Front Cover
 Title Page
 Table of Contents
 The biology and ecology of...
 A model predicting incidence of...
 Literature cited

Group Title: Bulletin University of Florida. Agricultural Experiment Station
Title: The biology and ecology of Nomuraea rileyi and a program for predicting its incidence on Anticarsia gemmatalis in soybean
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00026779/00001
 Material Information
Title: The biology and ecology of Nomuraea rileyi and a program for predicting its incidence on Anticarsia gemmatalis in soybean
Series Title: Bulletin University of Florida. Agricultural Experiment Station
Physical Description: v, 48 p. : ill. ; 23 cm.
Language: English
Creator: Kish, Leslie Paul, 1944-
Allen, George E
Publisher: Agricultural Experiment Stations, Institute of Food and Agricultural Sciences, University of Florida
Place of Publication: Gainesville
Publication Date: 1978
Copyright Date: 1978
Subject: Velvet-bean caterpillar -- Parasites -- Florida   ( lcsh )
Soybean -- Diseases and pests -- Florida   ( lcsh )
Fungi imperfecti   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
non-fiction   ( marcgt )
Bibliography: Includes bibliographical references (p. 47-48).
Statement of Responsibility: Leslie P. Kish and George E. Allen.
General Note: Cover title.
 Record Information
Bibliographic ID: UF00026779
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: ltuf - AHM1149
oclc - 04606150
alephbibnum - 001597019

Table of Contents
    Front Cover
        Front Cover
    Title Page
        Page i
        Page ii
        Page iii
    Table of Contents
        Page iv
        Page v
        Page vi
        Page 1
    The biology and ecology of nomuraea
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
    A model predicting incidence of nomuraea on velvetbean caterpillar in soybean
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
    Literature cited
        Page 47
        Page 48
Full Text
Bulletin 795 (Technical) March 1978

The Biology and Ecology of Nomuraea rileyi and a
Program for Predicting its Incidence on
Anticarsia gemmatalis in Soybean


Agricultural Experiment Stations
Institute of Food and Agricultural Sciences
University of Florida, Gainesville
F. A. Wood, Dean for Research

Leslie P. Kish and George E. Allen

Department of Entomology and Nematology
Florida Agricultural Experiment Stations
Institute of Food and Agricultural Sciences
University of Florida
Gainesville, Florida 32611

This public document was promulgated at a cost of $2820.77
or a cost of 94# per copy to present information on the
fungus Nomuraea rileyi, which serves as a biological con-
trol of the velvetbean caterpillar, the major insect pest on
soybean in Florida.

1. This bulletin is based in part upon a dissertation submitted by
the senior author to the faculty of the Department of Botany,
University of Florida, in partial fulfillment of the requirements
for the degree Doctor of Philosophy.
2. This publication was supported in whole or in part by the Na-
tional Science Foundation and the Environmental Protection
Agency, through a grant, (NSFGB-34718, later known as BMS
75-04223), to the University of California. The findings, opinions,
and recommendations expressed herein are those of the authors
and not necessarily those of the University of California, the
National Science Foundation, or the Environmental Protection
3. Mention of a pesticide or a proprietary product does not consti-
tute recommendation or endorsement by the University of



Basic biological and ecological aspects of the entomogenous
pathogen Nomuraea rileyi, (Fungi Imperfecti) were investigated.
Cultural isolates from cadavers of the velvetbean caterpillar
(Anticarsia gemmatalis) grew and sporulated on Sabouraud
Maltose Agar in both light and dark regimens. Optimum growth
and sporulation occurred between 150 C and 250 C, and between
80% and 100% relative humidity.
The time from exposure to the pathogen to death of the host
averaged 6 days. Infection occurred through the integument and
also by ingestion of conidia through feeding. The fungus rami-
fied throughout the host body via the hemocoel. Conidiogenesis
occurred in the 48 hours following death at relative humidities
above 70%. The number of conidia produced was proportionate
to host size at death.
An intricate and dynamic balance exists between Nomuraea
and the biotic community. Therefore a wide range of stimuli-
response relationships were studied separately and collectively.
Environmental data, as well as laboratory and field data con-
cerning the host and pathogen, were utilized in the development
of a pathogen model.
A predictive equation was formulated enabling investigators
to forecast fungal infection levels among velvetbean caterpillar
populations under any given set of conditions. Predicted infec-
tion levels were compared to those actually observed. Data were
analyzed on an overall, seasonal, and per plot basis.
Results of Chi square (X2) testing for each trial in 1975
indicate that predicted infection values were significant at the
20% level for 33 of 43 trials and at the 5% level for 24 of the
43 trials. Predicted values for 1976 were significant at the 29%
level for 19 of 24 trials and at the 5% level for 12 of the 24
Results of the coefficient of correlation analysis indicate a
high correlation between predicted and observed infection levels
both for 1975 (.76 for all trials) and 1976 (.82 for all trials).
The pathogen model will continue to be refined within its
present application; its apparent strengths and weaknesses as
well as expected utilization are discussed.


SUMMARY ............................................ iii

INTRODUCTION ........................................ 1

Growth and Development ............................... 2
General ........................................... 2
Effect of Light and Dark .............................. 3
Effect of Temperature ................................ 3
Conidial Germination ................................ 3
Infection and Humidity ............................... 4
Sporulation and Humidity ............................ 5
Infection and Temperature ............................ 5
Source of Contact .................................. 6
Mode of Infection .................................. 6
Ontogeny of the Pathogen ............................ 7
Conidial Production per Cadaver ....................... 8
Relative Density of Conidia in the Agroecosystem .......... 8
Epizootiology ........................................ 9
The Host Crop ..................................... 9
Sampling Pests and Monitoring Environment .............. 11
Life Cycle Relationships .............................. 12
Environment and Phases of the Fungal Life Cycle .......... 14
Conidiophore formation and sporulation ............... 14
Conidial density and dispersal ........................14
Irrigation as a rain simulator ......................... 16
Summary of Effects of Environment ..................... .17

Introduction .......................................... 18
Arithmetical Relationships Among Factors .................. 19


Inoculum Density ................................... 19
The Effect of Weather Conditions on Inoculum Density ...... 21
Relative Humidity ................................ 21
Rain ........................................... 21
Wind ........................................... 22
Holdfast Factor .................................... 24
Catch Factor ....................................... 24
Ultra Violet Factor .................................. 25
The Relationship Between Inoculum Density and
Infection Levels ................................... 25
Adjustments ......................................... 27
Relative Humidity ................................... 27
Wind ............................................. 27
Rain ............................................. 29
Non-susceptible Larvae .............................. 29
Trials of the Equation with Actual Field Data ................ 30
Statistical Analysis of Results ........................... 31
Chi Square ........................................ 31
Coefficient of Correlation ............................. 32

DISCUSSION .......................................... 33

ACKNOWLEDGEMENTS .................................. 46

LITERATURE CITED .................................... 47






The occurrence of entomogenous fungi has been noted many
times over the years. Since Metchnikoff infected the larvae of
Anisoplia austraca Hbst. with spores of Metarhizium ani-
sopliae (Metch.) Sor. in 1879, many attempts have been made
to control insect pests utilizing fungi. The potential of fungi for
biological control of insects has been actively investigated, but
practical implementation has met with only limited success. An
integrated approach to pest control, incorporating reproducible,
predictive control capabilities is yet to be formulated.
In recent years, the damaging effects of agricultural chemi-
cals have become apparent. As a result, stringent control or out-
right banning of some of our most effective control agents has
occurred. Obviously, pest control programs which are effective,
economical, and ecologically sound must be developed. The need
for increased food production for a continuously rising world
population dictates that biological control be developed to its
fullest potential, particularly for basic food crops.
A crop currently increasing in importance is soybean, Gly-
cine max (L.) Merr., which constitutes a basic source of protein
for man both directly as food and indirectly through its use as
a supplement for domestic animals. The increasing dependence
on this crop has brought recognition of the seriousness of the
developing insect control problems on soybean.
In the study reported here, an intensive effort was mounted
to develop the research information needed to design a pest man-
agement program that would curtail the use of insecticides.
Early in the research, the velvetbean caterpillar (VBC), Anti-
carsia gemmatalis Hiibner, was identified as the major insect
pest on soybean in Florida.
Fortunately, this pest has its own enemies. One of the most
significant is the entomogenous fungus Nomuraea rileyi (Far-
low) Samson, which under certain conditions in Florida can
decimate the caterpillar population, with mortality levels ap-
proaching 100% (Allen et al., 1971).
The fungus known today as Nomuraea rileyi was first de-
scribed as Nomuraea prasina by Maublanc (1903). Sawada
(1919) transferred Nomuraea prasina to the genus Spicaria. A
long series of misinterpretations concerning the concept of the
genus Spicaria, originating with Harz (1871) and given validity
by Saccardo (1886), culminated in the transfer of Botrytis rilcyi
to Spicaria rileyi by Charles (1936). After 1936, Spicaria existed


as an ambiguous taxon based on at least two generic concepts
until Kish et al. (1974) transferred Spicaria rileyi to Nomuraea
and placed Spicaria prasina into synonomy.
Natural control of pest populations by N. rileyi is not effec-
tive from a biological control standpoint for two reasons: (1)
man's lack of control over natural conditions which may or may
not be favorable for development of the disease, and (2) a mat-
ter of timing which allows for severe defoliation during a 2 to
3 week period between the peak of the pest population and the
occurrence of the fungus in epizootic proportions. The con-
sistency of occurrence and the high level of population control,
however, make Nomuraea a candidate to become one of the first
fungi incorporated into a successful integrated control program.
A logical approach would be to start with the basic pest pop-
ulation model and apply the numerical factors which increase
and decrease the population. Given a known, or preferably, a
predicted pest population, judgments could be made relative to
necessary control measures.
It is readily apparent that in order to do this, each factor
affecting the pest population must be modeled from predeter-
mined, quantitative relationships. The complexity of direct and
indirect influences of the physical environment on each organism
involved, through each phase of its life cycle, makes the prob-
lem difficult.
Under the sponsorship of the Integrated Pest Management
Program, efforts were initiated in Florida in 1971 to develop an
integrated program to manage VBC populations in soybean. As
each factor affecting VBC population was identified, work was
initiated to examine it for input into a population model. In
order to evaluate the potential of naturally occurring epizootics
of N. rileyi, it was necessary to determine the mechanism in-
volved in the outbreaks. It is the purpose of this paper to report
progress toward development of a pathogen simulator sub-model
for the effect of Nomuraea on VBC populations in soybean and
project its role in a total VBC pest management program.

Growth and Development
Living cultures of N. rileyi were inoculated on disposable
Petri plates and slants containing Sabouraud Maltose Agar with
1% yeast extract (SBY). The fungus was routinely maintained
in culture on this medium for most laboratory experimentation.


