept, and the concept changes
from market to market and from season to season. When tomatoes
are all good, the market is choosy as to quality; when they are all
mediocre, the market takes almost anything. AVhen prices are poor,
quality mils! be excellent if the fruit is to move. Wlien prices are
good, anything moves.

lirowii in 192N (2) seems to have made tlie only attempt to meas-
ure the effects of disease on yield of marketable fruit. Early in
August he surveyed 1,991 acres of canning tomatoes in several Indiana
counties and rated eacli field as to whether infection was slight, mod-
erate, or heavy. II<; obtained data from the canning factories on
tonnage j^urcliased. lie found (hat 1,199 slightly diseased acres
produced 3.47 tons of marketable fruit per acre; the 51!) moderately
diseased acres yielded 2.79 tons per acre, and the 27.'') heavily dis-

Defoliation on Tomatoes 193

eased acres yielded 2.26 tons per acre. Assuming 3.47 as a normal yield
of marketable fruit for the area for that year, it may be de-
duced that a medium attack reduced the yield by 19.3 per cent, and
that a severe attack reduced the yield by 34.5 per cent. This finding
is precisely in line with what farmers think in relation to the effect
of defoliation on yield of fruit that they can sell.

Defoliation induces or aggravates certain off-quality conditions,
such as flabbiness, cracks, sunscald, orange instead of red color, and
off-flavors. It also aggravates such diseases as stem-end rot, anthrac-
nose, and soil rot.

Relation to Type of Market. The effect of these factors on
marketability depends, of course, on the needs of the market con-
cerned. Flabbiness, cracks and fruit diseases are reflected in salabil-
ity in almost any market except the most bearish. Color, how-
ever, is of critical importance to the canner and roadside markets.
The canner constitutes a very critical market and he now buys large-
ly on U. S. grades which are based on color, presence of mold and size.

Poor Quality Types. Pickers often note that fruits on defoliated
plants are taore flabby than those on normal vines. Fruits on defoli-
ated plants also crack much worse than those on non-defoliated plants.
This reduces the marketability of the fruits in almost any market,
perhaps more in the local market than in the cannery, because the
canner may pare out the cracks. Cracks reduce marketability also
by permitting the entrance of rot-producing organisms that reduce
the fruit to a skin full of slime. The probable reason for the increase
in cracking of fruits on defoliated plants is that they do not have
the leaf tissue to soak up the extra water that the roots take in during
a rain. There may also be a reaction to light. Fruits grown in heavy
paper bags in 1940 cracked much less freely than those not so bagged.
Presumably this was a matter of light.

The exposure of fruits to strong sunlight when leaves fall often
results in sun scald which makes the fruit w;holly unsalable. This
may be a serious factor immediately following the loss of the leaves
if a hot spell occurs.

The flavor of fruits appears to diminish as the defoliation in-
creases. This factor is almost undefinable, but it is probably associat-
ed with a lowering of sugar content. The fruits seem to be insipid,
flat, or even mildly sour.

Probably the most important factor in lowered quality that
comes from defoliation is the poor color as first reported by Pritchard
and Porte (31). Fruits on severely defoliated vines seldom or never
attain a normal deep red color, but rather they reach an orange red
color that is not acceptable to a critical trade like a cannery or a
roadside market. However long such fruits remain in the field, they
remain a sickly orange, never becoming rich red. In one severely
diseased field in Connecticut in 1941 more than 50 fruits per plant
were left unpicked in the field because they would not "color up."

A study of some of the possible causes for this effect of defolia-

194 Connecticut Expenonent Station BnUetin -156

tion on color are interesting. In 1913 Diiggar (5) showed that the
red color (lycopene) in tomatoes is closely limited hy temperature.
Lycopene forms very slowly at temperatures below 55°F. This ac-
counts for poor coloration of fruits in the fall. Likewise, the color
is not formed if the fruit temperature rises much above S5°F. Rosa
(32) showed in. 1926 that the yellow pigment (largely carotin) forms
quite readily at a temperature of 85"F. or above. MacGillivray (26)
then showed in 1935 that the temperature of the fruits on defoliated
plants ma}' rise as much as 20°F. higher than that of fruits on non-
defoliated plants nearby. He concluded that these elevated temper-
atures encouraged the yellow color and discouraged the red color,
thus giving rise to orange colored fruits.

In 1936 Ora Smith (35) investigated the ellects of light on
tomato ripening in connection with his studies of artificial ripening.
He found that light favored the development of the yellow carotinoid
pigments and discouraged the developed of the red lycopene. In some
tests here in 191:0 fruits on staked vines were enclosed in heavy paper
bags. When they ripened the color was beautifully rich red instead
of the typical orange red of the fruits ripened as they hung from the
stakes in the sun. It follows that both light and temperature are
concerned in the ditt'erential coloration of fruits exposed in the sun
when the leaves die and fall away from them.

Diseases on Fruit. Loss of leaves, produced by disease or by
hand, appears to increase the susceptibility of fruits to anthracnose
or ripe rot caused by C olletot)'ichum pJw?no/(/es. Anthracnose ap-
pears to be on the increase in the Northeast, probably because de-
foliation is on the increase. The disease occurs as rounded sunken
spots with a smooth margin. They look as if they had been pushed
in by an index finger without a fiuiger nail. The sunken area is
ccjvered with minute pimples arranged in circles. The pimples usual-
ly turn dark in hite stages. Sometimes anthracnose is called nail-
head in Connecticut. This is a misnomer as that name was coined
for an entirely different disease found only in the South. The name
anthracnose or ripe rot is preferable.

Stem-end rot niay sometimes occur plentifully on (k^foliated
plants as it may be caused by Alternarla so/ctii, the fungus connnonly
associated at [)resent with (k'foliation in the Northeast. This disease
produce's a black sunken area around the stem sometimes spreading
irregularly out onto the shoulder of the fruit.

1 1 piohably attacks these ai'cas because the spores fall there and
HimI i-ondilions suitable for j)en('li-ation. It may sometimes attack
cracks as well. On occasion as in 193S it may cause widespread
(hopping of liiiits when the infection spreads to the pedicle and
kills it.

in lection oriiginaling fi-om soil borne organisms sonu^times si'cni
to be relal<'(l to defoliation but counts in New \(>rk in 193S and in
(Connecticut in 1910 failed to demonstrate the ellect. Otliei- fruit
diseases such as naillHsul, bacter'ial spot and blossoui-end rot do not
a|)|»c:ir to he iLLfiri';! \ ;i(c(| h\' i|cl'oli;il iori.

Defoliation on Tomatoes

195

Ripening

The effects of defoliation on ripening is a complex but exceed-
ingly interesting problem, that has been the subject of much specula-
tion in the literature. The problem arose out of an effort to explain
early results on the effects of sprajdng on yield of tomatoes. Two
opposing points of view have been evolved, one that spraying delays
ripening, the other that defoliation accelerates ripening. Although
some data have already been presented (20) on this subject, it will
be analj^zed here in more detail because more data have become avail-
able.

Lloyd and Brooks (24) in 1910, Boyle (1) in 1913, and Edgerton
(6) in 1914 to 1918 are chiefly responsible for the prevailing points
of view on ripening. They designed their experiments essentially
alike and they all obtained essentially the same results which have
been duplicated many times since (40, 43). They sprayed some
plants with bordeaux and kept some not sprayed. They picked the
fruit as it ripened and they examined the picking curves, expressed
either cumulatively or as frequencies.

One phenomenon is characteristic of all curves. The picking
curves for the sprayed plants are flatter in the beginning of the
season and they reach the peak of the harvest later in the season than
those for the non-sprayed plants.

80

1 \ \ \ \ A

J f -

/ i

/ i

60

f i

\ 1

40

f

i

1

1

20

— // -

y*/ CHECK r> ..n

y^ A SPRAYED® ®

0

^■Y 1 1 1 1

AUG. 18 25 SEPT 18 15 20
DATE OF HARVEST

Figure 4. Cumulative harvest curves
of tomatoes sprayed with bor-
deaux mixture and not spray-
ed in a mild disease year,
1937.

64

1

II II

-

^^

. — <i

48

32

-

/ /

-

/ / CHECK ° —

16

-

/ / SPFWED®--
/

II II

— -®-

\

<r^'^^

JULY 10 18 24 31 SEPT 8 14 7

DATE OF HARVEST

Figure 5. Cumulative harvest curves
of tomatoes sprayed with bor-
deaux mixture and not spray-
ed in a severe disease year,
1938.

196 Connecticut Experiment Station Bulletin 456

In addition, there are two distinct types of curves best expressed
as cumulative curves (Figures 4 and 5). In the first type the curve
for the sprayed plants remains below that for the non-sprayed plants
throughout the season. In the second type the curve for the sprayed
plants overtakes and finalh" passes that for the non-sprayed plants.

Both of these types have been obtained in the present research.
The picking curves obtained in 1937 Avhen disease was light illustrates
the first t3'pe. The curve for the non-sprayed plants remained always
ahead of that for the bordeaux-sprayed plants (Figure 4), so that by
the end of the season the non-spraj^ed plants had yielded more ripe
fruits than the jDlants sprayed with bordeaux mixture.

The picking curves obtained in 1938 (Figure 5) when disease was
heavy illustrate the second type. Although the curves for the non-
sprayed plants forged ahead of that for the spraj'ed plants early in
the season, it began to lose pre-eminence by early September. The
curve for the sprayed plants passed that for the non-sprayed between
September 8 and September 14.

The work on tomato defoliation almost invariably shows that
the sprayed plants retain more green fruits at frost time than non-
sprayed plants. This fact has been used as an indirect measure of
amount of disease (18).