Effect of Light and Dark
Culture plates containing SBY medium were inoculated with
conidia of N. rileyi. Five plates were inoculated and grown in
constant light; five plates were covered with aluminum foil and
grown in constant darkness; five plates were grown under a
mixed light regimen of 12 hours of light and 12 hours of dark-
ness. All cultures were held at 250 C.
All cultures grew and sporulated. It is therefore assumed
that photoperiod is not a limiting factor in the life cycle of the
Effect of Temperature
Plates of SBY were inoculated with conidia of N. rileyi, and
five replications were placed under each of the following con-
stant temperatures: 50 C, 100 C, 300 C, and 340 C. Inoculated
plates were grown in a 12-hour light regimen. Additional plates
were covered with aluminum foil and held in constant darkness
at 300 C and 340 C. Cultures grown in light were observed
periodically for several weeks; cultures grown in darkness were
observed only after termination of the experiment (3 weeks).
Growth and sporulation are affected by both high and low
temperatures. Nomuraea grew and sporulated slowly between
5 C and 150 C. Cultures grown at 300 C and 340 C remained
yeast-like and turned red. Conidiophores and conidia developed
within a few days following the removal of cultures from ele-
vated temperatures. Growth and sporulation appear consistent
and normal between 150 C and 300 C; beyond these limits sporu-
lation does not occur. However, since temperatures during the
soybean season rarely exceed these limits for any length of time,
temperature should.not limit growth and sporulation in Florida.
Conidial Germination
Eight Riddell mounts (Riddell, 1950) were placed in moist
chambers. Each contained one agar block, approximately one
square centimeter, of the following media: water agar (WA),
water agar with neopeptone (WAN), water agar with maltose
(WANM), water agar with maltose and yeast (WAMY), and
Sabouraud Maltose agar with yeast (SBY). Each was inoculated
with conidia of Nomuraea and placed on a laboratory counter
under ambient conditions of light and temperature. Mounts were
examined hourly for conidial germination during the first 6
hours and daily thereafter for 9 days. Conidia were placed on
two additional SBY Riddell mounts in a moist chamber, covered


with aluminum foil, and placed overnight in a closed cabinet.
The following day, mounts were examined for conidial germi-
Germination took place in less than 4 hours in darkness and
in light. Conidia did not germinate on WA. Maltose alone with
water agar and neopeptone alone with water agar would not
support growth beyond an occasional germination.
Germination levels approached 100% on SBY and WAMY
over the period of observation. Since conidia did not germinate
on WA (or in distilled water), and the fungus has never been
observed to grow saprophytically except on mycological agar,
we assume that the lepidopterous larva is absolutely necessary
for its growth and sporulation. Larvae and cadavers are the
only substrate involved. Additionally, it is assumed germination
levels of fresh conidia approach 100%.
Infection and Humidity
The effect of humidity on infection was investigated with the
aid of an environmental chamber providing control of tempera-
ture, photoperiod and relative humidity (R.H.).
Fifteen cups of rearing medium, each containing three larvae,
were dusted with an undetermined but roughly equivalent
amount of conidia at 0500, 0800, 1300, 1530 and 2000 hours, re-
spectively. After dusting, each cup of larvae was placed in an
environmental chamber, under a photoperiod corresponding
closely to day length during July in Florida, a daytime tempera-
ture of 320 C, and a nighttime temperature of 210 C. Humidity
was varied for each trial, which consisted of 225 larvae dusted
at the times previously cited. Trials were run at 90%, 70%, and
50% R.H. An undusted control group of 40 larvae was also
placed in the chamber at each corresponding variation of hu-
midity. Cumulative death counts were recorded for a period of
10 days after dusting. The number of days from dusting to
death of the first larva was also recorded.
The average infection at the 90% and 70% R.H. levels was
68%. No infection occurred at 50% R.H. Although some vari-
ation in infection levels among groups dusted at various times
did occur, no diurnal pattern was established. At 90% and 70%
R.H., the first larvae died 129 hours after dusting (5.4 days).
No deaths due to Nomuraea were recorded at any humidity level
in the control groups.
Results of this experiment indicate that the lower limit for
R.H. as a factor of infection is between 70% and 50%. It is
therefore assumed that infection does not occur below this level.


Since humidity levels in the trials remained constant and no
diurnal infection pattern was established, it is assumed infection
occurs uniformly above 50% R.H., regardless of diurnal con-

Sporulation and Humidity
Dead larvae from the "Infection and Humidity" experiment
were held under the regimen of R.H. at which they were in-
fected in order to determine the effect of R.H. on sporulation.
Cadavers were incubated at 90%, 70%, and 50% R.H. for 3 to
4 days after death. During this time, the appearance of conidia
imparted a green color to the cadavers. Sporulation was con-
sidered to be optimum when conidia were produced in such large
numbers that they settled around the cadaver, giving the appear-
ance of a ring of green dust. Production of fewer conidia was
termed partial sporulation.
No conidial production took place below 70% R.H., only
partial sporulation occurred at 70%, and total sporulation oc-
curred at 90% R.H.

Infection and Temperature
The effect of below average temperatures on infection was
investigated in the laboratory utilizing an environmental cham-
ber. The photoperiod was held constant on a diurnal sequence
of 16 hours of light. Relative humidity was held above 90%.
Twenty cups of rearing medium, each containing three
larvae, were dusted with conidia taken directly from a cadaver.
The number of conidia was undetermined, but roughly equiva-
lent amounts were dusted in each cup. The cups were capped
and placed in the chamber at 100 C. A control group of 46
larvae was prepared in the same manner, dusted with conidia,
but maintained at 250 C. Death due to the fungus was recorded
on a daily basis for both groups for 10 days. The dusted larvae
which had been subjected to the lower temperatures were re-
moved from the chamber and maintained under ambient con-
ditions for 10 additional days; death due to the fungus was re-
corded for the same group of larvae.
No infection of larvae dusted and maintained at 100 C was
noted after 10 days. A control group, held under ambient tem-
peratures, suffered 56% mortality after the same time period.
When the test group of larvae was removed from the lower tem-
perature, 88% of them died of Nomuraea infection within the
following 10 days. The control group suffered an additional 2%


mortality on the eleventh day, after which time all control
larvae pupated without additional mortality.
These data show that infection does not occur at or below
100 C; but temperature should not be a factor in Florida.

Source of Contact
A sample of laboratory-reared larvae were fed leaves col-
lected randomly from a field that had undergone an epizootic
of N. rileyi the previous year, in order to determine if initial
infection could result from conidia borne on leaves. Leaves were
collected twice weekly in the field, transported to the laboratory,
and placed in rearing cups to serve as the only diet for larvae
which initially fed on media. Leaf feeding began in the second
week of July, 1975, a few days before any infection was noted
in the field. Larvae were observed for fungal infection for 10
consecutive days.
Larvae that fed on the leaves did not become infected by
Nomuraea until 26 August. However, laboratory reared lots of
field collected larvae occasionally became infected beginning the
second week of July. At the time of first infection of the leaf-
fed larvae, the infection level in the field was 1%. This figure
was computed from laboratory based estimates of field col-
lected larvae. It is assumed that infection by Nomuraea can
occur via leafborne conidia (regardless of mode of infection,
i.e., via gut or via integument) and that infection resulting from
leafborne conidia correlates with conidial density per unit of
leaf surface. This type of correlation has been shown for
Nomuraea and Trichoplusia ni by Ignoffo et al. (1975, 1976).

Mode of Infection
Young larvae were isolated in a weight boat and lightly
touched with a camel's hair brush containing conidia. The larvae
were maintained overnight without food in a closed container
kept moist by wet paper toweling covering the bottom. The fol-
lowing day, the larvae were removed from the container, washed
vigorously with running tap water, and placed in a new con-
tainer on fresh rearing medium.
A second group of larvae were placed on medium containing
conidia. After feeding overnight, larvae were removed from the
containers, washed vigorously with running tap water, and
placed on medium without conidia. Control larvae were not
dusted. All larvae were reared for 10 days, and infection was
recorded within the three groups. Three replications were run,


using a total of 60 larvae. Further evidence of mode of infection
was obtained by feeding 10 larvae one microdrop of an aqueous
conidial suspension, administered directly into the mouth with
a finely drawn glass pipette. Ten larvae were inoculated with a
microdrop of the conidial suspension placed on their backs. Con-
trol groups were fed a drop of water without conidia and ad-
ministered a drop of ordinary water on their backs respectively.
No attempt was made to determine the number of conidia in
the dusting application or the suspension. All larvae were placed
in individual containers on a standard pinto bean diet and ob-
served for 10 days.
Within 7 days after dusting conidia on the integument of the
larvae, 87.5% had died of Nomuraea infection. After 10 days,
mortality was 100%. Those larvae which fed for 24 hours on
medium containing conidia suffered no mortality during the first
7 days, but evidenced a 40% level after 10 days. Larvae which
were force-fed a conidial suspension of Nomuraea suffered 30%
mortality in 8 days, the identical mortality level of those which
had the suspension applied to their backs. All control larvae
developed to maturity without disease incidence.
These data indicate that infection can occur via gut or in-
tegument and that the dominant mode is probably a function of
conidial density.

Ontogeny of the Pathogen
Larvae were fixed at varying lengths of time after dusting
with conidia. Cadavers were dehydrated in butanol and im-
bedded in paraffin. After microtoming, the sections were stained
in either congo red or lactophenol cotton blue. The sections were
examined for germinated and ungerminated conidia on the in-
tegument, penetration of hyphae, and localization and form of
the fungus inside the insect's body.
Conidia in contact with the insect integument germinated in
less than 8 hours. The integument surrounding the point of
penetration became blackened and ulcerous in many cases. Once
within the epicuticle, the fungus lysed large cavities in the
exocuticle and endocuticle but showed no marked inclination to
penetrate and grow perpendicularly to the surface of the epi-
cuticle. The fungus maintained filamentous growth while pene-
trating the integument. The mycelium fragmented and multi-
plied in a yeast-like fashion after entering the hemocoel. Four
days after penetration, the host became sluggish and moribund.
Squash mounts demonstrated the spread of the fungus through-
out the hemocoel, particularly in the posterior segments and host


tissues near points of penetration. The fungus did not appear
to have an affinity for the gut region and did not invade tissues
of the head until after death. A few hours before death of the
host, the fungal cells elongated and became fusiform.
During late stages of disease development, the insect often
climbed to the uppermost leaves of the host plant in the field.
A sticky secretion issued from the anus and firmly cemented the
insect to the plant. In many cases the insect died with the an-
terior portion of its body in an elevated position.
Conidiogensis commenced immediately after the cadaver
turned white with the maturation of conidiophores under con-
ditions of R.H. above 70%. Forty-eight hours after conidiophore
formation, the cadaver appeared green because of the large
numbers of conidia produced on the conidiophores.
The chronological sequence between germination and death
averages just over 6 days. Beyond this point, condiophore for-
mation and conidiogenesis occur in 3.5 days.
The period between germination and penetration is less than
15 hours. Conidia can effectively infect the host during the same
24-hour period they are produced, providing they are dispersed
and all other factors are favorable. It is assumed that ontogeny
of the pathogen from penetration to conidiophore formation is
largely independent of meteorological conditions.

Conidial Production per Cadaver
Kish and Allen (1976) reported that a definite relationship
exists between cadaver size and conidial production. The re-
gression equation is Y= -0.07544+0.00586(X) +0.0000259 (X2),
where Y equals conidia X 10" and X equals insect surface area
in square millimeters. Utilizing this equation, conidial produc-
tion for any size cadaver can be computed.
The authors found that conidial production takes place for
3 days, at which time the maximum number of conidia have
been formed and no further production can take place because
of depletion of nutrients (unpublished data). Conidial produc-
tion is assumed to be equivalent for each of the 3 days under
identical conditions.

Relative Density of Conidia in the Agroecosystem
Conidial trapping apparatuses (Hirst, 1952) were placed in
the field during the third week of July, 1974, approximately 6
weeks after the soybean were planted, in order to determine the
density of conidia over a soybean field on both a daily and hourly


basis. Conidial density was sampled continuously (hourly)
throughout the growing season until the first week in October.
Conidia were trapped on microscope slides covered (along the
entire length of one side) with a strip of clear dual adhesive
tape. Each slide contained conidia sampled over 24 hours, re-
corded in 24 bands on the sticky surface. Slides were recovered
at 48-hour intervals and transported to the laboratory in dust-
proof cases.
The slides were examined under a compound microscope,
and hourly conidial counts (or estimations when large numbers
of conidia were encountered) were made and plotted on a graph.
Each hourly count was averaged with the preceding hour's count
and with the 2 subsequent hours in order to minimize errors.
Hourly data were utilized in raw form (unaveraged) for cor-
relations of substantial hourly increases and/or decreases of
conidia with meteorological data.
Conidia of N. rileyi first appeared in the soybean field on 26
August 1974 in extremely small numbers and only during several
hours of the afternoon. The first larvae to die of Nomuraea in-
fection were collected on the same day. The relative airborne
conidial density (as sampled) remained low (less than 100 dur-
ing any given hour) until 5 September, when several conidial
counts surpassed 1,000 per hour. By 16 September, hourly coni-
dial counts over 16,000 were recorded, increasing to over 20,000
by 21 September, and to over 70,000 for a one-hour period on
27 September. Conidial density over the field gradually dimin-
ished to a daily high of just over 1,000 for several hours until
3 October, when the traps were removed from the field. Conidial
counts by day and hour throughout the period between 26
August and 30 September are presented graphically in Figure 1.
Airborne incidence'is discussed below in conjunction with cer-
tain environmental parameters.

The Host Crop
Bragg soybean were planted each year in 36-inch rows at
a rate of 12 seeds per row-foot. At the Agricultural Research
and Education Center farm in Quincy, Florida, cultivation was
employed to control weeds in 1973 and 1974 and was generally
completed twice before flowering. Soybean plots at the horti-
cultural farm and experiment station grounds on the University
of Florida campus were treated with Lasso herbicide at a rate
of 25 pounds per acre.


T 0.17 0.59 0.09 T 0,14 2.27
1 I I I I I I I I D

" 9


x A

o Sy- A
8/26 8/27 8/28 8/29 8/30 8/31 9/1 2 9/3 9/4 9/5 9/6

V 0.,44 0,32 0.58 0.09 T 0.42 0.45
I l I I I I D

J9 1' I \-'V B


Jf A
It + I v A W A

9/7 9/8 9/9 9/100 9/11 9/2 9/13 9/14 9/15 9/16 9/17 9/18


zI T 0.06 0.08 0.44

B3 B


9/19 9/20 9/21 9/22 9/23 9/24 9/25 9/26 9/27 98 9/29 9/30
Tui (AYS)

Figure 1.-A chronology of (A) airborne conidial densities,
(B) air movement, (C) foliage wetting and, (D) precipitation.
Chronology begins at 1200 hours, 26 August. Time is indicated in
12-hour increments. Broken lines after 4 September indicate miss-
ing conidial density and air movement data. The black horizontal
bars indicate periods of dry vegetation in both the upper (Ci) and
lower (C2) canopy. Shading indicates missing data. The vertical
bars at the top of the graph mark the beginning of a period of pre-
cipitation which resulted in the amounts shown above each bar.
T, trace.

Sampling Pests and Monitoring Environment
Population sampling of the VBC during the soybean growing
season was accomplished by the shake cloth method (Barnes
and Jones, 1970). Two hundred row-feet per acre were sampled
in a pattern determined to be a significant sample size and shake
pattern by computer analysis (Personal communication, 1974,
W.W. Menke, Department of Industrial Management, Clemson
University, Clemson, South Carolina). Sampling was conducted
bi-weekly, and pest populations were computed on a row-foot
A sample of larvae, collected in the field by the shake cloth
method, was retained for laboratory observation in order to
determine the percentage of fungal infection in the host popu-
lation. The larvae were retained for 5 days to insure that in-
fection occurred naturally in the field and not as a consequence


of collection methods or laboratory contamination. Mortality in
the sample due to Nomuraea was tabulated, and infection levels
were computed for the population in the field.
Field experiments conducted in Quincy during 1973 and
1974 were monitored through the facilities of the Environmental
Study Services Center of the National Oceanographic Atmos-
pheric Administration in cooperation with the University of
Florida's Institute of Food and Agricultural Sciences. Daily
data were provided on (a) solar radiation, (b) pan evaporation,
(c) soil temperature, (d) rainfall and intensity, (e) dew dura-
tion, (f) wind speed and direction, and (g) atmospheric pol-
The following microenvironmental monitoring equipment and
facilities of the Department of Fruit Crops, University of
Florida, were utilized to monitor specific parameters: (a) sensi-
tive wind speed and direction sensors having a stall speed of
less than one mile per hour, (b) Atkin's lithium chloride sensors
for dew point determination, (c) dew duration sensors, (d) air
and soil temperature thermocouples, (e) Esterline Angus re-
corders for wind speed and direction, (f) Esterline Angus re-
corders for dew duration sensors, (g) a Leeds and Northrop
24-point thermocouple for temperature, and (h) a Honeywell
recorder for dew point sensor temperature. Environmental
sampling was conducted continuously or at hourly intervals.
Environmental monitoring was conducted during 1975 at
the University of Florida's horticultural farm in Gainesville.
Data for the 3 years were reduced and analyzed. Various
averaging systems were employed to correlate meteorological
data with infection and sporulation inputs and to make the data
reflect trends during time periods considered significant for

Life Cycle Relationships
Selected life cycle relationships within the soybean agroeco-
system are depicted in Figure 2.
The life cycle of the soybean is basic to all other cycles in
the system. The growing season in Florida extends from plant-
ing, usually in early June, to early October, when the plants
become senescent. The soybean may exist in the field only as
seed from this time until the next June, with the exception of
some scattered germination, which usually progresses only to
the cotyledon stage of development before being turned with
the soil during winter or killed by freeze.


SC Germination



5 Dissemination


End epizootic sept *)

Begin epizootic

Fungus--long cycle

Soybean--growing season

C.I. Soybean--seed and detritus cycle

Fungus--short cycle
SVBC cycle

Figure 2.-Selected life cycle relationships in the soybean


The VBC adults usually appear in early July, though some
moths can occasionally be seen much earlier. The adult females
oviposit beneath the leaves from mid-July through September.
Between September and the following spring, all forms of the
insect are absent from the field, but larvae are sometimes found
on alternate leguminous host plants not killed by freeze.
The appearance of Nomuraea in the system is tertiary to
that of the plants and lepidopterous larvae. Seemingly dependent
on host density, the fungus makes its appearance on a relatively
few larvae in early August. It can become fairly well established
and infect nearly 100% of the VBC larvae by the third week in
August. Through abundant infection, sporulation, and dissemi-
nation, the fungus perpetuates itself through what may be
depicted as a short cycle, lasting until mid-October when the
crop is harvested. The manner in which the fungus overwinters
in unknown, but it is possible that it maintains a low population
level on alternate hosts and VBC in more southern regions of
Florida. Adult moths, migrating north in the spring, may carry
the conidia on their legs as they do during the growing season
(Kish, 1975).

Environment and Phases of the Fungal Life Cycle
Conidiophore Formation and Sporulation.-Approximately
5 row-feet of soybean in a field undergoing an epizootic were
enclosed by a cheesecloth tent. Over a period of 3 days, VBC
larvae present on the plants were tagged upon dying. The time
of death was recorded. Cadavers were observed each morning
and evening thereafter through conidiophore formation and
conidiogenesis or until they were lost to physical or biotic fac-
tors. Times of formation of conidiophores (white stage) and
conidia (green stage) were recorded on the tags. These data
were correlated with environmental conditions.
A total of eight VBC larvae died during the observation
period. Three of the eight cadavers were lost to predation or un-
known causes. All larvae except one died during the night, and
the conidiophores were conspicuous by 0900 hours the next
morning. Relative humidity approached 100% each night during
the observation period. Conidiophore formation took place in
less than 15 hours in all cases in which the larvae died at night,
and all cadavers were green within 36 hours after dying.
Conidial Density and Dispersal.-Conidial density over the
field was directly related to (1) the infection level in the host


population, (2) the daily hours of foliage wetting and drying,
(3) precipitation, and (4) air movement (Figures 1, 3).
Host density rose to approximately one larva per row-
foot on 26 August 1974. This was the first day infection was
noted among field collected larvae and also the first time conidia
were. collected in the spore traps. A total of five conidia were
recovered over a period of 13 hours. On 1 September, the larval
population reached four larvae per row-foot with an infection
level of less than 6%. No conidia were recovered on that day.
On 10 September, the larval population peaked at more than
six larvae per row-foot, while the infection level rose to 39%. On
19 September, the infection level reached 98% of the larvae in
the field, and the conidial count was well over 100,000 for a 24
hour period extending into 20 September.
Daily airborne conidial densities were also directly related
to the duration of foliage wetting. Conidial densities demon-
strated a periodicity of relatively low counts at night, when the
foliage was wet, and higher counts during the day, when the
foliage was dry, with the exception of rainfall.
Precipitation data (Figure 1), when correlated with conidial
densities, indicate almost without exception that rain was fol-
lowed by lower airborne conidial densities (7, 9, 10, 12, and 16
September). The notable exception to this observation was the
effect of a trace to several hundredths of an inch of rain, which

0 6


i I I I I I I I
8/5 B/10 8/15 8/20 8/25 8/30 9/5 9/10 9/15 9/20 9/25 9/30 10/5 10/10
Time (days)

Figure 3.-VBC population and infection levels at the Quincy
Station, 1974. Shaded area indicates the number of larvae per row-
foot infected with Nomuraea.


seemed to be followed by increased conidial counts, (26, 27, and
28 September). These data indicate that rain has a cleaning
effect on the air, removing conidia and depositing them on the
Air movement, like foliage wetting, demonstrated a periodicity
during the observation period (Figure 1). Air movement was
consistently higher during the daylight hours than at night.
High conidial counts correlated with dry foliage. Peak hours of
conidial density usually occurred during periods of gusty winds
associated with thunderstorms of passing weather fronts. There
were eight such fronts during September, 1974, at the Quincy
station. Of particular note was the early afternoon of 27 Sep-
tember (1300 hours), during the passage of a warm front ac-
companied by high winds, when conidial counts were estimated
at over 150,000 between 1200 and 1400 hours and contributed to
the highest daily total recorded in 1974.
These data suggest that air movement, precipitation, vegeta-
tive wetting and/or R.H., and infection levels in the field are
definite factors which must be taken into account in any attempt
to numerically define disease progression.