Upon examining these two types of curves Lloyd and Brooks
(24) concluded that the spraj^ caused the plants "to continue growth
rather than ripen early fruits." Boyle (1) concluded that defolia-
tion of the non-sprayed vines caused the plants "'to ripen their
fruit quicker" than bordeaux-sprayed vines, while Edgerton had dif-
ficulty in deciding between the two possibilities saying that "This
partial defoliation of plants causes a more rapid development of the

fruit Holding this foliage by means of sprays produces greater

vegetative growth and slower development of the fruit." Edgerton
apparently leaned toAvard the latter explanation because he titled his
paper "Delayed ripening of tomatoes caused by spraying with bor-
deaux mixture." This explanation has been retained even up to 1940
(40, 41), despite W. H. Martins conclusions in 19-20 (2.")) : "It is

not believed that tlie ])resence of boi-deaux mixture on tlie plant

has any direct influence on the ripening period."

It is well to examine the major premise in the arguments. The
major premise is that the slope of the picking curve is a function only
of the late of ripening. The alternative premise is not consideretl —
that th<; slope of the ])icking curve is a function of the rate at which
blossoms and then fruits are produced. 'Jlie latter assuni])tion is so
simple that it seems sti'ange that the formei' could h;i\»' hccn adoi)te(l
at all without disj)ro\ing the latter, at least.

'\\\K\ sitiiiition pci-haps arose because of the point of view that
many faimei-s hold. When a farmer picks (uc lield ahead of another
he often says that i( lijx'iied <'arlier. I'rohahly all he is aware of is
that he i)icl<s moi'e ii[)e fruil in th<> eai'ly pait of the si^ison from one
lield than fi'oni an(»tliei-. This is a confusion of the coiiceijt of the
ii|)ening or reildeniiig of a iiiiit with (he produit i<Mi oj' i-ipe fruit

Defoliation on ToTnatoes 197

from a field. It is then desirable in clarifying this matter to limit
the term, ripening, to the rate of maturity (reddening) of fruits,
not to the rate at which a field produces red fruits.

In the case of the problem in hand, it is imperative to decide
whether the difference between the picking curves is due to differences
in rate of fruit reddening or to differences in the rate at which fruits
are produced. Experiments must be designed to keep the two separ-
ate. At the same time the experimental design must also be capable
of keeping the effects of spraying separate from the effects of disease.

Rate of Ripening. The effects of spraying on ripening has to be
determined first in the absence of disease. The rate of ripening is
found most certainly by tagging blossoms. In 1936 drought was so
serious that disease never appeared in the plots in any quantity.
Bordeaux was applied all season. All the blossoms that appeared
were tagged on ten sprayed plants and on ten non-sprayed plants.
The number of days for the average fruit to ripen was 54.1 and 54.9,
respectively. Clearly the spray exerted no effect on rate of ripening
in the absence of disease. This agrees with Martin's conclusion (25).

The effect of defoliation in the absence of fungus was tested in
1937. From the 1936 data just presented, it was known that fruits
set after August 15 could not possibly be picked as ripe before frost.
Accordingly fruits were allowed to set normally until mid-Aujgust
on sprayed plants. Then half the leaves were removed by hand from
80 plants in four replicates of 20 plants each. Fruits were picked as
usual and curves were plotted (20). Since the curves could be super-
imposed, it follows that mechanical defoliation at least had no effect
on the slope.

It is commonly held that pruning to a single stem accelerates
ripening. Watts (39) tested this hypothesis experimentally by tag-
ging blossoms on pruned and non-pruned tomato plants growing in
the greenhouse. Fruits on pruned plants ripened in 43.3 days and
they ripened in 43.4 days on non-pruned plants. Here also it is clear
that defoliation had no effect on rate of ripening.

The possibility remains, however, that defoliation by disease
may act differently from mechanical defoliation in its effects on ripen-
ing. Accordingly, in 1941 blossoms were tagged on six bordeaux-
sprayed Victor tomato plants that lost less than 20 percent of their
leaves during the season and on six non-sprayed plants adjacent that
lost more than 80 percent of their leaves from Altemaria solani by
September 1. The 65 fruits tagged on the sprayed plants ripened
in 51.6 days and the 162 fruits on the defoliated plants ripened in
50.9 days.

Clearly the defoliation from disease effected the same results as
artificial defoliation or pruning. It had no measureable effect on the
length of time from pollination to ripening and hence no effect on rate
of ripening. Since neither spraying nor defoliation has exerted any
measurable effect on rate of ripening, it seems that the alternative
premise needs study — ^that picking curves are a function of rate of
production.

198 Connecticut Expenment Station Bulletin 456

Rate of Fruit Production. The effect of spraying on fruit pro-
duction has to be determined in the absence of disease. This subject
has been investigated in considerable detail (20) and it will be
summarized below. It is only sufficient for the purposes here to state
that bordeaux dwarfs young plants so that the rate of blossom pro-
duction is slowed. In the case of the 1936 experiment noted above for
tagged blossoms, the 10 sprayed plants produced 44 young fruits
dui-ing the peak bloom period of July 21 to 28. The 10 nearby non-
sprayed plants produced 164 young fruits, or four times as many.
The picking data 54 claj^s later showed that all 44 sprayed fruits were
picked between September 9 and 16, and that 157 of the 164 non-
spra3'ed fruits were picked. Clearly the higher picking rate of the
non-sprayed plants was due to a higher production rate, not to a
faster ripening rate. The assembl}^ line moves at the same rate of
speed for both sprayed and non-sprayed fruits, but the units are
spaced farther apart for sprayed plants than for non-sprayed plants,
so that fewei" units per week come off the end of the belt.

Harvest Peaks. The experimental work so far appears to ex-
plain adequately the smaller production of bordeaux-spraj-ed fruits
early in the season, but this work does not explain the fact that the
yield of bordeaux-spra3'ed plants often overtakes that of the non-
sprayed plants and may even surpass that of non-spraj^ed plants in
severe disease years as in 1938 (Figure 5).

Also the experimental work so far does not account for the fact
that sprayed plants pra-ctically alwaj^s produce more fruits than the
checks in the pickings at the end of the season. Frost often destroys
this portion of the crop. This fact has often been advanced in sup-
port of the argument that defoliation accelerates ripening. The
argument is that, if a larger proportion of the crop is picked ahead
of frost from the sick plants than from the healthy plants, then the
speed of ripening must be accelerated by defoliation.

Since the tagging experiments show that defoliation does not
accelerate ripening, some other mechanism must be sought to account
for this effect at the end of the season. If the cumulative picking
curves are converted to frequency curves, as those published by Ed-
gerton (6) and those obtained in this work in 1938 (Figure 5), it is
clear that the checks reach the peak of picking one or two weeks
aliead of bordeaux-sprayed plants.

It has already been shown (20) that tomato [)lants set fruit in
proportion to their size. Since non-sprayed plants grow faster tluiii
those sprayed with bordeaux, it follows that they will begin to set
fruits quicker; tliey will attain their maximum growth earlier in the
season and. arcoi-dingly. the pcalc of piclcing will o-'cur (Nii-li(M- in the
season.

On (he othei' hand disease usually begins to become serious by
the time the unspi'ayed plants ap])roa('h the peak of maxininm gi'owlh
and inaxiniuni Iruit setting. It not only hlows /growth down, but it
may actually cau.se tin; plant to lose green weight, ^^'hen that ocnii-s
the plant ceases abruptly to put on new fruits, so lli;i( the jiicking

Defoliation on Toniatoes 199

curve may be truncated. Plants with less disease are able to lay
on new tissue faster than the disease removes it. As a result they
continue to increase in total green weight and to set fruits. The
peak of the picking is thus displaced toward the frost end of the
season. This leads to the fact that frost destroys a larger proportion
of fruits on non-defoliated than on defoliated vines. In short, the
non-diseased plants continue in the normal fashion to produce fruits,
with the result that some are caught by frost. They would continue
to produce fruits indefinitely unless frost came. The defoliated plants
can no longer set a crop, however, and so frost catches but a small
proportion of the crop.

It follows that two factors account for a large proportion of
fruits at frost time on sprayed pl-ants : (1) dwarfing forces the
peak of production later in the season and (2) non-diseased plants
continue to add green weight and accordingly blossoms, while diseased
plants lose green weight and cease to set blossoms that might produce
fruits to be killed off.

Discussion. Since the effect of disease on ripening is such an
important subject in the tomato defoliation problem it will be sum-
marized and discussed here as a whole. These studies of the problem
indicate that neither disease nor sprajdng has any effect on the speed
at which fruits ripen. The slope of the picking curves is governed
wholly, it seems, by the rate at which young fruits are "set" by the
plants. The flatness of the slope of the picking curve for sprayed
plants early in the season and the lateness of the peak are due to the
dwarfing and defloration action of the sprays. Tiie flatness of the
picking curve for defoliated plants late in the season and the small
percentage of green fruits caught by the frost are due to failure
of the plants to continue to set fruits after disease becomes serious.

Pritchard and Porte (31) attempted to kill the theory of delayed
ripening at its inception by reporting data from many experiments
showing that spraying did not affect the picking curve. Their
criticism failed to register, however, because their data were of the
same type as that used to set up the theory, i. e. picking data. It
was only a question of one set of positive picking data against a
neigative set of picking data derived by Pritchard and Clark (30).
The fact was that the picking curves from sprayed plants are fre-
quently flatter than those of non-sprayed plants. No amount of tests
where this effect fails to appear can really disprove it. W, H. Martin
(25) almost solved the problem in 1920 by pointing out that bordeaux
itself probably had nothing to do with ripening, but that it maintain-
ed the leaves which shaded the fruit, "thus delaying its ripening."
Martin seemed to recognize that the "spray applications greatly in-
fluenced the production of ripe fruit." It seems unfortunate that he
did not explore the effect of sprays on production, not on ripening,
because his conclusions would have been different if he had.