Irrigation as a Rain Simulator.-The effect of overhead irri-
gation as a rain simulator on cadaver-borne conidia was investi-
gated. Larvae of various sizes were dusted with conidia and
held on rearing medium until death, at which time their surface
area was computed. The cadavers were placed in small, num-
bered weight boats and maintained in a moist chamber for 5
days to insure maximum conidial development. The number of
conidia per cadaver was computed utilizing a regression equa-
tion (Kish and Allen, 1976). Cadavers, which in all cases ad-
hered to the weight boat during conidiophore and conidial de-
velopment, were transported to the field in covered Petri plates
to avoid conidial loss to the air during transport. Once in the
field, the weight boats were removed and carefully taped to
rigid stems of soybean plants. Cadavers were placed in both an
irrigated and non-irrigated (check) plot.
The irrigation system was activated for 2 hours, resulting
in 0.72 inches of precipitation. The cadavers in both plots were
removed from the plants, placed in individual vials, and capped.
In the laboratory, each cadaver was subjected to the procedure
mentioned above for determination of the number of conidia
present. Total conidia on each cadaver was tabulated, and those
cadavers exposed to irrigation were compared with those of the
check plot.


Cadavers exposed to irrigation lost an average of 90% of
the conidia calculated to be present initially. During the same
time period, those cadavers not exposed to irrigation lost an
average of 49% of the conidia. Size of the cadavers ranged from
13.6 x 1.2 millimeters to 30 x 3 millimeters.
Field observations during irrigation indicated that most of
the conidia were dispersed by the water almost immediately
after irrigation began. Two consequences of this experiment
were apparent: first, only 90 % of the total conidia present were
removed through any natural occurrence (i.e. wind, rain, me-
chanical beating); second, the 90% conidia removed was prob-
ably dispersed in less than one hour. At present, we will assume
that even though approximately 41% of the conidia on the
cadavers in the check plot were removed by wind during the
experiment, the potential of these conidia as infective units was
not decreased to the extent of the 90% removed so quickly
that wind had no time to be a factor in their dispersal.

Summary of Effects of Environment
These data indicate that general concepts of earlier workers
concerning optimum conditions for an epizootic, such as gener-
ally wet, rainy conditions, must be qualified before being applied
as a general rule. Some conditions may even be highly detri-
mental to the progression of various phases of the disease.
Based on the results of three seasons of monitoring environ-
mental parameters and their relationships to the development
of N. rileyi epizootics, the following assumptions can be made.
1. Dry, windy conditions promote conidial dispersal. In-
creased conidial densities result from dry, windy conditions, but
are also dependent on host population density and infection
2. Dry windy conditions retard germination and infection,
as such, but promote infection if followed by humid conditions,
providing no excess of free water exists.
3. Fungal ontogeny within the host body, up to and includ-
ing death, is independent of weather conditions.
4. Conidiophores will form independently of fluctuations in
external humidity as long as the cadaver does not undergo rapid
5. The minimum R.H. at which conidial production takes
place is 70%. As R.H. increases, conidial production increases.
6. Rain and vegetative wetting promote conidiophore for-
mation and conidiogenesis.


7. Conidia on cadavers are washed to the ground by an ex-
tended light rain, a brief heavy rain, or long hours of heavy
vegetative wetting by dew.
8. An alternation of wet and dry conditions is necessary for
spread of infection.
9. An excess of free water during the height of an epizootic
has little net effect on the course of the epizootic, but an excess
of free water in the early stages of an epizootic (infection level
less than approximately 10%) may retard the spread of in-
fection, if it follows conidial formation but precedes conidial
10. The alternation of short periods of vegetative wetting
and high humidity with longer periods of dry conditions with
light winds favors the increase and spread of infection.


The most useful information concerning the control of soy-
bean pests is their population in the field and how this population
correlates with crop damage. Judgments are then made on the
necessity of control, and the type of control measures to be taken.
The grower usually follows variations of the following prac-
tices: (1) waiting until large, damaging populations are ob-
served before spraying or (2) spraying on a schedule, regardless
of pest populations. Heavy damage can result before control
measures can be initiated with the first approach. The second
approach could result in unnecessary control, environmental
pollution, and excessive costs. Neither approach takes into ac-
count natural control factors which might suppress the popula-
tion at any given time. For example, farmers often apply in-
secticides early in the season when defoliation is first noted.
Such applications eliminate valuable predators and parasites
that normally control later-occurring pests such as podworms,
stinkbugs, and armyworms. Secondary outbreaks of these pests
are created by attempting to control VBC populations, thus
necessitating further chemical applications throughout the sea-
son. A more accurate evaluation of actual or projected pest popu-
lations could be made if natural control could be predicted in ad-
Menke (1973) and Menke and Greene (1976) published a
pilot pest population model for VBC. This preliminary model


takes into account such factors as oviposition, hatching, pupation,
and mortality in order to project VBC population levels. A quan-
titative analysis of mortality due to predators, parasites, and
pathogens was unknown at the time of these publications.
Separate investigations were initiated on the effects of pre-
dators, parasites, and pathogens in order to make such population
models more accurate. The factors of prime importance to us as
pathologists were (1) the principal pathogens involved and (2)
the effects of these pathogens on pest population density.
The first question was easily answered by observation and the
literature (Watson, 1915, Allen et al., 1971). The second question
is answered, in part, by this paper.

Arithmetical Relationships Among Factors
The basic problem is to predict deaths due to Nomuraea for
a given day. This can be solved by determining the number of
infections taking place 6 days prior to the day in question, since
the time between infection and death is approximately 6 days.
The following must be known in order to tabulate the number
of infections taking place on a given day: (1) inoculum density,
(2) weather conditions and their effects on the inoculum and,
(3) the relationship between inoculum density and infection
levels. A minimum pest density threshold is also suspected to
affect disease progression, but no significant data on this subject
is available at present. The first three factors will be discussed
independently as the model is constructed.

Inoculum Density
Inoculum density on a given day is dependent on several
factors. Previously, conidial production was discussed as a func-
tion of cadaver surface area. Given a cadaver of known surface
area, conidial production progressed for 3 days under optimum
conditions (R.H. greater than 70%) according to the regression
equation. Ontogenetic studies demonstrated that after the larvae
die, they undergo a period of conidiophore development before
actual conidial production takes place. This stage of fungal de-
velopment, known as the white cadaver stage (because of the
cottony-white appearance of the cadavers) was the subject of
a paper by Kish et al. (1976). In this work, the authors identi-
fied a correlation factor between the number of white cadavers
which fell on the shake cloth during population sampling and the
number of cadavers actually on the plants. The relationship


was one to three: that is, three white cadavers on the plants for
every one cadaver 'observed during sampling.
Inoculum production is directly related to available substrate.
If cadaver number and size could be recorded during population
sampling, the theoretical, maximum inoculum production could
be calculated. To make the task somewhat easier, cadavers were
sampled for 200 row-feet per acre and assigned to a large,
medium, or small size range. A median conidial production figure
was then calculated for each size range. In order to calculate
cadaver numbers and conidial production on a per acre basis,
the factor B/C is introduced as being the number of row-feet
per acre, (B), relative to the number of row-feet sampled per
acre, (C). Since the maximum production takes place over a
3-day period, the factor 0.33 is introduced to put maximum
conidial production on a daily basis. The equation for maximum
conidial production per day per acre based on white cadaver
counts takes this form:
Maximum production=3 [(A1 (B/C x 5.6 x 108 x .33))
+ (A2 (B/C x 1.5 x 10" x .33))
+ (A3 (B/C x 3.18 x 109 x .33)) ]
Ai=the number of small cadavers, (max. 16 x 1.5 mm)
A = the number of medium cadavers, (max. 25 x 2.5 mm)
A= the number of large cadavers, (max. 34 x 3.5 mm)
5.6 x 108=conidia produced on a median sized, small
1.5 x 10=conidia produced on a median sized, medium
3.18 x 109=conidia produced on a median sized, large
B=the number of row-feet per acre
C=the number of row-feet sampled per acre
0.33=factor for determination of 1/3 of maximum pro-
duction which occurs over a 3 day period.
3=factor to compensate the fact that only 1/3 of the
cadavers present on the plants fall to the shake cloth.
The factors for determining maximum conidial production
for one day, 0.33, and the factor compensating for the recovery
of 1/3 of the cadavers present, 3, equate to 1 and can therefore
be excluded from the equation. Given a number of cadavers of
known size, the maximum number of conidia that could be pro-


duced in one day on one acre can be calculated. At this point,
there is the potential for a known conidial density in the field.
Other factors must now be taken into consideration. The above
equation calculates the maximum conidial potential which rarely
can be realized. Since this potential is the highest number of
conidia that could be produced, most factors which actually act
on conidial production, dispersal, catch, viability, and virulence
can be entered as zero factors (no effect) or negative factors
(those which subtract from the available potential). These fac-
tors can now be evaluated and applied to the equation.

The Effect of Weather Conditions On Inoculum Density
Relative Humidity.-Laboratory experiments indicated that
no conidial production occurred below 70% R.H. Therefore, any
period of R.H. below 70% should result in some percentage de-
crease in the maximum potential. A linear relationship among
time, R.H., and conidial production is assumed. The percentage
of time that the cadaver was influenced by R.H. below 70%
would result in that percentage fewer conidia. Working on a
24-hour basis, this relationship could be expressed mathemati-
cally as:
Conidial production=maximum potential x D/24
where D=the hours of R.H. above 70%.
If D=24, then the effect of R.H. would be a zero factor in
conidial production.
Relative humidity below 50% results in no infection. A survey
of R.H. in the Gainesville area during the soybean season demon-
strated that it rarely dropped below this level for any appreci-
able time. Therefore, this area need not be concerned with this
aspect of R.H.; elsewhere, it may be significant and would have
to be applied to the equation.
Rain.-Experiments demonstrated that rain had a detriment-
al effect on the inoculum load by washing conidia to the ground.
It was assumed by experimentation and observation that 0.36
inch of rain would wash away approximately 90% of the conidia.
To simplify calculations, these figures were converted to 63%
of the conidia being washed away by 0.25 inch of rain. The effect
of rain is thus expressed as:
Loss due to rain=E/.25 x .63
where E= rainfall over a 24-hour period, in inches.


When applied to the equation, the loss due to rain is subtracted
from 1 to arrive at a percentage of the total conidia left over
after the rain factor is entered. The entire equation to this point
thus becomes:
Inoculum load= [(At (B/C x 5.6 x 108))
"+ (A2 (B/C x 1.5 x 10o))
"+ (A3 (B/C X 3.18 X 109))]
(D/24) (1-(E/.25 x .63))
Wind.-This factor is one of the most difficult to understand
and integrate into the equation.
Unlike previous factors in the equation, wind is never a zero
factor. Air movement of certain limit disperses the conidia from
the cadavers so that maximum dispersal and minimum loss is
incurred. Anything below this limit results in fewer conidia
being disseminated. Anything over this limit results in conidia
being blown away. The dispersal aspect is treated here as a
separate factor from catch, which is discussed below.
Air movement in the environmental monitoring system was
recorded in miles per day. Further calculations were based on
several broad assumptions:
1. Air movement as monitored was strictly horizontal.
2. Air movement was uniform throughout the canopy.
3. Conidia settled according to the law of gravity regardless
of what was assumed to be "horizontal air movement".
McCubbin (1943) investigated and reported rates of fall of
fungal spores in relationship to their dimensions. None of the
fungi could be related in dimension directly to Nomuraea for a
number of reasons. First, it has been noted in conidial counting
of spore trap samples that part of the conidiophore, with its load
of conidia, usually broke off the cadavers (approximately one-
half). Thus, observing individual conidia was not the rule. Also,
the rate of fall of a portion of a conidiophore with its load of
conidia would be different (slower) than a solid spore of ap-
proximately the same size.
A simple experiment was designed to measure the rate of
fall of Nomuraea conidia and was then compared to McCubbin's
For the experiment, sporulating cadavers were tapped lightly
while held at the top of a glass cylinder containing a thin film
of oil on water. As the conidiophores and conidia fell the meas-
ured distance, their fall was timed. Conidiophores hitting the oil


surface were observed both directly and by the change in color
refraction patterns on the oil-water surface. The average rate
of fall was 39 inches per minute or about 16 millimeters per
second. If the average sized portion of a conidiophore impinging
on the oil surface was 80 by 20 microns (about one-half of a
conidiophore as described by Kish et al. (1974), it would be put
in the same dimension range as Helminthosporium savitum,
which had a rate of fall of 20 millimeters per second, according
to McCubbin.
It was known that the inoculum settled at the rate of 39 inches
per minute. The distance from the mean attached conidial load to
the ground was then needed. This distance was measured from
the ground to the middle of the leafy portion of the canopy or
about 39 inches for Bragg variety when fully grown.
Conidia falling at 39 inches per minute would reach the
ground in one minute. During this minute, the conidia would be
moved across the field by the horizontal air movement. It was
apparent that a certain percentage of the airborne conidia would
be displaced from the air mass directly over the acre where they
were produced and that this horizontal displacement was directly
related to air movement during the minute the conidia were
settling. An acre of soybean surrounded on all sides by soybean
would conceivably gain as many conidia via air movement as it
lost. On the other hand, as the conidia-laden air mass moved off
an isolated acre, the air mass replacing it would contain no
conidia. Therefore an isolated acre was used for simplicity. As-
suming that the conidial load of an air mass over such an acre
would be evenly dispersed, the percentage of the air mass dis-
placement off the field would be equal to the percentage of inocu-
lum lost. In other words, horizontal air movement of 207 feet
per minute would displace the entire air mass, (or most of it
depending on wind direction) and a proportional number of
conidia would be lost. The whole relationship is expressed as:
F/24 x 5280/60 + 207
F/24=miles air movement per hour, and 5280/60 con-
verts this information into miles air movement per
Subtracting this quantity from 1, to arrive at a percentage
of the total conidia remaining after loss due to wind is entered,
(1- (F/24 x 5280/60 207))


Since air movement beyond a certain limit could result in a
negative figure in the equation, a limit of 50 miles air movement
per day was determined. An air movement beyond 50 miles per
day is entered as 50.
Air movement below a certain limit has the same net effect
on dispersal and subsequent infection as an air movement above
the limit. That is, the conidia are not dislodged from the cadaver
and therefore are not dispersed. In order to compensate for this
lack of dispersal, a floating scale was formulated which equated
low air movement values to high movement values. This formu-
lation is an adjustment to the entry of air movement values below
20 miles per day and will be discussed below under Adjustments.

Holdfast Factor
Field and laboratory experiments indicated that only 90% of
the conidia produced on a cadaver could be removed by any
natural physical factor. Expressed in the equation, the number
0.9 accounts for the 10% that cannot be dispersed.

Catch Factor
Given a conidial load dispersed in a soybean field, the per-
centage of these conidia caught on the leaves and the percentage
lost to the ground had to be determined. Again, as with wind,
catch is never a zero factor.
As indicated earlier, catch was not related to wind, but in-
stead, correlated with leaf area and leaf area indices. The leaf
area index gives the ratio of leaf surface area to the area of an
The probability of a conidium being caught on the leaves or
falling to the ground would share the same relationship as the
leaf area index ratio. The number of airborne conidia with a po-
tential for being caught is then expressed as:

1- (1/G)

where G=leaf area index.

Various factors have been applied to a potential number of
conidia, resulting in a quantity of conidia in the field which have
a high probability of being successfully caught on the leaves,
where they will come into contact with the pest. The equation to
this point is:


Conidia= [(Ai(B/C x 5.6 x 10))+ (A2(B/C x 1.5 x 109))

+(A3(B/C x 3.18 x 109))] (D/24)(.9)

(1- (E/.25 x 6.3)) (1- (F/24 x 5280/60 + 207))

(1- (1/G))

Ultra Violet Factor
The final factor to take into account is the effect of UV on
conidial viability. Ignoffo et al. (1976, 1977) reported that one-
half of the conidia of N. rileyi are rendered non-viable over a
one day period. The factor 0.5 will therefore be entered in the
The inoculum load as derived above can now be looked at in
relationship to infection levels.

The Relationship Between Inoculum Density and Infection Levels
Ignoffo et al. (1975) determined infection levels of Tricho-
plusia ni by Nonmraea as a function of conidial density per unit
of leaf area using bioassay techniques. In the same paper, they
reported that the VBC is less susceptible to infection by
Noruaraea. No bioassay curve has been worked out for the VBC
and Nomnuraea; therefore the T. ni curve was used for the pre-
diction of infection levels. Since the bioassay curve plots conidia
per unit of leaf area, one more factor must be entered into the
equation. The total leaf area of soybean in the field (in square
millimeters) must be divided into the total conidial load as
derived above. This is expressed as:

2G(3.98 x 10")

G=leaf area index

(3.98x109) =the number of square millimeters in one acre.
The bioassay technique employs both upper and lower leaf sur-
faces. Thus, the leaf index must be multiplied by two. The final
form of the equation computes conidial density per square milli-
meter of leaf surface in the acre. This value is applied directly
to the bioassay curve to determine the anticipated infection level.
The simplest expression of the equation is:


Conidia/mm2= [ (A(B/C x 5.6 x 108))+(A2(B/C x 1.5 x 109))

+(A3(B/C x 3.18 x 109))] (D/24) (.9)

(1- (E/.25 x .62)) (1- (F/24 x 5280/60 207))

(1- (1/G)) (.5)

(2G) (3.98 x 109)

The bioassay curve is presented in Figure 4.
In summary, the number and sizes of white cadavers ob-
served to fall on the shake cloth during population sampling
were entered into an equation to tabulate their maximum po-
tential for conidial production. Factors affecting production,
loss through environmental factors, and rules of holdfast and
catch were applied. Finally, the number of conidia remaining
was divided by the leaf area in the field, yielding conidial density
per square millimeter of leaf area, which could then be applied
directly to a bioass'ay curve. The value of this equation is that
actual field data from one day would be utilized to predict an





0 10 20 30
Conidia per Unit Leaf Area (mm2)

Figure 4.-Infection levels relative to conidial density (Bioassay
curve) (after Ignoffo et al., 1975).


infection level 6 or 7 days later. Factors not incorporated to this
point include temperature and R.H. as related to infection,
neither of which are limiting factors in Florida.
A number of adjustment factors pertaining to the manner in
which data are entered, were incorporated to provide a better
"fit" as actual testing of the equation began. These will be dis-
cussed before presenting the trial data. The adjusted equation
is presented in Figure 5.

Relative Humidity
During application of the equation to the 1975 data, it was
noted that in every case where the day was preceded by rain,
the prediction of infection was low when compared to actual ob-
served levels. The entire significance of this phenomenon could
not be determined with the 1975 data because R.H. from the field
was not monitored in hourly averages. Therefore, R.H. data from
past years was checked to estimate the average number of hours
per day with R.H. above 70%. Using this approach, 20 hours of
humidity above 70% was programmed for each day. The error
for days preceded by some quantity of rain was corrected by
adding hours of R.H. on a sliding scale for a certain amount of
rain. For rainfall between 0.1 and 1 inch, 2 hours of R.H. above
the limit were added; for 1.01 to 2 inches, 4 hours of R.H. were
added; for 2.01 to 3 inches, 6 hours of R.H. were added; and 12
hours of R.H. were added for greater than 3 inches of rain the
previous day. Simplifying the scale and applying it to the hu-
midity factor in the equation, it can be expressed as:
(D+2r) /24
where r= rain in inches falling on the previous day.
Using this approach, the error for these days was significantly
reduced. The same factor was necessary for the 1976 data even
though hourly R.H. readings were accurately taken.
The final adjustment to the humidity entry was prompted by
the experimental observation that only very sparse conidio-
genesis took place at 70 % R.H. Accordingly, the R.H. limit for
maximum sporulation was raised to 80%.