Smith and Zimmerley (34) plot harvestiiiig curves and they
recognize an effect of spraying on time of ripening. If they had
examined the "time of ripening" on the basis of time of fruit set, in-

200 Connecticut Experiment Station Bulletin 456

stead of rate of ripening, it would not have been necessary to re-
investigate the problem 20 years later.

J. D. Wilson of Ohio (42, 43, 44) has made many observations of
the effects of sprays on vegetables. He has explained several of
these as instances of delayed ripening. It is interesting to inspect
these observations in the light of the hypothesis that spraying has no
effect on speed of ripening, but rather that the effects are due to
differences in production rate. "Wilson (41) reports that "net in-
creases in yield due to spraying are finalh' obtained if, and frequently
onh' if. defoliation of the untreated controls is severe enough to cause
a considerable decrease in A'ield."

'"This injury trend had to be offset by the beneficial effects

of disease control before any net increase due to spraying could be
obtained. The existence of this injury zone fails in some instances
to give a net increase in yield over similar but untreated plots."
One of the classic examples of reduced yields due to spraying involves

tomatoes Septoria leaf spot was severe enough (in 1938 and

1939) "that many of the spray materials used gave sufficient disease
control to offset the injury factor of spraying, with the result that
sprayed and dusted plots consistently produced a greater quantity of
fruit than untreated ones." As usual, however, the picking curves
were Hatter for the sprayed plants than for the unsprayed plants
and, as usual, this fiat slope could easil}^ be explained ]3y the injurious
effect of the spray on production of llowers and fruit set. ^^'ilson
feels that "ripening of the fruits was delayed long enough that pro-
duction from some of the sprayed plots did not exceed that of the
untreated controls until near the end of the picking season."

Wilson (41) offers as another example of delayed ripening of
muskmelons, the fact that materials change ranking as the season
advances. In the case of a material such as Compound A that yields
poorly at the beginning of the season and high at the end Wilson
feels that the ripening is delayed.

It is worth Avhile to study the comparative sliifts between Com-
pound A and Cupro K. Compound A ranks low early in the season
and high late in the season. Cupro K ranks high early in the season
and .low late in the season. If compound A delays ripening, then
does not Cupro K accelerate ripening I

Either conclusion is more easily explained on another basis.
Both materials are oxychlorides and both materials possess approxi-
mately equal tenacity. The copi)er in Coni]:)ound A has a higher spore
killing powci" than that in Cupro K and it is more injurious on copper
sensitive folia,ge such as liuia bean according to Wilson's (hita. Com-
poiiiid A ;ilso has a higher protective coefficient than Cupro K.

I he, theory of rediu;ed fmit production would give the following
cxplanalion: since Compound A is more injurious to tomatoes than
('u|)io K it would give lower early yields, but since it is relatively
more |M(.t('c(i\(' against disease, it woukl give liigher late yields.

ir llic woids. (ribasic copper sulfate, .-iic sultstitutcd for Cupro

Defoliation on Tomatoes 201

K in the preceding discussion, the general picture is the same and it
fits the facts equally well.

Wilson (41) offers still another interesting case of delayed ripen-
ing. Since cucumbers for seed are picked at the end of the season,
there is no question of criterion of ripening and the yielding poten-
tialities are all realized at the end of the season. In one experiment
the yield of cucumber fruits was somewhat higher on Compound
A-sprayed plants than on tribasic copper sulfate-sprayecl plants, but
the Compound A-sprayed plants had distinctly fewer seeds than the
tribasj.c-sprayed plants.

The explanation is offered that the seed production was halted in
mid-season by mosaic and since "fruits in the plot which had been
treated with Grasselli Copper A were the greenest .... they produc-
ed the smallest amount of good seed per pound of fruits .... It seems
likely that the premature arresting of seed devolpment in the fruits
on affected plants may have increased the vegetative growth sufficient-
ly to account to some extent for their greater weight at harvest time."

The major premise here is that mosaic arrested seed development.
It seems equally plausible that the copper in Compound A killed
more pollen, as it does more spores, than the copper in tribasic sul-
fate, and hence the fruits carried fewer seeds at the end of the season.

Even if mosaic were accountable for the low seed yield an alter-
native explanation is possible. On account of differential injury,
the tribasic plots were carrying mostly big early fruits with seeds
already set when mosaic struck. The Compound A plots were carry-
ing mosth^ late set small fruits with immature seeds when mosaic
struck. This explanation involves later production and later picking,
not later ripening.

Finally, the theory of delayed ripening suggested to Wilson
(42) that a farmer spray one portion of a tomato field with tribasic
sulfate and one with copper Compound A in order to spread his
picking load. On the theory of differential production, not differen-
tial ripening, the farmer would apply Compound A heavily early
in the season to one portion of his field, to kill off a lot of blossoms
and give good protection. This portion of the field would reach
peak production late. To the other portion of the field he would
apply Compound A lightly and late. He would pick this portion
early.

One aspect of this ripening that has not been emphasized in the
text, because no data on the point are available from these tests,
is the effect of defoliation in elevating the temperature of the fruit
and the effect of this temperature elevation on speed of coloration.
Rosa (32) picked fruits at comparable stages and stored them at
various constant temperatures. The rate of ripening in days was
24 for 11°C., 16 for 16°C., 8 for 25°C., and 11 for 30°C.

MacGillivray (26) has ample data to show that defoliation ele-
vates the fruit temperature, even as much as 10°C. The data at
New Haven in 1941 show that fruits on defoliated plants ripened no
faster than those on non-defoliated plants. The fruits were yellow

202 Connecticut Expeniiient Station Bulletin 456

red and not red, of course, but they could not be picked for peak
color, any sooner, on account of it. How does this fit with the above
data?

The expLination would appear to be that the fruits on defoliat-
ed plants were generally warm enough to ripen at almost maximum
speed. Any advantage of the extra warmth derived from the sun
in ripening the fruit Avas offset on defoliated plants by the possibility
of overheating them. Overheating reduces the speed of ripening
according to Rosa.

COMBATING DEFOLIATION DISEASES WITH FUNGICIDES

During the period of this research on tomato defoliation an
extensive study of fungicides has been made. It was obvious to
begin with that bordeaux was depressing the growth and yield of
tomatoes. One prime objective was to investigate the causes for this.
Much of the work on this objective has already been published, but
it will be summarized and supplemented herein.

The second objective was to develop new fungicides to combat
the disease without injury. As a part of this objective it was neces-
sary to investigate the properties of fungicides and learn whj' they
perform as they do.

Injuriousness of Fungicides

The research on injury has been confined to copper fungicides.
The effects of bordeaux mixture on ripening have just been discussed,
but the reasons why bordeaux depresses fruit set and therefore pick-
ing were not included.

The timing of sprays is related to injury. Measurement of in-
jury, of course, is complicated by the effects of disease control that
tends to parallel injur3^ It has been shown that the amount of green
fruit at frost time measures disease control (18). (xreen fruit, there-
fore, cannot serve as a good measure of injur3\ Likewise, the last
one or two ripe harvests tend to measure disease control rather tlian
injury.

It has been decided, tlierefurc, to use as a measure of injury the
cumulative ripe yield up to the date when the non-spraj^ed plants
reached the peak of their picking curve. In 1936, the plants reuiain-
ed essentially free of disease on account of drought, but the 'drought
exaggerated s[)ray injury. Tliroe materials were compared: bor-
deaux mixture, red copi)ei- oxide, and red c()j)])er oxide jilus cotton-
seed oil emulsion. In one series four applications were a])plied prior
to commencement of blooming on June 1^6. and in the other series
12 applications were applied all season ending September 1.

The checks i-eached llicir pcid^ of rijx'nin.t:- on September 14 and
the yield, up to that dale, was --'.Tf) pounds pfi- phint. Bordeaux aji-
plied all season reduced tlie yield to 0.72 ])ounds. If the ap])lications
were all a|>[)lied alifail of blooming, the yield was niluceil only to
2.6o jxiunds per plant.

■Defoliation on Tonatoes 203

The timing experiment was reversed in 1938. Four applications
were made late in the season, after August 1, instead of early, ending
June 26. The all-season sprays were used for comparison. The
checks reached their picking peak on September 8, 1938, and the
yield up to that date was 4.22 pounds per plant. Bordeaux ap-
plied ail season reduced the yield to 3.56 pounds per plant but,
when applied after August 1, the yield was 4.0 pounds per plant.
It is clear that withholding- the applications until the middle of the
season essentially eliminates the injury. From these two timing-
experiments, it follows that sprays applied either before blooming
begins or after bloomiuig ends are less injurious than those applied
all season.

The causes of this depressing action are not far to seek. The
comparative performance of bordeaux and red copper oxide imme-
diately indicates lime (22) because lime is the outstanding difference
between bordeaux and red copper oxide. The effects of lime have been
investigated extensively both on tomatoes (14) and on cucurbits (15).
lime, especially hydrated lime, appears to be definitely deleterious
when applied to foliage of these plants. If cuticles are thin, it
appears to saponify them (14) so that water escapes readily. This
effect may be minimized if cuticles are old and hard, however.

Lime appears to enter the tissues and to make them tough and
harsh. In 1939 Dr. K. F. Suit of the New York State Experiment
Station made puncture tests of sprayed tomato fruits using a Joly
balance with a flat-tipped needle 50 microns in diameter. The needle
was 100 microns in diameter 100 microns from the tip. He made four
punctures in each of 20 fruits for each treatment at 8 A. M. The
average pressure to puncture was 11.65 grams for the non-sprayed
fruits, 11.71 for red copper oxide-sprayed fruits, and 13.93 for bor-
deaux-sprayed fruits. The difference between the bordeaux and the
other two was statistically significant by analysis of variance.

It is suggested that the hardening of tissue occurs because cal-
cium hardens the pectin of the middle lamella as suggested by Ker-
tesz et al. (23) in researches on calcium in canned tomatoes.