The previous discussion of wind noted that air movement
below a certain level resulted in fewer conidia being dispersed



(AI(B/C x 5.6xD8))+( (B/C x 1.5x10))+(A3(B/C x 3.18x109))] ((2R)/24).9(1-(E/.25 x .62))(1-(F/24 x 5280/60 + 207))(1-1/G).5

(2G) (3,98x109)


Figure 5.-Equation for prediction of conidial loads of Nomuraea in a soybean field.

-*' 1I

and that for purposes of compensation, low air movement was
equated to high movement. During the trials, it was noted that
air movements below 20 miles per day resulted in considerable
error. As air movements decreased below 20 miles per day, a
sliding scale was devised to equate them to successively higher
air movements. The complete scale is as follows:
For air movements of: 6 and 50 enter 50
7 enter 49
8 enter 48
9 enter 47
10 enter 46
11 enter 45
12, 13 enter 44
14, 15 enter 43
16, 17 enter 42
18, 19 enter 41
20 enter 40
21, 49 enter as is.
This adjustment compensated satisfactorily for low air move-

According to the experimental data, 0.36 inch of rain would
wash off 90% of the conidia on a cadaver (all that could be re-
moved). It appeared that this figure should be the limit set for
rain. As trials progressed, rain amounts above 0.30 inch re-
sulted in successively larger error. Therefore, the limit of rain
that could be entered was lowered to 0.30 inch. Any amount per
day over 0.30 inch is entered as 0.30.

Non-susceptible Larvae
It is known that large larvae do not become infected by
Nomuraea. The fungus attacks only the larval stage and routine-
ly infects the first, second, and third instars only. Field popula-
tions of VBC have a certain percentage of large, non-susceptible
larvae which will not undergo the predicted infection at any
conidial density. To compensate for this factor, the percentage
of large larvae in the field population must be subtracted from
the predicted infection level. For conidial densities above one/
square millimeter (infection level at approximately 18%), the
adjustment factor gave the predictions a much better "fit" and
is therefore incorporated into the trials. For predicted low in-


fection levels (less than 18% or a conidial density of less than
one/square millimeter), the adjustment became a source of error
because more than 18% of the larval population usually are in
the fourth or fifth instar stage. The adjustment was therefore
deleted for prediction levels in this range.

Trials of the Equation with Actual Field Data
Two years of field data were tested in order to validate
the program. The first year test (1975) consisted of four plots
of soybean originally planted to test the effect of selected pesti-
cides on the development of Nomuraea. The plots and the treat-
ments in each plot are outlined below:
Plot 1-1/3 acre, treated on 12 August and 26 August 1975
with a mixture of Benomyl and Methyl Parathion.
Plot 2-1/3 acre, treated on the same dates as Plot 1, with a
mixture of Benomyl and Carbaryl.
Plot 3-1/3 acre, treated on the same dates with Benomyl
Plot 4-1 acre which served as the control and was un-
The plots were oriented north to south, with Plot 1 on the
northern end and Plot 4 on the southern end. All plots were
bounded on the west-southwest, (direction of the prevailing
wind) by a small buffer strip of soybean (untreated). Beyond
this strip were 2 acres of soybean in which the pest population
was completely suppressed during the season. Meteorlogical con-
ditions monitored in the field included: air movement in miles
per day; R.H., four times daily; and daily precipitation. Pest
population sampling, cadaver sampling, and determination of
infection levels among laboratory held lots of field collected
larvae were accomplished in each plot, twice weekly in con-
nection with the pesticide experiments.
Data were assembled for each day that infection levels based
on laboratory observations of field collected larvae were known.
The equation was programmed into a microcomputer and field
data inputs entered for each day. Predicted infection levels were
compared to actual observed levels. Forty-three trials were run
covering 14 different days during a period from 11 August to
18 September 1975.
The field test for the second year was carried out on 1/3 acre
of Bragg soybean planted the first week of July, 1976, at the
Florida Agricultural Experiment Stations grounds in Gainesville.
No fungicide or insecticide treatments were applied. The beans


were bordered in the direction of the prevailing wind by pasture
and a small plot of field corn.
Meteorological conditions monitored were the same as those
monitored in 1975. Relative humidity was monitored hourly.
Pest population sampling, cadaver sampling, and determination
of infection levels among laboratory held lots of field collected
larvae were accomplished daily commencing 8 September and
ending 1 October. Leaf area index of the plot was determined
at the end of the period by a DENK Area Meter and was deter-
mined to be 2.71, inclusive of an estimated 30% defoliation.
Weather and sampling data were assembled for each day
during the sampling period. The equation was programmed into
a microcomputer, and the field data inputs were entered for each
day. Predicted infection levels were compared to actual observed
levels. Twenty-four trials were run for the 24 consecutive days
for which data was available. Results of the trials and statistical
analysis are presented in Tables 1 through 7a.

Statistical Analysis of Results
Data were analyzed on an overall, seasonal, and per plot
basis. These data are discussed below.

Chi Square
Results of Chi Square (X2) testing for each trial are given
in Table 3a through 7a. An analysis of X' values indicates the
(1) Predicted values for 1975 were significant at the 20%
level for 33 of 43 trials (76%); at the 10% level for 27 of 43
trials (62%); at the 5% level for 24 of 43 trials (55%); and
at the 1% level for 17 of 43 trials (39%). The average Chi
Square value for 43 trials was 10.68.
(2) Predicted infection levels for the last day that infection
was noted in each plot during 1975 were significantly erroneous.
(3) Predicted infection levels for days on which no living
larvae were collected in the field (Plot 2 on 25, 29 August and
1 September 1975) were significantly erroneous.
(4) Predicted infection values for Plots 1, 2, and 3 between
19 August and 11 September 1975 were generally less significant
than the predicted values in the control plot for the same time
period. This period extended from one week after initiation of
pesticide testing until 2 weeks after its completion.
(5) Predicted values for 1976 were significant at the 20%
level for 19 of 24 trials (79%); at the 10% level for 16 of 24


trials (66%); at the 5% level for 12 of 24 trials (50%); and at
the 1% level for 7 of 24 trials (29%). The average Chi Square
value for the 24 trials was 5.79.
(6) Overall, 52 of 67 trials (77%) completed over the 2
years were significant at the 20 % level or better; 36 of 67 trials
(53%) were significant at the 5% level.
Chi Square values for various combinations of trials are
presented in Table 1.

Coefficient of Correlation
The results of the coefficient of correlation analysis are pre-
sented in Table 2. Analysis of the coefficient indicates a high

Table 1. Chi Square analysis of disease incidence predictions made
in soybean at the University of Florida Agricultural Experi-
ment Station, Gainesville, Florida, 1975 and 1976.
Sums of X2 X2(Average)

All trials 1975 (43) 448.92 10.68
All trials 1976 (24) 138.96 5.79
All trials 1975 and 1976 (67) 587.88 8.77
All trials in untreated beans 1975-76 (36) 197.81 5.49
All trials in treated* beans 1975 (31) 390.04 12.58
Plot 1, all trials (1975) 66.96 6.69
Plot 2, all trials (1975) 241.44 24.14
Plot 3, all trials (1975) 81.64 7.42
Plot 4, all trials (1975) 58.85 4.90
"* Treated with insecticide and/or fungicide. See text for rates, times and
frequency of application.

Table 2. Coefficient of correlation analysis of disease incidence pre-
dictions made in soybean at the University of Florida Agri-
cultural Experiment Station, Gainesville, Florida, 1975 and
r r2

All trials 1975 (43) .76 .57
All trials 1976 (24) .82 .67
All trials 1975 and 1976 (67) .72 .51
All trials in untreated beans 1975-76 (36) .82 .67
All trials in treated* beans 1975 (31) .72 .51
Plot 1, all trials (1975) .89 .79
Plot 2, all trials (1975) .57 .32
Plot 3, all trials (1975) .76 .57
Plot 4, all trials (1975) .87 .75
"* Treated with insecticide and/or fungicide. See text for rates, times and
frequency of application.


correlation between predicted and observed infection levels both
for 1975 (.76 for,all trials) and 1976 (.82 for all trials). The
combined years data have a coefficient of 0.72. The coefficient
for the untreated plots during the two seasons (Plot 4, 1975 and
all of 1976) is 0.82; the coefficient for the combined trials in the
treated plots (Plots 1, 2, and 3, 1975) is 0.72.

Although there have been attempts in the past to predict
incidence of plant disease (Cook, 1949; van der Plank, 1963;
Waggoner and Horsfall, 1969), there are no such precedents for
predicting disease incidence among insect pest populations. Re-
cently, modeling of pest populations has received increased at-
tention (Menke, 1973; Menke and Greene, 1976; Waddill et al.,
In most systems including the VBC in soybean, mortality
factors play a prominent role in the dynamics of population
fluctuations and must be explained quantitatively. Our attempts
to predict insect disease have followed an approach similar to
the formulation of plant disease models such as EPIDEM (Wag-
goner and Horsfall, 1969). It is anticipated that our proposed
model will provide a starting point for modeling in other systems
even though it may require further modification as additional in-
formation becomes available. The strengths and weaknesses of
the present model are discussed in this section.
Results of the 43 trials in 1975 showed that the predictive
error for some trials can be readily explained. For instance,
insecticide treatments killed most of the larval population in
Plot 2 between 25 August and 2 September 1975. Observed in-
fection levels were derived from the tabulation of deaths in a
laboratory-held sample (usually 100 larvae) of field-collected
larvae. Since no living larvae were found in the field, there could
be no infection. The error due to these trials can be excluded
from the results.
Another source of error is the trials involving Plots 1, 2, and
3 between 19 August and 11 September 1975. As was stated
earlier, the soybean plots which furnished the data input for the
trials that year were planted primarily to test the effects of cer-
tain pesticides on the development of Nomuraea. The hypothesis
of this experiment was that the Benomyl, alone or in combination
with insecticide, would have a deleterious effect on the progression
of the disease. Johnson et al (1976) reported that significant
differences in infection levels at the 5% level or less (Chi


Table 3. Input data for Plot 1, Experiment Station Farm, Gainesville, Florida, 1975. Plot was treated on 12 August and
26 August with a combination of Methyl Parathion and Benomyl. Sampling rate: 100 row-feet.

No. No. Rain Hours On Sample Date
White Living Previous Relative Air
Sampling Cadavers Larvae Day Humidity Rain Movement
Date S M L S M L (in.) >80% (in.) (mpd)

14 Aug 0 1 0 0 1 9 0 20 0 16
21 Aug 1 0 0 0 1 3 0 20 .59 16
o 29 Aug 1 0 0 77 23 1 0 20 .03 35
S 1 Sept 0 1 0 110 81 46 0 20 .02 19
9 Sept 1 4 1 162 156 76 0 20 1.17 25
11 Sept 4 6 3 147 169 52 .27 22 0 13
15 Sept 8 13 4 194 109 30 .20 22 0 74
18 Sept 11 6 2 110 48 6 .20 22 .39 20
22 Sept 11 6 2 57 58 2 0 20 0 40
25 Sept 2 3 1 66 27 8 3.30 32 0 29
S Small
M Medium
L Large
mpd Miles per day

Table 3a. Predicted and observed infection levels for Plot 1, Experiment Station Farm, Gainesville, Florida, 1975.

No. Unadjusted NSL Chi
Sampling Conidia Infection Adjusted Predicted Observed Square
Date per mm2 Level Factor Infection Infection Value

14 Aug .54 15 90 15 0 15.00
21 Aug .04 2 75 2 0 2.00
S29 Aug .28 13 1 13 0 13.00
S 1 Sept .56 15 19 15 15 0
9 Sept 1.87 30 19 11 20 7.36
11 Sept 7.41 50 14 36 36 0
15 Sept 7.17 50 9 41 48 1.19
18 Sept 2.43 36 3 33 38 0.75
22 Sept 9.08 53 1 52 50 0.07
25 Sept 9.72 54 7 47 83 27.60
NSL Non-susceptible larvae

Table 4. Input data for Plot 2, Experiment Station Farm, Gainesville, Florida, 1975. Plot was treated on 12 August and
26 August with a combination of Carbaryl and Benomyl. Sampling rate: 100 row-feet.

No. No. Rain Hours On Sample Date
White Living Previous Relative Rain Air
Sampling Cadavers Larvae Day Humidity (in.) Movement
Date S M L S M L (in.) 280% (mpd)

11 Aug 0 1 0 193 209 139 0 20 .42 18
25 Aug 0 1 0 0 0 0 0 20 0 33
S29 Aug 3 0 1 0 0 0 0 20 .03 35
1 Sept 2 0 0 0 O 0 0 20 .02 19
9 Sept 2 2 0 136 95 3 0 20 1.17 25
11 Sept 8 1 2 190 185 12 .27 22 0 13
15 Sept 28 21 11 239 102 19 .20 22 0 74
18 Sept 15 11 8 146 123 7 .20 22 .39 20
22 Sept 18 9 4 64 67 7 0 20 0 40
25 Sept 1 2 2 52 19 18 3.30 32 0 29
S Small
M Medium
L Large
mpd Miles per day

Table 4a. Predicted and observed infection levels for Plot 2, Experiment Station Farm, Gainesville, Florida, 1975.

No. Unadjusted NSL Chi
Sampling Conidia Infection Adjusted Predicted Observed Square
Date per mm2 Level Factor Infection Infection Value

11 Aug .14 5 25 .6 2 2.66
25 Aug .88 20 0 20 0 20.00
29 Aug 2.44 36 0 36 0 36.00
1 Sept .42 15 0 15 0 15.00
9 Sept .79 19 1 19 48 44.26
11 Sept 4.40 45 3 42 74 24.38
15 Sept 16.05 65 5 60 60 0
18 Sept 5.70 47 2 45 45 0
22 Sept 15.31 64 5 59 47 2.44
25 Sept 10.96 56 20 36 95 96.69
NSL Non-susceptible larvae

Table 5. Input data for Plot 3, Experiment Station Farm, Gainesville, Florida, 1975. Plot was treated on 12 August and
26 August with Benomyl. Sampling rate: 100 row-feet.

No. No. Rain Hours On Sample Date
White Living Previous Relative Rain Air
Sampling Cadavers Larvae Day Humidity (in.) Movement
Date S M L S M L (in.) >80% (mpd)

11 Aug 0 1 0 200 196 103 0 20 .42 18
19 Aug 2 1 0 65 43 48 0 20 0 28
21 Aug 0 6 0 52 40 12 0 20 .59 16
25 Aug 0 3 0 92 61 33 0 20 0 33
29 Aug 3 0 1 52 68 26 0 20 .03 35
3 Sept 0 4 1 123 41 26 2.85 26 0 20
9 Sept 10 5 3 127 109 50 0 20 1.17 25
11 Sept 10 4 0 135 130 54 .27 22 0 13
15 Sept 15 9 7 240 88 30 .20 22 0 74
18 Sept 7 10 4 106 61 16 .20 22 .39 20
22 Sept 3 4 0 107 93 13 0 20 0 40
S Small
M Medium
L Large
mpd Miles per djy

tt 'I -(

Table 5a. Predicted and observed infection levels for Plot 3, Experiment Station Farm, Gainesville, Florida, 1975.
No. Unadjusted NSL Chi
Sampling Conidia Infection Adjusted Predicted Observed Square
Date per mm2 Level Factor Infection Infection Value

11 Aug .14 5 20 5 5 0
19 Aug 1.87 30 30 0 2 -
21 Aug .81 19 11 19 7 7.57
25 Aug 2.66 39 17 22 0 22.00
29 Aug 2.44 36 17 19 0 19.00
3 Sept 5.03 46 13 33 12 13.36
9 Sept 4.35 45 17 28 33 0.89
11 Sept 13.80 61 16 45 54 1.81
15 Sept 8.63 52 8 44 38 0.81
18 Sept 3.58 43 8 35 24 3.45
22 Sept 3.68 44 6 38 16 12.75
NSL Non-susceptible larvae

Table 6. Input data for Plot 4, Experiment Station Farm, Gainesville, Florida, 1975. Plot was untreated. Sampling rate:
200 row-feet.

No. No. Rain Hours On Sample Date
White Living Previous Relative Rain Air
Sampling Cadavers Larvae Day Humidity (in.) Movement
Date S M L S M L (in.) 180% (mpd)

14 Aug 0 7 0 128 197 200 .32 22 0 22
19 Aug 3 1 2 144 117 104 0 20 0 28
21 Aug 28 25 8 207 126 55 0 20 .59 16
25 Aug 56 4 4 135 135 57 0 20 0 33
29 Aug 36 33 14 97 92 36 0 20 .03 35
1 Sept 34 23 10 158 49 21 0 20 .02 19
3 Sept 41 12 5 235 26 8 2.85 26 0 20
9 Sept 61 33 5 269 194 9 0 20 1.17 25
11 Sept 64 43 3 350 137 23 .27 22 0 13
15 Sept 68 19 12 335 87 6 .20 22 0 74
18 Sept 74 21 0 275 10 1 .20 22 .39 20
S Small
L Large
M Medium
mpd Miles per day

Table 6a. Predicted and observed infection levels for Plot 4, Experiment Station Farm, Gainesville, Florida, 1975.

No. Unadjusted NSL Chi
Sampling Conidia Infection Adjusted Predicted Observed Square
Date per mm2 Level Factor Infection Infection Value

11 Aug .07 5 22 5 17 28.80
14 Aug 4.97 46 38 8 9 0.12
19 Aug 3.41 43 28 15 20 1.66
21 Aug 3.57 43 14 29 31 0.13
25 Aug 14.84 63 17 46 48 .08
29 Aug 28.71 84 16 68 63 .36
1 Sept 16.10 65 9 56 56 0
3 Sept 15.59 65 2 63 71 1.01
9 Sept 9.57 54 1 53 45 1.20
11 Sept 19.59 71 4 67 71 .23
15 Sept 10.23 54 1 53 59 .46
18 Sept 4.13 44 0 44 11 24.75
NSL Non-susceptible larvae

Table 7. Input data for validation experiments, 1976. Experiment Station Farm, Gainesville, Florida.

No. No. Rain Hours On Sample Date
White Living Previous Relative Rain Air
Sampling Cadavers Larvae Day Humidity (in.) Movement
Date S M L S M L (in.) >80% (mpd)

8 Sept 0 0 0 113 75 21 0 23.5 1.35 18
9 Sept 1 0 0 162 62 15 1.35 27.5 .92 10
10 Sept 3 0 0 204 89 25 .92 25 .30 15
11 Sept 0 0 0 238 95 14 .30 26 .09 25
12 Sept 1 0 0 238 59 14 .09 24 0 49
13 Sept 1 1 0 212 43 3 0 24 .50 49
14 Sept 8 0 0 176 42 9 .50 26 .30 46
15 Sept 0 0 0 183 33 10 .30 26 0 19
16 Sept 0 2 2 241 46 16 0 24 0 16
17 Sept 9 2 0 291 13 2 0 16.5 0 11
S18 Sept 14 5 0 292 23 4 0 17 0 6
19 Sept 6 2 0 378 55 15 0 16.5 0 10
20 Sept 14 2 0 385 43 12 0 18 0 13
21 Sept 23 4 0 439 33 20 0 20.5 .33 8
22 Sept 9 2 0 565 55 24 .33 26 .02 5
23 Sept 7 0 0 707 76 29 .02 23 0 10
24 Sept 11 4 0 402 50 15 0 24 0 9
25 Sept 7 2 0 321 30 14 0 24 .08 18
26 Sept 32 1 0 415 63 33 .08 23.5 0 20
27 Sept 37 11 0 466 138 44 0 19 0 25
28 Sept 33 3 0 330 66 5 0 22 0 20
29 Sept 39 23 0 346 26 4 0 18.5 .01 17
30 Sept 30 5 0 291 26 8 .01 24 0 21
1 Oct 49 5 1 220 23 2 0 18 0 36
S Small
M Medium
L Large
mpd Miles per day

I L I # 4* V '

Table 7a. Predicted and observed infection levels for validation experiments, 1976. Experiment Station Farm, Gainesville,

No. Unadjusted NSL Chi
Sampling Conidia Infection Adjusted Predicted Observed Square
Date per mm2 Level Factor Infection Infection Value

8 Sept 0 0 10 0 10 -
9 Sept .02 3 6 3 3 0
10 Sept .09 6 7 6 16 17
11 Sept 0 0 4 0 24
12 Sept .14 7 4 7 9 1
13 Sept .12 7 1 7 18 17
14 Sept .42 15 3 15 22 3
15 Sept 0 0 4 0 14
16 Sept 1.46 25 5 20 6 10
17 Sept 2.20 33 1 32 31 0
S18 Sept 2.38 35 1 34 30 0
19 Sept 1.54 28 3 25 19 1
20 Sept 3.44 43 3 40 19 11
21 Sept 1.12 22 4 18 16 0
22 Sept 1.80 30 4 26 14 6
23 Sept 1.32 25 3 22 44 22
24 Sept 3.90 44 3 41 29 4
25 Sept 2.90 41 4 37 45 2
26 Sept 10.63 55 6 49 44 1
27 Sept 31.49 90 7 83 58 8
28 Sept 11.77 57 1 56 43 3
29 Sept 20.79 73 1 72 53 5
30 Sept 29.27 87 2 85 40 24
1 Oct 19.87 71 1 70 50 6
NSL Non-susceptible larvae