The fact that the depressing action of bordeaux occurs chiefly
on young plants ahead of and during blooming suggests two factors,
dwarfing and defloration. Both of these have been shown to be im-
portant factors in the field (20). The dwarfing appears to result
from the hardeniuig of the cells so that the expansion phases of growth
are interfered Avith. The explanation for defloration has not yet been
derived.

In any case both dwarfing and defloration, reduce fruit set and
this reduces the load of pickable fruits with its interesting results on
the slope of the picking curves.

The results on the nature of bordeaux dwarfing suggested im-
mediately the use of lime-free copper compounds. Red copper oxide
was first used experimentally as a dust for tomatoes in the summer
of 1932. Later work with the material and with other so-called
fixed copper compounds has showm that the materials are all some-

204

C onnecticut Experiment Station

Bulletin 456

u

Q

QJ C C
Q. <U 01

^ o o

o <u

eg C^l O O <^j tr> <^- 00 so O -^ CC ^ -^

t^r^'^C\"^00'O00C<lOC00— 1

OO-^-— 'COO»— iO»-<0«-<C3COt— (
^CJO'OOOOOOOOOOO

; Tt CNl r^, "^^ 1^1

! ^ \d o --< t^
. Lo -^ rg eg 00

■ O -^ O CN] --^ c^ oo

■ c\ t^ '^ -rt" *"! iW "S

. Tj- 'i- (N (N "M ro eg

s a s

SEE

o c c

(J o u ;i; _y I;;:; ■-

.5 (u .=: .5 .2 4; u £ li li .5 .i: .5 :i

u;2;uucj:2;;2;fiHPHPH

o

P ii s

ct

:; rt ^

;_;

•— "^ 2

ty.

e

-s •■'

3

.b'.5 .-f

iJ

a; u o

(fl

rt rt r;:

u

UUcTj

1- rt 1- p:

'-« O r- c

=j E c oi; Si;

ni 1- u p p ""

"S & & --^ '-^ s.

2 ,, ,, o o I-
K CQ eq U U UJ

= .h; .ii a

t- h C. C C C rt

2 iU''J«J*i<Uij2

ty (y w (y .

i_ rH ."5 rH ."H "^ '

G rs ."T 'o 'o 'G 'o _ t

X ,*. »^ /^ - r~ ,- ,- U V

O rt
CQOQ

aacaac-ay g
OOOP3333 h;? >$

£0

Defoliation on Tomatoes 205

what injurious to tomatoes, but that they are much less injurious
than bordeaux mixture. It seems probable that the copper acts in
somewhat the same fashion as calcium in producing injury, but that
it is much less active.

Performance of Fungicides

If the protective action of bordeaux on foliage diseases is to be
duplicated or improved without injury, it becomes imperative to
investigate how and why fungicides act as they do and how new
materials can be fitted into the knowledge thus learned.

Many lime-free copper compounds (Table 3) have been put for-
ward as bordeaux substitutes. Since these perform differently in the
field, as might be expected their properties had to be determined
and studied piecemeal.

The ability of a fungicide to protect plant parts in the field
has been defined as protective value (21). The two prime factors
that govern protective value are fungicidal value (i. e. spore-inhibit-
ing power) and tenacity (i. e. resistance to weathering). These two
factors can be investigated easily in the laboratory where many of
the extraneous factors encountered in the field can be controlled.

A considerable number of researchers have measured fungicidal
(fungus killing) value in the laboratory for various materials and
have attempted to correlate it with the protective value in the field.
The partial lack of correlation has lead some pathologists to feel
that laboratory testing is worthless.

The problem has been studied extensively for several plant dis-
eases. The underlying technical considerations are being published
by Dimond et at. (4) as an accompanying bulletin. The usefulness
of these considerations in the control of the defoliation diseases of
tomatoes with copper materials will be discussed here. Some of the
sources of error in studying fungicides on tomatoes will be discussed.

Fvngicidal Value. The first prerequisite to a study of the spore
inhibiting properties of insoluble fungicides was to develop a pre-
cision laboratory sprayer and the correlative precision techniques
(19). Briefly, the materials are suspended in water and sprayed
under standard conditions of humidity, time and distance to a stand-
ard surface of cellulose-nitrate on glass. Spores of the test fungus,
Macrosforiiim sarcinae forme.) are applied in standard concentrations,
in standard amounts with a standard pipette in a standard fashion.
This assures known and reproducible numbers of spores in relation
to known and reproducible amounts of toxicant.

After incubation under standard conditions the spore inliibition
is determined microscopically and expressed as percent. Using the
Wilcoxon and McCallan (45) simplification of the Bliss statistics,
the percentage inhibition is readily plotted against dosage on loga-
rithmic probability paper and the amount of material to inhibit
50 or 95 percent of the spores is read off directly.

Because of biologic variation and experimental error these values
vary from test to test. Spore concentration is such a variable, as it

206 Connecticut Experiinent Station Bulletin 456

cannot be regulated veiy easily. Variability can be reduced but
not eliminated by using a ratio of performance between the test
material and a standard such as bordeaux mixture to give a bordeaux
coefficient (19) as follows:

Dosage of standard bordeaux tor 95% inhibition
Bordeaux coefficient = Dosage of test material for 95% inhibition

A bordeaux coefficient of 1.00 means that the test material is
just as active as bordeaux. If below 1.00 the activity is less and,
if above 1.00, the activity is greater.

The fungicidal value so determined for a series of copper fun-
gicides is given in Table 3. The wide diiferences are interesting
and significant. Detailed studies have shown that some of these
differences can be exj)lained by differences in particle size. Yellow
cuprous oxide contains much smaller particles than red cuprous oxide
(10) and of course is more potent. The study of this relation showed
that the wave-length of reflected light was related to potenc}'. Particle
size decreased and potency' improved as the wave length shortened
and the color shifted from red to yellow.

It was then found that the basic copper compounds reacted simi-
larly. As the wave length shortened (green throuigh blue to violet),
potency increased (17). It is not known whether this phenomenon
is associated with particle size or not. It is true, however, that when
green Basicop was ballmilled wet for two weeks it became more blue
and the potency improved. Since this information was published the
manufacturers have taken the green grade off the market and have
substituted a blue grade.

It is noteworthy that the curves for cuprous oxides and cupric
salts cannot be superimijosed. For equal potenc}^ the copper as
cuprous oxide reflects a longer wave than the copper as a cuprio salt.
This is evidence on the hypothesis that cuprous copper is more potent
as a fungicide than cupric copper (i21).

It must not 1)e concluded that protective value is a function of
color of the copper material, because tenacity must be considered in
iicld jjerformance.

Tenacity. Fungicidal dej)()sits must not only have spore killing
jH-operties; they must maintain these properties in the face of drastic
washing. They must cling to plant surfaces while being buffeted by
wind and rain. Heuberger (5) has devised a laboratory test for
tenacity and a tenacity coefficient compai-able with bordeaux co-
eflicit^nt for converting the raw data to usable form. In order to
speed tlie work and to simulate the swaying action of rain-lashed
leaves, deposit-beai'ing slides are ])lace(l hack (o back and ])assed
rapidly through water for a standai'd 'iO strokes. 10 forward and 10
backward, the slides being raised from the water and sliaken after
each stroke.

The tenacity cocHicient is ihe perceritiige (if iniliiil toxic load
thill i> not \\a>he(| oil' hy the s(;iiid;ird lest. Tiie ruiii:ii> indicator

Defoliation on Tomatoes 207

measures not only the quantity removed, but also the spore inhibiting
properties of the deposit that is left.

The tenacity coefficients for the series of copper materials used
on tomatoes are given in Table 3, correct up to 1940.

It has been shown (7) that the tenacity test in the laboratory
gives results that are in essential agreement with field results with
several of the copper materials.

Protective Value. The enormous variability of field results is
impressive when comparing the protective values of fungicides
whether on tomato, apple or other foliage. The field results are
often so variable that a set of materials seldom arranges itself in the
same order from test to test. This source of error accounts for some
of the discrepancy between laboratory and field results. Methods
were needed for reducing this variability^, or of understanding it.
Considerable progress had been made in designin.g methods for re-
ducing just such variations in fungicidal value as determined in the
laboratory (19). These findings have been applied to problems of
measuring protective value in the field and they appear to be exactly
homologous as recently discussed (4).

It is apparent immediately that field comparisons have been
based on the control for equal dosages of materials, whereas labora-
tory comparisons have been based on dosage for equal response. It is
also apparent that no use has been made of the performance of a
standard fungicide in the field to give a figure comparable to bor-
deaux coefficient for fungicidal value data.

In 1940 (4) it was learned that dosage-response data for field
tests give straight lines when plotted on logarithmic-probability
paper. Trials were made of the effects of using dosage for equal
control as it is used in the laboratory. It soon developed that dosage
for equal control is more sensitive and more informative than control
for equal dosage (13). Control for equal dosaige has a low order of
sensitivity because the control scale is limited by a practical ceiling
of 100 percent, whereas, the dosage scale is unlimited.

The control scale is less informative than the dosage scale.
Moreover the use of the control scale is based on an assumption that
response in percent is linearly related to dosage — ^that unit change
in dosage produces unit change in response. The fact, however, is
that the relation is sigmoid. In everyday language this means that
change in dosage produces at first a small change in response, then a
large change, and finally a small change in response.

The fallacy in the use of the control scale appeared in a practical
way in the development of yellow cuprous oxide as a fungicide. The
data on fungicidal value obtained in 1938 (10) indicated that ap-
proximately twice as much copper as red oxide was required to in-
hibit the same number of spores as yellow oxide. That is to say the
copper in yellow oxide was twice as potent as that in red oxide.

Since the tenacity of the two forms is approximately equal (see
Table 3) the protective value should follow in the same order. This
was found to be true when the two materials were compared as protec-

208 Connecticut ExpeHinent Staticni Bulletin 456

tants for pea seed. Twice as much copper as red oxide as yellow
oxide, was required to give equal protection of pea seeds against
damping-off.

The two materials were then compared a5 sprat's on tomato
foliage to protect it from Alteimaria solani. Following general prac-
tice for field work the materials were compared at equal doses, not
at equal control as in the earlier tests. The control obtained was 71
and 53 percent, respectively. This ratio is 1.34 to 1, not 2 to 1 as
would have been expected. Was the discrepancy due to difference
in relative performance of the two materials or to difference in
technique of measurement?

This question was approached in 1941 when the two materials
were compared in a dosage series on the protection of muskmelon
foliage against bird's eye leaf spot caused b}' Macrosporium cucwner-
inum. The J&rst comparison is dosage for equal control. If the 80
percent control level is chosen, it appears that 15.5 pounds of copper
as red oxide is required, but only 7.5 pounds as yellow oxide. This is
a ratio of 2.06 to 1, as would have been expected from the data on
fungicidal value and data on pea seed protective value. It should be
stated that the ratio between the two remains 2 to 1 irrespective of
what level of control is chosen. This shows that the dosage scale
provides an invariable measure of performance.

The other comparison is control by the same dosage. The control
is 83 and 90, respectively, for red and yellow oxide for 20 pounds of
copper per acre for the season. This is a ratio of 1.09 to 1, not 2.06
to 1. Furthermore this ratio changes with the dosage level chosen.
At ten pounds per acre the ratio is 1.11 to 1, and at five pounds per
acre it is 1.19 to 1.

From this experiment it is clear that the dosage scale is more
sensitive because it spreads the materials farther apart than the con-
trol scale, and it is more informative and accurate because it gives re-
producible results. It is also more useful practically because it re-
duces the error of field experimentation.

Derivation of Protective Coefficient. The fact remains, how-
ever, that copper materials used as tomato sprays have been compared
up until quite recently through the control by equal dosages. Since
these data are all that are available they must be used for the present
in measuring the protective value of the materials in the field despite
the sources of errors and the mathematical inconsistencies in the
design of the experiments.

Tlie measurement of protective value in the field is beset witli all
the difficulties that occur in the laboratory, and more besides. In
addition to errors introduced by the fungus, the errors that come from
iiia(l(!(iuate s])rayers, soil hetei-ogeneity and inclhod of taking data
are inipdrtaiit. l^'iiially, thci'c arc variations introduccMl by the
weathei'.

Of these variables only the weather afl'ec-ts the action of the de-
posit after it is on the leaf. The other variables sinq^ly complicate
the measurements of the protective value of the deposit in the same

Defoliation on Tomatoes 209

way as they complicate the measurement of spore inhibiting power
of fungicidal deposits in the laboratory. It therefore seems probable
that the effects of these other variables can be reduced by calculating
a protective coefficient in terms of a standard fungicide, as in the case
of bordeaux coefficient for the laboratory (19). This calculation is
based here also on the assumption that all sources of error except
weather tend to operate on the test material and standard alike.

The only difference in procedure is that the calculations of field
data for the present must be based on the response scale rather than
on the dosage scale. The amount of disease control on plants sprayed
with the test material is divided by the amount of disease control
on plants sprayed with the standard material (4-4-50 bordeaux).
The quotient must serve for the present as the "protective coefficient,"
pending the accumulation of data on dosage for equal control. If
the quotient is greater than unity, the material has a better protective
value than bordeaux mixture ; if it is less than unity, the test material
is inferior to bordeaux.

There is experimental evidence to indicate that protective co-
efficient appears to cancel out variations in the methods of recording
the amount of disease (18). In 1938 four methods were used for
measuring disease on the same power-sprayed plots of tomatoes : per-
centage defoliation as counted, percentage of diseased fruits, index
of disease and the reciprocal of green weight per plant. The protec-
tive coefficients for red copper oxide obtained from these four kinds
of data were 0.77, 0.74, 0.81 and 0.81. For copper oxychloride the
coefficients were 0.51, 0.52, 0.58 and 0.41, respectively.

In the laboratory bordeaux coefficient reduces the effect of spore
load. Experimental evidence is available for field data likewise show
that protective coefficient reduces the variation due to inoculum
potential (i. e. disease producing power of the environment). It so
happens that red cuprous oxide has been compared for nine seasons
with bordeaux, but during the nine seasons the inoculum potential
has varied widely. When protective coefficient was plotted against
inoculum potential (expressed as percentage defoliation in the checks),
a scatter diagram was obtained showing that inoculum potential
bears no relation to protective coefficient, and that results in different
plots or in different years will not be influenced by variations in the
incidence of disease.

Another bit of data (Table 5) confirms this conclusion. Stem-
end rot counts were made on five picking dates in 1938. The per-
centage infection increased on the checks from 6.8 percent to 60.1
percent between August 18 and September 15 as the inoculum poten-
tial increased. Likewise the percentage of stem-end rot increased on
the plots sprayed with red copper oxide and bordeaux, but the pro-
tective coefficient remained approximately constant. At least the
variation in the protective coefficient bore no relation to the variation
in inoculum potential.

From these various studies it follows that test to test variation

210

Connecticut E:c])€rlment Station

BuUetin 456

^J

o

o

a.

U

,_,

n

QJ

j-'

c.

o

iu|

c p

'ii c
tu o

, :

u

0)

.^_

o

<D

i:

Q-UI

o,'^

(2u

i^« u~, 3 LC

C: CC CC CO

:>' ^i '^i •— ■

cc r>! r^ '^i

^1 Tf ^ CO

(LI 4/ v^

r- = " ."2 ."S rs ."2 •^ o

.:iT3^00Cl.i-ui-i-l-_g ^

■A t. c aa.o.a.ao.0. c- _ <li

rtOoOOOS33331-Oj=

tt K U U O U U O O O U H N U

O

H

;^,

u

!/;

O

<

U

^

r-^

ui

V.

Pi

■^ Jjj

u

Cii

£^

O

Cn

^-

O

^

o

,— 1

LO

OS

tj

O

oo

r^

c

^O

to

■^

c

X

rt

p

Q.

O

u

o

u

T)

O

lU

'A

P3

«

Defoliation on Tomatoes 211

in protective value of fungicides in the field can be reduced by using
(a) dosage for equal control, (b) protective coefficient.

The protective coefficients as obtained for a series of copper
materials durin,g the past several years are given in Table 4. The
protective coefficients for the copper fungicides show a wide range
among the materials, but it is difficult to see at first that any relation
exists between the performance as measured in the field as protective
value and the performance a& measured in the laboratory as bordeaux
coefficient.

Perhaps the most striking hiatus is that for red copper oxide.
This material stands low in the list of bordeaux coefficients, but it
stands high in the list of protective coefficients. Conversely, Com-
pound A stands high in the list of bordeaux coefficients and only
medium in the list of protective coefficients.

Although a complete explanation for these changes in rating is
not yet available, inspection of the tenacity coefficients offers con-
siderable help in explainiuig them. Eed copper oxide stands high in
tenacity and it stands high in protective value. Compound A stands
medium in tenacity and medium in protective value.

These interrelations of tenacity and fungicidal value show how
the trend of research on copper materials has been determined.
The fungicidal value of red cuprous oxide was raised by reducing
the particle size. Yellow cuprous oxide was the result, and it is now
much more widely used than red cuprous oxide.

Most of the so-called fixed copper materials are low in tenacity.
Research on them is directed toward artificial stickers, and soybean
flour, oils and resins are being investigated for these.

Timing. Apple scab control is the classical example of the neces-
sity for precision in timing applications of fungicides. The neces-
sity for early spring applications of fungicides for this disease has
been reflected in the thinking on other diseases. As a result the chief
emphasis on control of tomato defoliation has been early and mid-
season applications despite three important considerations: (a) spray
damage is more severe on young than on mature plants (20) ; (b)
the disease is a mid and late season disease and (c) as early as 1920
Martin (25) reported experiments indicating that delayed sprays
were essentially as effective as early sprays.

Timing experiments were made in five years when, fortunately,
there was enough disease to separate the effect of the various ap-
plications. The early tests in 1929 and 1932, made hy omitting pro-
gressively the late applications, gave j)reliminary indications (Figure
6 that sprays applied ahead of July 10 in western New York were
essentially valueless in disease control. The 1938 and 1939 tests, com-
paring all-season with late sprays, indicated that early August was
somewhat too late to make the first application.

On the basis of these four years' trials, it was obvious that the
critical period lay between July 10 and August 1.

A more elaborate timing test was designed in 1940 to test in Con-
necticut the schedules of various lengths; early, mid-season and late

212

Connecticut Experiment Station

Bulletin 456

Oh ^

u.

H

W

fNJ

o c

c^ c\ o o ir-< ~ o — . LO c: o o o

c if

■Jr

S3 \d vd OS f< o ^ -^ cvi "~^ o c^j o

O m

on

cou-jmo'OiO'— 1 T— (T^co

f^-j

o

fN

nj

^t^-*-^cooo-*r^u-i-^oo

Q)

C£

^■—^foONK^oKoJoKdod

Q

<

'^u-ivO^VCOvO'^'^vCiO-^ro

rr\

(N

tio

PS r>S rS 1^ ?N rS

3

<

VD

00

^^XXXX X

<

On

op

X'^XXX X

<

fN

<

XXXXX X

VO

CN

c
o

^

XXXXX X

3

ITJ

o.

Q.

<

o

"o

XXXXX X

o

Q

(N

>.

xxxy.x

_D

U^

^

XXXXX

on

rsi

0)

XXXXX

c

3

IN

0)

X XX XX

c

"^

•d

*•

c

1 •

u

,_ ,^, ro ^ lO O «^ CO 0\ O -^ '^1 O.

nj2

0)

H

5

X

Defoliation on Tomatoes

213

(8). Each treatment consisted of four replicate plots of ten plants each
of Scarlet Dawn tomatoes, sprayed at 250 pounds pressure with a
single nozzle, applying 4.8 pounds of yellow copper oxide in 200
gallons of water per acre.

Spraying began June 21 and continued weekly until August 23.
In one series of plots sprays in pairs were dropped from the end of
the season to study the effect of early sprays only. In a second series,
sprays in pairs were dropped from the beginning of the season to
study the effect of late sprays only. In a third series various plots
were given two applications one week apart in the middle of the
season.

Disease control data (Table 6 and Figure 7) clearly confirm
previous conclusions that the critical first application should be ap-
plied in mid-July. On the basis of the 1940 data at least, it seems
that applications should begin perhaps a week earlier in Connecticut
than in western New York.

86

-

1

1 1

I932G-
1929®-

1 1
o ^

P

i/

z

jf

y 74

a.

"

fi "

1

,

1

o
q:

^62
o

-

//

o

if

UJ

//

w

'/

<

i

UI

/

(O

/

5 50

\

^■

-

,/

38

-V

1 1

1 1

JUNE 10 24 JU4 16 28 AUG .9 21
DATE OF LAST SPRAY

Figure 6. Relation between date of
last spray and control of de-
foliation of tomatoes in 1929
and 1932.

Date when applications begun

30

JUNE20JU.2 14 26 AUG.7
SPRAYING DATES

19 31

Figure 7. Effect of timing sprays
of yellow copper oxide on
control of defoliation caused
by Alternaria in 1940.

Practical adoption of the theory of delayed spraying admits the
possibility that disease may obtain a start before work begins. It is
of interest, therefore, to investigate the effect that such a start has
on final disease control. This point was investigated in 1938, Two
fields infected with A. solani were chosen for spraying on August 3.
One showed approximately 5 percent, and the other 20 percent, de-
foliation (by number of leaves).

At the end of the season the unsprayed portion of the fields

214

Connecticut Experi/ment Station

Bullet in 456

showed 66.4 and 89.4 percent defoliation (by number), respectively.
When sprayed with bordeaux, thej' showed 30.1 and 55.0 percent de-
foliation, respectively. Plainly, the field that was severely diseased to
begin with lost more leaves than the slightly diseased field whether
sprayed or not. Evident!}^ bordeaux did not freeze the defoliation
at its initial level.

Coverage. If spraying is to be delayed until the last possible
moment when the fungus may be already established, it is plain
that the protective load of fungicide must be so applied as to cover
adequately all susceptible tissue, especially the old somewhat senes-
cent tissue at the base of the plant and inside the foliage crown.
Coverage of ground plants would seem to be more difficult than
coverage of staked plants.

There appear to be three variables in the application of fungi-
cides by spraying: (1) pressure, (2) nozzle aperture and (3) spray-
ing time. A study, incomplete as yet, is being made of the effect of
these variables on unstaked tomatoes (Scarlet Dawn).

In 1940 an initial attempt was made to improve coverage by
holding pressure constant and by varying the gallonage per acre of
spray fluid. The gallonage was increased hy increasing the nozzle
aperture and the spray time. It was expected that increasing the
nozzle aperture would increase the velocity of the spray stream at
the nozzle and that this would force the stream farther through the
crown of leaves toward the important inner and basal ones. The
plants were sprayed by directing a single nozzle to all parts of the
outer crown of leaves, occasionallj' pushing the nozzle inside. A
Myers wheelbarrow poAver sprayer provided the yellow copper oxide
at 250 pounds pressure. Four applications were made between July
24 and August 23. Disease readings were made on September 7
(Table 7).

Table 7.

Effect of Coverage by Yellow Cuprous Oxide Spray on Control of
Defoliation of Tomatoes Caused by Alternaria solaiii.

Nozzle

Spray time

Spray applied

Copper applied as metallic

Disease

Aperture
in.

output

control

gal./min.

sees. /plant

gals /acre

lbs / 100 gal.

lbs/acre

percent

5/64

0.73

13

375

4.0

15.00

54.8

5/64

0.73

13

375

2.0

7.50

48.2

5/64

0.73

13

375

1.0

.\7S

44.4

5/64

0.73

13

375

0.5

1.M8

34.6

4/64

0.5H

(>

150

-1.0

().00

2S.7

4/64

0.58

()

150

_'.o

3.00

22.0

4/64

0.5H

()

150

1.0

1.50

11.8

4/64

0.58

(>

150

0.5

./.I

11.8

3/64

0.43

3

(.0

4.0

_'.40

12.4

3/64

0.43

3

(,()

-Ml

1.20

13.4

3/64

0.43

3

(lO

1.0

O.dO

2.5

3/64

0.43

3

(jO

0.5

0.30

3.4

none

0.0

0.0

1.2

Defoliation on Tomatoes

215

Data were plotted (Figure 8) on log-probability paper using
dosaige as pounds of copper per acre. Alternaria attacked early and
heavily and spraying began somewhat late. On this account no
treatment gave very good control. In the first analysis of the data
the effect of the three gallonages was determined on the basis of
dosage for equal control (13). The level of control that fits all
three the best is 25 percent. This level of control was provided by

80

g60
o
cc
u

_r

o
q:

20

TT

1 — I I M MM

/

/

/

./ 5/64 nozzle o o

/• 4/64" nozzle @ @

3/64"nozzle • •

I Mini

I III

.6 1.0 3 6

DOSAGE.LBS. CU./ACRE.

10

Figure 8. Eflfect of coverage with
yellow copper oxide spray by
using various sizes of nozzles
on control of defoliation of
tomatoes by Alternaria solani.

1 1

\ llll

1 1 i 1 1 llll

80

-

60

^^

®

<

1 — '

13

^M

40

/

/

Stents /

y

y

/

20

-

^

^ / ® ^

b

—

J?

(-''■''^ /■

c

r^

^

1 0 sees® •®

/

Leaves

<\

\ III!

llll Mill

0.5 I 2 5 10

DOSAGE, LBS. CU./ACRE

Figure 9. Relation of spraying time
(seconds per plant) to cover-
age by yellow copper oxide
spray on the control of de-
foliation of tomatoes caused
by Alternaria solani.

0.48 pounds of copper per acre per application when applied in 375
gallons of water through the large aperture. The requisite dosage in-
creased by ten fold to 4.5 pounds when it was applied in 150 gallons
of water with the medium disc. The requisite dosage increased fur-
ther to 6.4 pounds when it was applied in only 60 gallons of water
with the smallest disc.

An unexpected result appeared in the data. The slope of the
dosage response curve became flatter as the coverage improved (Fig-
ures 8 and 9). Dimond (3) has shown that this slope is a linear func-
tion of coverage and he suggested that the slope of the curve offers a
convenient measure of coverage.

This experiment in 1940 was interesting and probably significant,
but it involved a confusion of the effects of nozzle aperture and spray
time. In 1941 a similar experiment was conducted except that the

216

Connecticut Experwient Station

BuUetin 456

nozzle aperture (3/64 inch) and pressure (250 pounds) were both held
constant. Spraying time was varied.

Alternaria attacked very heavily. Leaf disease readings were
made on September 2 as usual (Table 8), but readings were made on
stems as well, since these were heavily attacked also. In the case
of stems, the groupings were made on the proportion of area covered
by spots on the lower foot of stems.

Table S. Relatiox of Spraying Time to Coverage by Yellow Cuprous Oxide
Spray on the Control of Defoliation of Tomatoes Caused by Alternaria solaui.

Spraying time

Amount of spray

applied

gals. /acre

Amount copper expressed as
metallic

Disease control percent

sees. /plant

lbs./ 100 gals.

lbs /acre

Leaves

stems

20

400

4

16

16.7

85.2

20

400

9

8

22.0

77.0

20

400

1

4

6.7

65.5

20

400

0.5

2

4.2

55.7

10

200

4

8

19.7

79.0

10

200

2

4

19.2

73.0

10

200

1

2

9.5

52.7

10

200

0.5

1

4.5

45.5

5

100

4

4

14.2

67.7

5

100

2

-y

9.7

60.5

5

100

1

1

6.7

60.0

5

100

0.5

0.5

3.0

45.7

2.5

50

4

2

9.2

60.5

2.5

50

2

1

6.5

49.8

2.5

50

1

0.5

4.5

46.5

2.5

50

0.5

0.25

3.0

44.0

Data (Figure 9) w^ere plotted as usual on the basis of pounds
of copper per acre per application. To save cluttering the graph only
the first and third spray times are plotted (2.5 and 10 seconds per
plant). Data for the foliage and the stems are in excellent agree-
ment. As the spraying time per plant increases, the slope of the
dosage-response curve increases. If slope measures efficiency, as de-
duced from the. 1940 data, it follows here that the long spray times
were relatively less efficient than the short spray times per unit of
copper per acre.

This seems reasonable. Increasing the spray time increases the
run-off, and this means that nmch of the copper applied Avith a long
si)ray time rims off onto tlio ground wiiei'e it cannot ])rotect foliage.

On tlie basis of this discnissioii it follows that the improved
efficiency witli increasing gallonage noted in 1940 was due rather
to the u.se of larger nozzles than to increased gallonage directly. The
larger the nozzle aperture, the greater the nozzle velocity of a spray
s-tream, other things being efjual as they were here. The higher the
velocity of tlie spray stream, the more the outer leaves will be
l)nsli('<l aside, so tliat tlie inner ones may be covered.

Search of the litcralnrc lias i-evealcd little that is :ii)r(>i)()s here.
'J"hc circcts of pic.s.^nic :iii(| nn//|c ;i|t('it iiic li:i\t' Ix'i'n studied, but

Defoliation on Tomatoes 21Y

not in connection with the dosage-response technique. Farmers are
certainly trending toward larger nozzles with fewer gallons per acre,
and away from small nozzles and much gallonage per acre. Smith
and Zimmerley (34) constitute a typical case. They tested the effect
of pressure in spraying tomatoes in 1922. Their data are difficult to
evaluate in this connection because they report no disease readings
and because they have not distinguished dosage of copper per acre
from pressure.

Morris, Klotz and Sokoloff (29) published a paper late in 1941
giving results with bordeaux in the control of brown rot on citrus.
They applied only two concentrations and two gallonages but, when
their data are calculated as amount of copper per tree and plotted on
log-probability paper, two curves appear. The higher slope of the
curve for the gallonage is the flatter of the two as would be expected
if it had given improved coverage. These writers do not state, how-
ever, how they obtained the larger gallonage per tree.

PRACTICAL ASPECTS

Since Agricultural Experiment Stations are the research labor-
atories of farmers, the final objective of its research must be the prac-
tical application. The defoliation problem is intensely practical.
The question here is whether it can be helped by the present investi-
gation.

To Spray or Not to Spray

The first question is "to spray or not to spray." This question
like many others in science cannot be answered categorically. To
that end it is well to summarize here the factors that can help an
individual farmer to decide that for himself.

The farmer who sells in a quality market and who is troubled
with much defoliation will find a marked improvement in quality
for spraying. Although spraying will probably not greatly improve
the total tonnage, it will improve the tonnage of salable fruit. Spray-
ing will do much to guarantee rich red tomatoes that sell well in a
quality market. It will reduce the amount of stem-end rot, anthrac-
nose, cracking, blossom-end rot and flabbiness.

The farmer who sells extra early fruit cannot afford to spray,
because most of the fruit that he sells at a fancy price is sold before
disease becomes serious. Spraying will tend to reduce yield slightly,
and disease is seldom serious enough to offset this in the early crop.

On the contrary, the farmer who sells to the extra late market
often finds that defoliation is serious enough to reduce his yield of
marketable fruit and often forces him to reduce even his quality
standards in order to have any fruit at all available for sale. Judi-
cious spraying in late July and August will maintain sufficient foliage
in late September to produce quality fruit that is salable.

MateHals to Use. As yet no effective agent other than copper
is available for use on tomatoes, although organic compounds were

218 Connecticut Experinwx.t Station Bulletin 456

investiirated somewhat in 1940 and 1911. Sulfur seems to dwarf the
phints. Accordingly copper fungicides must receive the bid. bmce
lime is distincth' deleterious to tomatoes, bordeaux mixture can hard-
\\ be expected to be chosen, except as noted below.

This narrows the field to the so-called insoluble or "fixed" (41)
copjDer materials. These have been studied in considerable detail.
Some of these have shown themselves distinctly inferior to others in
disease controlling powers on tomatoes. These are Cupro K. Z-0,
Metrox, hj^drated cupric oxide and Cuprocide 54.

Others have shown themseh'es invariabl}^ toward the top of the
list as follows: Yellow copper oxide (Cuprocide), red copper oxide
(Cuprocide) and copper oxychloride (Compound A). In one year
when it was tested Tennessee Tribasic copper sulfate stood toward
the top of the list.

The other materials have not been tested sufficiently to arrange
them with too much certainty, but the only one that looks as if it
would find a place high in the list is Coposil, and it, unfortunatel}',
is too injurious to tomatoes.

In Connecticut Avhere flea beetles are a problem, it seems that a
rotenone or dry Pyrocide dust would be useful in keeping them down.
This would prevent the eating injuiT where the fungus spores may
gain entrance to the leaf.

When to Spray. It seems clear now that attention to timing
may save on materials and add to the value of the applications. In
the past, applications have gone on the plants earW in the season when
they were most injurious and least required. It seems better to apply
them later in the season Avhen the}' are least injurious and most re-
quired.

The generalization seems sound. Its application to specific cases
raises many knotty problems. Timing tests were made in New York
in 1929, 1932, 1938 and 1939 in plants set out the third week of May.
The tests in 1929 and 1932 showed that sprays ahead of July 10 were
worthless. In 1938 and 1939 an arbitrary date of August 1 was set,
but in both years it was somewhat too late. It would appear that
the critical date was betAveen July 10 and August 1.

A timing test was made at New Haven in 1940 using plants set
out about May 23. The results indicated that Jul}'' 10 was early
enough for the first application. If a similar test had been made in
1941 this date would have been close to optimum.

From these various timijig tests in two areas with similar grow-
ing seasons, it would seem that the first spray should be a[)plied
about July 10 I'or inaxiniuin disease controh 'I'liis (hite is basetl on
work with crops plantctl into the field about May 20. In IDll the
crop was set about May 10 and spraying had to begin alioul July 1.

This suggests that timing siiould be based on (he stage of the
I)laiit rathei- than calcnchir' (hitc as in the case of apph' scih. Aclually,
it seems from experience lliat the sprays should be iii)i)li(Ml just ahead
of tlie "brcalv" sja.trc, i. c, tiic stage when the weight ol the i)lants
begins to hrciik- thcui o\«-r. so th;it the inner leaves begin to l)e shaded

Defoliation on Tomatoes 219

and protected from drying out. An application at this stage is easy
to apply because the lower leaves are still exposed, and the plants
are not spread out over the ground between the rows.

The number of delayed applications is also important. Here
again the final answer is not in, but it seems probable that three at
ten-day intervals is enough. In some years one good application
would be enough, because growth slows down after Aujgust. If the
spray has good tenacity, the susceptible foliage would remain covered
long enough.

How to Apply Materials. The methods of applying fungicides
raise many unsolved problems. Vegetable growers prefer dusting,
although spraying has given the best disease control so far in this
research. Since the program calls for delay until the last nio-
ment, the fungus may get such a start as to make it imperative to
use the best possible procedures. Spraying is, therefore, preferred.
Additional research is now under way to improve dust mixtures and
methods of application.

Whatever the machine used, the problem remains of getting
through the fields after the plants have filled the rows. Farmers
in cannery areas report some success with airplane applications.
Fields are too small for this in Connecticut.

Other growers lay out the roads for picking earlier than usual,
throwing the vines together. Sprayers with long booms are driven
through these roads. Other growers make the roadways farther
apart and carry a very long boom by hand. If the number of ap-
plications can be trimmed to one or two this might be a feasible pro-
cedure.

From two years' results it seems that the problem of covering
the inner lower foliage of ground plants is critical, and it now seems
clear that large holes in the nozzles giving a strong drive to the
spray stream should give better success than small holes giving a
misty spray. On the basis of present information it is suggested to
at least three pounds of copper (as metallic) should be applied per
acre per application in a minimum of 200 gallons of water with
5/64 inch discs.

Varieties

As far as can be determined, no tomato variety shows any marked
resistance to the Alternaria defoliation. In a variety trial, the entries
show large differences in defoliation, but careful study shows that
these differences are associated with fruit load. Early varieties set
fruit early and become defoliated early. Late varieties set fruit late
and become defoliated late.

Fertilizers

Information on the relation of fertilizers to Alternaria defolia-
tion is yet insufficient to make definite statements, but evidence now
available points to an influence of nitrogen. Low levels of nitrogen
nutrition encoura/ge disease. Increasing the nitrogen nutrition is liable

220 Connecticut Experiment Station Bulletin 456

to reduce fruit set and. of course, to reduce total yield. If disease at-
tacks, however, a high level of nitrogen nutrition might permit the
field to pull through a marketable crop that might otherwise show
such poor quality from disease as to be almost unpickable.

Air Drainage

Other things being equal, tomato fields on slopes, especially
southern and western slopes, probably have less disease than those
without as good air drainage. Staking, of course, improves air drain-
age and reduces severity of the defoliation disease.

Miscellaneous Suggestions

Since the disease is seed-borne, the seed should be from certified
sources and it should be soaked in New Improved Ceresan 1-1000,
dried and dusted with red copper oxide. To prevent development
of disease in the seed bed, the seedlings should be sprayed at weekly
or t«n-day intervals with the material to be used in the field.

The fungus also lives over winter in field refuse. Accordingly,
a rotation of at least two years will keep down this source of inoculum.
Finally, it may be spread from plant to plant if plants are picked or
cultivated when they are wet.

SUMMARY

1. A study has been made during 12 seasons of the foliage and fruit
diseases of tomatoes with the objective of exploring the whole
field of defoliation diseases of tomatoes. Particular emphasis has
been devoted to solving the paradox of reduced jaelds from sprays
despite disease control. This paper reports data on the effects
of disease on the plant and the interacting effects of sprays and
disease on yields.

2. The problem has been attacked by studying plants in various
stages of disease and by studj^ing the varying control obtain-
ed by different sprays.

3. The primary cause of defoliation in the Northeast is Altet^nm'ia
solani, but since this fungus is not what may be called a vigorous
parasite, optimum conditions must prevail before attack sets in.

4. Optimum conditions for the disease include: (a) crowded plants,
(b) maturity of leaves, (c) heavy fruit load, (d) above normal
rainfall and dew and (e) sliading. Disease, of cout'so. may a]>-
pear when one or more of these conditions ai-c doI fnllilltMl. but
they all seem to l)la3^ a part.

5. A .special study was made of the ii'hition of IViiit jimii and age
of tissues to .susceptibility iind it upjx'ais that any factor such as
pruning, low nitrogen nntiition (»r lieavy rci)rodu('ti()n tends to
increase susceptibility. This is ('si)i'cially striking in tlie case of
fiMiit load. 'I'he ]f)ng('i- the plant I'cinains fi-ee of fruit, the
long(;r it remains fn-c of Ahci-nai-ia : ami llw more fruit it sets,
the more susceptible it Itcionics.

0. An extensive study of the cUci'is of disease and ^j)raying on

Defoliation on Tomatoes 221

ripenin>g has been made. No evidence can be found that indicates
any effect on the maturity of fruits, i. e. ripening. Many factors
such as disease, dwarfing and defloration from sprays reduce
fruit load. These factors affect, of course, the number of fruits
picked and thus they affect the shape of the picking curve.

7. In studying yields the problem arises of what constitutes ripen-
ing. Ripening is defined as reddening. Accordingly, many fruits
that have been picked ag ripe on defoliated plants were not ripe
because they were orange in color and never would have become
red.

8. Since this point, was not clarified until after the completion of
the current research, many fruits have been picked on defoliated
plants as ripe when they were not ripe in the same sense as those
on plants not defoliated. As a result picking data have tended
to favor defoliated plants unduly.

9. In studying the disease-controlling properties of fungicides, a
protective coefficient has been devised for reducing the variance
between tests that is due to inoculum potential, spraying techni-
que, method of recording disease and kind of disease. Although
this statistic has some weaknesses, it serves the useful purpose
of eliminating the effect of many confusing variables. It is the
quotient obtained by dividing the amount of disease on plants
sprayed with a standard by that on the test material. It is based
on the assumption that as extraneous factors affect the unknown
they also affect the standard.

10. One or more tests have been made of copper-containing bordeaux
substitutes. Insufficient data are available to rate them all with
precision, but three groups seem possible : good, intermediate and
poor. Those in the "good" group appear to be yellow copper oxide
(Cuprocide Y), bordeaux, red copper oxide (Cuprocide G), Com-
pound A, Coposil and Tenn. 34. Those in the "intermediate"
group are Basicop, Hydro 40 and Cuprocide 54. Those in the
"poor" group are Metrox, ZO, Cupro K and hydrated cupric
oxide.

11. Timing of tomato sprays is of critical importance in economical
control of defoliation. Since no spray is completely non-injur-
ious, and since injury is most pronounced on small plants, the
applications should be delayed as long as possible. On the other
hand, the longer the sprays are delayed, the less effective they
can be in stopping an outbreak. Consideration of all the data
suggests that the first application in Connecticut should be ap-
plied just as the plants break over.

12. Coverage becomes an important factor in spraying ground toma-
toes because the lower and inner leaves are the most susceptible
of any to defoliating fungi. Although evidence is somewhat lim-
ited, it appears now that insufficient attention has been paid
to size of nozzle orifice. This should be as large as possible so
that the spray stream will be hard enough to push aside the
outer crown of leaves.

2"2*2 Connecticut EayperiTnent Station Bulletin 456

LITERATURE CITED

1. Boyle, J. G. Tomato investigations. Indiana Agr. Exp. Sta. Bui. 165. 1913.

2. Brown, H. D. Loss caused bv Septoria leaf spot in the tomato canning crop

of Indiana— 1928. Plant Disease Rep. 13:164-165. 1929.

3. DiMOND, A. E. Measuring inoculum potential and coverage index of spravs.

Phytopath. 31:7. 1941.

4. , HoRSFALL, J. G., Heuberger, J. W., and Stoddard, E. M. Role

of the dosage-response curve in the evaluation of fungicides. Connecticut
Agr. Exp. Sta. Bui. 451. 1941.

5. DuGGAK, B. M. Lycopersicin, the red pigment of the tomato, and the effect

of conditions upon its development. Washington L'niv. Studies 1 :22-45,
1913.

6. Edgerton, C. W. Delayed ripening of tomatoes caused by spraying with

bordeaux mixture. Louisiana Agr. Exp. Sta. Bui. 164: 1918.

7. Heuberger, J. W. A laboratory biological assay of tenacitv of fungicides.

Phytopath. 30:840-847. 1940.

8. Timing and dosage of tomato spravs for Altcrnaria control.

Phytopath. 31:11. 1941.

9. and Dimond, A. E. Relation of flea beetle control to control

of Alternaria solani on tomatoes. Plant Disease Rep. 25:415-418. 1941.

10. and HoRSFALL, J. G. Relation of particle size and color to fun-
gicidal and protective value of cuprous oxide. Phytopath. 29:303-321. 1939.

11. and Maintaining quality of tomatoes by delayed

spraying. Phytopath. 30 : 9. 1940.

12. HoRSFALL, J. G. Meadow crop diseases in New York. Cornell LTniv. Agr.

Exp. Sta. Mem. 130. 1929.

13. and Dimoxd, A. E. Comparing fungicides at dosage for equal

control. Phytopath. 32:10. 1942.

14. and HARRisoiN, A. L. Effect of bordeaux mixture and its var-
ious elements on transpiration. Jour. Agr. Res. 58 :423-443. 1939.

15. . Hervey, G. E. R., and Suit, R. F. Dwarfnig of cucurbits

sprayed with bordeaux mixture. Jour. Agr. Res. 58:911-928. 1939.

16. and Heuberger, J. W. Delayed spraving of tomatoes. Phyto-
path. 29:11. 1939.

17. and Relation of color to fungicidal value of in-
soluble copper compounds. Phytopath. 30:11. 1940.

18. and . Measuring incidence of a defoliation disease

of tomatoes. Phytopath. 32: 1942. (In press)

19. , Heuberger, J. W., Sharvelle, E. G., and Hamii.ton. J. M.

A design for laboratory assay of fungicides. Phytopath. 30:545-563. 1940.

20. , Magie, R. O., and Suit, R. F. Bordeaux injury to tomatoes

and its effect on ripening. New York State Agr. Exp. Sta. Tech. Bui. 251.
1938.

21. , Marsh, R. W., and Martin, H. Studies upon the copper fun-
gicides. IV. The fungicidal value of the copper oxides. Ann. Appl. Biol.
24:867-882. 1937.

22. and Suit, R. F. Tiie lime factor in lionlr.iux injnrv. Phyto-
path. 28:9. 1938.

23. Kertesz, Z. I., Tbi.MAN, T. G., Leconti, J. D., and Ruyde, E. H. The ap-

plication of calcium in the commercial canning of wliolc tomatoes. New
York State Agr. ]«:xp. Sta. Tech. Ihil. 252. 1940.

24. Li-OYD, J. W. and Brooks, I. S. (irowing Ifmiatocs for oarlv market. Illinois

Agr. Exp. Sta. Bui. 144. 1910.

25. Martin, W. H. Studies on tomato kaf-spot cdiUrol. Ww Irrsev Agr. Exp.

Sta. Bui. 345. 1920.

26. MacGilmvkay, J. II. The variation in temporalure of tomatoes and their

color development. Anur. Soc. Ilort. Sci. 32:529-531. 1935.

27. McKinney, H. H. Intliu'nee of soil moisture and soil temperature on infec-

tion of wiu-at seedlings l)v I friiiiiiilhosjuiriiiiii .ui lint in. lour. Agr. Res. 26:
195-217. 1923.

Defoliation on Tomatoes 223

28. MooRE^ W. D. Results of tomato disease investigations in Georgia, 1937-1938.

Canning Age 21:124. 1940.

29. Morris, H. I., Klotz, L. J., and Sokoloff, V. P. Brown rot control and copper

injury. Calif. Citrogr. Aug., 1941.

30. Pritchard, F. J. and Clark, W. B. Effect of spraying on early ripening of

tomato fruit. Phytopath. 9:289-291. 1919.

31. and Porte, W. E. Use of copper soap dust as a fungicide.

Phytopath. 11:229-235. 1921.

32. Rosa, J. T. Ripening and storage of tomatoes. Proc. Amer. Soc. Hort. Sci.

23:233-240. 1926.
Z2). Samson, R. W. Defoliation of deflorated, late-set and variously fertilized
tomato plants by Septoria lycopersici. Purdue Univ. Agr. Exp. Sta. Rept.
. 53 :41-42. 1940.

34. Smith, L. B. and Zimmerley, H. H. Relation of pressure to effectiveness in
spraying tomatoes. Virginia Truck Exp. Sta. Bui. 33 and 34. 1921.

35. Smith, Ora. Effects of light on carotinoid formation in tomato fruits. Cor-

nell Univ. Agr. Exp. Sta. Mem. 187. 1936.
2)6. Steier, H. C. Effect of Date of Planting on Tomato Yields and Quality.

The Canner. pp. 14-16. Mar. 26, 1938.
2)7. Thomas, H. R. Collar-rot infection on direct-seeded tomatoes. Plant Disease

Rep. 24:8-10. 1940.

38. . Nitrogen nutrition of tomato seedlings affects susceptibility

to collar rot. Purdue Univ. Agr. Exp. Sta. Rept. 52:56. 1939.

39. Watts, V. M. Growth and fruiting responses toi pruning and defloration of

tomato plants. Arkansas Agr. Exp. Sta. Bui. 347. 1937.

40. "Whipple, O. C. and Walker^ J. C. Tomato sprays help under some condi-

tions. Wisconsin Agr. Exp. Sta. Bui. 449:35-36. 1940.

41. Wilson, J. D. Certain injurious effects of spraying vegetables with the fixed

coppers. Ohio Agr. Exp. Sta. Bi-mo. Bui. 25:36-43. 1940.

42. . Spraying versus dusting of canning tomatoes with earh^ and

delayed applications. Ohio Agr. Exp. Sta. Bi-mo. Bui. 25:76-84. 1940.

43. '■ and Runnels, H. A. Five years of tomato spraving. Ohio

Agr. Exp. Sta. Bi-mo. Bui. 22:13-18. 1937.

44. and . Some detrimental effects of spraying tomatoes

with bordeaux mixture. Ohio Agr. Exp. Sta. Bi-mo. Bui. 18:4-15. 1933.

45. WiLCOxON, F. and McCallan, S. E. A. Theoretical principles underlying

laboratory toxicity tests of fungicides. Contr. Boyce Thompson Institute
10:329-338. 1939.

University of
Connecticut

Libraries

39153029045137