Square) between the treated plots (1, 2, and 3) and the un-
treated control (Plot 4) occurred between these dates. It is clear
that Benomyl supressed development of the fungus. No infection
was noted for seven trial dates in the treated plots. Infection
in the check plot was consistent, and the highest X2 value for
any trial between 19 August and 11 September 1975 was 1.66.
These trials should be taken into consideration during evaluation
of the results.
For reasons yet unknown, final predictions of the 1975 sea-
son for each plot were consistently in high error. Overall, how-
ever, analysis of the 1975 data indicates the predicted values
are significantly correlated with observed infection levels.
Sampling in 1976 was more frequent (daily) than in 1975,
although the time period over which the sampling was accom-
plished was shorter. In addition, the R.H. and leaf area index
were measured accurately for input into the equation. In compari-
son with the 1975 sampling format, 1976 was definitely a sterner
test. Significantly, the results were consistent, if not somewhat
better. Certainly the overall correlation was better; however, as
pointed out earlier, the coefficient of correlation was lowered in
1975 by the trials involving plots treated with pesticide and/or
fungicide. Comparing untreated plots (Plot 4, 1975, and all of
1976), predictions were somewhat more accurate in 1975 (0.87
for 1975 versus 0.82 for 1976). The number of trials for 1976,
however, were twice that of 1975.
The chief mystery to date is why our predictions, based on
conidial densities on a given day, accurately predict a cumula-
tive infection level 5 days after sampling rather than the actual
infection level for that sixth day. Several possibilities are of-
fered as plausible explanations.
(1) As mentioned earlier, reduced susceptibility of the VBC
reported by Ignoffo et al (1975) may be an important factor.
However, predictions based on daily conidial counts, reduced
100-fold, were negatively correlated with daily infection levels
observed on the sixth day.
(2) Small and extremely small larvae die of Nomuraea more
quickly (2-3 days) than medium and large sized larvae, (labora-
tory observation). Usually, there is a higher proportion of these
small larvae than the larger larvae in the samples, and the death
of the smaller larvae which occurs earlier in the 5-day period
would lower the death count on the sixth day considerably.
(3) Infection levels do not normally change drastically or
rapidly over a 5-day period; a roughly fivefold miscalculation
(too high) in the manner in which daily conidial densities are


tabulated could result in predictions roughly five times higher
that the average daily mortality level during the 5-day period
in which the counts are taken.
Research is continuing to refine the parameter assessments
in the questionable areas, most critically, air movement. We are
convinced at present, however, that results and analysis to date
support the approach used in developing the model.
The model applies to an area of particular dimensions and
orientation in the agroenvironment at the present time. It is not
known exactly how the equation may have to be altered as it is
applied to varieties under various cultural practices. This model
will continue to be refined within its present application, although
the mathematical portion will evolve as sliding scales and tables
(wind, R.H., etc.) are incorporated into the equation as mathe-
matical factors, and as the entire formula is examined for re-
duction, simplification, and replacement into computer form.
The soybean insect pest management concept in Florida has
focused on the development of different suppression techniques
for VBC populations. These techniques are being refined so that
an economical pest management package can be implemented.
The essential components of VBC population suppression
will be integrated into an overall pest management system. Com-
ponents of particular importance are the effects of other pest
and beneficial insect species, pesticide usage, and environmental
conditions in the soybean ecosystem.
The key to the pest management system is the ability to
monitor the fungal entomopathogen N. rileyi. Using the fungal
pathogen model, mortality levels can be predicted on a daily
basis, and the subsequent increased accuracy of management
decisions can reduce the use of pesticides for VBC control by
60% to 100%.
The pest management system is expected to result in a com-
puter-based model for soybean which would have applicability
to Florida and other soybean producing states. It is expected that,
when implemented, the model system will eliminate' or greatly
reduce the current insecticide usage on soybean.
The pest management delivery system will be composed of a
two-way, computerized network which will permit the input and
acquisition of data at various hierarchal levels. An intermediate
computer, connected to a large regional computer complex lo-
cated in Gainesville, Florida, will serve as the center for the sys-
tem. Intelligent computer terminals will be located at the various
Agricultural Research and Education Centers (AREC's) through-


out the state. Extension center inputs will be coordinated
by a pest management specialist, supported by a research team
at each of the major centers. This specialist will be responsible
for coordinating and training other pest management specialists
located in the offices of county agents in his region. County pest
management specialists, in turn, will be responsible for the or-
ganization and training of local growers, scouts, and private
consultants. Basic data concerning phenological parameters, pest
and beneficial insect populations, insect disease incidence, and
local environmental conditions will be collected and transmitted
to data banks at Gainesville by trained personnel at the farm
level. Information will flow to AREC's and directly to the inter-
mediate computer in Gainesville. Environmental and weather
data will be collected automatically by the intermediate computer
at Gainesville from computerized weather stations located at
strategic AREC's in the state and from the National Aero-
nautics and Space Administration (NASA) and the Regional
United States Weather Bureau Center at Auburn, Alabama.
Data analyses from simulation model programs will return
pest population projections via the computer network to county
level personnel for rapid and accurate decision-making.


The authors gratefully acknowledge the technical assistance of
Dr. Jon Bartholic, Department of Fruit Crops, University of Flori-
da, the editorial assistance of Janet Snell, Susan Kynes, and Jose-
phine Burson; and critical reviews by Dr. Carlo Ignoffo, USDA, ARS,
Columbia, Missouri; Dr. Imre Otvos, Canadian Forestry Service,
St. Johns, Newfoundland; Drs. Leland Shanor and James Kim-
brough, Department of Botany, University of Florida; and the De-
partment of Entomology Review Committee, University of Florida.
The senior author expresses a special thanks to Dr. Leland Shanor,
Supervisory Committee Chairman, under whose direction the basic
biology of the pathogen was investigated.


Allen, G. E., G. L. Greene, and W. H. Whitcomb. 1971. An epizootic of
Spicaria rileyi on the velvetbean caterpillar, Anticarsia gemma-
talis, in Florida. Fla. Entomol. 54:189-91.

Barnes, G. and B. F. Jones. 1970. Control insects on soybeans. Ark. Agr.
Exten. Serv. Leaflet 193. 5 pp.

Charles, V. K. 1936. The synonomy of Botrytis rileyi Farlow. Mycologia
Cook, H. T. 1949. Forecasting late blight epiphytotics of potatoes and
tomatoes. J. Agr. Research 78:545-63.
S Harz, C. 0. 1871. Einige neue Hyphomyceten Berlin's und Wein's nebst
bertrage zur systematik derselben. Bull. Soc. Imp. Nat. Moscou.
Hirst, J. M. 1952. An automatic volumetric spore trap. Ann. Appl. Biol.
Ignoffo, C. M., D. L. Hostetter, P. Sikoroski, G. Sutter, and W. M. Brooks.
1977. Inactivation of representative species of entomopatho-
genic viruses, a bacterium, fungus and protozoan by simulated
sunlight. Environ. Entomol. (Submitted).
Ignoffo, C. M., N. L. Marston, D. L. Hostetter, B. Puttler, and J. V. Bell.
1976. Natural and induced epizootics of Nomuraea rileyi in soy-
bean caterpillars. J. Invertebr. Pathol. 27:191-98.
Ignoffo, C. M., B. Puttler, D. L. Hostetter, and W. A. Dickerson. 1975.
4 Seasonal incidence of the entomopathogenic fungus Spicaria rileyi
associated with noctuid pests of soybeans. J. Invertebr. Pathol.
Ignoffo, C. M., B. Puttler, D. L. Hostetter and W. A. Dickerson. 1976.
Susceptibility of the cabbage looper, Trichoplusia ni, and the
velvetbean caterpillar, Anticarsia gemmatalis, to several isolates
of the entomopathogenic fungus, Nomuraea rileyi. J. Invertebr.
Pathol. 28:259-62.
Johnson, D. W., L. P. Kish, and G. E. Allen. 1976. Field evaluation of
selected pesticides on the natural development of the entomo-
genous fungus, Nomuraea rileyi, on the velvetbean caterpillar in
soybean. Environ. Entomol. 5:964-66.
Kish, L. P. 1975. The biology and ecology of Nomuraea rileyi. Ph.D.
Dissertation, University of Florida. 83 pp.
Kish, L. P. and G. E. Allen. 1976. Conidia production of Nomuraea
rileyi on Pseudoplusia includes. Mycologia 68:436-39.
Kish, L. P., G. L. Greene, and G. E. Allen. 1976. A method of deter-
mining the number of potential conidia-forming cadavers of
Anticarsia gemmatalis infected with Nomuraea rileyi in a soy-
bean field. Fla. Entomol. 59:103-06.


Kish, L. P., R. A. Samson, and G. E. Allen. 1974. The genus Nomuraea
Maublanc. J. Invertebr. Pathol. 24:154-58.

Maublanc, A. 1903. Sur quelques especes nouvelles de champignons
inferieurs. Bull. Soc. Mycol. France 19:291-96.

McCubbin, W. A. 1943. Relation of spore dimension to their rate of fall.
Phytopathology. 34:230-34.

Menke, W. W. 1973. A computer simulation model: The velvetbean
caterpillar in the soybean agroecosystem. Fla. Entomol. 56:92-102.
Menke, W. W. and G. L. Greene. 1976. Experimental validation of a
pest management model. Fla. Entomol. 59:135-42.
Riddell, R. 1950. Permanent stained mycological preparations obtained
by slide cultures. Mycologia 42:265-70.
Saccardo, P. A. 1886. Hyphomycetes. Sylloge Fungorum 4:166.

Sawada, K. 1919. Catalogue of Formosan fungi. Agr. Expt. Sta. Govt.
of Formosa.
van der Plank, J. E. 1963. Plant diseases: Epidemics and control. New
York, Academic Press.
Waddill, V. H., B. M. Shepard, J. R. Lambert, G. R. Garner, and D. N.
Baker. 1976. Computer simulation model for populations of
mexican bean beetles on soybeans. So. Carolina Agr. Exp. Sta.
Bull. SB 590.

Waggoner, P. E. and J. G. Horsfall. 1969. EPIDEM-A simulator of
plant disease written for a computer. Bull. Conn. Agr. Exp. Sta.
New Haven, Conn.
Watson, J. R. 1915. Entomogenous fungi. Florida Agr. Exp. Sta. Ann.
Rept. 1914, pp. 46-48.


University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs