Title: Florida Entomologist
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00098813/00076
 Material Information
Title: Florida Entomologist
Physical Description: Serial
Creator: Florida Entomological Society
Publisher: Florida Entomological Society
Place of Publication: Winter Haven, Fla.
Publication Date: 1994
Copyright Date: 1917
Subject: Florida Entomological Society
Entomology -- Periodicals
Insects -- Florida
Insects -- Florida -- Periodicals
Insects -- Periodicals
General Note: Eigenfactor: Florida Entomologist: http://www.bioone.org/doi/full/10.1653/024.092.0401
 Record Information
Bibliographic ID: UF00098813
Volume ID: VID00076
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: Open Access

Full Text

(ISSN 0015-4040)


(An International Journal for the Americas)

Volume 71, No. 3 September, 1988


Preface ............................................... ............................................ 223
GUTHRIE, W. D.-The Contributions of Ferdinand F. Dicke ....................... 225
PASHLEY, D. P.-Current Status of Fall Armyworm Host Strains ................. 227
Preference, Mating Compatibility, and Development of Two Fall
Armyworm Strains ......................................................................... 234
WISEMAN, B. R., AND D. J. ISENHOUR-Feeding Response of Fall Armyworm
Larvae on Excised Green and Yellow Whorl Tissue ofResistant and Suscep-
tible Corn ................................................................................ 243
HRUSKA, A. J., AND S. M. GLADSTONE-Effect ofPeriod and Level ofInfestation
of the Fall Armyworm, Spodoptera frugiperda, on Irrigated Maize Yield 249
LYE, B-H., AND C. M. SMITH-Evaluation of Rice Cultivars for Antibiosis and
Tolerance Resistance to Fall Armyworm (Lepidoptera: Noctuidae) ........ 254
MIHM, J. A., M. E. SMITH, AND J. A. DEUTSCH-Development of Open-Polli-
nated Varities, Non-Conventional Hybrids and Inbred Lines of Tropical
Maize with Resistance to Fall Armyworm, Spodoptera frugiperda
(Lepidoptera: Noctuidae), at CIMMYT .......................................... 262
ALL, J. N.-Fall Armyworm (Lepidoptera: Noctuidae) Infestations In No-Tillage
Cropping Systems ........................................................................... 268
CASTRO, M., H. PITRE, AND D. MECKENSTOCK-Potential for Using Maize as
a Trap Crop for the Fall Armyworm, Spodoptera frugiperda (Lepidoptera:
Noctuidae), Where Sorghum and Maize are Intercropped on Subsistence
Farms .................................................. ............................... 273

Preface ..................................... .......................................................... 279
RINDERER, T. E.-Computer Assisted Identification of Hybrid Strains of West-
ern Honey Bees ........................................................................ 281
FINNERTY, V.-Ribosomal DNA Probes for Identification of Member Species of
the Anopheles gambiae Complex .................................................. 288
HALL, G. H.-Distinguishing African and European Honeybees Using Nuclear
DNA Restriction Fragment Polymorphosisms .................................. 294
Distinguish Sibling Species of the Anopheles quadrimaculatus Complex 299
NARANG, S. K., AND J. A. SEAWRIGHT-Electrophoretic Methodfor Recognition
of Sibling Species of Anopheline Mosquitoes. A Practical Approach ....... 303
KAISER, P. E.-Cytotaxonomy as a Tool for Identification of Siblings of the
Anopheles quadrimaculatus Complex ............................................... 311

Continued on Back Cover

Published by The Florida Entomological Society


President ........................................................ ................... R. S. Patterson
President-Elect ........................... .. .. .. ............... ...... J. E. Eger
Vice-President ........................................................ .................... J. F. Price
Secretary ........................................................ ......................... J. A. Coffelt
Treasurer ...................................................... .................... A. C. Knapp

J. L. Taylor
C. 0. Calkins
Other Members of the Executive Committee .............. F. Bennett
J. E. Pefia
N. Hinkle
M. F. Antolin
J. R. McLaughlin


Editor ................................................................................ J. R. McLaughlin

Associate Editors
Arshad Ali Carl S. Barfield Ronald H. Cherry
John B. Heppner Michael D. Hubbard Lance S. Osborne
John Sivinski Omelio Sosa, Jr. Howard V. Weems, Jr.
William W. Wirth

Business Manager .................. .. .. ... ..................... A. C. Knapp

FLORIDA ENTOMOLOGIST is issued quarterly-March, June, September, and De-
cember. Subscription price to non-members is $30 per year in advance, $7.50 per copy.
Membership in the Florida Entomological Society, including subscription to Florida
Entomologist, is $25 per year for regular membership and $10 per year for students.
Inquires regarding membership, subscriptions, and page charges should be addres-
sed to the Business Manager, P. O. Box 7326, Winter Haven, FL 33883-7326.
Florida Entomologist is entered as second class matter at the Post Office in DeLeon
Springs and Winter Haven, FL.
Manuscripts from all areas of the discipline of entomology are accepted for consider-
ation. At least one author must be a member of the Florida Entomological Society.
Please consult "Instructions to Authors" on the inside back cover. Submit the original
manuscript, original figures and tables, and 3 copies of the entire paper. Include an
abstract in Spanish, if possible. Upon receipt, a manuscript is acknowledged by the
Editor and assigned to an Associate Editor who sends it out for review by at least 3
knowledgeable peers. Reviewers are sought with regard only for their expertise; Soci-
ety membership plays no role in their selection. Page charges are assessed for printed
Manuscripts and other editorial matter should be sent to the Editor, JOHN R.
MCLAUGHLIN, 4628 NW 40th Street, Gainesville, FL 32606.

This issue mailed September 30, 1988

Fall Armyworm Symposium

Fall Armyworm Symposium


NOTE: Reprints of each paper of the Fall Armyworm Symposium are available from
the respective authors.
The 1988 Fall Armyworm Symposium is the fifth formal conference held in conjunc-
tion with the annual meeting of the Southeastern Branch of the Entomological Society
of America. We gratefully acknowledge the assistance of Wayne Gardner, University
of Georgia, Dept. of Entomology, Griffin, GA with reviewing and editing of papers.
This symposium is dedicated to Ferdinand F. Dickie. Mr. Dickie's long and distinguished
career has provided significant contributions to entomology in the areas of host plant
resistance, biology and ecology, and control methodologies.
Topics included in this year's symposium dealt with host plant resistance, population
dynamics, biology, management tactics, and genetics of the fall armyworm, Spodoptera
frugiperda (J. E. Smith). The annual symposium reflects the varied aspects of research
and the mutual sharing of data by scientists studying the fall armyworm. Allozyme loci
and mtDNA restriction enzyme profiles were reported to differ between two fall ar-
myworm strains. Strains also differed in physiological and developmental traits and
resistance to insecticides. The impact of these findings on research related to the fall
armyworm is far reaching and could significantly alter our thinking on the biology and
ecology of this pest and screening plants for insect resistance.
The continued interest and efforts of researchers in the Americas and Caribbean
Basin reflect the importance of the fall armyworm as a significant pest of agronomic
crops throughout this region.

Sharron S. Quisenberry David J. Isenhour
Department of Entomology Department of Entomology
Louisiana State University University of Georgia
Baton Rouge, LA Tifton, GA


Guthrie: Fall Armyworm Symposium



Corn Insects Research Unit
Agricultural Research Service, U.S. Department of Agriculture
Ankeny, Iowa 50021


Ferdinand F. Dicke's contributions to entomology have spanned more than sixty
years. Mr. Dicke has made significant contributions to entomology in the areas of host
plant resistance, biology and ecology, disease transmission by insects, chemical and
biological control, and methods for breeding for plant resistance to insects. Mr. Dicke's
long and distinguished career in the field of entomology is truly outstanding and worthy
of our recognition.


Por mas de sesenta afios Ferdinand F. Dicke ha contribuido a la Entomologia. El
sefor Dicke ha hecho contribuciones significantes en las areas de resistencia de plants
hospederas, en la biologia y ecologia, en la transmisi6n de enfermedades por insects,
en el control qufmico y biol6gico, y en el fitomejoramiento de plants resistentes a
insects. Mr. Dicke ha tenido una larga y distinguida carrera en el campo entomol6gico
que ha sido sobresalitente y digna de nuesto reconocimiento.

Mr. Ferdinand F. Dicke was born at New Bremen, Ohio, August 25, 1899. He
graduated from New Bremen High School in 1917. He received a B. Sc. degree from
Ohio State University in 1927 with a major in entomology and botany. From 1938
through 1942, he attended graduate school (at night) at George Washington University,
majoring in plant physiology and mycology.
Mr. Dicke's research experience with the United States Department of Agriculture
began at Monroe, Michigan (1927-1929) where he worked on chemical control, cultural
control, and varietal resistance of the European corn borer. From 1930-1933 he worked
on biology and ecology and chemical control of the corn earworm at Charlottesville,
He worked at Arlington, Virginia, and Beltsville, Maryland, on several research
problems from 1933-1942. Research on the corn earworm included biology and ecology
and varietal resistance of dent corn, disease transmission by insects, vectors and trans-
mission of Stewart's bacterial wilt in corn, biology of corn flea beetles, potato leafhopper
on alfalfa and peanuts, entomogenous fungi, sericulture, and collection of varieties of
silkworms in the U.S.
Mr. Dicke worked on varietal resistance to the European corn borer at Toledo, Ohio
from 1942-1950. He was project leader in the development of corn genotypes for resis-
tance to the European corn borer at Ankeny, Iowa from 1950-1963. During this period
he was also Professor, Department of Zoology and Entomology at Iowa State University
and provided counsel and direction for many graduate students. On July 6, 1963 (after
37 years of research) he retired from USDA and started a 20-year career with Pioneer
Hi-Bred International, Inc., Johnston, Iowa. He served as a consultant on insect and
diseases in Pioneer's domestic and overseas research program to improve varieties of
corn, sorghum, soybeans, alfalfa, cotton, wheat, and sunflowers. His major activities
with Pioneer consisted of evaluating crop germplasm for insect and disease resistance.

Florida Entomologist 71(3)

September, 1988

From 1984 to the present, he has been Research Collaborator with USDA-ARS-Iowa
State University at Ankeny, Iowa.
During Mr. Dicke's service with the U. S. Department of Agriculture he made
significant research contributions in the areas of host plant resistance to insects, biology
and ecology, disease transmission by insects, chemical control, biological control, and
methods of breeding for plant resistance to insects. His primary research has been on
cereal and forage crop insects. Some specific accomplishments are as follows: (1) Early
contributions on the effectiveness and phytotoxicity of insecticides when used against
the European corn borer, and reactions of the corn borer to a variety of host plants.
(2) Biology of the wheat jointworm with studies on the development of gall tissue. (3)
New interpretations on the seasonal populations of the corn earworm and factors affect-
ing winter survival. (4) Identification and evaluation of relative resistance of inbred
lines of corn against the corn earworm. Several of the lines found to be resistant are
now used in hybrids and are also used in developing synthetic varieties. (5) Identification
of new sources of resistance against the European corn borer through the development
of new resistant inbred lines from single crosses that had susceptible lines in their
parentage, demonstrating transfer and intensification of resistance; several of the lines
have been released by the Iowa and Minnesota stations and are now used in commercial
hybrids approved on a regional basis. He has made important contributions to an under-
standing of the genetics of resistance. (6) Important contributions to the large scale
production of corn borer egg masses in the laboratory and their efficient use for artifi-
cially infesting corn in resistance and other experimental work. (7) Detailed studies on
the biology of the corn borer on susceptible and resistant strains of corn leading to the
development of important time-saving visual methods for identifying and evaluating
resistance before pollination allowing effective selection in segregating plant populations
and greatly reducing pollination work and progeny testing. These methods have signific-
antly enhanced the studies on the genetics of resistance and breeding for resistance. (8)
Important contributions in a biological and statistical study demonstrating a highly
significant correlation between visual numerical leaf damage ratings, sheath and midrib
lesions, stalk cavities, and surviving larvae and the use of these criteria in connection
with corn breeding as well as with chemical, biological and other research. This work
has brought about significant savings in field work and has permitted an increase in the
scope of the program.
Mr. Dicke received the U. S. Department of Agriculture Superior Service Award
on June 5, 1956, while employed at the European Corn Borer Research Laboratory,
USDA-ARS, Ankeny, IA. This award was given for his outstanding contributions in
the field of agriculture, especially in connection with his work on corn borer resistant
varieties of corn. He was featured in 1957 in a Northrup-King and Co., Minneapolis,
MN, publication because of his accomplishments in developing corns that are resistant
to the European corn borer. He received the Alumni Member Award of Merit from
Gamma Sigma Delta, Iowa Beta Chapter, April 1975. On March 24, 1982, he received
the North Central Branch ESA Award of Merit. The CIMMYT Maize Program and the
HPR Symposium participants dedicated the International Symposium on Methodologies
for Developing Resistance to Maize Insects (March 16-20, 1987) to Mr. Dicke in recog-
nition of his long and distinguished career in the field.


Pashley: Fall Armyworm Symposium 227


Department of Entomology
Louisiana State University
Baton Rouge, Louisiana 70803


Research on two host strains of the fall armyworm (FAW), Spodoptera frugiperda
[J. E. Smith], is reviewed. They differ at allozyme loci and mtDNA restriction enzyme
profiles, in their development on a variety of plants, and in their resistance to insec-
ticides. Data on interstrain matings indicate unidirectional incompatibilities in the labo-
ratory and strong preferences among males for females of their own strain in nature.
Further study is required to clarify precise strain affinities for the long list of FAW
host records. Information on phenologies and geographic distributions is presented but
clearly in need of in-depth study. Those studying FAW are recommended to collect from
a single host, to keep host records with laboratory colonies, to replenish colonies with
individuals from the same host, and to establish a set of voucher specimens for each
collection or study.


Se revisa el studio sobre dos tipos de hospederos del gusano cogollero, Spodoptera
frugiperda (J. E. Smith). Ellos difieren en el loci de alozyme y en la restricci6n del
mtDNA de las enzymas, en su desarrollo en various tipos de plants, y en su resistencia
a insecticides. Datos sobre cruce de razas indican incompatibdades no-direccionales en
el laboratorio, y una gran preferencia entire los machos hacia hembras de su propia raza
bajo condiciones naturales. Se require un studio adicional para aclarar las afinidades
de las razas hacia la larga lista de hospederos del gusano cogollero. Se present informa-
ci6n sobre la fenologia y distribuci6n geogrffica, pero esta claramente necesitada de un
studio a fondo. Se recomienda a aquellos que estudian el gusano cogollero, que lo
coleccionen de un solo hospedero, que mantengan un registro del hospedero de las
colonies de laboratorio, que reemplacen las colonies con individuos del mismo hospedero,
y que establescan un grupo de esp6cimenes justificante de cada colecci6n o studio.

The fall armyworm (FAW), Spodoptera frugiperda [J. E. Smith], has been sub-
divided recently into two strains on the basis of genetic differentiation associated with
larval host plants (Pashley et al. 1985, Pashley 1986). One strain feeds primarily on corn
(Zea mays L.) and the other on rice (Oryza sativa L.) and various forage and native
grasses. The strains may represent one of three types of taxa. They may be biotypes
in which genetic differences are due to a selectively-mediated polymorphism within a
single randomly-mating species. They may be host races in the initial stages of specia-
tion in which interbreeding is reduced due to host preference differences (Diehl & Bush
1984). Lastly, they may be sibling species that are either capable of hybridizing to a
limited degree or completely reproductively isolated.
In this paper, I review research relevant to the status of these strains, elaborating
on some of the more problematic areas. I also emphasize the impact that these taxa
have on our current thinking about FAW biology and end with some recommendations
for those studying FAW.

Florida Entomologist 71(3)


FAW has been reported on more than 80 plant species in 23 families (Table 1). It is
primarily regarded, however, as a pest of grasses (Poaceae). The two host strains
appear to utilize different sets of hosts (Fig. 1). The corn strain feeds primarily on corn,
cotton (Gossypium hirsutum L.), and sorghum (Sorghum vulgare Pers.), but occasion-
ally will feed on other plants either growing in proximity to primary hosts (such as
signal grass, Brachiaria platyphylla (Nash), [LA 8/85 in Fig. 1]), or in isolation
(LA(BH) 8/85 in Fig. 1). The rice strain has been reported on rice, signal, bermudagrass
(Cynodon dactylon [Pers.]), and Johnson grass (Sorghum halepense [Pers.]). It has
been suggested that early season FAW feed on corn and then shift to forage grasses
after corn is harvested in the south (Morrill 1978). Genetic data do not support this


1. Chenopodiaceae

2. Amaranthaceae
3. Portulacaceae
4. Polygonaceae
5. Malvaceae

6. Violaceae
7. Caricaceae
8. Cucurbitaceae

9. Brassicaceae

10. Rosaceae

11. Fabaceae

12. Myrtaceae
13. Vitaceae
14. Rutaceae
15. Apocynaceae
16. Solanaceae

17. Convolvulaceae
18. Asteraceae

19. Cyperaceae
20. Poaceae

Chenopodium album L.* (lambs quarters); Spinacia oleracea
L.* (spinach); Beta vulgaris L.* (sugar beets)
Amaranthus spp. including spinosus L.' (pigweed)
Portulaca oleracea L.* (purslane)
Fagopyrum esculentum Moench* (buckwheat)
Althaea rosea Cav.* (hollyhock); Gossypium hirsutum L. (up-
land cotton)
Viola spp. (violets)
Carica papaya L.b (papaya)
Citrullus vulgaris Schrad.* (watermelon); Cucumis sativus L.*
(cucumber); Cucurbita pepo L.c (pumpkin)
Brassica napobrassica Mill.* (rutabaga); B. oleracea L.* (cab-
bage, kale); B. rapa L.* (turnip)
Fragaria chiloensis Duchesue (strawberry); Prunus persica
Batsch* (peach); Malus pumila Mill.* (apple)
Arachis hypogaea L. (peanut); Cajanus cajun Millsp.* (pigeon
pea); Cicer arietinum (L.)* (chick pea); Glycine max Merr.*
(soybean); Medicago sativa L.* (alfalfa); Pisum sativum L.*
(pea); Pueraria lobata Ohwi* (kudzu); Stizolobium deerin-
gianum Bort.* (velvet bean); Trifolium pratense L.* (red
clover); T. repens L.* (white clover); Vigna unguiculata Wal-
pers* cowpeaa)
Eucalyptus camaldulensis Dehuh.*b
Vitis spp. (grape)
Citrus aurantium L.* (seville orange)
Plumeria rubra L.d
Capsicum spp.b (red peppers); Lycopersicon esculentum Mill.
(tomato); Nicotiana tabacum L. (tobacco); Solanum tuberosum
L. (potato)
Ipomoea spp. (wild morning glory)
Lactuca sativa L.*e (lettuce); Xanthium italicum Moretti
Cyperus rotundus L.* (nut-grass)
Agrostis hyemalis (Walt.) (spring bentgrass); A. stolonifera L.
(redtop bentgrass); Andropogon virginicus L. (broomsedge);
Avena sativa L.* (oats); Axonopus affinis Chase (carpet grass);

September, 1988

Pashley: Fall Armyworm Symposium

TABLE 1. (Continued)


21. Liliaceae

22. Iridaceae
23. Pinaceae

Brachiaria platyphylla (Nash)' (signal grass); Cenchrus
tribuloides L. sandspurr grass); Chloris gayana Kunth* (Rhodes
grass); Cynodon dactylon (Pers.)* (bermudagrass); Dactyloc-
tenium aegyptium (Beauv.)* (crowfoot grass); Digitaria san-
guinalis (Scop.)* (crabgrass); Hordeum vulgare L.* (barley);
Oryza sativa L.* (rice); Panicum maximum Jacq.* (guinea
grass); P. miliaceum L.* (broomcorn millet); P. purpurascens
Raddig (paragrass); P. texanum Buckl. (Texas millet); Pen-
nisetum glaucum (R. Br.)* (pearl millet); P. purpureum
Schum.*h (Napier grass); Phleum pratense L.* (timothy grass);
Poa pratensis L.* (Kentucky bluegrass); Saccharum of-
ficinarum L.* (sugarcane); Secale cereale L.* (rye); Sorghum
sudanense Hitch.* (sudan grass); S. vulgare Pers.* (sorghum);
S. halepense (Pers.)* (Johnson grass); Triticum aestivum L.*
(wheat); Zea mays L. (corn); Z. mexicana Reeves & Man-
gelsdorf (teosinte)
Allium cepa L.* (onion); A. sativum L.*e (garlic); Asparagus
officinalis L.* (asparagus)
Gladiolus spp.*b
Pinus caribaea Morelet' (Caribbean or slash pine)

aChereguino & Menendez 1975; bBruner et al. 1975; cPretto 1970; dSantiago-Blay 1983; eMcGuire & Crandall 1967;
fD. P. Pashley, pers. obs.; "Ashley et al. 1983; hpiedra 1974; 'Howell 1979


LA 8/85
PR 8/84


LA 8/86
PR 2/85 [CORN
LA 8/84
LA 8/83
LA(BH) 8/8a-
LA(LSU) s8/8
,PR 2/85 -RI
PR 8/84 __
SLA(HAM) 8/881


0.20 0.10


Fig. 1. Genetic relationships among wild FAW collected from different host plants.
Genetic distances are based on three polymorphic loci used to separate strains esterasee,
hydroxybutyrate dehydrogenase, and peptidase). PR = Puerto Rico, LA = Louisiana
(sites in parentheses are different locations in LA), HO = Honduras.


Florida Entomologist 71(3)

Data on the phenology of strains in Louisiana indicate that the rice strain is present
throughout much of the year (D. P. Pashley, T. N. Hardy & A. M. Hammond, unpub-
lished data). Populations remain at low density on various grasses until a rapid buildup
that generally occurs in late summer. Corn strain individuals are also present from at
least late spring through December but densities peak in early to mid-summer prior to
corn harvest. At present it is unclear where these individuals spend the rest of summer.
They may migrate northward following developing corn, shift to some other locally
available host, die because no hosts are available, or, although unlikely, enter some sort
of summer aestivation.
The geographic distribution of the corn strain is well documented (Pashley et al.
1985, Pashley 1986). It occurs throughout much of the western hemisphere, generally
in most places where corn is grown (Fig. 2). Less is known about the distribution of
the rice strain because very little sampling has been done on hosts other than corn. It
occurs sympatrically with the corn strain, often in neighboring fields, in Puerto Rico,
southern Florida, Georgia, and Louisiana. Reported attacks on forage grasses through-
out much of Latin America are more than likely due to the presence of the rice strain.
Future collaboration with scientists from these localities will allow a more precise
characterization of the rice strain's range.


Populations of FAW sampled from corn exhibit significant differences at five al-
lozyme loci from populations collected from rice or bermudagrass (Pashley 1986). Cross-
rearing experiments in which larvae were reared on their own and the other strain's
host indicated that differences were not due to different selective regimes of the host
environment (Pashley 1986).
Although it is possible to use genotypes from multiple allozymes to type individuals
to strain with a high degree of accuracy, there are no diagnostic loci. Because of that,
an independent genetic marker has recently been examined for use in strain identifica-
tion. Mitochondrial DNA (mtDNA) has been isolated from individuals of each strain and
restriction enzyme patterns at 35 enzymes have been examined (D. P. Pashley, unpub-
lished data). Preliminary data indicate that major strain differences exist at two restric-
tion enzymes (Table 2), and possibly at two others (MboI and MboII; D. P. Pashley,
unpublished data). Like the allozyme data, genotypes are strongly associated with hosts;
the small amount of overlap (e.g., 2 rice collected individuals with AA and 2 corn
collected individuals with BB in Table 2) is probably due to a small number of individuals
within each strain that use the other strain's host. Further analysis and geographic
sampling are required to determine if differences in mtDNA are diagnostic. Host-as-
sociated differences at these two very independent markers (allozymes and mtDNA)
are strong evidence for the existence of sibling species.
Many developmental differences have been reported between the strains (Pashley
et al. 1987a, Pashley 1988, Whitford et al. 1988). Rates of development and pupal
weights tend to differ consistently. The only adult character that differs significantly is
preoviposition period (D. P. Pashley, T. N. Hardy & A. M. Hammond, unpublished
data). In addition, the rice strain appears to be physiologically adapted to its hosts. It
performs well on bermudagrass and rice but poorly on corn, whereas the corn strain is
less influenced by larval host. The rice strain is probably the more specialized of the
two strains.
Reproductive incompatibilities have been reported in interstrain matings performed
in the laboratory (Pashley & Martin 1987). Rice strain females mated successfully with
corn strain males but matings in the reciprocal cross (corn females with rice males) did
not occur. In backcrosses using individuals from the successful cross, only F, hybrid


September, 1988

Pashley: Fall Armyworm Symposium

'U '.r

d ~

Fig. 2. Collection sites of FAW strains. Circles indicate locations where the corn
strain has been collected. Stars indicate locations where both corn and rice strains

males mated with parentals while Fi hybrid females would not. In a subsequent study
(Whitford et al. 1988), strains successfully mated in both directions. In the two exper-
ments reported by Pashley & Martin (1987), material was relatively freshly collected
(first and second lab generations in one case and fourth and fifth in the other). Whitford
et al. (1988) used colonies at least three years old. Incompatibilities may be due to
behavioral characteristics that are altered during the selection process that accompanies
colonization. Future studies should focus on newly collected material to clarify this

232 Florida Entomologist 71(3) September, 1988


Hinfl BstNI Hinfl/BstNI

Corn 16 2 14 2 1 1 14 1 1 2
Rice 2 13 2 13 0 0 2 0 0 13

The detection of incompatibilities in laboratory mating studies is informative because
it probably reflects conditions in nature (particularly when intrastrain breeding is
routinely successful under laboratory conditions). However, when taxa are interfertile
in a laboratory environment, it cannot be assumed that mating occurs in nature (Mayr
1970, Diehl & Bush 1984). Therefore, definitive tests of interbreeding must involve
studies of cross-attraction and mating in the field. Preliminary data from several ongo-
ing studies suggest that although males are attracted preferentially to females of their
own strain, some cross-attraction occurs. In Louisiana and Georgia, females of each
strain have been placed in traps and captured males identified to strain using elec-
trophoresis. In Louisiana, most rice strain males (119/166 = 72%) were attracted to
rice strain females and most corn strain males (12/16 = 75%) were attracted to corn
females (D. P. Pashley, A. M. Hammond & T. N. Hardy, unpublished data). Sample
sizes are smaller in Georgia but roughly the same percentages, 75% fidelity for the rice
strain and 86% for the corn strain, were observed (J. E. Carpenter & D. P. Pashley,
unpublished data). Similarities between these separate studies suggest that either a
pheromonal or calling time difference facilitates assortative mating within strains. Cur-
rent efforts are being focused on the analysis of captured field-mated pairs and observa-
tions at mating tables to determine if the attraction of males to females of the opposite
strain actually results in interstrain mating (J. E. Carpenter & D. P. Pashley, unpub-
lished data).
Data have been gathered or studies initiated on several other features of the biology
of the two strains. They have been reported to exhibit differences in resistance to
several insecticides (Pashley et al. 1987b) and differences in survival and development
on selections of bermudagrass (Pashley et al. 1987a). Morphological differences at traits
in both larval and adult stages have been detected (A. M. Hammond, unpublished data;
D. P. Pashley & E. R. Taylor, unpublished). So far, however, no differences have been
found in pheromone chemistry (A. M. Hammond & H. Fescmeyer, unpublished data).


Differences have been detected between strains at almost every character examined
to date. It is unlikely that selective regimes on the two hosts could be maintaining
differences at all of these independent characters. The strains are probably not inter-
breeding to a large degree or differences would quickly disappear. Therefore, these
taxa are not biotypes and represent either host races or sibling species. Because the life
history characteristics of FAW are not in any way similar to those proposed for host
races (Bush 1975a, 1975b), they are probably sibling species. When studies of reproduc-
tive isolation in nature are complete, the correct taxonomic status of FAW host strains
will be resolved.

Pashley: Fall Armyworm Symposium


It should be clear that much is yet to be learned about the biology of these two
host-associated strains. Their taxonomic status is only a technical point. Differences
exhibited at many different traits should cause those involved in research on these taxa
to be extremely cautious. Material for study should be collected from a single host plant
species and records of those collections should be kept with laboratory colonies. Colonies
should not be replenished with collections from other hosts. Voucher specimens should
be set aside for every collection made in the event that morphological differences can
eventually be used for strain identification. If there is access to an ultra-cold (-70C)
freezer, a sample should be frozen for electrophoretic analysis, if necessary. If these
suggestions are followed, research results will not be invalidated if the strains are
concluded to be separate species.


Research on FAW biology was supported by USDA Grant 86-CRCR-1-2027. The
manuscript was improved by comments from David Pashley and Bruce McPheron.


Parasitization of fall armyworm larvae on volunteer corn, Bermuda grass, and
paragrass. Florida Entomol. 66: 267-271.
BRUNER, S., L. SCARAMUZZA, AND A. OTERA. 1975. Catalogo de los insects que
atacan a las plants economics de Cuba. Academia de Ciencias de Cuba. La
Habana, Cuba. 393 pp.
BUSH, G. L. 1975a. Modes of animal speciation. Annu. Rev. Ecol. Syst. 6: 334-364.
BUSH, G. L. 1975b. Sympatric speciation in phytophagous parasitic insects, pp. 187-
206. In P. W. Price (ed.), Evolutionary strategies of parasitic insects and mites.
Plenum, New York.
CHEREGUINO, R. S., AND A. L. MENENDEZ. 1975. Biologia y habitos del gusano
cogollero (Spodoptera frugiperda) en El Salvador. Memoria de la XXI Reunion
de PCCMCA. San Salvador, El Salvador. pp. 251-261.
DIEHL, S. R., AND G. L. BUSH. 1984. An evolutionary and applied perspective of
insect biotypes. Annu. Rev. Entomol. 29: 471-504.
HOWELL, H. N. 1979. Fall armyworm (Spodoptera frugiperda) in a pine (Pinus)
nursery in Honduras, C. A. Ceiba 22: 35-37.
LUGINBILL, P. 1928. The fall armyworm. USDA Tech. Bull. No. 34. 92 pp.
MAYR, E. 1970. Populations, species, and evolution. Harvard Univ. Press, Cam-
bridge, Mass.
McGUIRE, J. U., AND B. S. CRANDALL. 1967. Survey of insect pests and plant
diseases of selected crops of Mexico, Central America, and Panama. USDA/U.S.
Agency for International Development, Washington, D.C. 157 pp.
MORRILL, W. L. 1978. Georgia grasslands: reservoirs for beneficial and destructive
insects. Georgia Agr. Res. (spring issue): 25-28.
PASHLEY, D. P. 1986. Host associated genetic differentiation in fall armyworm: a
sibling species complex? Ann. Entomol. Soc. Am. 79: 898-904.
PASHLEY, D. P. 1988. Quantitative genetics, development and physiological adapta-
tion in sympatric host strains of fall armyworm. Evolution 42: 93-102.
PASHLEY, D. P., S. J. JOHNSON, AND A. N. SPARKS. 1985. Genetic population
structure of migratory moths: the fall armyworm (Lepidoptera: Noctuidae). Ann.
Entomol. Soc. America 78: 756-762.
PASHLEY, D. P., AND J. A. MARTIN. 1987. Reproductive incompatibility between
host strains of fall armyworm (Lepidoptera: Noctuidae). Ann. Entomol. Soc.
America 80: 731-733.


234 Florida Entomologist 71(3) September, 1988

PASHLEY, D. P., S. S. QUISENBERRY, AND T. JAMJANYA. 1987a. Impact of fall
armyworm (Lepidoptera: Noctuidae) host strains on the evaluation of Bermuda
grass resistance. J. Econ. Entomol. 80: 1127-1130.
1987b. Two fall armyworm strains feed on corn, rice and Bermuda grass.
Louisiana Agriculture 30: 8-9.
PIEDRA, F. 1974. Effects of different forage diets on the biology of Spodoptera
frugiperda (Lepidoptera: Noctuidae). Cuban J. Agric. 8: 99-103.
PRETTO, M. R. 1970. Evaluacion de dos insecticides para el control del gusano talad-
rador del cuello del maiz, Spodopterafrugiperda. Boletin Num. 9. MAG Panama.
13 pp.
SANTIAGO-BLAY, J. A. 1983. Plumeria rubra: A new host plant record of the fall
armyworm (Lepidoptera: Noctuidae) in Puerto Rico. Florida Entomol. 66: 359.
compatibility, ovipositional preference, and larval development of two elec-
trophoretically differentiated fall armyworm colonies. Florida Entomol. 71: 234-


Department of Entomology
Louisiana Agricultural Experiment Station
Louisiana State University Agricultural Center
Baton Rouge, Louisiana 70803


Host strains (corn and rice) of the fall armyworm (FAW), Spodopterafrugiperda (J.
E. Smith), were evaluated for oviposition preference, mating compatibility, and de-
velopment on artificial diet and four plant species. Both strains oviposited a greater
percentage of egg masses on corn (Zea mays L.), sorghum (Sorghum bicolor [L.]
Moench.), and bermudagrass (Cynodon dactylon [L.] Pers.) than on centipedegrass
(Eremochola ophiuroides [Munro] Hack). The corn strain oviposited preferentially on
corn and sorghum, while the rice strain preferred bermudagrass. The number of fertile
intrastrain and interstrain pairs ranged from 77 to 100%. Egg hatch was extremely high
for all crosses (>80%). The F1 (rice ? x cornd) interhybrid cross had fewer pairs mating
and a decreased number of egg masses per female than the F1 (corn ? x rice S) interhyb-
rid cross. The corn and rice strain progeny had similar larval weights and survivorship,
but the two strains differed significantly in pupal weights and rates of development on
modified pinto bean diet. The corn strain developed equally well on corn, bermudagrass
and sorghum, whereas the rice strain developed best on bermudagrass. On these hosts
the corn strain had significantly heavier larvae and pupae than rice strain feeding on
the same hosts. A discussion is provided that delineates problem areas associated with
bermudagrass/FAW resistance studies.


Se evaluaron razas hospederas (maiz y arroz) del gusano cogollero, Spodoptera
frugiperda (J. E. Smith), sobre su preferencia oviposicional, compatibilidad de
apareamiento, y su desarrollo en dieta artificial y en cuatro species de plants. Ambas

Whitford et al.: Fall Armyworm Symposium

razas pusieron un mayor percentage de las masas de huevo en el maiz (Zea mays L.),
sorgo (Sorghum bicolor [L.] Moench.), y en la hierba-bermuda (Cynodon dactylon [L.]
Pers.) que en la hierba-centipede (Eremochola ophiuroides [Munro] Hack). La raza del
maiz preferentemente oviposit6 en el maiz y el sorgo, mientras que la raza del arroz
prefiri6 la hierba-bermuda. El nfmero de parejas frrtiles dentro de la raza y entire razas
vari6 del 77 al 100%. La salida del cascar6n fue extremadamente alta en todos los cruces
(>80%). El F1 (arroz X maizd) del cruce entire hibridos tuvieron un menor ndmero
de apareamientos y una disminuci6n en el nfimero de masas de huevos por hembra que
el F, (maiz9 X arrozd) del cruce entire dos razas. La prole de la raza del maiz y del
arroz tuvieron larvas con peso y sobreviviencia similares, pero las dos razas fueron
significativamente diferentes en el peso de las pupas y en el grado de desarroollo en la
dieta modificada del frijol pinto. La raza del maiz se desarroll6 igualmente en el maiz,
la hierba-bermuda y en el sorgo, mientras que la raza del arroz se desarroll6 mejor en
la hierba-bermuda. En estos hospederos la raza del maiz tuvo significativamente larvas
mas pesadas que las razas del arroz comiendo del mismo hospedero. Se discuten prob-
lemas que delinean Areas problemAticas asociadas con studios de resistencia sobre la
hierba-bermuda y los gusanos cogolleros.

The fall armyworm (FAW), Spodoptera frugiperda (J. E. Smith), is an important
agricultural pest on graminaceous crops in the southeastern United States (Luginbill
1928). The annual influx of FAW moths in early spring is generally assumed to originate
from overwintering adults along the Gulf Coast or from long distance movements from
Central and South America (Mitchell 1979, Sparks 1979, Johnson 1988).
Entomologists have assumed that one interbreeding species of FAW exists through-
out the Western Hemisphere; however, Pashley (1986) presented electrophoretic evi-
dence that FAW populations were actually composed of two genetically differentiated
host-associated strains. One strain feeds primarily on corn (corn strain), while the other
feeds on rice or bermudagrass (rice strain). Pashley & Martin (1987) indicated that
pre-reproductive isolating mechanisms exist between the female of the corn strain and
the male of the rice strain. These reproductive isolating mechanisms suggest that the
one recognized species of FAW is actually composed of one or more sibling species
(Pashley & Martin 1987).
The biological criteria presently used to evaluate bermudagrass, Cynodon dactylon
(L.) Pers., genotypes for resistance to FAW are insect development, larval consumption
and utilization, and survivorship (Lynch et al. 1983, Quisenberry & Wilson 1985, Jam-
janya & Quisenberry 1988, Quisenberry & Whitford 1988). The host strain used in the
screening process can influence the selection of bermudagrass genotypes resistant to
FAW (Pashley et al. 1987). Physiological differences between the corn and rice FAW
strains have been identified and include differences in growth and development on
artificial diets (Quisenberry & Whitford 1988), on each other's host plants (Pashley
1988), and on bermudagrass genotypes (Lynch et al. 1983, Pashley et al. 1987, Quisen-
berry & Whitford 1988). The inadvertent use of different FAW strains in screening
trials makes biological data very difficult to assess. This further complicates our ability
to compare research results between different laboratories studying bermudagrass
genotype resistance to the FAW.
The purpose of this research was to investigate the oviposition preference, mating
compatibility, and development of intrastrain, interstrain, and interhybrid crosses on
artificial diet and four host plants of the two FAW strains (corn and rice).


Fall Armyworm Sources and Colony Maintenance. The two FAW colonies used in
this study had previously been identified electrophoretically as the rice (Pashley et al.

Florida Entomologist 71(3)

1987) and corn (Mason et al. 1987, Pashley et al. 1987) strains. The rice strain was
initially field collected as larvae on bermudagrass from East Feliciana Parish in
Louisiana (August 1984). The corn strain was established from eggs obtained from corn
by F. M. Davis, ARS-USDA, Crop Science Research Laboratory, Mississippi State,
Mississippi, and in 1986 was crossed with feral males obtained by rearing egg masses
collected on corn, Zea mays L. Both strains were maintained on a modified pinto bean
diet according to procedures by Perkins (1979) for one generation to remove any con-
founding effects of the artificial diet (Quisenberry & Whitford 1988). Strains were main-
tained in a growth chamber at 26.7 + 0.5 C, 14:10 (L:D) photoperiod, and > 50% RH.
Oviposition Preference Study. All plants were grown in plastic pots (1.89 liter) that
contained potting soil (Fisons Western, Vancouver, B.C., Canada). Corn and sorghum,
Sorghum bicolor [L.] Moench, plants were thinned after germination to 4 and 6 plants
per pot, respectively. Plants were in the 5-leaf stage when both trials were conducted.
Two wk prior to use, vegetatively propagated bermudagrass and centipedegrass,
Eremochola ophiuroides (Muro) Hack, were clipped (7-cm) and fertilized with N-P-K.
One pot each of corn, sorghum, bermudagrass, and centipedegrass was placed
equidistantly in a galvanized metal container (52 liter) with vermiculite used as a sub-
strate around the pots. A plastic pipe (13 mm diam.) cage frame (48 by 48 by 92 cm)
was inserted into the vermiculite, and a fine mesh screen (0.5 mm) placed over the
frame. Four replicates (cages) were established per treatment (FAW strain) and re-
peated in two trials.
Pupae were placed in paper cartons (3.78 liters) and allowed to emerge. Each carton
contained one hundred (trial 1) and 120 (trial 2) FAW (19:ld ratio) pupae from each
strain. Two days after the emergence of the first adult, one carton of adults from one
of the two strains was released into each cage. During the first trial, adult moths were
released on 19 June and 21 June for the rice and corn strain, respectively, and on 18
July (rice strain) and 19 July (corn strain) in the second trial. Releases were made
between 8 and 9 am. An adult diet (honey-beer: honey [150 ml] + distilled water [150
ml], + beer [335 ml] + ascorbic acid [12 g]) and water were provided to moths. After
a 4-day oviposition period, each plant was examined and the number of egg masses per
plant recorded.
A completely randomized design with four replicates per treatment (FAW strain)
was used in each trial. Data were analyzed using analysis of variance, and significant
means separated by orthogonal contrasts (SAS Institute 1985) and Duncan's multiple
range test (Duncan 1955). Data were analyzed within each trial and strain because adult
emergence differed between strains, and the number of adults used were different
between trials.
Mating Compatibility Study. Male and female pupae from the rice and corn strains
were placed individually in polystyrene cups (29.7 ml). The underside of the cup lid was
fitted with moistened cellulose cotton wadding to help maintain humidity. At
emergence, pairs (corn x cornd, corn x rice, rice? x rice rice? x cornd) of
virgin FAW adults were placed in oviposition chambers (946 ml) containing moistened
vermiculite and an ovipositional substrate (cheesecloth top). An adult diet (honey-beer)
and water were provided in wicked polystryrene cups (29.7 ml). Egg masses were
collected daily from the cheesecloth and weighed. Egg mass weights were adjusted by
subtracting the weight of the cloth material. The total number of eggs from three egg
masses in each of five weight categories (5-10, 11-15, 16-20, 21-25, 26-30 mg/egg mass)
was used to determine the relationship between the number of eggs per mass (y) and
egg mass weight (x). Regression analyses (y = a + bx) were performed separately for
each intrastrain and interstrain cross to estimate the number of eggs per egg mass.
Egg masses laid on the sides of the ovipositional chambers were counted but not
weighed. A completely randomized design was used with mating type as the treatments.

September, 1988

Whitford et al.: Fall Armyworm Symposium

Treatments were replicated 10 to 18 and 13 to 15 times in trials 1 and 2, respectively.
The parameters measured were the number of egg masses per female, egg mass weight,
percent hatch, and the number of eggs per mass. Data were subjected to analysis of
variance and significant means separated by least-squares means (SAS Institute 1985).
Treatments were maintained in a growth chamber at 26.7 0.5C, 14:10 (L:D) photo-
period, and > 50% RH.
F1 interhybrid larvae were placed individually in polystyrene cups (29.7 ml) contain-
ing a modified pinto bean diet. Pupae were sexed and held according to the procedures
described above. A completely randomized design was used with the treatments consist-
ing of two interhybrid crosses. Each treatment was replicated 11 and 18 (rice ? x corn )
and 5 and 26 (corn x riced) times in trials 1 and 2, respectively. Data collected were
egp mnass number per female, percent hatch, egg mass weight, and number of eggs per
mass. Data were analyzed using analysis of variance, and significant means separated
singg the F test (SAS Institute 1985).
Development on Artificial Diet. One hundred pupae (1 :1 l ratio) for each intrast-
-ain and interstrain cross were placed in individual paper cartons (3.78 liter) that con-
tained moistened vermiculite, an adult diet (honey-beer), and a cheesecloth top. Egg
masses from the cheesecloth top were collected daily, placed in plastic bags, and allowed
to hatch. Newly emerged neonate larvae of the four mating combinations were placed
individually on a modified pinto bean diet in polystyrene cups (29.7 ml). Treatments
were maintained in a growth chamber at 26.7 0.50C, 14:10 (L:D) photoperiod, and >
50% RH.
A completely randomized design with six replicates per cross was used in two trials.
A replicate consisted of measurements averaged within a subsample of five larvae.
Developmental parameters measured were 9-day larval weights, pupal weights, larval
and pupal duration, and survivorship to adult eclosion. Data were subjected to analysis
of variance and significant means separated by least-squares means (SAS Institute
Influence of Host Plant on Development. All grasses were grown in metal flats (50
x 35 x 9 cm) containing potting soil (Fisons Western, Vancouver, B.C., Canada). The
bermudagrass ('Grazer') and centipedegrass (common) were cut (7 cm) 2 wk prior to
feeding and fertilized with N-P-K. Hybrid field corn (Pioneer Brand 3055), and hybrid
grain sorghum (Funk's 522DR) seeds were planted 10 days prior to feeding. Plants were
maintained under greenhouse conditions at 29 5C, 16:8 (L:D) photoperiod, and >
50% RH.
Neonate larvae of the corn or rice strain were placed individually in a clear petri
dish (10 x 2 cm) that contained a moistened filter cellulose wadding and excised leaves.
Excised leaves of corn and sorghum (3-6 leaf stage) and bermudagrass and centipedeg-
rass (regrowth at 2 wk) were initially replaced after 2 days and renewed every day
thereafter. The experimental design was completely randomized with six replicates.
Treatments consisted of a FAW strain (rice or corn) and host plant (corn, bermudagrass
var 'Grazer', sorghum, or centipedegrass) combination. Each replicate consisted of five
FAW larvae, and all parameters tested were averaged within each subsample. Biolog-
ical measurements taken were 9-day larval weights, pupal weights, larval and pupal
duration, and survivorship to adult eclosion. Data were subjected to analysis of vari-
ance, and significant means separated by least-squares means (SAS Institute 1985).


Oviposition Preference. Corn and rice strain female oviposition preference for corn,
sorghum, bermudagrass, and centipedegrass is presented in Table 1. A greater percent-
age of egg masses for both strains was oviposited on corn, sorghum, and bermudagrass

Z38 Flordta Entomologist 71(3) September, 1988


Trial 1 Trial 2
Corn Rice Corn Rice
Grass Host strain strain strain strain

corn 41.7a' 24.0b 32.7a 29.2ab
bermudagrass 16.0c 40.5a 20.7b 20.5b
centipedegrass 16.5c 15.5b 14.7b 16.5b
sorghum 25.0b 19.7b 32.0a 34.0a
-------------.----------------- Contrast2------------------------------
Corn, sorghum
bermudagrass versus
centipedegrass ** ** **
Corn, sorghum versus
bermudagrass ** ** *
Corn versus sorghum ** ** ns ns

'Means within a column (grass hosts within strain) followed by the same letter are not significantly different (P
> 0.05, Duncan's [1955] multiple range test).
2NS = P > 0.05, = P < 0.05, ** = P < 0.01.

than on centipedegrass during both trials (P < 0.05; orthogonal contrasts). Centipedeg-
rass had previously been identified as unsuitable for FAW development (Wiseman et
al. 1982, Quisenberry & Whitford 1988). Corn strain females laid a greater percentage
of their egg masses on corn and sorghum than on bermudagrass (P < 0.05; orthogonal
contrasts). The corn strain females showed an ovipositional preference for corn over
sorghum during the first trial (P < 0.001; orthogonal contrasts). FAW moths have been
shown to prefer corn and sorghum rather than bermudagrass (Pitre et al. 1983). These
authors used FAW moths obtained from the USDA, ARS, Southern Grain Insect Lab-
oratory at Tifton, Georgia; a laboratory colony that was subsequently identified as a
corn strain by Mason et al. (1987) and Pashley et al. (1987).
Rice strain egg-laying preferences were predominately bermudagrass (trial 1) and
sorghum (trial 2) (Table 1). Pitre et al. (1983) indicated that as one preferred host
becomes less attractive for oviposition, FAW moths will switch to a second host. We
suggest that rice strain females in the second trial switched their oviposition preference
to sorghum and corn because the bermudagrass was not as suitable for oviposition (i.e.,
the regrowth was not as dense and plant heights were shorter) as the grass used in the
first trial. Our results indicate behavioral differences in oviposition between the two
host-associated strains.
Mating Compatibility. The percentage of corn and rice females mated with intrast-
rain or interstrain males was 77 to 100% (Table 2). The corn intrastrain matings resulted
in the highest number (12.2) of egg masses per female. In contrast, the rice intrastrain
and rice x cornd crosses tended to produce fewer egg masses in the first trial; how-
ever, during the second trial the rice intrastrain cross produced a higher number of egg
masses than the rice? x corn cross. Egg masses among the rice and corn crosses
weighed 10 to 15 mg per mass. The number of eggs per mass was highly variable;
however, the rice? x cornd cross produced a higher number of eggs per mass. Egg
hatch was extremely high for the intrastrain and interstrain crosses (ca. 90%).
The F1 (rice? x cornd) interhybrid cross produced fewer matings, decreased
number of egg masses per female (P = 0.06), and lower percent hatch than the F,
(corn x rice ) interhybrid cross. Pashley & Martin (1987) observed fertility rates of


_1 ~ __ 1 _____

Whitford et al.: Fall Armyworm Symposium

cz cl cd cz -! << M c se ni I < -!i
,- - LM q 1- t- M cc- c
M, C oM00 r-o M M tOtOcc 0O00"

,Iz 00 CO i0 CI CO!
S .c CO C 01
-00 4C -4l

0 010 N 01

. 00 CO 1 O 00 L 7 I COOO


O0 0 = 1t

>i s > >

+ +


>, >

11D 00L10CT 11 0
1-- -l
-4~ h4l4 -


>> i > >


>i >

cd M 't M - cd ce d M -
cooM o 00 o=Ct o occ Co co I .c
a -coi-4 4 ic r-i -4 c c6
-4-4-4q-I4-- -q4- -4 -4 -4 -4

01000) ccc 0
oi~rlcd cdd
- -

M Mi c3 "ci "co!
01- CO 0 CO1
1--4 t-4

C00co CO CON t MP-LO
00sooooC t- C) 02 =- V o
-I 1--4

00 O c 00 L Uo LO UotoiCM 00 COO
-4-4 - -4--4-4-4 011

8W ~ .0) 8
5 o G 5


x x
$. C.
S+ +



0 '0+0


u h

0 0-o>F C x

f-0 f-0
o o- .. o+ o+
t C)$G









a)" -^







I- g


TI s


" S
a a



ca o



a) "a)
" .0
4ot (

a) .




Florida Entomologist 71(3)

25% when F1 (rice 9 x corn ) were crossed. Additionally, they were not able to make
interhybrid Fi (corn? x rice) crosses because the corn? and rice were reproduc-
tively incompatible.
Development on Artificial Diet. The development ofintrastrain and interstrain prog-
eny on modified pinto bean diet is presented in Table 3. Larval weights at 9-days
exceeded 300 mg, except for the the corn intrastrain cross in the second trial. Rice
intrastrain pupae weighed significantly less than other crosses. Quisenberry & Whitford
(1988) also reported that pupae of the corn strain reared on modified pinto bean diet
were significantly heavier than those of the rice strain. The corn intrastrain cross took
ca. 0.4 to 1.4 days longer to complete larval development than the other crosses. Sur-
vivorship of neonate larvae to adult eclosion was high (87 to 100%) for all crosses. Thus,
the corn and rice strain progeny had similar larval weights and survivorship, but the
two strains differed significantly in pupal weights and rates of development on modified
pinto bean diet.
Influence of Host Plant on Development. Strain development was strongly influ-
enced by larval host (Table 4). Larvae that were fed centipedegrass produced signific-
antly smaller larvae and required a longer time to complete development than those on
corn, bermudagrass, or sorghum. Wiseman et al. (1982), Chang et al. (1985), and Quisen-
berry & Whitford (1988) found centipedegrass significantly reduced larval and pupal
weights and increased the time necessary to complete larval and pupal development.
Development of the corn strain was comparable on corn, bermudagrass, and sorghum.
The larvae and pupae of the corn strain on these hosts were hea,' _r than those of the
rice strain on the same hosts orthogonall contrasts; P < 0.0001). However, rice strain
larvae on bermudagrass had larval weights comparable to those of the corn strain on
corn, bermudagrass and sorghum. The rice strain larvae completed larval development
more rapidly on bermudagrass than on the other grass hosts. Pupal weights of the corn
strain individuals were heavier on corn and bermudagrass. In contrast, rice strain pupal
weights were significantly heavier on bermudagrass than on corn, centipedegrass, and
sorghum. The survivorship of the corn strain was high on all hosts (> 80%); however,
less than 80% of the rice strain survived to adult eclosion. These results, along with
Pashley (1988), confirm differences in development between the corn and rice strains.
In both studies, corn strain larvae and pupae were heavier, on average, than rice strain
individuals. The corn strain also developed equally well on corn and rice (Pashley 1988)
or corn and bermudagrass (Table 4). In contrast, the rice strain developed better on
rice or bermudagrass than on corn.


Weight Duration of stage
l Strain cross (mg) (days) Percent
Trial Percent
no. female male Larval Pupal Larval Pupal survival

1 corn corn 336.0b' 270.1a 13.1a 8.5a 97a
corn rice 339.7b 229.8b 12.7b 8.9a 93a
rice rice 322.5b 199.0c 12.5b 8.7a 90a
rice corn 381.2a 224.2b 12.7b 8.4a 100a
2 corn corn 239.9b 264.4a 14.3a 9.la 87a
corn rice 380.5a 245.6b 12.5bc 8.0c 83a
rice rice 380.9a 221.3c 12.4c 8.3b 83a
rice corn 362.2a 258.9ab 12.9b 8.6b 90a

'Means within a column by trial are not significantly different (P > 0.05; Duncan's [1955] multiple range test).

September, 1988


Whitford et al.: Fall Armyworm Symposium


weight Larval Pupal Pupal
at day 9 duration weight duration Survival
Strain/host (mg) (days) (mg) (days) (%)

Corn/corn 275.5a' 14.2c 205.7a 8.6ab 90a
Corn/bermudagrass 274.6a 14.3c 192.3ab 8.3bcd 93a
Corn/centipedegrass 30.7d 27.3a 167.6c 8.5a 83ab
Corn/sorghum 290.8a 14.3c 191.3b 8.6ab 90a
Rice/corn 203.9bc 14. Ic 147. ld 7.8d 77ab
Rice/bermudagrass 251.0ab 13.0d 166.3c 8.2bcd 77ab
Rice/centipedegrass 27.6d 24.7b 137.8d 8.9a 42c
Rice/sorghum 174.4c 14.4c 148.6d 8.0cd 60bc

'Means within a column followed by the same letter are not significantly different (P > 0.05; Duncan's [1955]
multiple range test).


The bermudagrass variety 'Tifton 292' (Leuck et al. 1968, Lynch et al. 1983, Chang
et al. 1985) has been identified as a resistant genotype to FAW; however, Jamjanya &
Quisenberry (1988) reported 'Tifton 292' to be only intermediately resistant to FAW
when compared with a susceptible variety ('Grazer'). It is apparent that factors other
than the plant mechanisms (Quisenberry et al. 1988) are confounding our ability to
accurately assess resistant bermudagrass genotypes to FAW. Researchers working
with host plant resistance to the FAW need to be cognizant of three potential problem
areas: 1) the influence of environment (field versus greenhouse-grown grass) on the
acceptability of the bermudagrass to FAW, 2) the effects of chemical composition vari-
ation in bermudagrass genotypes from spring to fall on utilization of plant material by
FAW, and 3) the host-associated FAW strain used in screening bermudagrass
Jamjanya (1987) reported bermudagrasses grown in the field were lower in protein,
higher in fiber content, and less acceptable to FAW than the same genotypes cultured
in the greenhouse. Furthermore, field-grown bermudagrass genotypes were initially
lower in protein content and total digestibility in June and declined more rapidly through
September. For example, FAW fed greenhouse 'Tifton 292' in June had lower larval
weights, increased larval duration, and lower survivorship than larvae fed the same
field grown genotypes during the same time period (Jamjanya 1987). Conversely, larvae
fed field-grown 'Tifton 292' in August had lower larvae weights and decreased develop-
mental rates than larvae fed greenhouse-grown plants. Consequently, 'Tifton 292' would
be classified as a susceptible genotype when grown in one environment (field) in the
spring and resistant in the same environment later in the season.
Pashley (1986) suggested that two strains of FAW can be identified. The detection
of corn and rice strains constitutes an important discovery and provides an essential
refinement to the studies of host plant resistance to the FAW. The biological differences
between the two strains are especially pronounced when 'Tifton 292' is evaluated for
FAW resistance. Lynch et al. (1983) found 'Tifton 292' to be extremely resistant (100%
larval mortality) to FAW. These authors used a laboratory colony that was subsequently
identified as a corn strain (Pashley 1986). Pashley et al. (1987) also found 'Tifton 292'
to be resistant when tested with the corn strain, but more importantly, it was suscep-
tible when tested with the rice strain.

Florida Entomologist 71(3)

Physiological (i.e., development) and behavioral (i.e., ovipositional preference) dif-
ferences were detected in this study between the corn and rice strains. Oviposition
preference and greater larval development by the rice strain on bermudagrass indicate
a degree of adaptation to bermudagrass. However, the corn strain appears to be more
of a generalist feeder due to greater acceptance of corn, bermudagrass, and sorghum.
These results confirm the conclusions by Pashley (1986, 1988) that two host-associated
strains exist. We suggest that further research needs to be conducted to clarify the
taxonomic status between the corn and rice strain and to determine if the concept of
host-associated strains/sibling species can be applied to FAW populations over its entire
geographic range.


The authors wish to express their gratitude to F. M. Davis (USDA-ARS, Crop
Science Research Laboratory, Mississippi State, MS) for supplying fall armyworm. We
also wish to thank D. Pashley, C. M. Smith, and J. B. Graves for their critical review
of the manuscript. Special thanks to S. Decoteau, K. Gotreaux, and R. Decoteau for
their assistance in data collection. Approved for publication by the Director of the
Louisiana Agricultural Experiment Station as manuscript number 88-17-2129.


CHANG, N. T., B. R. WISEMAN, R. E. LYNCH, AND D. H. HABECK. 1985. Fall
armyworm: expressions of antibiosis in selected grasses. J. Entomol. Sci. 20:
DUNCAN, D. B. 1955. Multiple range and multiple F tests. Biometrics 11: 1-42.
JAMJANYA, T. 1987. Consumption and utilization, biology, and economic injury levels
of fall armyworm, Spodopterafrugiperta (J. E. Smith), on selected bermudagras-
ses. Ph.D. dissertation. Louisiana State Univ., Baton Rouge, La.
JAMJANYA, T., AND S. S. QUISENBERRY. 1988. Fall armyworm (Lepidoptera: Noc-
tuidae) consumption and utilization of nine bermudagrasses. J. Econ. Entomol.
81: 697-704.
JOHNSON, S. J. 1988. Migration and the life history strategy of the fall armyworm,
Spodoptera frugiperda, in the Western Hemisphere. Insect Sci. Applic. 8: 543-
BOWMAN. 1968. Resistance in bermudagrass to the fall armyworm. J. Econ.
Entomol. 61: 1321-1322.
LUGINBILL, P. 1928. The fall armyworm. U.S. Dept. Agric. Tech. Bull. 34: 1-91.
mudagrass resistance to fall armyworm (Lepidoptera: Noctuidae). Environ. En-
tomol. 12: 1837-1840.
MASON, L. J., D. P. PASHLEY, AND S. J. JOHNSON. 1987. The laboratory as an
altered habitat: phenotypic and genetic consequences of colonization. Florida En-
tomol. 70: 49-58.
MITCHELL, E. R. 1979. Monitoring adult populations of the fall armyworm. Florida
Entomol. 62: 91-98.
PASHLEY, D. P. 1986. Host-associated genetic differentiation in fall armyworm
(Lepidoptera: Noctuidae): a sibling species complex. Ann. Entomol. Soc. America
79: 898-904.
PASHLEY, D. P. 1988. Quantitative genetics, development, and physiological adapta-
tion in host strains of fall armyworm. Evolution 42: 93-102.
PASHLEY, D. P., AND J. A. MARTIN. 1987. Reproductive incompatibility between
host strains of the fall armyworm (Lepidoptera: Noctuidae). Ann. Entomol. Soc.
America 80: 731-733.


September, 1988

Wiseman & Isenhour: Fall Armyworm Symposium

PASHLEY, D. P., S. S. QUISENBERRY, AND T. JAMJANYA. 1987. Impact of fall
armyworm (Lepidoptera: Noctuidae) host strains on the evaluation of Bermuda
grass resistance. J. Econ. Entomol. 80: 1127-1130.
PERKINS, W. D. 1979. Laboratory rearing of the fall armyworm. Florida Entomol.
62: 87-91.
PITRE, H. N., J. E. MULROONEY, AND D. B. HOGG. 1983. Fall armyworm (Lepidopt-
era: Noctuidae) oviposition: crop preferences and egg distribution on plants. J.
Econ. Entomol. 76: 463-466.
QUISENBERRY, S. S., AND H. K. WILSON. 1985. Consumption and utilization of
Bermuda grass by fall armyworm (Lepidoptera: Noctuidae) larvae. J. Econ. En-
tomol. 78: 820-824.
QUISENBERRY, S. S., AND F. WHITFORD. 1988. Evaluation of bermudagrass resis-
tance to fall armyworm (Lepidoptera: Noctuidae): Influence of host strain and
dietary conditioning. J. Econ. Entomol. 81: 1463-1468.
QUISENBERRY, S. S., P. CABALLERO, AND C. M. SMITH. 1988. Influence of ber-
mudagrass leaf extracts on development and mortality of fall armyworm
(Lepidoptera: Noctuidae). J. Econ. Entomol. 81: 910-913.
SAS INSTITUTE. 1985. SAS user's guide: statistics, version 5. SAS Institute, Cary,
SPARKS, A. N. 1979. A review of the biology of the fall armyworm. Florida Entomol.
62: 82-87.
WISEMAN, B. R., R. C. GUELDNER, AND R. E. LYNCH. 1982. Resistance in common
centipedegrass to the fall armyworm. J. Econ. Entomol. 75: 245-247.


Insect Biology and Population Management Research Lab, USDA-ARS
Tifton, GA 31793-0748

University of Georgia, Department of Entomology
Coastal Plain Experiment Station
Tifton, GA 31793-0748


Green and yellow whorl-stage foliage of resistant and susceptible corn, Zea mays
L., were fed fresh as excised foliage or mixed in meridic diets to fall armyworm,
Spodoptera frugiperda (J. E. Smith), neonates. MpSWCB-4 manifested its resistance
when fall armyworm larvae were fed fresh foliage (1985-87) but not when larvae were
fed on tissue in a meridic diet. Larvae fed the green foliage were larger than those fed
on yellow foliage tissue, irrespective of whether the foliage was resistant or susceptible.
The resistance appears to be more behavioral than has been reported earlier.


El follaje verde y amarillo del verticilo de plants resistentes y susceptibles del maiz,
Zea mays L., se le di6 como follaje fresco cortado o mezclado en dietas meridicas, a


244 Florida Entomologist 71(3) September, 1988

neonatos del gusano cogollero, Spodopterafrugiperda (J. E. Smith). MpSWCB-4 man-
ifest6 su resistencia cuando el gusano cogollero se aliment6 con follaje fresco (1985-87),
pero no cuando las larvas se alimentaron con una dieta meridica. Las larvas alimentadas
con follaje amarillo eran mas grande que aquellas que se alimentaron con tejido foliar
amarillo, aparte de que el follaje sea resistente o susceptible. La resistencia parece ser
debida mas al comportamiento que al que se ha reportado anteriormente.

The discovery of resistance to the fall armyworm (FAW), Spodopterafrugiperda (J.
E. Smith), in the Coastal Tropical Flint corn, Zea mays L., (Anonymous 1965, Wiseman
et al. 1966), has encouraged numerous studies in plant resistance to FAW (Wiseman &
Davis 1979). Mihm et al. (1978) developed a mechanical device to artificially infest
whorl-stage corn with FAW neonates. Wiseman et al. (1980) modified this device and
showed its usefulness in mass rearing programs, infestations in the greenhouse, and in
small or large field tests. Wiseman et al. (1981, 1983) also demonstrated through novel
techniques that two of the resistance mechanisms (antibiosis and nonpreference) are
present in MpSWCB-4 and Antigua 2D-118. Wiseman & Widstrom (1986) and Wiseman
et al. (1986) later studied silk resistance in corn (Zapalote Chico) and panicle resistance
in sorghum to the FAW through the use of a laboratory bioassay.
Extensive attempts have been made to develop a laboratory bioassay to explain the
biochemical basis for corn resistance to the FAW. We have experienced difficulty in
developing a laboratory bioassay giving consistent positive results using MpSWCB-4,
Pioneer X304C, and Antigua 2D-118 as known resistant plant material. Here, we report
laboratory tests to demonstrate a relationship between FAW larvae feeding on whorl-
stage tissue of resistant and susceptible corns.


Single-row bulk plantings (6.1 m long and 0.76 m apart) of Cacahuacintle X's (Susc.),
Pioneer XZ304C (Resist.), and MpSWCB-4 (Resist.) were made during 1985-87 at the
Insect Biology and Population Management Research farm in Tift County, Ga., using
agronomic practices common to the area. Whorls of 10-12 leaf stage corn were excised,
brought to the laboratory, and processed for use in the laboratory tests.
The whorl of each genotype was separated on the basis of green or yellow leaf tissue.
Approximately 7.6 to 10.2 cm of tissue in the middle of each leaf was discarded and only
green or yellow tissue was tested. The midrib of each leaf also was excised and dis-
carded. Two separate tests were conducted each year. In one test, only fresh leaf tissue
was fed to neonate FAW. In another test, the leaf tissue was oven-dried at 103F (10
d to 2 wk), ground to a fine powder, and processed into meridic diets for laboratory
bioassay as described by Isenhour et al. (1985), Wiseman & Widstrom (1986), and
Wiseman et al. (1986).
Fresh Leaf Tests: The neonate FAW used in testing were maintained on bean diet
as reported by Perkins (1979). Neonate FAW (100/large dish or 3/small dish) were fed
fresh leaf tissue of each genotype for 3 d. After the 3 d of initial feeding, individual
larvae were placed on excised leaf tissue (green or yellow) on pieces of moistened paper
towel in Growth Chambers (small petri dishes; Wiseman et al. 1981). Fresh leaf sec-
tions (ca. 5.08 cm2) were replaced every other day or as needed. Tests were maintained
in an environmentally controlled room at 80 2% RH and 27 2C. Weights of larvae
were recorded after 7 d in 1985, and after 7 and 10 d in 1986-87. The tests consisted of
a split-plot design with 16 replications in 1985 and 24 replications in 1986-87, with two
dishes or larvae/replicate. Whole plots were leaf tissue (green or yellow) and subplots
were genotype entries.

Wiseman & Isenhour: Fall Armyworm Symposium




4 ^
1- a:




s c




r( c



'0 c 0C



t- cq

" 0 ci

to 00- CO


C 00 N 0l

cc Occ (o
cc ci ci~


03 .0 U

Scc c c


co -^co ^

*3 ci



Florida Entomologist 71(3)

Bioassay Tests: Fresh leaf tissue (60 g) blended in 300 ml of bean diet and 100 ml of
distilled water was used in 1985. Oven-dried leaf tissue (20 g) blended in 300 ml of bean
diet and 150 ml of distilled water was used in 1986-87 as outlined by Wiseman (1988).
Tests were arranged as a split-plot design with 18 replications with two cups per repli-
cate. The weight of larvae was recorded on the 9th day after the cups of diets were
infested with neonates.
All data were analyzed by an analysis of variance and significant mean differences
were separated with the use of Waller-Duncan k ratio t test and k ratio = 100 and P
< 0.05 (Waller & Duncan 1969, SAS Institute 1982).


Table 1 presents the mean weight of 7-d-old (1985) and 7- and 10-d-old FAW (1986-87)
larvae fed fresh excised whorl tissue. In four out of five instances, FAW larvae fed on
the yellow whorl tissue were smaller than those fed green whorl tissue. But in three
out of five instances, FAW larvae were significantly smaller. In each case, larvae that
fed on fresh whorl tissue of MpSWCB-4 were significantly smaller than larvae fed on
Cacahuacintle X's. Larvae that fed on Pioneer X304C were intermediate in size, in that
for 50% of the cases, larvae were smaller than those fed on Cacahuacintle X's. The
resistance of MpSWCB-4 manifested itself in all test years irrespective of whether
larvae fed on green or yellow whorl tissue. The only significant whole plot-subplot
interaction involved the 10-d weight for larvae that fed on fresh foliage in 1986. The
interaction was due to the magnitude of weight differences for larvae fed on Cacahuacin-
tle X's and those fed on Pioneer X304C.
The resistance of MpSWCB-4 to feeding by larvae was not manifested in the larvae
that fed on plant materials in meridic diet mixtures (Table 2). No interactions between
whole and sub-plots occurred, but in all 3 years the larvae that fed on yellow whorl
tissue mixed in diet weighed significantly less than larvae fed on the green whorl tissue
mixed in diet. The results for weights of larvae fed on green or yellow tissue in diet are
consistent with results of larvae fed fresh whorl foliage. There were no significant
separations of weight of larvae that fed on resistant versus the susceptible foliage diet
mixtures. In fact, larvae that fed on the resistant foliage in diet media tended to be
larger at 9 d than larvae that fed on susceptible foliage in diet media; this was especially
true when compared to those fed the bean diet check.


1985 1986 1987
9 d wts 9 d wts 9 d wts
Genotype Green Yellow Green Yellow Green Yellow

Cacahuacintle X's 191a 180a 332a 273a 510a 404a
X304C 177b 162b 355a 257a 420b 436b
MpSWCB-4 199a 187a 357a 289a 531c 456c
Bean check 181ab 177ab 250b 256a 354a 347a
Mean 187 177 323 269 454 411

'1985 diets used fresh foliage and 1986-87 used oven-dried foliage. Means within a column not followed by the
same letter and gree-yellow means separated by are significantly different (k ratio 1.00; P 0.05; Waller-Duncan
t-test and least significant differences, respectively [Waller & Duncan 1969; SAS Institue 1982]).


September, 1988

Wiseman & Isenhour: Fall Armyworm Symposium

Therefore, the development of a laboratory bioassay such as described by Wiseman
(1988), Wiseman et al. (1986), and Wiseman & Widstrom (1986) needs more study. One
explanation for weight differences of FAW fed resistant and susceptible foliage may
be that the whorls of Cacahuacintle X's are much deeper and tighter than those of
MpSWCB-4 (Fig. 1). Whorls of MpSWCB-4 are short and more open, and thus, if larvae
of FAW fed higher in the whorl of Cacahuacintle X's and only on green tissue, this

Fig. 1. Comparison of whorl lengths of MpSWCB-4 (M) and Cacahuacintle (C) sus-







" r:
:5:\:, ..;.'

s; :~


Florida Entomologist 71(3)

feeding response would result in larger larvae. If FAW fed lower in the whorls of
MpSWCB-4, this probably would be in the green-yellow to yellow whorl tissue, which
would result in the production of small larvae. Then, the magnitude of differences
between the weights of larvae fed the resistant and susceptible probably would be
identified easily. However, if the resistance is due to nonpreference (Wiseman et al.
1983) such as occurs with Pioneer X304C and Antigua 2D, then forced-feeding tests as
reported here probably would not identify the resistance as was the case with the
Pioneer X304C (Table 1). In addition, the possible loss of leaf volatiles or surface waxes
during the laboratory processing of meridic diets could account for the lack of resistance
response. Also, enzymatic degradation may occur when the resistant plant material is
mixed with the hot bean diet, possibly resulting in a loss of resistance activity.
In summary, the resistance of MpSWCB-4 to feeding of FAW larvae was manifested
in the fresh whorl tissue test but not in laboratory diet bioassays. The resistance of
whorl-stage corn may be more of an interaction of larval behavior and physical charac-
teristics of the plant due to the feeding location of larvae within the whorls of resistant
or susceptible corns than was expected earlier (Wiseman et al. 1981). The laboratory
diet bioassay we use at present appears impractical for the FAW and stage of resistance.
Therefore, more study is needed to develop a consistent bioassay for laboratory inves-


We thank J. L. Skinner, Charles Mullis, Peggy Goodman, 1 David Atkins of this
laboratory for their technical assistance.


ANONYMOUS. 1965. The Rockefeller Foundation program in the agricultural sciences.
Annu. Rpt., 1964-65. The Rockefeller Foundation, New York. 262 p.
ISENHOUR, D. J., B. R. WISEMAN, AND N. W. WIDSTROM. 1985. Fall armyworm
(Lepidoptera: Noctuidae) feeding responses on corn foliage and foliage/artificial
diet medium mixtures at different temperatures. J. Econ. Entomol. 78: 328-332.
MIHM, J. A., F. B. PEAIRS, AND A. ORTEGA. 1978. New procedures for mass produc-
tion and artificial infestation with lepidopterous pests of maize. CIMMYT Re-
view. 138 p.
PERKINS, W. D. 1979. Laboratory rearing of the fall armyworm. Florida Entomol.
62: 87-91.
SAS INSTITUTE. 1982. SAS user's guide. Statistics. SAS Institute. Cary, N. C.
WALLER, R. A., AND D. B. DUNCAN. 1969. A bayes rule for the symmetric multiple
comparison problem. J. Am. Stat. Assoc. 64: 1484-1499.
WISEMAN, B. R. 1988. Technological advances for determining resistance in maize to
Heliothis zea (Boddie). In Methodologies used for determining resistance in
maize to insects. CIMMYT. (In press).
WISEMAN, B. R., AND F. M. DAVIS. 1979. Plant resistance to the fall armyworm.
Florida Entomol. 63: 425-432.
WISMAN, B. R., F. M. DAVIS, AND J. E. CAMPBELL. 1980. Mechanical infestation
device used in fall armyworm plant resistance programs. Florida Entomol. 63:
WISEMAN, B. R., F. M. DAVIS, AND W. P. WILLIAMS. 1983. Fall armyworm: larval
density and movement as an indication of nonpreference in resistant corn. Protec-
tion Ecol. 5: 135-141.
WISEMAN, B. R., R. A. PAINTER, AND C. E. WASSOM. 1966. Detecting corn seedling
differences in the greenhouse by visual classification of damage to the fall ar-
myworm. J. Econ. Entomol. 59: 1211-1214.


September, 1988

Hruska & Gladstone: Fall Armyworm Symposium


WISEMAN, B. R., H. N. PITRE, S. L. FALES, AND R. R. DUNCAN. 1986. Biological
effects of developing sorghum panicles in a meridic diet on fall armyworm
(Lepidoptera: Noctuidae) development. J. Econ. Entomol. 79: 1637-1640.
WISEMAN, B. R., AND N. W. WIDSTROM. 1986. Mechanisms of resistance in 'Zapalote
Chico' corn silks to fall armyworm (Lepidoptera: Noctuidae) larvae. J. Econ.
Entomol. 79: 1390-1393.
WISEMAN, B. R., W. P. WILLIAMS, AND F. M. DAVIS. 1981. Fall armyworm: resis-
tance mechanisms in selected corns. J. Econ. Entomol. 74: 622-624.


1IPM in Maize Project
Escuela de Sanidad Vegetal
Institute Superior de Ciencias Agropecuarias
Apartado 453
Managua, Nicaragua
2Department of Entomology
Box 7634
North Carolina State University
Raleigh, NC 27695


The effect of period and level of infestation by the fall armyworm (FAW), Spodoptera
frugiperda (J.E. Smith) (Lepidoptera: Noctuidae), on irrigated maize, Zea mays L.,
yield was determined. Two applications of the insecticide chlorpyriphos (336 g[ai]/ha)
applied directly to the whorl produced yields equivalent to treatments with three appli-
cations. FAW infestation of 100% caused 45% yield reduction. A linear regression model
(yield[g/plant] = 87.84 0.384 (% plants infested by FAW)) explained 46% of the yield
variation. The economic injury level (EIL) was calculated as 2% of the plants infested.
This low EIL is due to currently high subsidies provided pesticides in Nicaragua. The
impact of pesticide subsidies on EIL's is discussed.


Se determine el efecto del period y nivel de infestaci6n del gusano cogollero,
Spodoptera frugiperda (J. E. Smith) (Lepidoptera: Noctuide), sobre el rendimiento del
maiz, Zea mays L., de riego. Dos aplicaciones del insecticide chlorpyrifos (336 g ia/ha) ap-
licados directamente al cogollo resultaron en rendimientos iguales que tratamientos con
tres aplicaciones. Una infestaci6n de 100% por el gusano cogollero caus6 una reducci6n
de rendimiento de un 45%. Un modelo lineal de regresi6n (rendimiento [g/planta]=
87.84 0.384 [% plants infestadas por S. frugiperda]) explic6 el 46% de la variaci6n en
rendimiento. Se calcul6 el nivel de dafo econ6mico (NDE) como 2% de las plants
infestadas. Se debi6 este NDE bajo a los subsidies de insecticides en Nicaragua. Se
discute el efecto de subsidies de insecticides sobre NDE's.

Florida Entomologist 71(3)

In Central America, maize (Zea mays L.) is traditionally grown during the rainy
season (May-December), and two crops can be harvested per year. In an effort to
become self-sufficient in maize production, Nicaragua has added a third crop each year,
during the dry season, through the introduction of irrigation on the Pacific Plain. Be-
cause this new production system is only 3 years old, an integrated pest management
(IPM) program for the important insect pests of irrigated maize has not yet been de-
veloped. Because we already know the principle pests of irrigated maize, the first step
in developing an IPM program is to determine the critical periods of protection and
economic injury levels (EIL).
One of the principal pests of maize in Nicaragua is the fall armyworm (FAW),
Spodopterafrugiperda (J. E. Smith). FAW is present during all three cropping cycles,
but is much more abundant during the second natural rainfall season and in irrigated
maize. Pest populations are very high in irrigated maize in the Pacific Plain, where
100% of untreated plants are infested, with up to three 6th-instar larvae per plant.
The current recommendations issued by the Ministry of Agriculture to control FAW
in irrigated maize are to apply insecticides 5 times during the vegetative phase every
5 days, three of which are directed solely against FAW (MIDINRA, 1984). Some pro-
ducers apply insecticides up to 12 times per cropping cycle (personal observation).
Given the current control recommendations and practices used against FAW in
irrigated maize and as a first step in developing an IPM program, we carried out a
study with 3 objectives: 1) to determine the critical periods of protection from FAW
during the vegetative phase, 2) within the periods of infestation which cause yield
reduction, to determine the damage function between infestation and yield, and 3) use
the damage function to calculate the economic injury level for FAW in irrigated maize.

0 20 40 60 80

Fig. 1. Effect of FAW infestation at 23 DAG on Yield.


September, 1988

Hruska & Gladstone: Fall Armyworm Symposium


Open-pollinated maize variety NB-6 was planted at the National Center for Basic
Grain Research in Managua, Nicaragua, in March, 1987, with 7 plants/m and 80 cm
between rows. Due to the naturally occurring 100% infestation level of FAW in maize,
we created the treatments by applying insecticide to individual plants. Chlorpyrifos
(480 E) was mixed with sawdust and applied by hand directly into whorls (336 g[ai]/ha),
a commonly used practice against FAW in Central America.
A randomized complete block design with 20 treatments and 8 blocks was used.
Treatments were combinations of 3 periods and 4 levels of infestation. The three periods
were: 5-17, 17-31, and 31-45 days after germination (DAG), corresponding to seedling
and early whorl, mid-whorl, and late whorl to tassel. The four levels of infestation were
100%, 70%, 40%, and 0% of the plants infested with FAW larvae (Fig. 1). Infestation
by FAW was determined at 10, 23, 30, and 39 DAG, using the center row of each plot,
by visually examining the whorl for FAW larvae, feeding damage, or frass.
At 115 DAG the central row of each plot was harvested, and total number, number
broken, and number of lodged plants were counted. Mean yield of commercial grain at
15% humidity per plant was calculated for each plot.
To determine the effects of periods of infestation on yield, the 7 treatments that had
100% infestation during the periods were analyzed with an analysis of variance
(ANOVA). Differences among treatments were determined with Tukey's multiple com-
parison test (Steel & Torrie 1980). Regression models were used to determine the
relation between FAW infestation and yield. Stepwise regression was used to determine
which date of infestation was the best predictor of yield.


There was a significant effect of infestation period on yield (ANOVA, F = 3.53, d.f. =
6,38, P<.01). Treatments that were infested with FAW for 2 or 3 periods had a signific-
antly lower yield than treatments that were not infested, or were infested during one
of the periods (Table 1).
Using treatments that had infestation in two or three of the periods, a stepwise
regression analysis indicated that, of the 4 infestation dates, 23 DAG was the best
predictor of yield. Linear and quadratic regression models were compared, with the
linear model explaining a greater percentage of the variance in yield. The estimated
regression model is: Y= 87.84 0.384(PI), with r2 = 0.458, where Y= yield (g/plant)
and PI = % plants infested (Table 2).
The results indicate that there is no one period that is more critical in causing yield
reduction, although there is a non-significant tendency for infestation during later
periods to cause greater yield loss (Table 2).
The results demonstrate that the entire vegetative phase of irrigated maize does not
have to be protected from FAW to produce maximum yield. Two applications of chlor-
pyriphos produced yields that were not different from the treatment with three applica-
tions (Table 2). The effect of infestation by FAW on yield of irrigated maize appears
cumulative during the vegetative phase, but maize plants are able to compensate for
damage inflicted during any one period. Thus, the current insecticide recommendations
could be reduced from current levels of 3-8 applications from seedling until tassel stage,
to 2 during mid- to late-whorl stage, without reducing yield.
The shape of the damage function is different from the function obtained in similar
experiments under natural rainfall where the regression model was curvilinear and
yield did not decrease until 40% of the plants were infested (Hruska et al. 1987). Damage
functions may be very specific to environmental conditions.

Florida Entomologist 71(3)


Period and Level of Infestationi
Days After Germination
Treatment 5 17 31 45

1. I------100% ------ I
2. 1------ 70%------
3. ------ 40%------
4. I------100%.------
5. 1 I------ 70% ------I
6. I I------ 40% ------ I
7. I I------100%------I
8. I I------ 70% ------I
9. ------ 40% ------ I
10. I------100%------ ------100%------
11. 1------ 70%------ ------ 70%------
12. ------ 40%------ ------ 40%------
13. I ------100% ------ ------100% ------
14. ------ 70%------ ------ 70% ------
15. I I------ 40%------ ------ 40%------ 1
16. ------100%------ ------100%------ ------100%------
17. 1------ 70%------ ------ 70%------ ------ 70%------
18. 1------ 40%------ ------ 40%------ ------ 40%------
20. ------100%------ ------100%------ ------100%------I

'Lines indicate infestation by FAW and numbers the percentage of plants infested.


Period of Infestation
(Days After Germination) Yield'
Treatment 5 17 31 45 (S.D.)

1. ---------- 69.76 A
2. ---------- 64.95 A
3. 59.11 A
4. ---------- 58.73 A
5. ------------------- 51.72 B
6. ---------------------- 41.12 B
7. -------------------------------- 38.67 B

'Means followed by the same letter do not differ significantly according to Tukey's test (P>.05).


September, 1988

Hruska & Gladstone: Fall Armyworm Symposium 253

Yield was reduced by 45% when 100% of the plants were infested. This yield reduc-
tion is more severe than studies conducted under natural rainfall, where 100% infesta-
tion caused between 15-30% yield reduction (Obando 1976, Van Huis 1981, Hruska et
al. 1987).
The difference in the shape of the yield/infestation function and the greater reduction
in yield in irrigated maize may be due to greater water loss in plants under irrigated
maize than under natural rainfall. Irrigated maize grown in the hot dry season with
typically strong winds would undergo high rates of evapotranspiration. Equal amounts
of defoliation would be expected to produce greater water loss in the dry season than
under natural rainfall.
Using the estimated damage function, the economic injury level (EIL) was calculated
to be 2% of the plants infested. This is far lower than any other calculated EIL for
FAW in maize, where published values vary between 20 and 50% (Sarmiento &
Casanova 1975, Young & Gross 1975, Obando 1976, Van Huis 1981, Young & Gross,
EIL's change any time there is a change in the damage function, value of the crop,
cost of control, or efficacy of control changes (Hruska & Rosset 1987). An examination
of the EIL calcualtion revealed that the low value was due to the relatively low cost of
In Nicaragua, as in many developing countries pesticides are heavily subsidized by
the government (Repetto 1985). Nicaraguan farmers pay between 3-5% of the world
market cost for agricultural inputs. At the same time that pesticide prices are low, crop
prices are relatively high, due to a shortage of basic grains in the country. The relation
between the cost of insecticide and the value of maize is 0.52 (price of 3.78 liter of
chlorpy- to the producer/value of 45.4 kg of maize), whereas in the United States
the realtioclship 1, 13.45 (Hruska & Gladstone 1987).
If we use world prices for insecticides and maize the EIL is 20% of the plants
infested, similar to values reported in the literature (Sarmiento & Casanova 1975,
Obando 1976, Van Huis 1981).
The raison d'etre of the EIL is to eliminate uneconomical pesticide applications. The
decision criterion is set at maximizing profit for the individual farmer. What maximizes
individual profit, however, may not maximize social profit.
Pesticide subsidies are a double-edged sword. On the one hand they make agricul-
tural inputs readily available to farmers with limited resources, while on the other hand
they may encourage unnecessary pesticide use. The classical EIL calculation may not
provide a useful decision rule for farmers and countries with subsidized pesticides. A
new decision rule is necessary, one that estimates and incorporates the values of "exter-
nalities" in the calculation. Hruska & Gladstone (1987) suggest a method to calculate a
revised EIL and set pesticide subsidy levels to meet the goals of society.

We thank the Instituto Superior de Ciencias Agropecuarias and the Ministerio de
Desarrollo Agropeucario y Reforma Agraria for support of this work with material
resources. This research was supported in part by a grant from the Norwegian Ministry
of Development Cooperation.


HRUSKA, A. J., AND S. M. GLADSTONE. 1987. The economic threshold and the polit-
ical economy of pest management decisions in Nicaragua. Institute Superior de
Ciencias Agropecuarias. Managua, Nicaragua. 18 pp.

254 Florida Entomologist 71(3) September, 1988

HRUSKA, A. J., AND P. M. ROSSET. 1987. Estimaci6n de los niveles de dfio econ6mico
para plagas insectiles. Manejo Integrado de Plagas 5: 30-44.
HRUSKA, A. J., S. M. GLADSTONE, AND R. LOP Z. 1987. El period critic de
protecci6n para el gusano cogollero, Spodoptera frugiperda, en maiz de la prim-
era. Institute Superior de Ciencias Agropecuarias. Managua, Nicaragua. 12 pp.
OBANDO S., S. R. 1976. Umbrales permisibles de dafo foliar en maiz. XXII Reuni6n
Anual del PCCMCA, San Jos6, Costa Rica.
1984. Guia Thcnica para la Producci6n de Maiz con Riego. Managua, Nicaragaua.
REPETTO, R. 1985. Paying the Price: Pesticide Subsidies in Developing Countries.
World Resources Institute. Washington, DC.
SARMIENTO J., AND J. CASSANOVA. 1975. Busqueda de limits de aplicai6n en el
control del 'cogollero' de maiz. Spodoptera frugiperda S & A. Rev. Peruana
Entomol. 18: 104-107.
STEEL, R. G. D., AND J. H. TORRIE. 1980. Principles and Procedures of Statistics,
2nd ed. McGraw-Hill, New York.
VAN HUIS, A. 1981. Integrated pest management in the small farmer's maize crop in
Nicaragua. Mededelingen Landbouwhogeschool Wageningen. 81-6.


Department of Entomology
Louisiana Agricultural Experiment Station
Louisiana State University Agriculture Center
Baton Rouge, Louisiana 70803


The effects of antibiosis on the growth of fall armyworm (FAW), Spodoptera
frugiperda (J. E. Smith), larvae by rice, Oryza sativa L., seedlings at different plant
stages and the tolerance of seedlings to artificial defoliation at the 7-leaf stage were
evaluated in five plant introductions (PI) and the check cultivar 'Mars'. FAW larvae fed
5-leaf stage foliage had significantly lower larval, pupal, and adult weights than larvae
fed 3-leaf stage foliage. The effects of antibiosis on the growth of FAW were evident
in PIs 160842, 160827, 346833, and 346839, but not in PI 160823 or Mars. PIs 160842,
160827, and 160823 exhibited higher tolerance in plant regrowth after defoliation than
PI 346833, 346839, or Mars. PIs 160842 and 160827 exhibited both the antibiosis and
tolerance resistance.


Se evalu6 en cinco plants introducidas (PI) y en la variedad testigo 'Mars', el efecto
de antibiosis en el crecimiento de larvas del gusano cogollero, Spodoptera frugiperda
(J. E. Smith), en plants de semillero de diferentes etapas de crecimiento del arroz,
Oriza sativa L., y de la tolerancia de las plants a la defoliaci6n artificial en la etapa de
7-hojas. Larvas, pupas, y adults del gusano cogollero alimentadas con follaje de la
etapa de 5-hojas, tuvieron significativamente un peso menor que las larvas que se
alimentaron con follaje de la etapa de 3-hojas. El efecto de antibiosis en el crecimiento

Lye & Smith: Fall Armyworm Symposium

del gusano cogollero fue evidence en PI 160842, 160827, 346833, y en 346839, pero no
en PI 160823 o Mars. PI 160842, 160827, y 160823 tuvieron una tolerancia mas alta en
el crecimiento de la plant despu6s de la defoliaci6n que PI 346833, 346839, o Mars. PI
160842 y 160827, ambas dfemostraron resistencia de tipo de antibiosis y tolerancia.

The fall armyworm (FAW), Spodoptera frugiperda (J. E. Smith), is a polyphagous
insect which is an important pest on many crops (Luginbill 1928). However, cereals and
grasses are the most preferred among its host plants (Crumbs 1927). On rice, Oryza
sativa L., the FAW defoliates the leaf particularly at the seedling stage and is a sporadic
pest in the southern U.S.A. (Bowling 1978, Smith et al. 1986). Severe damage by FAW
to rice has been reported in Puerto Rico (Chandler et al. 1977), and several countries
in South America (Navas 1966, Gallego 1967, Machado 1978).
Plant resistance to FAW has been studied on many crops as reviewed by Davis
(1980). Pantoja et al. (1986) evaluated over 5,000 rice plant introductions for resistance
to FAW defoliation. In their study, 11 plant introductions (PI) were found resistant in
choice and no-choice experiments in the greenhouse and field.
The antixenosis, antibiosis, and tolerance resistance are all important to the im-
plementation of resistant cultivars into an insect pest management program. Both rice
stands and yield are reduced due to the damage caused by FAW feeding during the
seedling stage (Bowling 1978). For this reason, the ability of rice to regrow after damage
by FAW is an useful indicator for evaluating the tolerance resistance of rice germplasm.
Similarly, the defoliation of rice by FAW or by artificial methods has differential effects
on rice yield. Upland rice in Panama tolerates substantial artificial or FAW defoliation
without incurring significant yield loss (Navas 1976). Artificial defoliation of the upper
2/3 of the plants of a leafy rice variety, N.P. 130, increased yields by 10-34% depending
on the age of the plants at defoliation (Tripathi et al. 1973). Mukerji (1973) reported
that rice plants pruned at the seedling stage have slight reductions in yield.
The objectives of this study were to evaluate the effects of rice plant growth stage
on the growth of FAW larvae and to evaluate antibiosis and tolerance of five rice PIs
found to have antixenosis resistance by Pantoja et al. (1986).


Four resistant PIs (160842, 346833, 346839, and 160823), one susceptible PI 160827,
and check cultivar 'Mars' (PI 009945) were selected from the study of Pantoja et al.
(1986). Seeds were pregerminated in an oven at 30C for 3 d before being sown into
pots in a greenhouse.
In the antibiosis test, pregerminated seeds were sown into the center of a plastic
pot (5.1 by 5.1 by 5.1 cm) with ca. four to six seeds per pot. Baccto potting soil was
used, but no fertilizers were applied during the experiment. Pots were set in plastic-
lined benches in the greenhouse, and the water level in benches was maintained at a
height of 2-3 cm. Pots were moved to the laboratory when most of the plants reached
the 3-leaf stage (Yoshida 1981). All of the plants in each pot were confined in a transpar-
ent plastic tub (2.5 cm diam., 45.0 cm high). One FAW neonate larva per pot was
introduced onto plants inside the tube with a camel's hair brush. A total of 20 pots of
seedlings of each PI were used. Infested plants were placed inside an incubator at
28+2C, 14:10 L:D, and 70% RH. Larval weights were measured daily from the 8th
day until larvae became prepupae. Pupal weights were recorded within 24 h after the
pupation of each larva. The duration and mortality of the entire larval stage, the pre-
pupal stage, and the pupal stage were also recorded. Individual larvae on severely
consumed plants were transferred into cages with plants of the same stage to ensure
that enough fresh plants were provided throughout the larval stage.


Florida Entomologist 71(3)

The same procedures were repeated using rice seedlings at the 5-leaf stage of
growth. Effects of rice plant growth stage on growth of FAW larvae were compared
by a t test (SAS Institute 1985). Means of measurements were subjected to Duncan's
multiple range test (Duncan 1955) to compare antibiosis effects among PIs.
In the tolerance test, pregerminated seeds were sown into plastic pots (15 cm diam.,
15 cm high) with six to tn seeds per pot. The soil was a mixture of rice field soil and
Baccto potting soil at a 4:1 ratio. No fertilizers wer applied to the soil during this test.
Pots were set in plastic-lined benches in greenhouse, and the water level in the benches
was maintained at ca. 10 cm. The number of seedlings was thinned to two plants per
pot at the 4-leaf stage. Leaf blades were removed with scissors at rates of 0, 33, 66,
and 99% dfoliation at the 7-leaf stage, and pots were flooded after defoliation. This
experiment was conducted with a split-plot design with four replications. The six PIs
were the whole plots and the different percent defoliation levels were the subplots.
There were three pots for 0% defoliation (control) and two pots for each of the remaining
treatments in each replication. Four weeks after defoliation, plants were removed from
pots and soil was washed from the roots. The number of tillers, vegetation wet weight,
root wet weight, and root volume were measured in each pot. Plant dry weights were
measured after drying at 55C for 4 d.
Analysis of variance (ANOVA) was performed to compare the effects of different
defoliation levels on plant growth among and within rice PIs (SAS Institute 1985). The
relationships between plant growth measurements and rice PIs were quantified by
correlation analysis (SAS Institute 1985). The percent reduction in plant dry weight,
(Dry Wt. of Uninfested Plant Dry Wt. of Infested Plant)
x 100%,
Dry Wt. of Uninfested Plant
where plant dry weight = dry weight of vegetative parts + dry weight of roots, was
calculated to compare tolerance among PIs (Panda & Heinrichs 1983). The relationship
between antibiosis and tolerance among PIs was quantified by regrssing the last day
larval weight (maximum larval weight) at the 5-leaf stage (antibiosis index) on the
average percent plant loss in dry weight (tolerance index) (SAS Institute 1985). This
relationship was used to measure the relative contributions of the antibiosis and toler-
ance components of resistance in each of the tested PIs (Ortega et al. 1980).


Rice plant growth stage had a significant effect on the growth of FAW (Table 1).
Larvae fed 5-leaf stage foliage weighed significantly less and had a longer duration of
the larval stage than larvae fed 3-leaf stage foliage. The prepupal duration was longer
on 5-leaf stage foliage than on 3-leaf stage foliage; however, the duration of the pupal
stage was not different. Both pupal and adult weights of larvae fed on 5-leaf stage
foliage were also lower than weights of larvae fed on 3-leaf stage foliage. The overall
mortality of FAW was ca. 10% and showed no differences due to the rice plant growth
stage. No larval mortality was observed, with the exception of one larva fed 5-leaf stage
'Mars' foliage. Most of the mortality was observed in the pupal stage. Adults emerging
from larvae fed both plant growth stages appeared normal.
The effects of rice PI on larval and pupal weights and th duration of the larval and
pupal stage were significant at both the 3-leaf and 5-leaf stages (Table 2). Larvae
exhibited different trends in growth on different PIs tested. Therefore, the larval
weight at the last larval day (maximum larval weight) was used to compare antibiosis
effects. Maximum larval weights on PIs 346833 and 346839 were lower than those on
all other PIs in the 3-leaf stage of growth. The same trend appeared among weights of
larvae fed 5-leaf stage foliage, except that the maximum larval weight of FAW larvae
fed PI 160827 was the same as that of larvae fed PI 346833. The duration of larval


September, 1988

Lye & Smith: Fall Armyworm Symposium


Growth Stage t Test
FAW Growth Index 3-leaf 5-leaf t df P>F

8th day larval wt. (mg) 147.22 109.66 4.58 218 0.0001**
Last day larval wt. (mg) 424.17 388.92 4.62 233 0.0001**
Larval duration (day) 11.62 12.36 -3.91 196 0.0001**
Prepupal duration (day) 1.87 2.01 -3.10 226 0.0022**
Pupal duration (day) 7.30 7.49 -1.84 193 0.0665Ns
Pupal wt. (mg) 172.10 145.06 7.99 210 0.0001**
Adult wt. (mg) 81.93 67.30 6.63 207 0.0001**
Percent mortality 9.60 10.00 0.67 7 0.5188Ns

**highly significant at a = 0.05; Nsnot significant at a = 0.05.

development appeared to be correlated to the larval weight at the 8th day of develop-
ment. The lower the larval weight on day 8, the longer the larval duration and vice
versa (r = -0.805, P = 0.0001, n = 235). Pupal weights were not correlated to the duration
of development among larvae fed rice at the 3-leaf stage (r= 0.007, P=0.940, n= 115)
or the 5-leaf stage (r=-0.165, P=0.080, n= 113) of growth. However, pupal weights
were significantly correlated to maximum larval weights (3-leaf stage; r= 0.736,
P = 0.001, n = 115; 5-leaf stage; r= 0.755, P = 0.001, n = 113). FAW larvae had the high-
est pupal mortality when fed foliage of PI 160827 at both stages of rice growth.
In the tolerance test, the effects of PI were significant in relation to the number of
tillers (F=15.04; df=5,15; P=0.0001), vegetation wet weight (F=10.47; df=5,15;


PI Larval Weight (mg) Duration (d) Pupal
PI Pupal %
Number 8th Day Last Day Larva Pupa Wt. (mg) Mortality

3-leaf stage

160842 57.7 d' 458.6 a 13.2 a 7.6 a 179.4 ab 0
346833 105.0 c 400.0 bc 12.0 b 7.7 a 165.4 bc 5.0
346839 179.2 b 385.3 c 10.7 c 7.4 a 162.5 c 10.5
160823 202.6 ab 433.2 ab 10.9 c 6.6 b 173.9 abc 10.0
Mars 123.6 c 446.7 a 11.9 b 7.8 a 186.2 a 5.3
160827 218.8 a 438.0 a 10.9 c 6.4 b 164.9 c 31.6
5-leaf stage

160842 105.3 c 388.2 be 12.2 c 7.4 a 138.2 c 0
346833 66.5 d 354.4 c 13.7 b 7.7 a 131.8 c 0
346839 185.8 a 381.8 be 10.4 e 7.8 a 139.9 b 0
160823 145.6 b 448.0 a 11.3 d 7.0 b 181.3 a 11.8
Mars 42.7 e 406.0 b 14.7 a 7.8 a 154.2 b 5
160827 108.7 b 354.4 c 12.0 c 7.0 b 128.4 c 22.2

'Means followed by the same letters in each column for each plant growth stage are not significantly different (P
= 0.05; Duncan's [1955] multiple range test).


258 Florida Entomologist 71(3) September, 1988

P=0.0002), vegetation dry weight (F=14.50; df=5,15; P=0.0001), root wet weight
(F= 9.18; df= 5,15; P= 0.0004), root dry wight (F= 3.42; df= 5,15; P= 0.0294), and root
volume (F=3.58; df=5,15; P=0.0248). Defoliation also significantly affected the
number of tillers (F= 12.62; df=3,54; P=0.0001), vegetation wet weight (F=27.50;
df= 3,54; P= 0.0001) and dry weight (F=27.50; df= 3,54; P= 0.0001), root wet weight
(F= 8.69; df= 3,54; P= 0.0001) and dry weight (F= 6.21; df= 3,54; P= 0.0011), and root
volume (F= 9.70; df= 3,54; P = 0.0001). Interactions of PI and defoliation were not sig-
nificant across any rice plant growth measurements. Among the PIs tested, 346833 and
160823 had higher growth measurement values than the remaining PIs (Table 3), indi-
cating that these two PIs have a larger biomass than the other PIs tested.
The 33% defoliation rate did not induce a significant effect among PI on percent
reduction in plant dry weight, but the 66% and 99% defoliation rates did (a=0.10)
(Table 4). In general, PIs 160842 and 160827 had the lowest reductions among all rice
PIs tested. PI 160823 had a moderate reduction in plant dry weight, whereas the
susceptible cultivar 'Mars' had the highest plant dry weight reduction across all three
defoliation levels.
The linear relationship between the average percent reduction in plant dry weight
and the maximum weight of larvae fed foliage of plants at the 5-leaf stage of growth
(Fig. 1) had a very small slope. Among the five PIs tested, only the plotted values of
PIs 160842 and 160827 fell within the region in which both antibiosis and tolerance were
present. Values of PIs 346833 and 346839 fell in the region in which only antibiosis was
present. PI 160823 had the highest larval weight among the PIs tested, but still plotted
just within the tolerance quadrant. Values for the susceptible cultivar Mars fell within
the region in which neither antibiosis nor tolerance was expressed.


The use of whole rice seedling plants proved to be a useful bioassay technique to
evaluate the antibiosis effects of rice on the growth of FAW. During the experiment,
the first, second, and third instars fed exclusively on leaf blades while the fourth and
fifth instars fed randomly on leaf blades or leaf sheaths. Pupal cells were built success-
fully under the base of plants by pooling plant fragments, frass, and soil. No prepupal
mortality was observed during the test.
The significant differences in the growth of FAW fed rice plants of different stages
indicated that plant growth stage affects the growth of FAW on rice. Older 5-leaf stage
seedlings had more detrimental effects on the growth of FAW than younger 3-leaf


PI No. of Vegetation Weight Root Weight
PNo. of Root
Number Tillers Wet (g) Dry (g) Wt (g) Dry (g) Volume (ml)

160842 9.3 c2 65.5 b 8.6 c 96.7 b 17.6 be 86.4 bc
346833 13.5 b 94.4 a 12.2 a 181.9 a 35.7 a 135.1 a
346839 10.6 c 62.0b 7.5 c 71.1b 14.0 c 66.1 c
160823 17.8 a 84.2 a 11.8 ab 166.6 a 33.9 ab 127.4 ab
Mars 10.1 c 67.8 b 8.9 c 106.8 b 20.0 be 86.6 bc
160827 13.1b 71.9b 10.6b 89.7b 24.0 be 84.8 be

'Plants were defoliated at the rate of 0, 33, 66, and 99% at the 7-leaf stage of growth. Plant growth measurements
were taken 4 weeks after defoliation.
'Means followed by the same letters are not significantly different (P = 0.05; Duncan's [1955] multiple range test).

Lye & Smith: Fall Armyworm Symposium


Defoliation Rate
Number 33% 66% 99% Average

Mars 33.28 51.35 a' 83.16 a 56.02 a
346839 30.33 44.17 ab 56.07 b 43.47 a
346833 2.01 50.42 a 62.13 ab 36.00 ab
160823 27.75 5.62 b 54.95 b 31.61 b
160827 -0.95 22.17 ab 48.44 b 23.78 b
160842 -14.23 18.82 ab 50.41 b 19.70 b

F = 0.82 2.28 2.55 3.64
df= 5, 14 5, 15 5, 15 5, 80
P= 0.5544 0.0987 0.0733 0.0485

'Means followed by the same letters are not significantly different (P = 0.10; Duncan's [1955] multiple range test).


Antibiosis Present Mars Antibiosis Absent
Tolerance Absent Tolerance Absent



.Y=.Y 6.182 *0.074X
Antibiosis Present
r = 0.20 PI160823
Tolerance Present r =0.20 P1160823
I = 388.70
I Y = 35.10








Antibiosis Absent
Tolerance Present



Maximum Wt. (mg) of Larvae Fed 5-leaf Stage Rice Foliage

Fig. 1. Relationship between percent rice plant dry weight reduction (tolerance) and
maximum larval weight of fall armyworms (antibiosis) fed foliage of five rice plant
introductions and the cultivar 'Mars' at the 5-leaf stage of development.


Florida Entomologist 71(3)

stage seedlings. Hardy et al. (1986) reported that both neonate and fourth instar FAW
larvae feed more on new growth than on older growth of tall fescue grass. FAW larvae
also prefer young corn leaves over old leaves as a food source (Garner & Lynch 1981).
Similarly, the resistance level of rice plants may vary as the plants mature. In our
study, the maximum larval weight of FAW larvae on PI160827 was among the highest
when fed 3-leaf stage foliage, but became the lowest when larvae were fed 5-leaf stage
Larval weight has been used previously as an indicator of rice plant resistance to
insects. Oliver & Gifford (1975) studied the effects of seven rice lines on the larval
weight of the rice stalk borr, Chilo plejadellus Zincken, and identified two cultivars as
potential sources of stalk borer resistance. In our study, the maximum larval weights of
FAW fed 5-leaf stage foliage of PIs 346833 and 160827 were the lowest among the five
PIs tested, indicating that these two PIs are more resistant to FAW than the other
PIs. In general, PIs 160842, 346833, 346839, and 160827 had more antibiotic effects on
FAW growth than Mars and PI160823, based on the maximum larval weight of FAW
fed 5-leaf stage foliage.
The percent reductions in plant dry weight of the five rice PIs tested were lower
than that of the check cultivar 'Mars'. PI 160827, which did not exhibit antixenosis
(Pantoja et al. 1986), had an antibiotic effect on FAW growth and was tolerant to
defoliation as indicated by its low percent reduction in plant dry weight. These results
indicate that the resistance by one cultivar cannot be justified by a single category of
resistance. The relationships between antibiosis and tolerance (Fig. 1) demonstrate the
relative contributions of each type of FAW resistance in PIs tested. PIs 346833 and
346839 have antibiosis but no tolerance to FAW damage. PIs 160842 and 160827 appear
to be sources of antibiosis and tolerance resistance to FAW; however, PI160842 appears
to be the most promising PI to use for development of lasting resistance to FAW,
because it also has the antixenosis resistance reported by Pantoja et al. (1986).


The authors thank Drs. E. A. Heinrichs, J. R. Fuxa, and T. J. Riley for their critical
review of the manuscript. This research was supported by funds from the Louisiana
Agricultural Experiment Station and USDA Special Grant no. 85-CRSR-2-2621 and is
approved for publication by the Director of the Louisiana Agricultural Experiment
Station as manuscript number 88-17-2067.
Michael Smith is presently at the Department of Plant, Soil and Entomological
Sciences, University of Idaho, Moscow, ID 83843.


BOWLING, C. C. 1978. Simulated insect damage to rice: effects of leaf removal. J.
Econ. Entomol. 71: 377-378.
RAMIREZ. 1977. Cultivo intensive y perspectives del arroz en Puerto Rico Bol.
Univ. P. R. Est. Exp. Agric. 25.
CRUMBS, S. E. 1927. The armyworm (Lepidoptera: Noctuidae). Bull. Brooklyn En-
tomol. Soc. 22: 41-55.
DAVIS, F. M. 1980. Fall armyworm plant resistance programs. Florida Entomol. 63:
DUNCAN, D. B. 1955. Multiple range and multiple F tests. Biometrics 11: 1-42.
GALLEGO, M. F. L. 1967. Lista preliminary de insects de importanicia econ6mica y
secundarios, que afectan los principles cultivos, animals dom6sticos y al
hombre, en Colombia. Rev. Fac. Agron., Colombia. Rev. Fac. Agron., Colombia
65: 32-66.


September, 1988

Lye & Smith: Fall Armyworm Symposium

GARNER, J. W., AND R. E. LYNCH. 1981. Fall armyworm leaf consumption and
development on florunner peanuts. J. Econ. Entomol. 74: 191-193.
HARDY, T. N., K. CLAY, AND A. M. HAMMOND JR. 1986. Leaf age and related
factors affecting endophyte-mediated resistance to fall armyworm (Lepidoptera:
Noctuidae) in tall fescue. Environ. Entomol. 15: 1083-1089.
LUGINBILL, P. 1928. The fall armyworm. U. S. Dept. Agric. Tech. Bull. 34.
MACHADO, S. S. 1978. Alguns fatores limitantes da produtividade das lavoras de
arroz. Lavoura Arrozeira 307: 52-54.
MUKERJI, D. K. 1973. Effects of pruning rice seedlings on growth and yield. Sci. Cult.
39: 198-199.
NAVAS, D. 1966. Evaluaci6n del dano que ocasionan algunans plagas del arroz en
Panama, vol. XII, pp. 77-78. In Programa cooperative centroamericano para dl
mejoramiento de cultivos alimenticios. XII Reunion Anual, Managua, Nicaragua,
Marzo 28-Abril 2.
1976. Fall armyworm in rice. Proc. Tall Timbers Conf. Ecol. Anim. Control
Habitat Manage. 6: 99-106.
OLIVER, B. F., AND GIFFORD, J. R. 1975. Weight differences among stalk borer
larvae collected from rice lines showing resistance in field studies. J. Econ. En-
tomol. 68: 134.
ORTEGA, A., S. K. VASAL, J. MIHM, AND C. HERSHEY. 1980. Breeding for insect
resistance in maize. pp. 372-419. In F. G. Maxwell and P. R. Jnnings (ed.),
Breeding Plants Resistant to Insects. John Wiley and Sons. N.Y.
PANDA, N., AND E. A. HEINRICHS. 1983. Levels of tolerance and antibiosis in rice
varieties having moderate resistance to the brown planthopper, Nilaparvata
lugens (Stal) (Hemiptera: Delphacidae). Environ. Entomol. 12: 1204-1214.
PANTOJA, A., C. M. SMITH, AND J. F. ROBINSON. 1986. Evaluation of rice germ
plasm for resistance to the fall armyworm (Lepidoptera: Noctuidae). J. Econ.
Entomol. 79: 1319-1323.
SAS INSTITUTE. 1985. SAS user's guide: Statistics. SAS Institute, N.C.
pests of rice in Louisiana. La. Agric. Exp. Sta. Bull. 774. 24 pp.
Effect of pruning on the yield of paddy N.P. Sci. Cult. 39: 269-70.
YOSHIDA, S. 1981. Fundamentals of Rice Crop Science. International Rice Research
Institute, Los Bahios, Laguna, Philippines. 269 pp.

262 Florida Entomologist 71(3) September, 1988


'Entomologists and Maize Breeder, resp.
CIMMYT, Lisboa 27, Apdo. Postal 6-641
Deleg. Cuauhtemoc, Mexico D.F., Mexico 06600
2Maize Breeder, Cornell University
Dept. Plant Breeding and Biometry
252 Emerson Hall
Ithaca, NY 14853-1902


As part of the overall improvement of maize, Zea mays L., the International Maize
and Wheat Improvement Center (CIMMYT) has been working to develop tropical maize
materials with improved levels of resistance to fall armyworm (FAW), Spodoptera
frugiperda (J. E. Smith), and good agronomic qualities. Efficient mass rearing and
infestation techniques have been developed, many of which have been adopted or
adapted by other FAW host plant resistance workers. Our initial efforts to develop
FAW resistance consisted of a population improvement approach. Recently, CIMMYT's
maize insect resistance improvement program has been amplified to include screening
and development of new sources of resistance, international testing of these sources,
and inbred line and 'non-conventional' hybrid development, in addition to the continuing
development of open-pollinated varieties.


Como parte del program de mejoramiento de maiz, el Centro Internacional de
Mejoramiento de Maiz y Trigo (CIMMYT) ha trabajado para desarollar el germoplasma
de maiz con niveles de resistencia de la plant hu6sped contra al "cogollero", Spodoptera
frugiperda (J. E. Smith). Se han desarollado t6cnicas eficientes de crianza masiva e
infestaci6n artificial al nivel de campo. Muchas de esas tecnicas han sido adoptadas por
otros trabajadores en resistencia de la plant hu6sped para facilitar sus trabajos y
studios. Nuestros trabajos iniciales fueron sobre el mejoramiento de resistencia al
nivel de poblaciones de maiz. Recientemente, el program de entomologia sobre la
resistencia de las plants huespedes ha sido ampliado para incluir la evaluaci6n y de-
sarollo de fuentes de resistencia nuevas, ensayos internacionales para probar dicha
resistencia, y la formaci6n de variedades de polinizaci6n libre, lines endocriadas e
hibridos 'no-convencionales' con resistencia al cogollero.

For more than a decade, the maize improvement program of the International Maize
and Wheat Improvement Center (CIMMYT) has included improvement of host plant
resistance (HPR) to the fall armyworm (FAW), Spodoptera frugiperda (J. E. Smith),
as an integral part of its selection and improvement process in maize materials of tropical
lowland and subtropical adaptation (Mihm 1985). The FAW is a major destructive field
pest of maize in tropical and subtropical areas of the Americas, and in favorable years
even damages fields in temperate production areas. Enhanced levels of resistance to

Mihm et al.: Fall Armyworm Symposium


Category Composition Number

Pools 500 Half-sibs + 50
Populations 250 Full-sibs 40
Experimental varieties + 10 Full-sibs + 600
Inbred lines S3-S5 level + 5000
'Non-conventional' hybrids Variety & Topcross 500

FAW in maize cultivars is highly desirable and would have great utility to CIMMYT's
clientele, the National Agricultural Research Systems (NARS) and their farmers. In
many developing countries, where low input agriculture predominates, HPR is the only
practical means of managing pest populations and minimizing production losses resulting
from pest attack.
Because of its facilities for insect mass rearing, its personnel (breeders, en-
tomologists, and pathologists), its resource commitment to the improvement of maize
germplasm, and its well established ties to NARS for testing and production research,
CIMMYT is uniquely qualified to conduct FAW HPR research (Mihm 1983).
Before 1984, HPR improvement to FAW was carried out in selected CIMMYT
germplasm 'pools' and 'populations' (Table 1) of lowland tropical adaptation (Ortega et
al. 1980). The goal was gradual improvement of resistance on a population level, and
extract superior progenies to form open-pollinated experimental varieties with im-
proved resistance (Mihm 1985). When the work was begun in the late 1970s, such
varieties were the most appropriate intermediate product for CIMMYT to produce
because of the ease of maintenance, seed production, and distribution for most client
In 1984, it was apparent that the progress made was slow because of low gene
frequency and competitive selection concerns. Also, several NARS were developing
capabilities of viable hybrid seed production systems, especially for utilizing varietal or
topcross hybrids (termed 'non-conventional' by the CIMMYT maize program) (Table 2).
Nearly all the maize grown in developed countries is produced from single-, three-way,
or double-cross hybrids, which are formed from crosses of highly inbred maize lines,
and growers purchase new seed for each planting. In developing countries, it is difficult
to maintain inbred line purity, to produce good quality seed of hybrids, to get high
enough seed yields, and to get seed distributed on time to farmers at a price they can
afford. Many of these problems are minimized in the production of 'non-conventional'
hybrids. Although these are less uniform for many characteristics than conventional


Conventional Components 'Non-conventional' Components

Single cross AxB Topcross: Commercial hybrid x Line
Three-way (A x B) x C Family (half-, full-sib) x Line
Double cross (A x B) x (C x D) Variety x Line
Varietal: Family (half-, full-sib) x Family
Variety x Family
Variety x Variety

Florida Entomologist 71(3)

September, 1988

Step of operation

Fig. 1. Flow-chart showing operations and breeding methodology used in developing
the Multiple Borer Resistance (MBR) population, inbred line extraction, and the forma-
tion of experimental varieties and 'non-conventional' hybrids.

hybrids, this usually does not detract from their acceptability. Many developing country
farmers actually benefit from some heterogeneity in their cultivars, as their production
fields are often not uniform, and harvest is done by hand labor, not machines. Suitable
cultivars for most subsistence farmers need only be fairly uniform in grain color and
texture, maturity, yield, and postharvest storability (Deutsch 1988).
In order to make greater and faster progress in developing FAW resistance, and to
produce an array of germplasm products with greater utility to a wider range of maize


Winter 1984

Summer 1984

Winter 1985

Summer 1985

Winter 1986

Summer 1986

Winter 1987

Summer 1987

Winter 1988

Summer 1988

Mihm et al.: Fall Armyworm Symposium

improvement programs in both developing and developed countries, CIMMYT formed
a new maize population and began improving it for resistance to several maize field
pests. This population was named the Multiple Borer Resistance (MBR) population
(Fig. 1). It was formed from predominantly subtropical and temperate materials that
were reported to have high or intermediate levels of resistance to FAW and the follow-
ing subtropical or temperate region maize stem borers: Ostrinia nubilalis (Hubner),
Diatraea grandiosella (Dyar), D. saccharalis Fabricius, and Chilo partellus (Swinhoe)
(Benson 1986, Mihm 1986, Smith et al. 1988).
Progenies from this population were evaluated internationally for resistance in 1986.
Two hundred full sib families were sent to interested cooperators who had the capability
to artificially infest at least 10 plants per progeny row with FAW or one of the maize
stem borers to assess resistance to first and/or second generation damage by that pest
(Smith et al. 1988). Each cooperator infested 10-15 plants per row with 30-40 newly
hatched FAW larvae (Mihm 1983). Leaf feeding ratings were taken 14-21 days after
infestation using a 1 to 9 scale (1= highly resistant, 9 = highly susceptible. These, and
other data of interest were reported to CIMMYT. The number of families in each
resistance category from the Mississippi, Georgia, and Mexico FAW evaluations are
presented in Fig. 2. The results indicated fairly high frequencies of FAW resistance,
with most of the families showing high (1-3 ratings) or intermediate (4-6 ratings) levels
of resistance. Only a few families were classified in the susceptible category. These
results indicate this population is an excellent source of FAW resistance; most breeding
materials and commercial cultivars are found to be susceptible when artificially infested
with FAW.
In addition to taking resistance ratings, each cooperator selected the 20 to 30 most
resistant families (equal or superior to the best resistant check entry) with good ag-
ronomic characters and self pollinated the best plants in these families. The majority of
the seed harvested from these plants remained with the cooperators for use in their
HPR program. Most of the cooperators sent 50-100 seeds of each selected ear to CIM-
MYT, where they are being used to regenerate a new set of progenies for testing in
1988, and for line and variety development. For each test site that successfully
evaluated the population, there is the potential to develop an experimental variety
and/or extract inbred lines with potentially superior resistance. If the population was
tested for the same species at more than one site, there is the potential to develop an
across site experimental variety, using the 8-10 most resistant families across sites as
progenitors. Based on the results from the 1986 MBR international test, CIMMYT
decided to develop only Across 86 MBR-FAW for the following reasons: (1) the
cooperators from Mississippi and Georgia have the production of resistant inbred lines
as their goal; (2) both sites are similar in latitude and general adaptation to the Mexican
test site; and (3) most of the selected families were common across sites, therefore site
specific experimental varieties would not differ greatly.
Seed of selected Si's returned to CIMMYT from these and other sites which
evaluated for stem borer resistance was used to continue inbred line development and
to regenerate a new set of full sib progeny for international testing in 1988 (Fig. 1).
Some of the initial S1 lines used in developing the full sib families for the 1986 interna-
tional progeny testing trial (IPTT) were also testcrossed to a set of testers to determine
combining ability groups. These testcrosses were grown in a subsequent season under
both infested and protected conditions to identify their resistance reaction and yield
performance. The results were used to help form two insect resistant pools with good
potential combining ability. Information from the testcross performance also will be
used to identify lines for forming synthetics and experimental 'non-conventional' hybrids
with high yield potential and good resistance, primarily for developing country NARS,
but of possible value in developed country programs as well.


Florida Entomologist 71(3)

120 Mississippi








,.%. .*.
*.*. t.'

w; @cCzQ---- -


a n a




Damage rating

Fig. 2. Fall armyworm leaf-feeding ratings of 200 full-sib families of MBR population
at three locations.

In developing the MBR population and monitoring the sites where it was internation-
ally tested, two things were observed. First, although there was notable expression of
daylength sensitivity in this germplasm, cooperators were able to obtain viable seed
from testing sites more than 400 North latitude to 300 South latitude, indicating broad
general adaptation of the population. Second, there was variability for resistance to
common foliar diseases within the population. As we are aware of the importance of
resistance to plant pathogens as well as insect pests, and know that varieties will not
be acceptable to farmers unless they are resistant to both, we also screen and select
for resistance to leaf diseases, and ear and stalk rots.
Lines from the MBR population that are selected for insect resistance are being
screened for reaction to common rust, Puccinia sorghii Schw. and Helminthosporium
sp. leaf blights. The most resistant disease selections are then re-screened for insect
reaction and re-incorporated into the population.
Testing of the MBR indicated a more complex approach was needed for developing
resistant germplasm for lowland humid tropical areas, especially where the downy mil-
dew disease complex or maize streak virus diseases occur (although there were apparent
differences for streak reaction at the Kenya testing site). In order to develop insect

September, 1988

Mihm et al.: Fall Armyworm Symposium


Material Selections available Adaptation

Tuxpeno PR 8321, Exp'l. variety Tropical/subtropical
Antig.-Ver. 181 PR 8424, Exp'l. variety Tropical/subtropical
Pools 24, 26 Half- & full-sib families, Si lines Tropical/subtropical
MBR Across 86-FAW, Exp'l. var. Subtropical/temperate
Varios x Mp lines Full-sibs, S1 & S2 lines Tropical/subtropical
MIRT Population development Tropical/subtropical

resistant germplasm to serve in such areas, another population is being developed. It
is named the Multiple Insect Resistant, Tropical (MIRT) population (Mihm 1986, Smith
et al. 1988). The germplasm incorporated into this population is predominantly of low-
land tropical adaptation. Crosses with sources of downy mildew resistance from Thai-
land and streak resistance from the CIMMYT/IITA program in Nigeria were included.
Selections of FAW and stem borer resistance from CIMMYT pools and populations,
from 'Antigua' collections from the CIMMYT maize germplasm bank, and Mississippi
resistant lines and hybrids were added.
The MIRT population is still being formed, but we expect to have a set of full-sib
progenies ready for international testing in 1989. We hope to send it for insect resistance
screening to cooperators in countries with lowland humid tropical climates, where FAW
and other armyworms and borer pest complexes occur. Additional evaluation for disease
reaction will be useful though not essential at this time.
The best sources of FAW and stem-borer resistance available to the present, the
Mississippi-released inbred lines, are deficient in many agronomic characteristics (Wil-
liams & Davis 1988). However, Overman (1988) was able to use them to develop new
hybrids with multiple resistance to leaf-feeding and stalk-boring lepidopterous pests by
crossing them with susceptible, agronomically superior U.S. Corn Belt lines, then ex-
tracting new resistant lines. These lines, in hybrid combinations, are competitive in
yield with current commercial hybrids in the absence of pest attack, yet superior in
yield under insect infestation (Overman 1988). Some of the recently extracted CIMMYT
'Antigua' lines are equal to the Mississippi lines in FAW and stem borer resistance, and
superior to them in standability and leaf blight, stalk and ear rot resistance under
Mexican growing conditions. Their combining ability and yield performance are yet to
be tested.
Table 3 lists various materials that have been selected for FAW resistance at CIM-
MYT that are presently available in small quantities for testing. In the near future, we
expect to develop and have available new materials, in the form of open-pollinated
varieties, synthetics, inbred lines, and experimental non-conventional hybrids with po-
tential resistance to FAW and other insect and disease pests of maize from temperate
to tropical growing areas.


BENSON, D. L. 1986. Evaluation of stalk borer resistance mechanisms and the de-
velopment of a population for multiple stalk borer resistance in maize, Zea mays
L. Ph.D. Dissertation, Cornell University, Ithaca, NY.
DEUTSCH, J. A. 1988. Pre-release testing of insect resistant maize cultivars, seed
production and promotion to farmers. Proc. Int'l. Symp. on Methodologies for
Developing Resistance to Maize Insects. March, 1987. CIMMYT, Mexico. (In

Florida Entomologist 71(3)

MIHM, J. A. 1983. Efficient mass-rearing and infestation techniques to screen for host
plant resistance to fall armyworm, Spodoptera frJ'igl" ,l.iI CIMMYT, Mexico
D.F. Mexico. 16 pp.
MIHM, J. A. 1985. Breeding for host plant resistance to maize stem-borers. Insect Sci.
Applic. 6(3): 369-377.
MIHM, J. A. 1986. Multiple insect resistance development at CIMMYT. CIMMYT
Research Highlights 1985. Mexico, D.F. Mexico.
ORTEGA, A. O., S. K. VASAL, J. MIHM, AND C. HERSHEY. 1980. Chap. 16. Breeding
for insect resistance in maize. pp. 371-419, In Maxwell, F. G., and P. R. Jennings
(Eds.) Breeding Plants Resistant to Insects. J. Wiley & Sons, NY. 683 pp.
OVERMAN, J. L. 1988. A maize breeding program for development of hybrids with
resistance to multiple species of leaf-feeding and stalk-boring Lepidoptera. Proc.
Int'l. Symp. on Methodologies for Developing Resistance to Maize Insects.
March, 1987. CIMMYT, Mexico. (In Press).
SMITH, M. E., J. A. MIHM, AND D. C. JEWE:L. 1988. Breeding for multiple resis-
tance to temperate, subtropical and tropical maize insect pests. Proc. Int'l. Symp.
on Methodologies for Developing Resistance to Maize Insects. March, 1987. CIM-
MYT, Mexico. (In Press).
WILLIAMS, W. P., AND F. M. DAVIS. 1988. Breeding for resistance in maize to
southwestern corn borer and fall armyworm. Proc. Int'l. Symp. on Methodologies
for Developing Resistance to Maize Insects. March, 1987. CIMMYT, Mexico. (In

- -* --* - - - ---- *-- -- -*^


Department of Entomology
University of Georgia
Athens, GA 30602 USA

Field experiments comparing no-tillage and plow-tillage practices demonstrated that
infestations by the fall armyworm, Spodoptera frugiperda (J. E. Smith), ultimately
become similar in either cropping system. However, in certain no-tillage situations
where high mulch concentrations were present on the soil surface, oviposition and dam-
age were reduced. Significantly fewer egg masses and damage were sampled on corn,
Zea mays L., (3-leaf stage) while seedlings remained within no-tillage mulch. Oviposi-
tion quickly became similar to that observed in plow-tillage systems when the plants
grew above the mulch canopy of no-tillage. The number of egg masses on corn older
than 4 leaves was similar in either cropping system, and leaf injury at plant silking was
the same. In a comparison of corn, sorghum (Sorghum bicolor [L.] Moench.), and soy-
beans (Glycine max L.), the latter crop had no damage in either tillage system while
corn and sorghum were heavily infested. Efficacy of chlorpyrifos (0.56 kg [AI]/ha) for
controlling fall armyworm leaf damage was similar in corn and sorghum in either crop-
ping system.


Experimentos donde se compare el no labrar y el labrar con arado demostr4 que las
infestaciones por el gusano cogollero, Spodopterafrugiperda (J. E. Smith), al final, son
igual en cualquiera de los dos sistemas. Sin embargo, en ciertas situaciones donde no

September, 1988


All: Fall Armyworm Symposium

se ara y esta present una concentraci6n alta de capa vegetal en la superficie del suelo,
se redujo la oviposici6n y el daho. Significativamente menos masas de huevos y dafio se
encontr6 en muestras de maiz, Zea mays L., (etapa de 3-hojas), mientras las plants
de semilleros permanecieron dentro de la capa vegetal no arada. La oviposici6n se igual6
rapidamente a aquella observada en el sistema labrado con arado cuando las plants
crecieron por arriba de la capa vegetal cuando no se ar6. El ndmero de masas de huevos
en maiz mas viejo de 4-hojas, fue similar en cualquiera de los dos sistemas de labranza,
y el dafo a la hoja cuando la mazorca esta produciendo seda tambien fue igual. En una
comparaci6n entire el maiz, el sorgo (Sorghum bicolor [L.] Moench.), y la soya (Glycine
max L.), el iltimo cultivo no tuvo daho en ninguno de los sistemas de labranza, mientras
que el maiz y el sorgo fueron muy infestados. La eficacia de clorpirifos (0.56 kg [AI]/ha)
en controlar el daho a las hojas por el gusano cogollero fue igual en el maiz y el sorgo
en cualquiera de los sistemas de labranza.

Entomologists recognize that the crop environments created by no-tillage practices
are often greatly different from those of plow-tillage systems and are justifiably con-
2erned about enhanced biological potential of pests like the fall armyworm, Spodoptera
frugiperda (J. E. Smith), in no-tillage (All & Musick 1986). Increased damage by other
noctuids such as the armyworm, Pseudaletia unipuncta (Haworth), and the black cut-
worm, Agrotis ipsilon (Hufnagel), in no-tillage cropping systems has been reported
(Musick & Petty 1973, Wrenn 1975), and the fall armyworm infests many of the same
crops. Therefore, this paper reviews data from experiments designed to determine if
fall armyworm hazard is influenced by lack of tillage.


Field experiments were conducted during 3 years (1985-1987) near Athens and Grif-
fin, GA. The experiments had either a split plot or split-split plot design. The main plots
were preplanting land treatments consisting of some form of no-tillage compared with
plow-tillage. The designation 'no-tillage' consisted of treatments in which no plowing
operations were used immediately prior to planting of seed. This included planting into
fallow land with weed cover and/or debris from former crops. Also, no-tillage treatments
were used in double cropping systems where a field crop was planted into standing small
grains (including rye, Secale cereale L.; wheat, Triticum vulgare; and barley, Hordeum
sativum, or into stubble following harvest of the grain.
Plow-tillage treatments consisted of tilling the soil with either a moldboard plow,
disk harrow, or rotary tiller. Soil was tilled up to 5 times in some treatments to insure
a smooth seed bed for planting. The size of no-tillage and plow-tillage plots varied from
0.03-0.1 ha with 3-5 replications in different experiments.
The subplots in the experiments were selected crops (corn, Zea mays L.; sorghum,
Sorghum bicolor [L.] Moench.; and soybeans, Glycine max L.) and insecticide treated
or untreated. In the crop comparison test, 16 rows (20 m long) of corn, sorghum, or
soybeans were each planted on the same day in June plantings of replicated plow-tillage
or no-tillage plots. Chlorpyrifos was applied at a rate of 0.56 kg [AI]/ha at 10 day
intervals for 3 applications, beginning in the 3-leaf stage of development for corn and
sorghum, and for 1 application in the V6 stage of soybeans. The chlorpyrifos treatments
were 4 row plots x 8 m long and were paired with untreated plots in each of the corn,
sorghum, and soybean treatments.
Sampling of fall armyworm infestations was done by examining 5 consecutive plants
for egg masses at 3.3 m intervals, covering a field or plot uniformly. At least 500 plants
were examined in each experiment. Damage was assessed on the plants when they were
examined for egg masses by making a visual estimate of the degree of injury based on

270 Florida Entomologist 71(3) September, 1988

a scale of 0-7. Plants received a 0 rating when no injury was present; 1 was slight (<
10%) leaf damage with no feeding in the whorl as evidenced by the lack of gummy
excrement; 2 was moderate (10-20%) leaf damage and no whorl feeding; 3 was heavy
(> 20%) leaf damage and no whorl feeding; 4 was < 20% leaf damage and light whorl
damage with a slight amount of excrement; 5 was 20-40% leaf damage with a moderate
amount of excrement in the whorl; 6 was severe leaf feeding (> 40%) with a large
amount of excrement in the whorl; and 7 was plants that had leaves with only midribs
remaining and buds destroyed (plant was virtually destroyed).
One or more pheromone traps baited with (Z)-9-dodecen-l-ol acetate were placed in
the test fields to monitor fall armyworm adult seasonality patterns. Information on
adult flights into fields was used in timing planting operations and insecticide applica-
tions in the tests.
The data were analyzed using analysis of variance (SAS Institute 1985) with proce-
dures appropriate for a specified experimental design. Duncan's new multiple range test
(Duncan 1955) was used to separate means.


Table 1 shows results from 3 tests with no significant difference between plow-tillage
and no-tillage in % egg masses (except test 3) or foliar damage by the fall armyworm.
These results are representative of general observations in many experiments in corn
over several years (1980-1988) of fall armyworm infestations in fields with plants in the
6-leaf stage (25-30 days after planting) or older in comparisons of plow-tillage and no-til-
lage systems.
On several occasions, it was observed that damage to 2- and 3-leaf stage plants
(10-15 days after planting) was greatly reduced in no-tillage as compared with plow-til-
lage, especially when the no-tillage fields had a heavy mulch from previous crops. Figure
1 shows results from a test using corn planted on July 29 following harvest of rye. A
heavy rye mulch was present on the soil surface in the no-tillage plots. Corn germination
and development in the no-tillage plots was similar to that observed in plow-tillage. The
corn seedlings were shorter than the mulch height up to the early 3-leaf stage. Ovipos-
ition and damage were significantly less (P < 0.05) in no-tillage plots as compared with
plow-tillage. Leaves were first observed above the mulch in the late 3-leaf stage 16 days
after planting. This coincided with greater oviposition and, thus, at 20 days the number
of egg masses on plants was similar in the 2 tillage systems. Damage remained signific-
antly higher in the plow-tillage plots than in no-tillage. This trend also was evident after
25 days, but the differences were not significant. The number of egg masses that were
sampled at 32 days dropped substantially and may reflect sampling error due to diffi-


Till No-Till
% Egg Damage % Egg Damage
Test' Masses Rating Masses Rating

1 (20 da) 0.8 3.8 0.3 3.7
2 (30 da) 5.7 4.9 6.8 4.0
3 (40 da) 3.02 3.7 0.5 4.0

'Tests 1 and 3 were conducted near Griffin and test 2 near Athens GA. A minimum of 100 plants wer- examined
in each plot at a specified number of (days) after planting.
2Significantly different from plow tillage in analysis of variance (P < 0.05).

All: Fall Armyworm Symposium

Plant Development (leaves)

2 4 6 8 silking
7 5 D
% a
E 6 n
4 a
9 5 9
4 3e
a 3 \- a
a 2 i
8 2
e i n

0 0 g
13 20 25 32 83
Days After Planting
-- Till % Egg Masses NT % Egg Masses
-- Till Damage Rating NT Damage Rating

Fig. 1. Seasonal comparison of fall armyworm infestations in no-tillage and plow-til-
lage corn as expressed by sampling of unhatched egg masses on leaves and by visual
estimates of leaf and whorl damage on a scale of increasing severity of 0-7. Significant
(P < 0.05) differences occurred between tillage systems in egg masses and damage on
day 13 and in damage on day 20 after planting.

culty in finding egg masses on the heavily damaged leaves of the late 8-leaf stage plants.
Injury was not significantly different in the 2 tillage systems at 32 days, or when the
plants initiated silking 83 days after planting. In this test, the yield was significantly
(P < 0.05) higher in the no-tillage (5033.0 kg grain/ha) treatment as compared with
plow-tillage (2597.7 hg/ha).
In another test in which corn, sorghum, and soybeans were compared, there was no
significant difference in either fall armyworm damage or yield between the 2 tillage
systems in corn and sorghum (Table 2). No infestation occurred in any of the soybean
treatments. Significantly (P < 0.05) less damage occurred in the noninsecticide-treated
sorghum as compared with corn. Chlorpyrifos sprays provided significant (P < 0.05)
control of infestations in only the plow-tillage treatments of corn, but significant differ-
ences in yield occurred between insecticide and nontreated plots in both tillage systems
of corn and sorghum.
In conclusion, these data show that infestations of the fall armyworm may be reduced
in no-tillage systems under certain circumstances. Infestations on recently germinated
plants are reduced during the short period of 5-10 days that seedlings remain below the
mulch canopy of certain no-tillage systems. Less oviposition occurs on plants in these
situations and may be due to an inability of moths to locate the crop within the mulch.
When plants grow above the mulch, oviposition quickly becomes similar to that observed
in plow-tillage systems. The ultimate defoliation damage in later crop developmental
stages is similar in no-tillage and plow-tillage systems. The lag in the development of
fall armyworm infestations in no-tillage may have pest management value by reducing
the need for 1 or more insecticide applications to protect young seedlings during the
critical post-germination stages of plant establishment and growth.


Florida Entomologist 71(3)


Till No-Till
Treated' Untreated Treated Untreated

Damage Rating 1.3a 4.2b 2.0a 3.9a
Yield kg/ha x 100 73.2a 22.6b 68.0a 34.8c

Till No-Till
Treated Untreated Treated Untreated

Damage Rating 1.3a 2.1a 1.6a 2.1a
Yield kg/ha x 100 22.6a 11.5b 19.2a 13.4b

Till No-Till
Treated Untreated T Ad Untreated

Damage Rating Oa Oa Oa Oa
Yield kg/ha x 100 0.6a 0.6a 0.7a 0.5a

'Damage rating 60 days after planting was significantly different between crops, but not between tillage. Duncan's
multiple range analysis (P < 0.05) values are for tillage and insecticide treatments within a crop. Damage and yield
analyses are separate.

In comparisons of corn, sorghum, and soybeans, it is apparent that soybeans are
least susceptible. Soybeans would be preferred over the other crops in high hazard
situations for fall armyworm infestations, such as certain double cropping systems that
have later than normal planting dates. In situations where insecticides are required to
suppress fall armyworm infestations, data indicate that chemical control operations
utilized in plow-tillage systems are applicable to no-tillage.


ALL, J. N., AND G. J. MUSICK. 1986. Management of vertebrate and invertebrate
pests. pp. 347-387. In M. A. Sprague and G. B. Triplett (eds.), No-tillage and
surface-tillage agriculture: the tillage revolution. Wiley, New York.
DUNCAN, D. B. 1955. Multiple range and multiple F tests. Biometrics 11: 1-42.
MUSICK, G. J., AND H. B. PETTY. 1973. Insect control in conservation tillage sys-
tems. In Conservation tillage: the proceedings of a national conference. Soil Con-
serv. Soc. America, Ankeny, IA.
SAS INSTITUTE. 1985. SAS user's guide: statistics, Version 5 Edition. Cary, NC. 956
WRENN, E. 1975. Armyworms launch heavy attack on many corn fields in Virginia.
S. E. Farm Press 7: 5, 28.


September, 1988

Castro et al.: Fall Armyworm Symposium 273


Department of Entomology
Mississippi Agricultural and Forestry Experiment Station
Mississippi State University
Mississippi State, MS 39762

Department of Soil and Crop Science
Texas A&M University
College Station, TX 77843


The effectiveness of a maize trap crop in reducing fall armyworm (FAW), Spodoptera
frugiperda (J. E. Smith), larval infestations on sorghum, Sorghum bicolor (L.) Moench.
in areas where sorghum and maize, Zea mays L., are grown together in the same field
was investigated in small plots at Choluteca, Honduras in 1984 and at Starkville, Missis-
sippi in 1984 and 1985. Infestations of FAW larvae were similar on sorghum treatments
with and without the maize trap crop in Honduras and in one test in Mississippi, 1984.
Apparently, the area ratio of 2:1 for sorghum : maize plantings in close proximity did
not differentially restrict FAW activity on the preferred maize plants. However, sig-
nificantly higher FAW larval infestations and plant damage were observed on maize
than on sorghum in pure stand or when grown together with maize as the trap crop in
a second test conducted in 1985 in Mississippi. The results of these studies support
reported observations in greenhouse and field cages showing higher oviposition by FAW
moths on maize than on sorghum. The higher infestation of FAW on maize further
suggests the potential for use of maize as a trap crop for monitoring insect populations
and its possible use as a trap crop control tactic for FAW in areas of low rainfall where
sorghum is grown as a main crop and this insect is a serious pest.


Se investig6 la efectividad del mafz usado como trampa para reducir infestaciones
por larvas del gusano cogollero, Spodoptera frugiperda (J. E. Smith), en sorgo, Sor-
ghum bicolor (L.) Moench., en Areas donde el sorgo y el mafz, Zea mays L., se cultivan
juntos en el mismo campo, en pequefias parcelas en Choluteca, Honduras, en 1984, y
en Starkville, Mississippi, en 1984 y 1985. Infestaciones de larvas del gusano cogollero
fueron similares en sorgo tratados con y sin trampas de maiz en Honduras, y una prueba
en Mississippi en 1984. Aparentemente, la proporci6n de 2:1 de sorgo:maiz sembrados
en corta proximidad, significativamente no restringi6 la actividad del gusano cogollero
en las preferidas plants de maiz. Sin embargo se observaron significativamente mas
altas infestaciones larvales y daflo a las plants de mais que en las de sorgo, o cuando
cultivado junto con maiz como el cultivo de trampa en una sequnda prueba hecha en
1985 en Mississippi. El resultado de estos studios apoyan las observaciones que se han
reportado en invernaderos y jaulas en el campo que indicaron una mayor oviposici6n en
maiz que en sorgo. La alta infestaci6n de gusano cogolleros en el maiz sugiere ademAs
el potential uso del maiz como un cultivo de trampa para chequear las poblaciones de
insects y su possible uso como tActica de control contra el gusano cogollero en Areas de
pocas lluvias donde el sorgo es cultivado come un cultivo principal y este insecto es una
plaga seria.

274 Florida Entomologist 71(3) September, 1988

The fall armyworm (FAW), Spodoptera frugiperda (J. E. Smith), is a polyphagous
insect that attacks 50 non-economically and 30 economically important plants (Ashley
1979). The FAW acts as a cutworm on young plants, cutting plants at soil level, and on
more developed plants the larvae feed on foliage. When severe infestations occur on
maize or sorghum, all the foliage may be consumed except the midrib (Andrews 1984).
Yield reduction in maize by FAW feeding damage ranges from 20 to 87 percent (Hender-
son et al. 1966, Andrews 1980, respectively).
Fall armyworm infestations can be high on hosts in the grass family (Poaceae),
wherein more oviposition occurs on maize than on sorghum, even when the two crops
are grown together (Sifuentes 1967, Van Huis 1981). Resource management has re-
ceived little attention in the past but is perhaps the most feasible overall approach to
FAW control (Lewis & Nordlund 1980). Some research has been conducted using trap
crops to attract a pest species or to provide a more favorable habitat to increase natural
enemies. The interplanting of alfalfa strips in cotton fields is an example (Huffaker &
Messenger 1976) since Lygus hesperus Knight prefers alfalfa over cotton as long as the
alfalfa remains in a lush growing condition.
Because FAW prefer to oviposit on maize, we investigated the hypothesis that a
maize trap crop would concentrate FAW larval populations in sorghum production
fields. Thereby, insecticide applications for control of this pest could be limited to the
area occupied by the trap crop. This pest management tactic has application in high
technology agricultural production areas and would be of value to subsistence farmers
in developing countries, such as Honduras, where 93% of the sorghum is intercropped
with maize (Donaire 1982).


1984 Study I. In 1984 FAW larval infestations in sorghum grown with a maize trap
crop were compared to infestations in pure stand sorghum (adjacent plots) (Fig. 1A) at
Choluteca, Honduras. Sorghum and maize were planted on June 14 and June 21, respec-
tively. Maize, planted in a block in the middle of the sorghum, comprised 20% of the
total area in the trap crop treatment. Each treatment plot was 40 x 40 m arranged in
a randomized complete block design with four replications. Infestation by FAW larvae
and damage ratings (on a scale of 0-9; 0= no damage, 9= plant dead (Wiseman et al.
1966, Wiseman & Davis 1979) were recorded weekly, for six weeks, on 20 plants selected
at random in each treatment plot. To find larvae feeding in the whorl, plants were pulled
from the ground and the whorl leaves were separated. Using this destructive sampling
technique, the whole plant was searched for larvae. The data were analyzed by analysis

4m A 4m 4m B 4m 4m C

S 0s 2 7m
r s m s I N (' 3n ,o N C9

I m

u4 40m 6m 6m 6m 6m 6m

Fig. 1. Sorghum (S) and sorghum-maize (M) trap crop planting designs. A. 1984
Honduras; B. 1984 Mississippi, USA; C. 1985 Mississippi.

Castro et al.: Fall Armyworm Symposium 275

of variance and treatment means were separated by Duncan's multiple range test (Dun-
can 1955).
1984 Study II. Treatments included pure stand maize and sorghum and a trap crop
plot with one-half of the area of each crop separated by 3 m of uncultivated land (Fig.
1B). Each treatment plot was 6 x 15 m separated by a 4 m uncultivated alley. Sorghum
was planted on May 28 and maize on June 3 at Starkville, Oktibbeha County, Missis-
sippi. The delay in maize planting was not intended. The experimental design, sampling
of plants for FAW larvae, estimates of FAW larval feeding damage, and analysis of the
data were as described for the 1984 study I.
1985 Study. In this study, sorghum and maize in pure stands were compared with
a sorghum-maize trap crop where maize and sorghum occurred in an area ratio of 1 maize
: 2 sorghum (Fig. 1C). Sorghum was planted on May 28 and maize on June 3 to duplicate
the 1984 Study II in the same field. Maize and sorghum in the trap crop treatment plots
were separated by a 7 m uncultivated alley. The test design, sampling procedures, and
analysis of data were as described above.


Numbers of FAW larvae per plant or plant damage ratings did not differ significantly
(P>0.05) on any sample date in the 1984 study in Honduras and the 1984 study in
Mississippi. There was a trend for more larvae on maize than on sorghum in the Missis-
sippi study in 1984. However, numbers of larvae per plant (Fig. 2) and plant damage
ratings (Table 1) differed significantly on sorghum and maize in the different planting
systems on all sample dates in the 1985 Mississippi study. More larvae infested maize
than sorghum in pure stand or when grown together with maize as the trap crop. Fall
armyworm larval feeding damage was consistently greater on maize than sorghum in
these systems.


The effective size and location of a maize trap crop in small production areas was
not determined in these studies. Due to adjacent stands of maize and sorghum in the
1984 Study I, the FAW larvae and damage to both crops were distributed uniformly
within and among treatment plots. With a 1:1 ratio of area planted to each crop sepa-
rated by 3 m of uncultivated land (1984 Study II), the crops obviously attracted moths
into the test area, but due to the relatively large area planted to maize and the close
proximity of the crops, the FAW larval population and plant damage to maize and
sorghum were uniformly distributed and similar among treatments. Where sorghum
and maize were planted in a 2:1 ratio (1985 Study), there were significantly (P>0.05)
fewer larvae on sorghum than maize in the trap crop system. The larger numbers of
larvae on maize than sorghum in the pure stand and trap crop planting systems resulted
in significantly (P>0.05) more damage to the maize than the sorghum. Fall armyworm
larval infestations and damage on sorghum grown in pure stand did not differ signific-
antly (P>0.05) from sorghum grown with a maize trap crop. In the relatively small test
area, the FAW infestations on sorghum in adjacent treatments were uniform reflecting
the equal attractiveness of the sorghum treatments in the test, whereas maize appa-
rently attracted a higher population of FAW. The maize plantings in this study appeared
to serve as the trap crop for all sorghum plantings regardless of treatment design. The
close spatial arrangement of treatment plots, allowing for uniform dispersal of larvae,
negated the possible separation of treatment effects on FAW infestations on sorghum
in pure stand and with a maize trap crop.

Florida Entomologist 71(3)



1 1 I

September, 1988


Maize(t rap)

Sorghum (trap)



7/24 7/30 8/6


Fig. 2. Fall armyworm larvae on sorghum and maize in different treatment plots
[trap indicates that sorghum and maize (the trap crop) in 2:1 crop area ratio (See Fig.
1C)]. Starkville, Mississippi. 1985 Study. Means on each date followed by the same
letter are not significantly different [P>0.05; Duncan's multiple range test Duncan

Mean damage rating' per plant on dates
System 7/11 7/19 7/24 7/30 8/6
Trap Crop
Corn 1.15 b2 2.70 a 2.40 a 3.70 a 3.40 a
Sorghum 0.72 bc 1.70 b 1.80b 2.30 b 2.40 b
Pure Stand
Corn 2.08 a 2.70 a 2.60 a 3.60 a 3.90 a
Sorghum 0.68 c 1.70 b 1.90 b 2.10 b 2.20 b
1Damage rating: 0 =no damage, 9-dead plant.
'Means in a column followed by the same letter are not significantly different at [P>O.05; Duncan's multiple range
test Duncan 1955)].


Castro et al.: Fall Armyworm Symposium

Where sorghum is intercropped with maize in many developing countries, the infes-
tations of FAW on both the sorghum and maize plants would be expected to be high
and similar due to the closeness of the plants in the system, often in the same hill. Due
to general dispersal characteristics (Green & Morrill 1970, Morrill & Green 1973) of
FAW larvae and competition for space and food resources, the larvae can readily dis-
perse from the preferred maize plants to adjacent sorghum plants. Therefore, both
maize and sorghum in intercropped plantings can be damaged severely by high FAW
These data support observations by others (Van Huis 1981) of higher FAW larval
infestations on maize than sorghum as a result of the greater attractiveness of maize
for oviposition. A reduction in area planted to maize compared to that planted to sor-
ghum (less area planted to maize than the 2:1 sorghum : maize ratio used in the present
study) and a greater separation of maize from sorghum in a trap crop planting system
may improve the effectiveness of the trap crop management tactic by limiting FAW
damage to the sorghum. This would be especially important in Honduras and surround-
ing areas, for example, during the early part of the crop growing season when the plants
are attacked by a complex of lepidcpterous defoliators, but this needs to be investigated.
Additionally, the initial concentration of eggs and larvae on the maize in a relatively
restricted cropping area, the maize trap crop, would provide pest survey information
required for recommending the application of some pest control tactics. The early detec-
tion of FAW egg masses and small larvae should be made easier on the maize in the
area of the concentrated trap crop. The dispersal of larvae and infestation on the sor-
ghum crop can be reduced significantly by timely application and concentration of insec-
ticide on the limited area of the maize trap crop. This will result in restricting the use
of insectic' les to a relatively small area compared with the total area on which the crop
is produced and will represent a reduction in production cost to the farmer. The reduc-
tion in insecticide use will be less disruptive to the environment and particularly the
associated beneficial organisms in the total cropping system.
These benefits would be desired in crop production systems experiencing pest prob-
lems regardless of the level of technology involved in producing the crop. In developing
countries, where the subsistence farmer can ill afford the use of expensive production
practices, the planting of a preferred host plant, in a small area relative to the total
area of the main crop with little cost to the farmer to achieve an easy, deliberate method
of detection and control of lepidopterous defoliators on crops in early growth stages,
can be a valuable tool for the farmer. However, the use of this planting strategy to
achieve insect pest control with minimal economic sacrifice must be evaluated in the
field in large plots for effectiveness and acceptance.


This research was supported in part by grant AID/DSAN/XXI-G-0149 from the
United States Agency for International Development to the Sorghum and Millet Col-
laborative Research Support Program (INTSORMIL) and was conducted as partial
fulfillment of the memorandum of understanding between the Ministry of Natural Re-
sources of the government of Honduras and INTSORMIL, Acuerdo No. 152
Tegucigalpa, D.C., February 8, 1983. The views and interpretations in this publication
are those of the authors and should not be attributed to USAID. Mississippi Agricultural
and Forestry Experiment Station publication number J-6859.


ANDREWS, K. L. 1980. The whorlworm, Spodoptera frugiperda, in Central America
and neighboring areas. Florida Entomol. 63: 456-467.

278 Florida Entomologist 71(3) September, 1988

ANDREWS, K. L. 1984. El manejo integrado de plagas invertebradas en cultivos ag-
ronomicos, horticolas y frutales en la escuela agricola Panamericana E.A.P./
A.I.D. El Zamorano, Honduras. 91 pp.
ASHLEY, T. R. 1979. Classification and distribution of fall armyworm parasites.
Florida Entomol. 62: 114-123.
DONAIRE, R. E. 1982. Caracterizacion y relaciones ambiente-manejo en sistemas de
frijol y sorgo asociados con maiz en Honduras. Tesis Magister Scientiae, UCR-
CATIE, Turrialba, Costa Rica. 63 pp.
DUNCAN, D. B. 1955. Multiple range and multiple F tests. Biometrics 11: 1-41.
GREEN, G. L., AND WENDELL L. MORRILL. 1970. Behavioral responses of newly
hatched cabbage looper and fall armyworm larvae to light and gravity. J. Econ.
Entomol. 63: 1984-1986.
HENDERSON, C. F., H. G. KINZER, AND E. G. THOMPSON. 1966. Growth aInbred
yield of grain sorghum infested in the whorl with fall armyworm. J. Econ. En-
tomol. 59: 1001-1003.
HUFFAKER, C. B., AND P. S. MESSENGER. 1976. Theory and practice of biological
control, Academic Press, NY. 635 pp.
LEWIS, W. J., AND D. A. NORDLUND. 1980. Employment of parasitoids and pre-
dators for fall armyworm control. Florida Entomol. 63: 433-438.
MORRILL, WENDELL, L., AND G. L. GREEN. 1973. Distribution of fall armyworm
larvae. 1. Regions of field corn plants infested by larvae. Environ. Entomol. 2:
SIFUENTES A., J. A. 1967. Oviposicion de palomillas de cogollero y dano de las larvas
en plantulas de maiz y sorgo, en invernadero. Agriculture Tecnica en Mexico.
II: 311-314.
VAN HUIS, A. 1981. Integrated pest management in the small farmer's maize crop in
Nicaragua. Meded. Landbouwhogeschool Wageningen 81-6. The Netherlands.
221 pp.
WISEMAN, B. R., AND F. M. DAVIS. 1979. Plant resistance to the fall armyworm.
Florida Entomol. 62: 123-130.
WISEMAN, B. R., R. H. PAINTER, AND C. E. WASSOM. 1966. Detecting corn seedling
differences in the greenhouse by visual classification of damage by the fall ar-
myworm. J. Econ. Entomol. 59: 1211-1214.

New Technologies for Taxonomy




According to classical alpha-taxonomy a species is defined as a group of populations
with similarity in appearance. All the individuals of a species are considered to be
relatively uniform in anatomical features and clearly separable from other species. Type
specimens are used with the presumption that they possess all the essential features of
the species. Since the advent during the mid-1970s of electrophoretic methods of uncov-
ering hidden genetic variability, studies have established that natural populations are
highly variable. There is no single individual or a given population that represents the
entire gene pool of a species. Genetic variability among populations reveals geographi-
cal, temporal, clinal and other ecologically related patterns. Unidirectional or bidirec-
tional hybrid sterility has been demonstrated within natural populations that were al-
most morphologically inseparable. These observations led to the genetic concept of a
species. Genetical species can be defined as groups of actual or potentially interbreeding
populations reproductively isolated from other such groups. Thus populations of the
same species are actually or potentially capable of exhanging genes; i.e., they share a
common gene pool. By inference, populations of different species do not share a common
gene pool and by definition are said to be reproductively isolated. Under this concept,
morphologically indistinguishable, but genetically distinct species are called sibling
species. Laboratory hybridization tests and comparative studies of karyotypes have
traditionally been the methods for recognition of sibling species. Recently, the entire
area of taxonomy is undergoing revolution in conepts and methodologies. Electrophore-
tic, chemical, and recombinant DNA techniques aided by computer programs,. have
greatly facilitated the recognition of new genetical species. This symposium endeavors
to address these new technologies for taxonomic identification of arthropods.


Rinderer: New Technologies for Taxonomy 281


USDA-ARS, Honey-Bee Breeding
Genetics & Physiology Laboratory
1157 Ben Hur Road
Baton Rouge, Louisiana 70820


Standard, generally-agreed-upon principles of taxomony guide the naming of groups
as species. Chief among these principles is that species-specific characteristics can exist
because members of different species do not freely produce fertile hybrids.
The history of the nomencalture of subspecies of Western honey bees exemplifies 3
additional taxonomic principles. First, subspecies-specific characteristics have not been
found in honey bees. Second, subspecies are distinguishable as groups by the mul-
tivariate techniques of principal component analysis or discriminant analysis. Third,
subspecies identifications are statistically based and hence probabilistic. These charac-
teristics of subspecies identification are ideally suited to computer applications. Com-
puter systems can be used to collect data and then used to apply complex statistical
procedures to provide identifications.
Such techniques have been used to analyze the honey-bee populations of the
Americas. The honey bees in the Americas constitute populations resulting from sub-
species hybridization. Populations of Africanized bees from South America are mor-
phometrically quite distinct from populations of A. m. scutellata in South Africa. In
every respect, the Africanized bees clearly show the genetical influence of their Euro-
pean parents. Additionally, F, progeny of Africanized and European honey bees provide
further evidence of the hybrid nature of Africanized bees. It is clear that Africanized
bees in the Americas constitute a "hybrid swarm". That is, they are "a continuous series
of morphometrically distinct hybrids resulting from hybridization".


Las normas, de acuerdo comfin, de los principios de taxonomia son la guia en el
nombramiento de grupos como species. Primero dentro de estos principios es que
caracteristicas especificas a la especie pueden existir porque miembros de diferentes
species no produce hibridos f6rtiles libremente.
La historic de la nomenclatura de sub-especies de la abeja melifera Occidental hace
ejemplo de 3 principios taxon6micos diferentes. Primero, caracteristicas especificas a
sub-especies no se han encontrado en la abeja melifera. Segundo, sub-especies son
distinguidas como grupos por las t6cnicas multivariantes de analisis de component
principal o analisis discriminate. Tercero, identificaciones de sub-especies son basadas
estadisticamente y asi probabilisticas. Estas caracteristicas de identificaci6n de sub-es-
pecies se prestan idealmente para la aplicaci6n a computadoras. Systems de computa-
ci6n pueden ser usados para recolectar datos y entonces aplicarles procedimientos de
estadistica complejos para identificarlos.
Tales tecnicas se han utilizado para analisar la poblaci6n de la abeja melifera en las
Americas. Las abejas meliferas en las Americas se constituyen de poblaciones resul-
tantes de hibridizaci6n de sub-especies. Poblaciones de abejas Africanizadas de America
del Sur son morfometricamente muy distintas a la poblaci6n de A. m. scutellata en Sur
Africa. En todo respect las abejas Africanizadas muestran claramente la influencia
gendtica de sus padres Europeos. Adicionalmente, cria F1 de abejas Africanizadas y
Europeas proporcionan mAs evidencia de la indole hibrida de las abejas Africanizadas.
Una comparaci6n de abejas Africanizadas de Argentina y de Venezuela demuestra que

Florida Entomologist 71(3)

estos dos grupos son diferentes morfometricamente y proporciona m6s evidencia de
hibridizaci6n. Queda claro que abejas Africanizadas en las Americas constituyen un
"enjambre hibrido." 0 sea, son "una series continue de hibridos morfomutricamente
distintos resultando de la hibridizaci6n."

This discussion is first intended to describe the special taxonomic problems attending
the subspecies identification of Western honey bees. Secondly, this paper will describe
a developed system of identification that is reliant upon computer-assisted measure-
ments of body parts and computer comparisons of measured data to baseline data
through the application of multivariate discriminant analysis technologies.


A brief review of classical taxonomy reveals some rather clear concepts. First, the
"species is the basic unit of taxonomy and evolution." (Mayr 1970). In order to name a
group as a species, certain generally accepted questions are asked: (1) do similar groups
have allopatric or sympatric occurrences; (2) do similar groups freely hybridize or not;
and (3) do groups have species-specific characters?
Certainly a group that has species-specific characters which does not hybridize with
similar groups and has an allopatric distribution with other groups is deserving of
species designation. The more clearly the information at hand provides these answers
the more certain it is that a species designation is correct. The difficult problems of
species-level taxonomy, for example, undescribed ecoclines and sibling species, are
problems of incomplete information. As more information becomes available, clearer
and more correct applications of the principles of taxonomy can be made.
Different taxonomic problems arise with the Western honey bee (Apis mellifera).
In contrast to other Apoidea where congeneric sympatry and limited distributions are
typical, the Western honey bee occupies an amazingly large and varied area. Naturally,
it ranges from near the Arctic circle in Europe to the Cape of Good Hope to the south
and from the North and South Atlantic coasts of Europe and Africa to the Ural Moun-
tains in the north and the Arabian peninsula's coast of Oman in the south. Populations
of Western honey bees freely interbreed with opportunity and produce fertile offspring.
Indeed, honey bees may have only occupied this area since the last ice age of the late
Pliocene (Ruttner 1987). In addition to this natural distribution, man in the last 300
years has spread honey bees throughout the New World areas of North and South
America, Australia and New Zealand. Although clearly one species, A. mellifera shows
substantial variation in morphology, behavior and physiology (Cornuet & Louveaux
1981). This subspecific variation has been the primary focus of honey-bee taxonomists
within this century. Historically, rules governing the description and naming of varing
populations of honey bees at the subspecies level have been non-existent. Guides to such
nomenclatural issues have frequently included political and sociological elements as well
as biological ones. As recent as 1953, Maa presented a list of 146 subspecific names of
honey bees which he considered acceptable. This list comprised a revision of 600 named
groups of Apis mellifera used previously. The difficulties of subspecies honey-bee
taxonomy are obviously entangling. If it were not important for achieving various social
and economic goals having to do with the public health and commercial characteristics
of honey bees, subspecies taxonomy of honey bees would probably have collapsed.
Nonetheless, biological differences among honey bees that reflect adaptations to
ecological differences do exist. A long tradition of biometry (Buttel-Reepen 1906, Al-
patov 1929, 1948, Goetze 1930, 1940, 1964, Dupraw 1965a, 1956, Daly & Balling 1978,
Ruttner 1986) has attempted to describe these differences. Biometrical applications of


September, 1988

Rinderer: New Technologies for Taxonomy

modern statistical methods have lead to certain suggestions for taxonomic decisions at
the subspecific level with A. mellifera. Ruttner (1986) suggests the "operational
taxonomic units" called for by Sneath & Sokal (1973) can be the means of standard
individual measures. The application of standard multivariate methods which are capa-
ble of describing the magnitudes of differences between populations, such as discrimin-
ant analysis or principal component analysis provides a basis for the acceptance of
groups as subspecies. Applying these principles Ruttner (1986) described 23 subspecies
from the Old World range of the species. In doing so, he observed that variation among
groups was quantitative not qualitative and frequently occurred within narrow limits.
The application of modern computer technology to the collection and analysis of
morphometric honey-bee data was begun by DuPraw (1965a, 1965b) and refined by Daly
& Balling 1978 and Daly et al. 1982.


Data collection was according to procedures adapted to the morphometric analysis
of honey bees by Daly et al. (1982). Body-part images from wings, legs and sternites
mounted on slides were projected with a standardized magnification from a projecting
microscope to a digitizing tablet connected to a micro-computer (Fig 1, A and B). Vari-
ous points on the projected images were identified by the operator with the aid of an
electronic "mouse" or "cursor". The positions of the electronically marked points were
transmitted to the micro-computer where they were converted to values representing
actual lengths angles and counts of the 25 characters measured (Daly & Balling 1978).
In the numerical analysis of a sample, reductionist statistical procedures as used to
calculate and then compare descriptive values of the sample to similar values for baseline
groups. The analysis provides probabilities of a sample belonging to groups, based on
these comparisons. At least two and probably more of the major suppliers of statistical
software provide discriminant analysis packages that are appropriate for use in honey
bee taxonomy.
Many interesting problems in honey bee classification are found among the honey
bees brought to the New World. These bees have come from a variety of sources and
the populations of bees found in the Americas cannot be said to be members of identifi-
able sub-specific groups. Nonetheless, differences do occur among American populations
of bees. Those populations derived in part from parents from the Highland population
of East Africa (A. m. scutellata) are considered economically undesirable because of
excess stinging (Collins et al. 1982) and inferior honey production (Rinderer et al. 1984,
1985). This parental stock was introduced to apiaries near Piracicaba Brazil from Africa
in 1957 (Kerr 1967). Because of economic concerns, the identification of Africanized bees
has become important.
As examples of the value of the discriminant analysis of morphometric measures,
the results of two separate analyses are presented. The first compares a sample of a
population of Africanized bees to a sample of the parental population in Africa. The
second compares Africanized and European honey bees and F, progeny of crosses be-
tween these two.


One use of multivariate discriminant function analysis is to compare two or more
groups on the basis of a variety of quantitative measures. Such an analysis was done
in a description of morphometric differences between South American Africanized and
South African (Apis mellifera scutellata) honey bees (Buco et al. 1987). This study
compared the two groups for each of the univariate measures with the use of t-tests.

Florida Entomologist 71(3)


I (

Fig. 1. Photographs of the data collection hardware. A. The projection of body part
images onto the digitizing tablet. B. The data collecting assembly: projecting micro-
scope, digitizing tablet and micro-computer.


September, 1988


Rinderer: New Technologies for Taxonomy 285

Nineteen of the 24 characteristics differed between the groups (P = 0.05 or less).
However, a full appreciation of the overall difference between the groups depended
upon the multivariate discriminant further analysis.
The differences found in comparisons of individual characteristics were reflected in
the multivariate discriminant function analysis. The two populations were completely
separated by the discriminant function (Fig 2). The posterior probability of group mem-
bership for all cases indicated a correct classification of P = 0.9984 or greater using the
derived discriminant function.
The feral bees of South America constitute a population which is morphometrically
quite different from the parental population of A. m. scutellata in Southern Africa.
Comparisons of individual characteristics, as well as the composite analysis represented
by the discriminant function, indicated that Africanized bees are generally larger than
their African parental stock.
The most likely explanation for these differences is that Africanized bees in the
Americas are a subspecific hybrid swarm (King 1968) which has resulted from matings
with European bees and then subsequent crossing and back-crossing. The generally
intermediate morphometric character of Africanized bees suggests that a large number
of loci are involved in the determination of the honey bee's morphometric make up.
Predictable Mendelian events associated with crossing and then backcrossing between
African and European bees would produce the Africanzed bees that were measured.
Such bees clearly show the influence of both their European and their African paren-
A different use for multivariate discriminant function analysis is to describe popula-
tions in terms of functions and use the functions to classify unknown samples. This was
originally lone for Africanized and European bees for the Americas by Daly & Balling
(1978). However, the experimental production of Fi progeny of European and Af-

o African



-6.0 -4.0 -2.0 0.0 2.0 4.0 6.0

Discriminant Axis

Fig. 2. Histogram of the results of the multivariate discriminant function analysis
of African and Africanized bees.

Florida Entomologist 71(3)

ricanized parents made possible the development of discriminant functions for the three
groups (Rinderer et al., unpublished).
The multivariate discriminant analysis correctly identified all samples as Af-
ricanized, European or F1 hybrids (Fig. 3) Most classifications were at P = 1.00 and
all but one were at P s 0.090. The remaining sample, an F1 colony, was correctly
classified at P = 0.55. As expected, the discriminant analysis procedure selected the
F, category as the second most likely classification for all Africanized and European
colonies. Interestingly, the second most likely category for the F1 samples was Euro-
pean for all but 3 Fi samples. Thus, the procedure was conservative in its classification
of Fi colonies in ways appropriate to regulatory needs. Only clearly Africanized or Fi
colonies would be so classified.
The identification of unknown samples by comparison to this baseline can be ex-
pected to correctly classify samples of Africanized and European worker bees and sam-
ples of worker bees that are completely or mostly F1 hybrids. According to established
practice, such identifications would be made with posterior probabilities of group mem-
bership of P 0.90. Samples which are a mixture of bees would likely either be classified
with a low probability of group membership (P<0.90) or the range for the majority of
measurements would span 2 or 3 of the 3 group ranges.

U Ur



AA e

* *
* *
; 4o


A Africanized
U Hybrids
-5 European
-6 -i --
-7 -6 -5 -4 -3

-2 -1

0 1 2 3 4 5 6

Discriminant Function One

Fig. 3. Scatterplot of the results of the multivariate discriminant function analysis
of Africanized, European and their F1 hybrid bees.

-2 -


September, 1988


Rinderer: New Technologies for Taxonomy

Such samples would warrant close inspection. The classification of individual bees
may provide evidence of mixed samples. Data supporting the classification of individual
bees have been analyzed and discriminant functions of identifying individual bees have
been developed.
Scientific and commercial issues regarding identifying Africanized bees and their
hybrids have been well served by the computer-assisted discriminant function analysis
of morphometric data. The process has permitted rather difficult taxonomic problems
to be resolved clearly with numerical procedures. The success of the techniques with
honey bee groups even at the subspecific level, suggests that other identification prob-
lems are within the capabilities of this tool to solve.


In cooperation with the Louisiana Agricultural Experiment Station.


ALPATOV, W. W. 1929. Biometrical studies on variation and races of the honey bee
(Apis mellifera L.). Quart. Rev. Biol. 4: 1-58.
ALPATOV, W. W. 1948. The races of honey bees and their use in agriculture. In
Russian. Sredi prirody, v.4. Moscow Sec. Res. Nat., Moscow.
TER, AND R. M. CREWE. 1987. Morphometric differences between South
American Africanized and South African (Apis mellifera scutellata) honey bees.
Apidologie 18: 217-222.
BUTTEL-REEPEN, H. V. 1906. Apistica. Mitt. Zool. Mus., Berlin: 3: 117-201.
defense by Africanized and European honey bees. Science 218: 72-74.
CORNUET, J. M., AND J. LOUVEAUX. 1981. Aspects of genetic variability in Apis
mellifera L. pp. 85-93. In Biosystematics of Social Insects. Howse, P. E. and
Clement, J. L. eds. Academic Press, New York.
DALY, H. V., AND S. S. BALLING. 1978. Identification of Africanized honey bees in
the Western Hemisphere by discriminant analysis. J. Kansas Entomol. Soc., 51:
DALY, H. V., K. HOELMER, P. NORMAN, AND T. ALLEN. 1982. Computer-assisted
measurement and identification of honey bees (Hymenoptera: Apidae). Ann. En-
tomol. Soc. America. 75: 591-94.
DUPRAW, E. J. 1965a. The recognition and handling of honeybee specimens in non-
Linnean taxonomy. J. Apic. Res. 4: 71-84.
DUPRAW, E. J. 1965b. Non-Linnean taxonomy and the systematics of honeybees.
Syst. Zool. 14: 1-24.
GOETZE, G. K. L. 1930. Variabilitats und zuchtungsstudien an der honigbiene mit
besonderes berucksichtung der langrusseligkeil. Arch. Bienenkunde 11: 185-236.
GOETZE, G. K. L. 1940. Die Beste Biene. Liedloff, Loth und Michaelis. Leipzig.
GOETZE, G. K. L. 1964. Die Honigbiene in naturlicher und kunstlicher Zuchtauslese.
Teil I, II. Monoger. angew. Entomol. 19: 1-120; 20: 1-92.
KERR, W. E. 1967. The history of the introduction of African bees to Brazil. South
African Bee J. 39: 3-5.
MAYR, E. 1970. Populations, Species and Evolution. Belknap press, Cambridge, Mas-
foraging characteristics of Africanized and European honey bees in the neot-
ropics. J. Apic. Res. 23: 70-79.
RINDERER, T. E., A. M. COLLINS, AND K. W. TUCKER. 1985. Honey production
and underlying nectar harvesting activities of Africanized and European honey
bees. J. Apic. Res. 23: 161-167.


Florida Entomologist 71(3)

RUTTNER, F. 1986. Geographical variability and classification, pp. 23-52. In: Bee
Genetics and Breeding, T. E. Rinderer, ed. Academic Press. Orlando, Florida.
SNEATH, P., AND R. SOKAL. 1973. Numerical Taxonomy. Freeman Co., San Fran-
cisco, California.


Emory University
Atlanta, Georgia 30322

Centers for Disease Control, Malaria Branch
Chamblee, Georgia 30333


Many insects belong to species complexes wherein member species are morphologi-
cally indistinguishable. Yet, in numerous cases, both economic and medical exigencies
require that the species of such individuals be reliably determined. The specific problem
addressed here concerns the Anopheles gambiae mosquito complex which includes An.
gambiae and An. arabiensis, currently the two major African malaria vectors. The
present work uses a ribosomal DNA gene probe to differentiate member species of this
complex. The method is shown to be extremely useful and sensitive because it can easily
test just portions of a single dried adult. The rationale for using ribosomal DNA clones
to provide such diagnostic probes to distinguish among other morphologically identical
insect species is also discussed.


Muchos insects pertenecen a un complejo de species donde los miembros son mor-
fol6gicamente indistinguibles. Sin embargo, en muchos casos, exigencias econ6micas y
m6dicas requieren que la especie sea identificada con seguridad. El problema especifico
tratado aqui le concierne al complejo del mosquito Anopheles gambiae, que incluye a
An. gambiae y An. arabiensis, que son actualmente los dos vectores principles de la
malaria Africana. El present trabajo usa un gene ribosomal de DNA como sonda para
diferenciar species de este complejo. Se demostr6 que el m6todo es extremadamente
ditil y sensitive porque facilmente puede probar solo porciones de un solo adulto secado.
Se discute tamabi6n la racional del uso de clones ribosomal de DNA para proveer sondas
diagn6sticas para distinguir entire otras species de insects morfol6gicamente id6nticas.

Malaria is the most debilitating disease in the world today; among the 92 million new
cases each year, there are close to 1 million deaths, nearly all of which are young
children (Service 1985). Although malaria is found world-wide, the problem is most
acute in subsaharan Africa where the disease presents an enormous obstacle to social
and economic development. Although malaria was eradicated from most of its original


September, 1988

Finnerty & Collins: New Technologies for Taxonomy 289

temperate range in the 1960s, its incidence in the tropics continues to increase, due in
part to insecticide resistance in the mosquito vector and chemoprophylactic drug resis-
tance in the malaria parasite (Service 1985).
Two of the principal African malaria vectors, An. gambiae and An. arabiensis,
belong to the Anopheles gambiae species complex, which contains six member species.
All of the species are morphologically indistinguishable, and two (and in some areas as
many as four) of the species are often sympatric in many parts of Africa (White 1974).
The species' two principal vectors, An. gambiae and An. arabiensis, differ in behavior
and preferred habitat. Moreover, there is abundant evidence suggesting that the two
major vector species may not be equally involved in malaria transmission, depending
upon the season and location (Coluzzi et al. 1977). Therefore, one of the problems for
epidemiological studies of these insect vectors is to determine whether an individual
female mosquito is infected with the malaria parasite, and also, to what species does
she belong. The latter consideration is most pressing for ecological studies of habitat
and reproductive behavior, which then provide information essential for the design of
various control strategies.
We have focused our work upon two areas: One involves development of a rapid,
accurate and easy means of distinguishing the species of single mosquitoes. The other
involves what will ultimately become a fairly detailed analysis of the ribosomal RNA
genes from the perspective of population structure and evolution of the species complex.
In developing a rapid means of distinguishing species we chose to use some feature
of DNA because of its stability in dried tissue. Dessication is the easiest way todeal
with field specimens, and further, a rapid, highly sensitive assay was already in use for
detecting the presence of the malaria sporozoite antigen in single dry mosquitoes (Col-
lins et al. 1984). We chose therefore to focus on the ribosomal DNA gene family because
they are most likely to be present in hundreds of copies per genome, thus increasing
the sensitivity of a probe derived from this region. More important, the ribosomal genes
contain a highly conserved coding region as well as a spacer region, and although the
latter is transcribed, its sequences are processed out, and it is apparently subject to
fairly rapid divergence (Beckingham 1982). Thus, when used as probes, sequences from
portions of the coding region flanked by spacer sequence are expected to reveal species-
specific restriction fragment length polymorphisms (RFLP). Such characteristic RFLP
have been used in numerous instances to distinguish among closely related species
(Jeffreys 1981, Avise et al. 1979, Langley et al. 1982).


A genomic library from An. gambiae was constructed in EMBL-3 as described in
Collins et al. (1987). A variety of clones containing ribosomal coding sequences were
isolated from this library as previously described (Collins et al. 1987), then restricted
with EcoRI and Sall or with EcoRI and BamHI and subjected to Southern analysis.
The blots were probed with a ribosomal DNA clone isolated from Sciara coprophila
(Renkawitz et al. 1979). Since the coding regions tend to be highly conserved they were
expected to hybridize with the Sciara clone. However, the non-vector non-hybridizing
fragments were expected to derive from spacer or intron regions, and we therefore
chose these fragments for further study.
Southern analysis and probe preparation were carried out as described in Collins et
al. (1987). An rDNA (pBC2) clone from S. coprophila was provided by S. Gerbi. The
18S and 28S coding region clones, derived from Calliphora erythrocephala, were pro-
vided by K. Beckingham (Beckingham 1980). Mosquitoes were supplied by the Malaria
Branch, Centers for Disease Control, Atlanta. DNA extraction is a modification of a
method reported by Livak (1984). All of the techniques used for Southern analysis are

290 Florida Entomologist 71(3) September, 1988

fully described by Maniatis (1982); certain modifications are described in Collins et al.


Preliminary characterization of clone XAGrl2. In order to be certain of the nature
of the fragments we would ultimately use as probes to reveal a species-characteristic
RFLP, we chose one representative clone for further study. A preliminary restriction
map of this clone is shown in Figure 1. The coding regions were roughly delineated by
Southern analysis, again using separate Calliphora 18S and 28S clones as probes. The
arrangement left to right of 28S-18S-non-coding-28S is as expected for insect ribosomal
genes, and this clone evidently contains one entire repeat unit. The fragments of interest
were those which include both coding and adjacent spacer sequence.
Identification of species-diagnostic RFLP. The interesting fragments (identified ac-
cording to the procedure outlined in Methods) were those which did not appear to
hybridize to the heterologous rDNA probe, pBC2. Figure 1 shows the position of two
fragments, pAGrl2A and pAGrl2B, within the rDNA colone. Such fragments were
either subcloned or eluted from a gel and then used individually to probe blots containing
genomic DNA from different members of the An. gambiae ccomplex. One such blot is
shown in Figure 2A. This fragment (pAgrl2A), when used as a probe, reveals a charac-
teristic RFLP which, as demonstrated in the figure, can distinguish all five of the
member species. Another fragment (pAGrl2B) was sued to probe a similar blot, reve-
aling a different RFLP which is shown in Figure 2B. The RFLP revealed in Figure 2A
most likely reflects the changed position of the underlined EcoRI site (Figure 1). The
RFLP shown in Figure 2B may reflect the changed position of the underlined HindIII
site shown in Figure 1, although variation in the HindIII site to its left is also possible.
Ribosomal DNA probes are particularly useful because of their great sensitivity.
Since the rDNA genes of many insects are present in 200-500 copies per nucleus, they
are readily visualized with an overnight exposure of Southern blots hybridized with a
moderately hot (107 cpm/[g) probe. The blot shown in figure 2 contains an amount of
DNA equivalent to one fourth of a mosquito per lane.


Thus far, species identification in the An. gambiae complex relied almost exclusively
on cytogenetic criteria (Coluzzi et al. 1979). This method requires a great deal of tech-

pAGri2B pAGri2A


1 KB

Fig. 1. Simplified restriction map of the An. gambiae rDNA clone, XAGrl2. The
dark bars represent transcribed regions, not all of which is coding region. The colone
contains one complete intronless repeat unit. S = SalI, E = EcoRI, B = BamHI, H
= HindIII.

Finnerty & Collins: New Technologies for Taxonomy

6-8 KB [

4-6 KB

3.7 KB -"N

2.0 KB -

1.3 KB -

Fig. 2A. Autoradiograph of a Southern blot from an overnite exposure containing
an EcoRI digest of the indicated mosquito DNAs probed with pAGrl2A (107 cpm/p.g).
Shown from left to right are An. arabiensis, An. gambiae, An. guadriannulatus, An.
merus, and An. melas. One quarter of a single mosquito DNA extract was used per

nical expertise, as well as freshly blood fed mosquitoes. Alternative methods based
upon differences in several dehydrogenase isozyme patterns have other disadvantages,
among which is the requirement for fresh or frozen specimens (Miles 1978, Marchand
& Mnzava 1985). Cuticular hydrocarbon profiles can be used to reliably distinguish dried
specimens, but the method is not practical for large numbers of specimens or for field
use (Carlson & Service 1980, Hamilton & Service 1983). Recently, we demonstrated
that the RFLP revealed by rDNA probes were as reliable as either cytogenetic or
isozyme methods by direct comparison of results from the same individuals or siblings
from an isofemale line (Collins et al. 1988, Collins et al., unpublished). The DNA method
has several clear advantages, one being that dried individuals of virtually all life stages
can be used. Further, only part of the specimen is required for DNA analysis, so that

Florida Entomologist 71(3)

6.5 KB -

3.2 KB -


2.5 KB

1.5 KB
1.4 KB

Fig. 2B. Autoradiograph of a Southern blot containing a HindIII digest of the indi-
cated mosquito DNAs probed with pAGrl2B. All other conditions are as described for
figure 2A.

assessment of the blood meal source as well as the presence of sporozoite antigen can
readily be done on the remainder of the specimen. However, it must be emphasized
that the DNA method currently requires DNA extraction and Southern blotting. A
more useful version of the test would emerge if the necessity of running and blotting a
gel could be eliminated in favor of a "dot blot." Such a dot blot assay would require a
series of species-specific probes. We have therefore begun to determine whether there
are sequences other than the spacer within the ribosomal DNA genes which may have
diverged sufficiently between species to the extent that they no longer behave as if they
are homologous under the stringent conditions we use for Southern analysis. One obvi-
ous candidate is an intron region, and many Diptera contain a mixture of 28S gene
types, some with a single intron, and others not having the intron, such as XAGrl2,
shown in Figure 1. Since mosquito rDNA genes may or may not bear such introns, the
intron itself might be even more likely (than the transcribed spacer) to contain se-
quences which differ significantly between species (Gerbi 1985). Therefore we have
examined our collection of An. gambiae rDNA clones in order to focus upon those which

September, 1988



Finnerty & Collins: New Technologies for Taxonomy


might contain an intron in the 28S region. One such "species-specific" sequence has been
subcloned from an An. gambiae intron. Preliminary analysis of this fragment indicates
that it is perfectly useful as an An. gambiae-specific probe. Currently we are searching
for analogous fragments in rDNA clones from the other members of the complex which
could prove to be useful as species-specific probes. In contrast to the highly conserved
18S and 28S coding regions, insect rDNA is generally found to contain both spacer and
intron sequences which appear to diverge more rapidly between even closely related
species. The rDNA gene family is therefore an excellent candidate for providing (as it
has in the present case) a number of fragments, some of which may be used to reveal
a species-specific RFLP, as well as other fragments which could behave as species-spe-
cific probes.


Ne thank J. M. Clark for preparing the manuscript. Research was supported by
S. Army contract DAMD 17-85-C-5184 to VF and by U.S. Agency for International
Development, PASA BST-0453-P-HC-2086-02 to FHC.


AVISE, J., R. LANSMAN, AND R. 0. SHADE. 1979. Use of restriction endonucleases
to measure mtDNA sequence relatedness in natural populations. I. Peromyscus.
Genetics. 92: 279-295.
BECKINGHAM, K. 1980. The ribosomal DNA of Calliphora erythrocephala: An
analysis of hybrid plasmids containing ribosomal DNA. J. Mol. Biol. 136: 349-
BECKINGHAM, K. 1982. In: The Cell Nucleus, X. H. Busch and L. Rothblum, eds.
Academic Press, New York, pp. 205-263.
CARLSON, D. A., AND M. W. SERVICE. 1980. Identification of mosquitoes of
Anopheles gambiae species complex A and B by analysis of cuticular components.
Science. 207: 1089-1091.
AKOH, AND R. S. NUSSENZWEIG. 1984. First field trial of an immunoradiometric
assay for the detection of malaria sporozoites in mosquitoes. American J. Trop.
Med. Hyg. 33: 227-230.
SANSKY, AND V. FINNERTY. 1987. A ribosomal RNA gene probe differen-
tiates member species of the Anopheles gambiae complex. American J. Trop.
Med. Hyg. 37: 37-41.
NETT, J. SANDE ODERA, AND V. FINNERTY. 1988. Comparison of DNA-probe
and isoenzyme methods for differentiating Anopheles gambiae and Anopheles
arabiensis (Diptera: Culicidae). J. Med. Entomol., in press.
divergences between mosquitoes with different inversion karyotypes in polymor-
phic populations of the Anopheles gambiae complex. Nature. 266: 832-833.
differentiation and adaptation to human environments in the Anopheles gambiae
complex. Trans. R. Soc. Trop. Med. Hyg. 73: 483-497.
GERBI, S. 1985. Evolution of ribosomal DNA. In: Molecular Evolutionary Genetics.
R. J. MacIntyre, ed. Plenum Corp., New York, pp. 419-517.
HAMILTON, R. J., AND M. W. SERVICE. 1983. Value of cuticular and internal hydro-
carbons for the identification of larvae of Anopheles gambiae Giles, Anopheles
araiensis Patton and Anopheles melas Theobald. Ann. Trop. Med. parasitol. 77:

Florida Entomologist 71(3)

JEFFREYS, A. J. 1981. Recent studies of gene evolution using recombinant DNA. In:
Genetic Engineering, II. R. Williamson, ed. Academic Press, New York, pp.
map variation in the AdH region of Drosophila. Proc. Nat. Acad. Sci. USA. 79:
LIVAK, K. 1984. Organization and mapping of a sequence on the Drosophila
melanogaster X and Y chromosomes that is transcribed during spermatogenesis.
Genetics. 107: 611-634.
MARCHAND, R. P., AND A. E. P. MNZAVA. 1985. A field test of a biochemical key
to identify members of the A. gambiae group of species in northeast Tanzania.
J. Trop. Med. Hyg. 88: 205-210.
MILES, S. J. 1978. Enzyme variations in the Anopheles gambiae Giles group of species
(Diptera, Culicidae). Bull. Entomol. Res. 68: 85-96.
RENKAWITZ, R., S. A. GERBI, AND K. H. GLATZER. 1979. Ribosomal DNA of the
fly Sciara coprophila has a very small and homogeneous repeat unit. Molec. Gen.
Genet. 173: 1-13.
SERVICE, M. W. 1985. Anopheles gambiae: Africa's principal malaria vector, 1902-
1984. Bull. Ent. Soc. America. Fall 1985: 8-12.
WHITE, G. B. 1974. Anopheles gambiae complex and disease and transmission in
Africa. Trans. R. Soc. Trop. Med. Hyg. 68: 278-301.


Department of Entomology and Nematology
University of Florida
Gainesville, Fl 32611


African-derived honeybees, expected to enter the United States within two years,
will adversely affect the honey and pollination industries and will be an environmental
hazard to the public. Control measures are dependent upon a reliable identification
method, heretofore unavailable. Nuclear DNA restriction fragment polymorphisms are
being found that successfully distinguish African and European bees as well as hybrids
between the two. Beyond identification for regulatory purposes, DNA markers will
enhance studies on African-European bee genetic interactions and population dynamics.


Se espera que abejas africanizadas entren los Estados Unidos dentro de dos aflos y
que adversamente afectaran las industrial de la miel y de la polinizaci6n, y tambi6n serin
un peligro al public. Las medidas de control dependent de un m6todo de identificaci6n
de confianza, hasta ahora no disponible. El polimorfismo de la restricci6n de fragments
de DNA nuclear se ha encontrado que distinguen con 6xito entire abejas Africanas y
abejas Europeas, asi como entire hibridos de las dos. Despues de la identificaci6n por
prop6sitos regulatorios, marcadores de DNA mejoraran los studios sobre la in.eracci6n
gendtica y el dinamismo de poblaci6n de las abejas Africanas-Europeas.


September, 1988

Hall: New Technologies for Taxonomy

European honeybees were imported to the American continents at the time of coloni-
zation. They are now vital to the pollination of many agricultural crops at an estimated
US annual value of $19 billion (McDowell 1981). African honeybees (Apis mellifera
scutellata; Ruttner 1976), accidently released in 1957 from experimental hives in Brazil
(Kerr 1967, Woyke 1969, Michener et al. 1972, Michener 1975), have undergone a phe-
nomenal migration with the replacement of the extant European bees throughout much
of South and Central America (Taylor 1977). The bees are now in Mexico and are
expected to enter Texas en masse within two years and the panhandle of Florida within
six years (Taylor 1985). In Floridia's tropical climate, African bees could reach saturat-
ing densities (Taylor & Spivak 1984). Excessive stinging, excitability and swarming
make management of African bees difficult. If they cannot be kept from apiaries,
beekeepers will discontinue their business. Competitive European bee populations will
be eliminated, and numbers of commercial colonies available for pollination will be re-
duced. Large populations of African bees will increase stinging incidents, and resulting
publicity will likely impact tourism in Florida. The threat to human and animal popula-
tions has been sensationalized but is, nevertheless, real (Taylor 1986). Proposed
methods of control include quarantine, extermination, and certification of breeding
stock. These are dependent upon diagnostic identification of African honeybees (Page
& Erickson 1985). However, current methods are inadequate. Morphometric statistical
analysis, now most commonly used to identify African bees (Daly & Balling 1978, Daly
et al. 1982), is subject to error due to environmental influences and overlapping distri-
butions. The composition of cuticular hydrocarbons (Carlson & Bolten 1985, Smith 1988)
has considerable promise for identification, but the inheritance of these compounds is
not understood. Honeybees show limited allozyme variation which is generally charac-
teristic of hymenopteran insects. Alleles of only four proteins are known to have signif-
icant frequency differences between African and European bee populations, but none
are diagnostic. Therefore, allozyme analysis cannot make certain identifications, espe-
cially of hybrids after several generations (Rinderer & Sylvester 1981).
Honeybee subspecies are distinguished here through DNA fragments generated by
restriction endonucleases. A number of studies have demonstrated the value of this
approach in establishing the genetic relatedness of organisms (Brown et al. 1982, Cann
et al. 1984, Ferris et al. 1982). DNA restriction fragment polymorphisms do not neces-
sarily result in, nor their detection depend upon, functional changes subject to selection.
Thus, they can provide allele distinction at many loci within natural populations. In the
initial study with honeybees, greater DNA polymorphism was found than all the protein
differences reported so far (Hall 1986).
Restriction enzymes recognize and cut DNA at specific short sequences, most com-
monly consisting of four to six nucleotides. Alterations in the sequences, due to genetic
diverence, are manifested in the resulting DNA fragments, separated according to size
by electrophoresis. When the total genome is digested, thousands of different-size frag-
ments are generated. To visualize only a small portion at a time and to compare the
homologous regions between individuals, use is made of DNA probes (Southern 1980,
Maniatis et al. 1982). These short pieces of cloned radioactive DNA recognize homolog-
ous sequences, binding through complementary base pairing, i.e. hybridization. Thus
genetic distinction is based on a small percentage of nucleotides comprising the restric-
tion sites within the probe region (Jeffreys 1979).


As a source of probes, random fragments of honeybee nuclear DNA were cloned in
E. coli bacteria, by insertion into plasmid pBR322 (Hall 1986). To have probes suffi-
ciently large to overlap a number of sites, clones carrying inserts of honeybee DNA

296 Florida Entomologist 71(3) September, 1988

greater than four kilobases were also selected. Clones with low copy number sequences
were distinguished from those with repetitive sequences.
DNA was isolated from uncapped honeybee larvae. Each sample came from a single
colony, thus represented a combination of progeny from a single queen mated with one
to a dozen drones. European bees were provided by N. Gary (Dept. of Entomology,
Univ. of Calif., Davis) established by R. Page (Dept. of Entomology, Ohio State Univ.)
and E. Erickson (USDA-ARS, Tucson) as members of a closed breeding population
from stocks across the United States. African samples from Costa Rica and Venezuela
were provided by O. Taylor, M. Spivak (Dept. of Entomology, Univ. of Kansas) R.
Hellmich, A. Collins and T. Rinderer (USDA-ARS, Baton Rouge). DNA can be isolated
from any stage, but the soft larval tissue facilitates DNA isolation. Also, the use of
larvae ensures the colony origin of samples, since adults can drift among hives. Each
DNA sample was digested separately with different restriction enzymes that recognize
short nucleotide sequences (four). Short sequence enzymes generate more fragments,
increasing the chances of finding differences. Each enzyme digest was placed in a sepa-
rate lane in an agarose gel, and the fragments were separated by electrophoresis and
blotted onto membranes of derivatized nylon. Probes were radioactively-labelled with
32P deoxycytidine by nick translation. The blots were hybridized with the denatured
labelled probes, one at a time, washed and exposed to X-ray film (Southern 1980,
Maniatis et al. 1982).


DNA Differences:

DNA from a few European and African samples was digested with nine separate
restriction enzymes. Of sixteen probes hybridized to these samples, three fourths re-
vealed polymorphisms. Pairwise comparison of the fragments clearly showed that Euro-
pean samples from the U.S. were more distantly related to an African sample from
Costa Rica than they were to each other (Hall 1986). Some of the first polymorphisms
have promise as diagnostic markers. Enzyme-probe combinations that revealed differ-
ences among a few samples were tested against multiple samples of European bees from
different U.S. locations and African bees from Costa Rica and Venezuela. So far three
of the probes reveal restriction fragments present in all members of one group and
absent in all members of the other group (Hall 1988 a,b). Other fragments are totally
present or absent in one group but variable in the other. Although many more samples
need to be tested, these results strongly indicate that numerous diagnostic differences
will be found. By a rough estimation, the probes tested so far together represent only
about 1/2000th of the genome.
Since the probes were derived from random fragments of DNA, presumably they
represent loci scattered throughout the genome. Although clones to low copy number
sequences were selected, some polymorphisms seen in very intense bands may,
nevertheless, be of repetitive sequences. These could be advantageous, since each probe
would represent many loci. To form the discrete bands needed for analysis, fragments
would have to be derived entirely from within the repetitive sequence. Those spanning
the repetitive sequence and different flanking DNA would be of many different sizes,
forming an unresolvable smear.
Restriction site variations in nuclear DNA sequences show Mendelian inheritance
as codominant alleles (Cavenee et al. 1982, Gusella et al. 1984) which enables the iden-
tification of heterozygous loci. A single fragment characteristic of a subspecies will be
present in the homozygotes and heterozyotes but will be absent only in homozygotes of
the other subspecies. Differences detected by two thirds of the probes in the samples

Hall: New Technologies for Taxonomy

from South and Central America are represented by the lack of fragments, indicating
that the loci are homozygous for African alleles (or at least not heterozygous with any
of the alleles in the European samples). Interestingly, among Venezuelan bees, greater
homozygosity is found in feral swarms. Samples from managed apiaries show varying,
although minor, amounts of European-characteristic fragments, which probably repre-
sent a minority of colony members sired by European drones.

Use of DNA Markers to Study Honeybee Population Dynamics

DNA restriction fragment polymorphisms will enable studies that have been limited
due to the lack of protein variation in natural bee populations. The term "Africanized
bees" is a presumption that the honeybees in South and Central America are hybrids
but further implies that they are mostly European with introgression of African genes.
This was a logical assumption at the time the bees were released when, as it has been
reported, there were only twenty-six African queens compared to the entire extant
European population (Kerr 1967). However, the extent to which African and European
bees freely interbreed and the degree of hybridization is not well documented and has
become a controversial subject (Taylor 1985, Rinderer 1986).
Finding loci homozygous for African DNA alleles is consistent with the predomi-
nance of African morphological and behavioral characteristics. This may be solely a
consequence of tropical environmental selection against hybrids of European bees. How-
ever, other mechanisms may exist that allow retention of the African genotype (Taylor
1985). The extent to which the South and Central American bees are hybrids can be
clearly resolved with further DNA restriction fragment analysis. Possible genotype
retention mechanisms can be followed in natural populations without need to introduce
genetic markers that would disrupt the bees' genetic integrity.

Use of DNA Markers for Regulatory Control of African Bees

African bee identification through morphometrics is rapid, simple and can be used
to test many colonies but with limited certainty. DNA and morphometric analyses
together would make an effective combination for regulatory purposes. Restriction frag-
ment analysis would be used for a smaller number of colonies when a high degree of
reliability and hybrid identification are necessary. With diagnostic DNA alleles at many
loci and codominant expression of restriction fragments, hybrids can be recognized after
many generations and multiple segregational events. This will increase the efficacy of
control measures (Stibick 1984) and will reduce the chances of misidentification leading
to unnecessary regulatory steps, costly to the government and beekeepers. A quaran-
tine operation to eliminate an isolated introduction in California (Cobey and Lawrence
1985) cost over $1 million. Regular requeening of apiary colonies with certified stocks
would likely be a central control tactic. Presently, DNA analysis is the only promising
approach for genetic certification. Simplification of the restriction fragment analysis,
likely to be forthcoming, will make it more practical for routine use.


BROWN, W. M., E. M. PRAGER, A. WANG, AND A. C. WILSON. 1982. Mitochondria
DNA sequences of primates: tempo and mode of evolution. J. Mol. Evol. 18:
CANN, R. L., W. M. BROWN, AND A. C. WILSON. 1984. Polymorphic sites and the
mechanism of evolution in human mitochondria DNA. Genetics 106: 479-499.


Florida Entomologist 71(3)

CARSON, D. A., AND A. B. BOLTEN. 1984. Identification of Africanized and European
honeybees using extracted hydrocarbons. Bull. Ent. Soc. America 30: 32-35.
Isolation and regional localization of DNA segments revealing polymorphic loci
from human chromosome 13. American J. Human Genetics 36: 10-24.
COBEY, S., AND T. LAWRENCE. 1985. Status of the Africanized bee find in California.
American Bee J. 125: 607-711.
DALY, H. V., AND S. S. BALLING. 1978. Identification of Africanized honeybees in
the western hemisphere by discriminant analysis. J. Kansas Ent. Soc. 51: 857-
DALY, H. V., K. HOELMER, P. NORMAN, AND T. ALLEN. 1982. Computer-assisted
measurement and identification of honeybees (Hymenoptera, Apidae). Ann. Ent.
Soc. America 75: 591-594.
FERRIS, S. D., R. D. SAGE, A. C. WILSON. 1982. Evidence from mtDNA sequences
that common laboratory strains of inbred mice are decended from a single female.
Nature 295: 163-165.
CONNEALLY. 1984. DNA markers for nervous system diseases. Nature 225:
HALL, H. G. 1986. DNA differences found between Africanized and European hon-
eybees. Proc. Natl. Acad. Sci. USA 83: 4874-4877.
HALL, H. G. 1988a. Characterization of the African honey-bee genotype by DNA
restriction fragments. Proc. Intern. Congress on Africanized Bees and Bee
Mites. pp. 287-293.
HALL, H. G. 1988b. Genetic characterization of honey bees through DNA analysis.
In Breed, M. D. and D. J. C. Fletcher (eds.), The African Honey Bee. Westview
Press. (In Press).
JEFFREYS, A. J. 1979. DNA sequence variants in the G -, Ay-, and 8-, and p-globin
genes of man, Cell 18: 1-10.
KERR, W. E. 1967. The history of the introduction of African bees to Brazil. S. Afr.
Bee J. 39: 3-5.
MANIATIS, T., E. F. FRITSH, AND J. SAMBROOK. 1982. Molecular cloning. A labora-
tory manual. Cold Spring Harbor Laboratory.
MCDOWELL, R. 1984. The Africanized honey bee in the United States. What will
happen to the U.S. beekeeping industry? USDA Agricultural Economic Report
No. 519.
Final Report: Committee on the African honey bee. NAS-NRC. Washington,
MICHENER, C. D. 1975. The Brazilian bee problem. Ann. Rev. Entomol. 20: 399-416.
PAGE, R. E. JR., E. H. ERICKSON JR. 1985. Identification and certification of Af-
ricanized honeybees. Ann. Ent. Soc. America 78: 149-158.
RINDERER, T. E. 1988. Africanized bees: the Africanization process and potential
range in the United States. Bull. Ent. Soc. America 32: 222-227.
RINDERER, T. E., AND H. A. SYLVESTER. 1981. Identification of Africanized bees.
American Bee J. 121: 512-516.
RUTTNER, F. 1976. The Races of Bees in Africa. Proc. 25th Int. Congress on Apicul-
ture. Apimondia.
SMITH, R. K. 1988. Europeanization of honey bees in South Africa. American Bee J.
128: 329-330.
STIBICK, J. N. L. 1984. Animal and plant health inspection service strategy and the
African honeybee. Bull. Ent. Soc. America 30: 22-26.
SOUTHERN, E. 1980. Gel electrophoresis of restriction fragments. Methods Enzymol.
68: 152-176.
TAYLOR, O. R. 1977. The past and possible future spread of Africanized bees in the
Americas. Bee World 58: 19-30.


September, 1988

Cockburn et al.: New Technologies for Taxonomy

TAYLOR, O. R. 1985. African Bees: Potential impact in the United States. Bull. Ent.
Soc. America 31: 14-24.
TAYLOR, O. R. 1986. Health problems associated with African bees. Ann. Internal
Med. 104: 267-268.
TAYLOR, O. R., AND M. SPIVAK. 1984. Climatic limits of tropical African honeybees
in the Americas. Bee World 65: 38-47.
WOYKE, J. 1969. African honeybees in Brazil. Amerian Bee J. 9: 342-344.

o ,e a 0 a a -a -a


Insects Affecting Man and Animals Research Laboratory
U.S.D.A., Agriculture Research Service
Gainesville, Florida 32601


Isozyme electrophoresis has traditionally been the method of choice for analyzing
the genetic structure of populations. It can also be useful for distinguishing morpholog-
ically identical sibling species. More recently mitochondrial DNA restriction analysis
has become popular as an alternative for studying population structure; and the isolation
of species-specific DNA probes has become the method of choice for distinguishing
closely related species. We have evaluated the usefulness of these DNA based tech-
niques in mosquitoes and discuss their advantages and disadvantages.


La electroforesis de isoenzyma tradicionalmente ha sido el metodo preferido para
analizar la estructura gen6tica de poblaciones. Tambien puede ser utilizada para distin-
guir entire species hermanas morfol6gicamente identicas. Mas recientemente, analysis
de la restricci6n de la mitocondria del DNA se ha hecho mas popular como una alter-
nativa para estudiar la estructura de la poblaci6n; y el aislamiento de sondas de DNA
especificas a especie ha sido el m6todo preferido para distinguir species cercamente
relacionadas. Nosostros hemos evaluado la utilida de estas tecnicas basadas en el DNA
de mosquitos y se discuten sus ventajas y desventajas.

The A. quadrimaculatus species complex consists of at least four sibling species:
species A, species B (Lanzaro 1986), species C and species D (unpublished data). We
have used species A, B, and C of this complex to investigate the usefulness of DNA
techniques in mosquito evolution research. Our first aim was an estimate of the degree
of diversity between different populations (in this case the three sibling species) and
the second was a rapid identification tool for wild-caught mosquitoes. These are quite
different questions and have involved different approaches.
DNA restriction analysis has been widely used in vertebrate population and evolu-
tion studies (Brown 1985), and is becoming more popular in entomology (DeSalle et al.
1986, DeSalle & Giddings 1986, Hale & Beckenbach 1985, Hale & Singh 1986, Harrison
et al. 1987, Hall 1986, LaTorre et al. 1986, Powell 1983, Shah & Langley 1979, Solignac


Florida Entomologist 71(3)

et al. 1986a, 1986b, Solignac & Monnerot 1986). The basic approach is to isolate DNA
from individual insects, cut it at specific sites using a restriction endonuclease, separate
the fragments by size by agarose gel electrophoresis, and then probe to identify specific
fragments of interest. Except for the restriction endonuclease step, this is analogous to
the technique of isozyme electrophoresis. The resulting data is even very similar, con-
sisting of a series of bands in lanes on a gel.
DNA probes have been used to identify species or strains of microorganisms when
no reliable phenotypic key was available. The approach is to identify a DNA sequence
that is present in only one species or strain, clone that sequence in bacteria, and then
use the cloned DNA as a probe against DNA of an unknown specimen. This is similar
to the use of antibodies for serotyping. Techniques such as this are not in general use
for identifying higher eucaryotes, though several groups have used them to identify
individual insects infected with parasites (e.g. Kirkpatrick et al. 1987).


Restriction analysis

The most widely used DNA for restriction analysis is mitochondrial DNA (mDNA).
There are several reasons why this molecule is preferred: it is relatively small (about
16,000 base pairs in mosquitoes), it is present in thousands of copies per cell, and it can
be isolated away from the bulk of the DNA in the nucleus (Brown 1985). Mammals are
large enough that it is possible to isolate enough mDNA from a single individual to
analyze it by staining. That is not possible in mosquitoes because of their small size, so
we were forced to take a different approach to identifying the mDNA.
DNA fractionated on a gel can be transferred to a nitrocellulose membrane and
hybridized to a radioactive probe sequence to detect specific fragments. We investigated
various DNAs to use to detect the mDNA fragments. Heliothus zea mDNA (courtesy
of S. Miller) hybridized to most of the A. quadrimaculatus mDNA fragments, but about
20% of the molecule did not hybridize. In addition, the sensitivity of this probe was
rather low and it would have been difficult to detect the mDNA of a single individual.
We devised a procedure to isolate A. quadrimaculatus mDNA and used this as a
probe. All of the fragments were easily detectable. Two problems were that contaminat-
ing nuclear DNA gave minor background bands and in some cases the patterns were
too complex to easily analyze. As an alternative we used three cloned fragments of
Aedes albipictus mDNA (Dubin et al. 1986, HsuChen et al. 1984) as probes. These
hybridized as intensely as the A. quadrimaculatus probe, gave no background, and lit
up only a subset of the bands (which made interpretation easier).
The sensitivity of the technique was studied by serial dilutions of the DNA from a
single mosquito. The mDNA fragments could be scored in the 1/64 dilution. Therefore
theoretically 64 different enzymes could be scored for each mosquito.
Different methods of preservation of material for DNA preparation were tried. Live
mosquitoes gave the highest yields, but mosquitoes stored for several days in ethanol
also produced DNA, and the banding patterns were unchanged (though less intense).
This method of preservation could be of great utility when collecting field material in
areas where dry ice or liquid nitrogen is unavailable.
Analysis of the three A. quadrimaculatus sibling species with thirteen different
restriction endonucleases showed ten differences among the three mitochondrial
genomes. Species C was most divergent but species A and B also differed. This work
was done on pooled samples, so only the major forms would have been detected. Further
work on individual samples will be necessary to determine if these variants are fixed
in the different species.


September, 1988

Cockburn et al.: New Technologies for Taxonomy 301

The ribosomal RNA genes (rDNA) (Gerbi 1985) could also useful for restriction
analysis. They are also repeated hundreds of times per cell and cloned probes are
available. We probed the same blots used to analyze the mDNA with an A. gambiae
rDNA probe (courtesy of V. Finnerty) and found that the sensitivity was about the
same. The results when the rDNA patterns of the A. quadrimaculatus species complex
were compared were similar to the mDNA results. With the thirteen restriction en-
donucleases tested, six differences were seen among the three species. This is slightly
less variability than in the mDNA, but the A. gambiae probe only hybridized to the
coding sequences, not the non-transcribed spacer region. If this region had been in-
cluded the differences probably would have been about the same. This is similar to the
situation in Drosophila (Powell et al. 1986) but different from mammals (Brown 1985)
wEl re the mitochondrial DNA evolves at a much faster rate than the nuclear genes.
T. probing the filters was little additional effort and generated a great deal of additional
information, so it is highly recommended that both probes be used rather than just one.
Restriction enzyme analysis has to compared to isozyme analysis to determine its

Disadvantages could be:

1. More time consuming.
2. Slightly more expensive.
3. Requires new expertise.

Advantages could be:

1. DNA is invariant in all cells (no artifacts based on comparing different life stages).
2. DNA is much more stable than enzymes.
3. Only one type of gel and DNA isolation are used for all insects, therefore results can
be directly compared between different workers.
4. Restriction endonucleases are commercially available.
5. Polymorphisms usually result in changed numbers of bands, not slight changes in
mobility, making analysis much simpler.

Species specific probes

In applied entomology it is often important to be able to identify field specimens
rapidly and cheaply. For example, one member of a species complex might be a serious
disease vector while another is not. Chromosome analysis, isozyme analysis, and other
techniques that have been useful in research situations are slow and require expert
We have developed a simple means of isolating cloned DNA sequences that are
species-specific and used it to obtain aclone (Arpl) that is specific to A. quadrimaculatus
species A. Arpl can be used as a probe to identify DNA from individual mosquitoes of
the species in question. Since the assay is for presence or absence of the Arpl sequence,
it is not necessary to digest the DNA or run gels. Hundreds of "bug blots" can be
analyzed by a single person in a day.


BROWN, W. M. 1985. The mitochondrial genome of animals, pp. 95-30. in: MacIntyre,
R. J. (ed.), Molecular Evolutionary Genetics. Plenum, New York.

302 Florida Entomologist 71(3) September, 1988

DESALLE, R., AND L. V. GIDDINGS. 1986. Discordance of nuclear and mitochondrial
DNA phylogenies in Hawaiian Drosophila. Proc. Natl. Acad. Sci. USA 83: 6902-
DESALLE, R., L. V. GIDDINGS, AND K. Y. KANESHIRO. 1986. Mitochondrial DNA
variability in natural populations of Hawaiian Drosophila. II. Genetic and
phylogenetic relationships of D. silvestris and D. heteroneura. Heredity 56: 87-
DUBIN, D. T., C. C. HSUCHEN, AND L. E. TILLOTSON. 1986. Mosquito mitochondrial
transfer RNAs for valine, glycine, and glutamate: RNA and gene sequences and
vicinal genome organization. Curr. Genet. 10: 701-707.
GERBI, S. A. 1985. Evolution of ribosomal DNA, pp. 419-517. In MacIntyre, R. J.
(ed.), Molecular Evolutionary Genetics. Plenum, New York.
HALE, L. R., AND A. T. BECKENBACH. 1985. Mitochondrial DNA variation in
Drosophila psuedoobscura and related species in Pacific Northwest populations.
Canada J. Genet. Cytol. 27: 357-364.
HALE, L. R., AND R. S. SINGH. 1986. Extensive variation and heteroplasmy in size
of mitochondrial DNA among geographic populations of Drosophila melanogas-
ter. Proc. Natl. Acad. Sci. USA 83: 8813-8817.
HALL, H. G. 1986. DNA differences found between Africanized and European hon-
eybees. Proc. Natl. Acad. Sci. USA 83: 4874-4877.
HARRISON, R. G., D. M. RAND, AND W. C. WHEELER. 1987. Mitochondrial DNA
variation in field crickets across a narrow hybrid zone. Molec. Biol. Evol. 4:
HSUCHEN, C. C., R. M. KOTIN, AND D. T. DUBIN. 1984. Sequences of the coding
and flanking regions of the large ribosomal subunit RNA of mosquito mitochon-
dria. Nucl. Acids Res. 12: 7771-7785.
Cloning and detection of DNA from a nonculturable plant pathogenic mycop-
lasma-like organism. Science 238: 197-200.
LANZARO, G. C. 1986. Use of enzyme polymorphism and hybridization crosses to
identify sibling species of the mosquito, Anopheles quadrimaculatus Say. Ph.D.
dissertation, University of Florida.
LATORRE, A., A. MOYA, AND F. J. AYALA. 1986. Evolution of mitochondrial DNA
in Drosophila subobscura. Proc. Natl. Acad. Sci. USA 83: 8649-8653.
POWELL, J. R. 1983. Interspecific cytoplasmic gene flow in the absence of nuclear gene
flow: evidence from Drosophila. Proc. Natl. Acad. Sci. USA 80: 492-495.
POWELL, J. R., A. CACCONE, G. D. AMATO, AND C. YOON. 1986. Rates ofnucleotide
substitution in Drosophila mitochondrial DNA and nuclear DNA are similar.
Proc. Natl. Acad. Sci. USA 83: 9090-9093.
SHAH, D. M., AND C. H. LANGLEY. 1979. Inter- and intraspecific variation in restric-
tion maps of Drosophila mitochondrial DNAs. Nature 281: 696-699.
SOLIGNAC, M., M. MONNEROT, AND J. C. MOUNOLOU. 1986a. Concerted evolution
of sequence repeats in Drosophila mitochondrial DNA. J. Mol. Evol. 24: 53-60.
SOLIGNAC, M., M. MONNEROT, AND J. C. MOUNOLOU. 1986b. Mitochondrial DNA
evolution in the melanogaster species subgroup of Drosophila. J. Mol. Evol. 23:
SOLINGNAC, M., AND M. MONNEROT. 1986. Race formation, speciation, and introg-
ression within Drosophila simulans, D. mauritiana, and D. seychellia inferred
from mitochondrial DNA analysis. Evolution 40: 531-539.

Narang & Seawright: New Technologies for Taxonomy 303


Insects Affecting Man and Animals Research Laboratory
U.S.D.A.. Agricultural Research Service
Gainesville, Florida 32604


Laboratory hybridization tests, comparative study of the banding patterns of
polytene chromosomes, and electrophroetic techniques for isozyme analysis are the
methods that are usually employed for the recognition of sibling species of anopheline
and some other mosquitoes. Hybridization tests under laboratory conditions are suffi-
cient for measuring postmating reproductive isolation, but premating mechanisms are
generally undetectable. Difficulty in obtaining good quality preparations tends to limit
the usefulness of the chromosomal approach for the identification of sibling species.
Generally, isozyme analysis is more accurate and reliable in uncovering cryptic vari-
ation, providing meaningful information on genetic relatedness, and in the identification
of sibling species. We have presented a step-by-step analysis of electrophoretic data for
the identification of both allopatric and sympatric sibling species.


Pruebas de hibridaci6n en el laboratorio, studio comparative de los patrons de las
bandas de los cromosomas de polytene, y t&cnicas electrofor6ticas para el analisis de
isoenzyma son los m6todos que usualmente se emplean para reconocer species her-
manas de anopheline y algunos otros mosquitos. Pruebas de hibridaci6n bajo condiciones
de laboratorio son suficientes para medir la reproducci6n en aislamiento despu6s del
acoplamiento, pero mecanismos antes del acoplamiento no son generalmente detecta-
bles. Dificultad en obtener preparaciones de buena calidad tienden a limitar la utilidad
del metodo usando los cromosomas para la identificaci6n de species hermanas. General-
mente el analisis de isoenzyma es mis exacto y confiable en descubrir variaciones crip-
ticas, en proveer informaci6n sobre la afinidad gen6tica, y en la identificaci6n de species
hermanas. Nosotros hemos presentado paso a paso un analisis de datos electrofor6ticos
para la identificaci6n de species hermanas alopatricas y simpatricas.

Sibling species are genetically distinct, but these closely related types are quite
often indistinguishable morphologically. Since speciation is a gradual process, sibling
species are thought to represent an incipient level of evolutionary divergence in com-
parison to morphologically distinguishable species. The lack of clear morphological dif-
ferences creates a serious problem for taxonomists and other scientists, especially con-
trol specialists who are attempting to exploit biological and genetic strategies for insect
control. Before the development of electrophoretic methods that were suitable for
analysis of isozymes (Hubby & Throckmorton 1965, 1968), sibling species were identified
by laboratory hybridization tests and comparative study of karyotypes (especially the
banding patterns of polytene chromosomes in species of Diptera). Both these approaches
have serious limitations. For the most part, only postmating reproductive isolation
mechanisms are detected in hybridization tests under laboratory conditions, because
.he abnormal situation of the laboratory can obscure premating isolation.For some
species of anopheline mosquitoes, the polytene chromosomes are of good quality and

Florida Entomologist 71(3)

can be used for the purpose of correct identification of sibling forms, but there are
problems in obtaining good preparations for some species, even to the extent that this
technique can not be used on a routine basis. Such limitations do not exist in the
electrophoretic analysis of gene-enzyme systems. A comprehensive discussion on the
applications of hybridization, karyotypes, and electrophoretic studies for evolutionary
and taxonomic inferences in the study of sibling species in mosquitoes can be found in
review articles by Bullini & Coluzzi (1982) and Narang & Seawright (1988). Throughout
the text, we have used three terms to designate a given sibling species in order to
highlight the taxonomic tool used for its identification. A sibling species identified on
the basis of diagnostic allozymes will be referred to as "allospecies", in contrast to
"cytospecies" based on chromosomal identification, and "biological species" based on
hybridization tests.
During the last few years, we have used electrophoretic methods for analysis of
gene-enzyme systems to identify sibling species in natural populations of the Anopheles
quadrimaculatus complex (Narang et al. 1988). A dichotomous electrophortic key was
prepared (Narang et al. 1988) and can be used accurately for the identification of field-
collected adults of the A. quadrimaculatus complex. In the early stages of this work,
studies on ovarian nurse cell polytene chromosomes and laboratory hybridization tests
were used to complement and correlate the identification of the members of the complex
(Kaiser et al. 1988). In this paper, we present a step-by-step approach to the analysis
of electrophoretic data for recognition of sibling species, both in sympatry and allopatry.
Electrophoretic data of Narang et al. (1988) will be used to illust- "e these steps.


The following text describes the materials and methods used for identification of
sibling species of A. quadrimaculatus complex. This general approach should be useful
for recognizing sibling species in other species complexes of mosquitoes.
Adult mosquitoes of A. quadrimaculatus were collected from about 30 localities
throughout the distribution range of this species in the eastern half of the U.S. Wild-
caught adults were stored frozen at -70 C until used for electrophoresis. Horizontal
starch gel electrophoresis was performed according to Steiner & Joslyn with a few
modifications (Narang et al. 1988). Gel trays measuring 13 X 20 cm were used. Indi-
vidual mosquitoes were homogenized in 30ul of grinding buffer (10mM Tris, 1mM
EDTA, 1mM 2-mercaptoethanol, pH 7.0) in a multiple sample grinding block. Homoge-
nate of each adult was absorbed with three 3 X 10 mm wicks (Whatman 3 mm paper)
and analyzed on three 10 mm thick starch gels (one wick from each mosquito per gel).
Homogenates of thirty-two adults including equal numbers of mosquitoes from two
localities and three samples of a standard reference strain were run on each gel as
described by Narang et al (1988). After electrophoresis, each gel was cut into six 1.5
mm thick slices. These 18 slices from three gels were stained for different enzymes.
Samples of 30-50 adults from each locality were analyzed for preliminary electrophoretic
test for recognition of genetic substructuring. Electromorphs (enzyme bands) at thirty
four presumptive loci of 16 enzyme systems were scored using bands of the reference
strain as controls as described earlier (Narang et al 1988).
Statistical analysis of electrophoretic variability data was performed by using the
computer program, Biosys 1 (Swofford & Selander 1981). The tests included: (1) chi-
square test for conformance of observed electromorph frequencies to those expected
under Hardy-Weinberg equilibrium; (2) Selander's (1970) coefficient for deficiency or
excess of heterozygotes for each polymorphic locus; (3) genetic identity and genetic
distance (Nei 1978); and (4) the diagnostic value of a locus (Ayala & Powell 1972).

September, 1988

Narang & Seawright: New Technologies for Taxonomy 305


The following text, divided into 5 steps, describes a sequential approach to the
analysis of electrophoretic data for recognition of sibling species in natural populations
of A. quadrimacultus. We have used some data from previously reported studies on A.
quadrimaculatus (Narang et al. 1988) to serve as an example for this paper. Elec-
trophoretic data on populations from four localities, Shell Mound, Levy Co. FL (LEV),
Bear Bay Swamp, Dixie Co. FL (BBS), Noxubee, Noxubee Co. MS (NuX) and
Panasoffkee, Sumter Co. FL (PAN), will be used for illustrations.

Step I. Recognition of genotypes with allelic clusters:

The first level of evidence for the occurrence of presumed sympatric sibling species
(allospecies) can be obtained from visual observation of zymograms of polymorphic loci
as shown in Figures 1-5. For example, in the LEV samples (lanes 5-8 in each figure),
adults can be grouped into two classes. In the first (lanes 5 & 6 in each figure), there
is a cluster of alleles such as: Had-1 (92), Had-3 (45), Got-2 (38), Pep-2 (110), Pgi (95),
Idh-2 (162), Idh-1 (100). The second, (lanes 7 & 8 in each figure) lacks these alleles at
the respective loci. These results give the erroneous indication of strong linkage among
these loci. In addition, without formal statistical analysis, there are obvious deficiencies
of heterozygotes for alleles at these loci. These results suggest that the two classes of
adults are reproductively isolated from each other. Similarly, clustering of Idh-1 (86)
and Idh-2 (162) electromorphs to certain individuals is indicative of sympatric sibling
species in NOX. On the contrary, no allelic clustering is apparent in samples from PAN
and BBS; therefore, presumably at this point it appears that pure populations were
collected at those locations.

Step II. Chi-square test for Hardy-Weinberg equilibrium and Selander's D coefficient:

The chi-square test helps to define statistically the presence of sympatric sibling
species. This test is based on the assumption that a population is in Hardy-Weinberg
equilibrium, i.e., the observed frequencies of alleles can be used to calculate expected
genotypic frequencies that are not significantly different from the observed genotypic
frequencies. A significant chi-square value is indicative of genetic substructuring, i.e.,
sibling species. As shown in Table 1, there is a lack or significant deficiency of heterozy-
gotes for alleles at certain loci in LEV (Had-1, Had-3, Got-2, Pep-2, Pgi, and Idh-2),
and in NOX (Idh-1, Idh-2 and Had-1). Therefore, these results indicate that at both
localities, there are sympatric allospecies. Furthermore, on the basis of the number and
types of loci for which there a deficiency of heterozygotes, one of the species in LEV
is different from the two species in NOX. The chi-square values for the BBS and PAN
populations were not significant. Selander's D coefficient also indicates whether a pop-
ulation is in equilibrium, but this statistic also provides information on the proportion
and direction (either positive or negative) of an unusual frequency of heterozygotes at
each locus. For the data in Table 1, there was general agreement, as expected, between
the values for Selander's D coefficient and chi-square. By Selander's D coefficient, there
was no evidence for the occurrence of sympatric allospecies in the populations from BBS
and PAN.

Step III. Significant differences in allelic frequencies:

Comparison of allelic frequencies at various loci will help recognize presumed allos-
pecies in allopatry. As shown in Table 2, samples from PAN and BBS are either fixed

Florida Entomologist 71(3)








--- -- ----- -- ---- --

I I0

^ - - -- - -



1II '0

41 I .

0 1 0 N
-I -
8- 8 +

September, 1988


- +

- --- --

Narang & Seawright: New Technologies for Taxonomy


Observed Exptected Selander's
Locus Locality heterozygotes heterozygotes' X2 D














*Significant chi-square values (p < 0.05).
'Based on Hardy-Weinberg equilibrium.

for different alleles or show significant differences in allelic frequencies (most common
allele in PAN:BBS shown in parenthesis) at loci: Had-1 (100:92), Had-3 (100:45), Got-2
(100:38), Pep-2 (100:110), Pgi-2 (100:95) and Idh-2 loci (132:162). It is also clear that
alleles at these loci in BBS are similar to alleles in some of the adults of LEV, indicating

Figs. 1-5. Schematic representations of zymograms of adult samples from PAN,
BBS, LEV and NOX. Individual adults were run on one gel and gel slices stained for
various enzymes. The sample order is same in all zymograms. From left to right samples
no. 1-2 (PAN), 3-4 (BBS), 5-8 (LEV), 9-12 (NOX). Notice that samples no. 3, 4, 5 and
6 have different bands for most loci (species C specific cluster of alleles). Sample 9, 10
(species B) differs from others at Had-1, Idh-1 and Idh-2.


Florida Entomologist 71(3)


Locus (NOX) (LEV) (PAN) (NOX) (LEV) (BBS)












































September, 1988

Narang & Seawright: New Technologies for Taxonomy 309

that the species in BBS is most likely the same that occurs in sympatry in LEV. Table
3 shows the diagnostic values (Ayala & Powell 1972) for these loci, which can be used
for taxonomic identification of sibling species.

Step IV. Partitioning of genotypic frequency data of samples containing allospecies in
sympatry into data corresponding to respective presumed allospecies (using information
from step I and II). Hardy-Weinberg equilibrium test on partitioned data set:

The results in Table 1 and 2 (steps I and II) indicated that the NOX population was
composed of two presumed allospecies which differ at the loci for Idh-1, Idh-2, and
Had-1. Adults with the allelic cluster at these three loci, Idh-1 (86), Idh-2 (162) and
Had-1 (95) can be tentatively assigned to one allospecies (species B). Adults of NOX
lacking this cluster can be assigned to a distinct allospecies A (see alleles of A at three
loci in Table 3). At this point the chi-square test for conformance of each data subset
(of allospecies A and B) to Hardy-Weinberg equilibrium should be done. If the par-
titioned genotypic frequency data of each subset is in equilibrium (i.e.. the chi-square
value is not significant), this can be considered adequate to identify and characterize
sibling species A and B. When partitioned, the LEV population was found to be com-
posed of two presumed species, (1) allospecies A similar in genotypes at Idh-1, Idh-2
and Had-1 loci to that of A in NOX and allospecies C characterized by a cluster of alleles
such as: Had-1 (92), Had-3 (45), Got-2 (38), Pep-2 (110) Pgi (95) and a characteristic
two-locus genotype, Idh-2:Idh-1 (162/100).

Step V. Genetic Identity (I) and genetic distance (D).

The coefficients of genetic relationships (Nei 1978). i.e. genetic identity (I) and gene-
tic distance (D, -log I), can be very useful indicators for recognition of allopatric sibling
species. As shown in Table 4 the intra-specific I is 0.95 to 0.97 in species A, 0.98 in
species B and 0.98 in species C. The interspecific I ranged from 0.80 to 0.85 between
A and B, 0.54 to 0.64 between A and C, and 0.54 between B and C. Allopatric popula-
tions of PAN and BBS, which were both in Hardy-Weinberg equilibrium for genotypes
at all loci showed I value of 0.61 (N in each locality was 50) confirming the two popula-
tions were composed of two distinct allopatric allospecies.



Locus A:B A:C B:C

Idh-1 98.82 99.97
Idh-2 95.03* 95.01* -
Had-1 97.66* 99.98 99.97
Had-3 99.92 99.92
Got-2 99.19 99.75
Pep-2 99.07 99.36
Pgi-1 99.99 99.65

'Diagnostic values were calculated by the method of Ayala and Powell (1972). Frequencies of electromorphs of
species A from pooled samples of PAN and NOX. B was from NOX and C from Levy Co. and BBS were used.
2A locus is considered diagnostic, if the probability of correct diagnosis of an individual to a given taxon is 99%
or higher. Loci marked with asterisk (*) can be used only as two-loci diagnostics.

Florida Entomologist 71(3)


Populations 1 2 3 4 5 6

1. PAN species A 0.973 0.948 0.825 0.578 0.539
2. NOX species A 0.027 0.924 0.818 0.541 0.563
3. LEV species A 0.053 0.079 0.799 0.642 0.610
4. NOX species B 0.193 0.201 0.442 0.544 0.587
5. LEV species C 0.548 0.614 0.442 0.610 0.964
6. BBS species C 0.618 0.574 0.494 0.533 0.037 -

An alternative approach to identify sympatric sibling species directly from step I
consists of examining chromosomal banding patterns in adults showing allelic clusters.
Ovaries from individual wild gravid females are removed for study of ovarian nurse cell
polytene chromosome banding pattern and the remaining part of each mosquito is used
for electrophoretic analysis of selected gene.enzyme systems (steps I). This simultane-
ous chromosome-electrophoretic analysis is helpful to correlate the identity of allos-
pecies with that of cytospecies. Our results (Narang et al. 1988, Kaiser et al. 1988)
showed that each of the three allospecies, A, B and C had indeed diagnostic
chromosomal complements. Chromosomes of gravid females of allospecies A from NOX
and LEV were similar to those of the pure population of allospecies A from PAN.
Similarly, females of allospecies C from LEV were similar to those of pure population
of species from BBS. In order to designate these species as true biological species (more
so for species in allopatry), hybrid breakdown among these taxa must be established
by hybridization crosses in laboratory condition. Crosses among A, B and C showed
hybrid sterility to different degrees (Kaiser 1988).


AYALA, J., AND J. R. POWELL. 1972. Allozymes as diagnostic characters of sibling
species of Drosophila. Proc. Nat. Acad. Sci. USA 69: 1094-96.
BULLINI, L., AND M. COLUZZI. 1982. Evolutionary and taxonomic inferences of
electrophoretic studies in mosquitoes. In: "Recent Developments in the Genetics
of Insect Disease Vectors." W. M. M. Steiner, W. J. Tabachnick, K. S. Rai and
S. Narang (eds), pp. 465-82. Stipes Publishing Co.Champaign. II.
HUBBY, J. L., AND L. H. THROKMORTON. 1965. Protein differences in Drosophila.
II. Comparative species genetics and evolutionary problems. Genetics 52: 203-
HUBBY, J. L., AND L. H. THROKMORTON. 1968. Protein differences in Drosophila.
IV. A study of sibling species. Am. Natur. 102: 193-205.
KAISER, P. E. 1988. Cytotaxonomy as a tool for identification of siblings of the
Anopheles quadrimaculatus complex. Florida Entomol. 71: 311-323.
Hybridization of laboratory strains of sibling species A and B of Anopheles quad-
rimaculatus. J. Am. Mosq. Cont. Assoc. 4: 34-38.
NARANG, S. K., P. E. KAISER, AND J. A. SEAWRIGHT. 1988. Dichotomous elec-
trophoretic key for the identification of sibling species A, B and C of the
Anopheles quadrimaculatus (Say) complex (Diptera:Culicidae). J. Med. En-
tomol. (accepted for publication).
NARANG, S. K., AND J. A. SEAWRIGHT. 1988. Genetic differentiation among mem-
bers of species complexes in Anopheline mosquitoes (Diptera:Culicidae). Book

September, 1988

Kaiser: New Technologies for Taxonomy

Chapter, Commemorative Volume "The Eukaryotic Structural and Functional
Aspects." (in press)
NEI, M. 1978. Estimation of average heterozygosity and genetic distance from a small
number of individuals. Genetics, 89: 583-90.
SELANDER, R. K. 1970. Behavior and genetic variation in natural populations. Amer-
ican Zool. 10: 53-66.
STEINER, W. W. M., AND D. J. JOSLYN. 1979. Electrophoretic techniques for the
genetic study of mosquitoes. Mosq. News 39: 35-54.
SWOFFORD, D. L., AND R. B. SELANDER. 1981. Biosys-1: A FORTRAN program
for comprehensive analysis of electrophoretic .-ta in population genetics and
systematics. J. Hered. 72: 281-283.


Insects Affecting Man and Animals Research Laboratory
U.S.D.A. Agricultrual Research Service
Gainesville, Florida 32604


Wild populations of Anopheles quadrimaculatus were collected in the southeastern
United States for cytogenetic studies. Two approaches, i.e., polytene chromosome com-
parisons and hybridization studies were used. Observations on the ovarian nurse cell
polytene chromosomes indicated two distinct X chromosomes, one with a standard band-
ing arrangement and the other with a fixed inversion. Three polymorphic inversions on
the right arm of chromosome 3 were associated with the normal X (cytotype A). Indi-
viduals with the inverted X contained a polymorphic and a fixed inversion on both 3R
and the left arm of chromosome 2 (cytotype B). The two types shared a small inversion
on 2L and a complex arrangement on the left arm of chromosome 3 that included two
distinct homokarotypes. The ovarian polytene chromosomes of the F, hybrids (A x B
and reciprocal) demonstrated complete asynapsis in the X chromosome and 3L, and
extensive asynapsis in the other arms. Differential mortality was expressed in Fi hy-
brids depending on the cross; some crosses produced only females, some produced only
males, and others resulted in normal sex ratios. The reproductive organs of F1 adults
were always abnormal, ranging from atrophied to complete absence. These aberrations
were expressed in backcrosses where very high embryonic and larval mortality were
observed. These studies proved that the chromosomally differentiated types A and B
are distinct sibling species of the Anopheles quadrimaculatus complex.


Se colectaron poblaciones salvajes de Anopheles quadrimaculatus en el sudeste de
los Estados Unidos para studios citogeneticos. Se usaron dos aproches, comparaciones
de cromosomas de politeno, y studios de hibridaci6n. Observaciones de los cromosomas
de politeno de la c6lula nodriza del ovario indic6 dos cromosomas X distintos, uno con
unas bandas patrons, y otro con una inversion fija. Tres inversiones polim6rficas en el
brazo derecho del cromosoma 3 fueron asociadas con la X normal (citotipo A). Individuos
con la X invertida contenian una inversi6n polim6rfica y fija en ambos 3R y en el brazo

312 Florida Entomologist 71(3) September, 1988

izqierdo del cromosoma 2 (citotipo B). Los dos tipos comparten una pequefia inversi6n
en el 2L y un complejo arreglo en el brazo izquierdo del cromosoma 3, que incluye dos
distintos homokarotipos. Los cromosomas de politeno del ovario de hibridos F1 (A X B
y el reciproco) demostraron una asinapsis complete en el cromosoma X y 3L, y una
extensive asinapsis en en los otros brazos. Mortalidad diferencial se express en hibridos
F1 dependiente del cruce; algunos cruces producieron solo hembras, algunos producieron
solo machos, y los otros resultaron en una proporci6n normal de los sexos. Los 6rganos
reproductivos de adults Fi siempre fueron anormales, variando de atrofiados a com-
pletamente ausentes. Estas observaciones se expresaron en el cruce de la progene del
F1 de vuelta a los padres, donde se observ6 una alta mortalidad embri6nica y larval.
Estos studios probaron que los tipos A y B de cromosomas diferenciados son species
hermanas pero distintas del complejo de Anopheles quadrimaculatus.

The use of cytology in conjunction with classical taxonomy is not a new idea, as it
has been 40 years since Frizzi used salivary gland polytene chromosomes to distinguish
members of the European Maculipennis sibling species complex (Frizzi 1947). Twenty
years later Coluzzi & Sabatini (1967, 1968, 1969) published a series of papers that
differentiated the members of the Anopheles gambiae complex according to differences
in their salivary gland polytene chromosomes. Polytene chromosome maps have been
prepared for most of the medically important anophelines, and these maps have been
used to uncover sibling species complexes in Anopheles '- hensi, Anopheles
culicifacies, Anopheles balabacensis and Anopheles farauti.
Kitzmiller et al. (1967) reviewed the cytological relate :nshi' )f the important North
American anophelines and determined that many similarities of ethnological significance
existed. Although the adults of the different species were morphologically distinguish-
able, many homologies in the DNA banding patterns were revealed. They used these
homologies and the results of hybrid crosses to determine that Anopheles freeborni
Aitken, Anopheles aztecus Hoffman, Anopheles occidentalis Dyar and Anopheles earlei
Vargas were more related within the nearctic Maculipennis complex. Anopheles
punctipennis Say and Anopheles quadrimaculatus Say were also related to this com-
plex, but to a lesser degree. Kaiser et al. (1988) indicated that inversion sequences in
A. earlei, A. punctipennis and A. quadrimaculatus demonstrated relatedness between
these three eastern U.S.A. species. This, perhaps, should not be surprising because
they occupy overlapping ranges. Therefore, it is clear that cytotaxonomy can be utilized
to distinguish between morphologically similar species (siblings), and to demonstrate
similarities among related species. This paper shows how cytology was used in a re-
search program on A. quadrimaculatus, the goal of which was to determine if a sibling
species complex existed in this species.

BACKGROUND ON A. quadrimaculatus

Anopheles quadrimaculatus was of medical importance because it was the primary
vector of malaria in the eastern United States when it was endemic in this country.
Today it is a serious pest in rice-growing areas and in the Tennessee Valley Authority
(TVA) reservoir system, where pest management programs are monetarily and environ-
mentally expensive. Permanent bodies of water with large amounts of aquatic vegeta-
tion provide excellent breeding habitats for the immatures, so the establishment and
proliferation of exotic aquatic weeds in the eastern United States has enhanced the
breeding potential of this mosquito. The broad distribution of A. quadrimaculatus ex-
tends from southern Florida and Texas to southern Canada; the northwestern range
extends through southern Minnesota. The fact that the distribution of this mosquito
covers such a large geographic area precludes the notion that environmental

Kaiser: New Technologies for Taxonomy 313

homogeneity could exist throughout its breedirig range. Therefore, we postulated that
evolutionary processes may have occurred that resulted in geographic and, eventually,
reproductive isolation (this is not meant to discount the possibility of sympatric specia-
tion). Morphological changes may not have occurred or may have been minor because
the speciation process was relatively new. Or, perhaps there was no advantage in
morphological change. At any rate, the result of this speciation process would be a
sibling species complex in A. quadrimaculatus.
To test this hypothesis we collected adults from throughout the eastern United
States and returned them to our laboratory for study. Standard taxonomic keys from
Carpenter & LaCasse (1955) and Darsie & Ward (1981) were used to verify that the
adults were A. quadrimaculatus; females were blooded and incubated at 27C for 24
hr and then transferred to a refrigerator (50C) pending dissection. The ovaries were
removed from these females and the ovarian nurse cell polytene chromosomes were
prepared by the method of Kaiser et al. (1988a). Chromosomes were observed under
phase-contrast and compared to the standard ovarian polytene map for A. quad-
rimaculatus prepared by Kaiser & Seawright (1987).
Males and females from some of the populations were mated to adults representing
other populations; the wild adults would not mate freely in the laboratory so a forced
copulation technique similar to that of Baker (1964) was employed. We used these
matings to observe F1 hybrid mortality, polytene chromosomes and sterility.


Cytological techniques for differentiating sibling species can be divided into two
main categories.

Poytene Chromosome Comparisons

In anophelines, polytene chromosomes usually occur in the salivary glands of 4th
instar larvae and in adult ovaries. The quality of the chromosome preparations varies
from tissue to tissue and from species to species, so the various options should be
considered. Polytene chromosome maps have been prepared for most of the important
North American anophelines (Kitzmiller et al. 1967), but if not available they must be
prepared. Although there was a salivary gland chromosome map available for A. quad-
rimaculatus (Klassen et al. 1965), the polytene chromosome preparations from the
salivary glands were usually of very poor quality. Therefore, it was necessary to draw
a map from the ovarian polytene chromosomes (Kaiser & Seawright 1987), the quality
and consistently of which were good, even in wild females. The map was prepared from
permanent slides, photographs of standard enlargement, and camera lucida drawings.
When suitable polytene maps are available, field populations can be analyzed by
observing the chromosomes of random samples of larvae and/or adults for a) polymor-
phic and b) fixed inversions. Polymorphic (floating) inversions are important in regard
to frequency and kind. Polymorphic inversions may be shared by sibling species, in
which case variations in frequency may indicate separate mating populations (Bedo 979,
Bryan et al. 1982). Or, more typically, they may be unique for a particular species.
Inversion breakpoints are usually easy to identify and inversion loops are distinct fea-
tures in polytene chromosome preparations that can be used to distinguish species with
a high level of competency. Inversion freqencies of conspecific polymorphic inversions
may vary, geographically and temporally, between populations, but this does not change
their status as a diagnostic tool (Stalker 1964, Dobzhansky & Powell 1975, Coluzzi et
al. 1979). In North American Anophelines, the right arm of chromosome 3 contains the
most inversion polymorphisms; the X chromosome and the left arm of chromosome 2
may also contain inversion polymorphisms (Kitzmiller et al. 1967).

314 Florida Entomologist 71(3) September, 1988

Fixed inversions, which are inversion homozygotes that are fixed in a mating popu-
lation, i.e., no inversion polymorphism occurs, are often used to separate sibling species
(Green & Hunt 1980, Rothfels 1981). Diagnosis may initially be difficult, but if camera
lucida drawings of sections of the chromosome arms are compared to the respective
sections in photographs, they can usually be detected. Then, prominent groups of bands
within the inversion can be used as a diagnostic feature, by noting the order of the
bands, to easily discern the presence of the homozygote. Fixed inversions may occur
on the X chromosome or on either of the autosomes.

Hybridization Studies

Adults from different populations can be mated to each other and their progeny
observed for a) aberrant polytene chromosomes, b) sterility and c) mortality. Polytene
chromosomes of hybrids from interspecific crosses commonly have regions that are
asynaptic, i.e., unpaired zones of a diploid chromosome (Kitzmiller et al. 1967). These
asynaptic regions, which may have homosequential banding patterns, presumably have
undergone changes at the molecular level that affect their normal affinity for each other.
A general rule of thumb is that the degree of asynapsis is directly proportional to the
degree of relatedness of the two parents. Another important reason for observing hy-
brid polytene chromosomes is that fixed inversions, which may be difficult to distinguish
in normal chromosome complements, produce the standard inversion loop in their effort
to pair with their homologous partner. An example of this can be seen in Figure lb,
where a very small region of a chromosome has been inverted and inserted at a new
location. Although obvious here, this rearrangement would have been difficult to detect
in a normal complement.
Hybrid sterility is the second method of assaying the compatibility of parents, and
is usually a more accurate measurement of relatedness than asynapsis. Hybrid sterility,
which in effect is reproductive isolation, is expressed in three ways. First, the testes
and ovaries of hybrids may undergo structural changes; these changes can vary from a
slight reduction in size, to atrophy, to total absence (Davidson 1964, Kreutzer & Kitzmil-
ler 1971). The size reduction usually results in a decrease in fecundity for the females.
The amount of sperm in testes is reduced and many of the sperm and spermatozoa are
deformed. The last two expressions of hybrid sterility are pre- and post-zygotic mortal-
ity. When hybrid females and males are outcrossed to either of the parental types,
fertility is usually reduced and mortality may be as high as 100%. The survival rate of
the progeny that do hatch is less than normal (Curtis 1982, Kaiser et al. 1988).
Mortality of F1 hybrids may occur in one or both sexes, and the mortality may be
unidirectional or in both directions (Kreutzer & Kitzmiller 1972, Miles 1981). Unidirec-
tional mortality occurs when species P x species Q produces only females (or males) but
sp. Q x sp. P results in a normal sex ratio. The mortality usually occurs in the larval
or pupal stage.


The results of observing the ovarian polytene chromosomes of adults from several
different populations and comparing them to the standard map (Kaiser & Seawright
1987) are described below.

X Chromosome

A small fixed inversion, In(X)a, occurred in a portion of the adults from several of
the populations. The normal arrangement was also present, however, we failed to detect

Kaiser: New Technologies for Taxonomy

$ 1

In(2. *.
-n A .,,T ^ ril l,,; i* t

ris r e .
R-, -,ae Ow
..iCt .
~ ~ ~ o v l = ,,. ,.,-l .


E i
'. :C.,
*- ll II
-- 39

^y S~io" "'t-f *';
a* \r"..,,I ...Ol [ rr I
-.) ,..

t ,* : ' *'*4


8!*---4 SB


rIF;1 r. *i -
t%. **
$t t.

Fig. 1. The ovarian nurse cell polytene chromosomes of hybrids of an interspecific
cross between species A and species B of the Anopheles quadrimaculatus complex. A)
Left arm of chromosome 2. B) Left arm of chromosome 2 that shows the fixed inversion,
In(2L)c. C) Right arm of chromosome 3 that shows the fixed inversion, In(3R)e. D) Left
arm of chromosome 3; this homokaryotype (3L1) is one of two that have been described.
E) The homokaryotype 3L2 (unpublished data).

any individuals that were heterozygous for this inversion. This was an indication that
two separate mating populations, reproductively isolated, existed sympatrically at some
of the collection sites. We provisionally called the new sibling species B. The two X
chromosomes can be seen in Figure 2; the fact that the puff was included in the inversion
made detection fairly easy.


316 Florida Entomologist 71(3) September, 1988


P .X, Xp

B K- .-,

Fig. 2. The X chromosomes of Anopheles quadrimaculatus. A) Species A. B) Species
B. C) Asynaptic X chromosomes of species A x species B hybrid (unpublished data).

Chromosome 2

Three inversions were observed on the left arm of chromosome 2. By using the X
chromosome for species determination (A or B), we were able to associate the autosomal
inversions with their respective species. A small inversion, In(2L)a, was located near
the centromere and was seen in both species A and B. The two remaining inversions
were found only in species B. The polymorphic inversion, In(2L)b, and the fixed inver-
sion, In(2L)c, were located in the middle of the arm where they shared a common
breakpoint; In(2L)c was also inserted into a new position on the distal side of In(2L)b
(Fig. Ib). No inversions were found on the right arm of chromosome 2.

Chromosome 3

The left arm of chromosome 3 is very unusual in that two karyotypes, 3L1 and 3L2,
exist in both species A and B. The heterokaryotype was observed much more often than
either of the homokaryotypes (the ratios did not fit Hardy-Weinberg in any of the
populations); most of the heterokaryotype is asynaptic which gives it a distinctive ap-
pearance. Two small fixed inversions have taken place on 3L2, and these are the only
regions that pair in the heterokaryotype.
The right arm is typically the most polymorphic arm in North American anophelines.
It is also the longest of the arms which makes identification easy. We have found six
inversions in this arm, 4 in species A and 2 in species B (Figure 3). The polymorphic
species A inversions are: In(3R)a, which is a small inversion near the distal end of the
chromosome.; In(3R)b is a somewhat larger inversion occurring in the middle of the
arm. This inversion sometimes includes a smaller inversion, i.e., included inversion.;
In(3R)c is a small inversion that is adjacent to the centromere.; In(3R)f is the small
inversion included in In(3R)b that occurs infrequently by itself. The two species B
inversions are In(3R)d, which is polymorphic, and In(3R)e, which is fixed. They occur
in the middle regions of 3R (Fig. 3d,e).

Kaiser: New Technologies for Taxonomy

a \ ), I'
S, .V'

S . i .t
'. \ "/?' i'r ; ; / '- 1 |

.. ^ 14.y-
C '.- ""

Fi .; 3. T r a" of s' e .hee u) A

inversion (e) (Kaisr et 3).
-t I'{ -I.-

I i\ .... ,"I|
. *

Fig. 3. The right arm of chromosome 3 of Anopheles quadrimaculatus. A) A stand-
ard 3R of species A that shows the inversion breakpoints for species A (above) and
species B (below). B) 3R of species A with 3 homozygous inversions. C) 3R of species
A with 3 inversion heterozygotes. D) 3R of species B with homozygous inversion (d)
and fixed inversion (e). E) 3R of species B with heterozygous inversion (d) and fixed
inversion (e) (Kaiser et al. 1988a).

The frequencies of the conspecific floating inversions varied between populations,
i.e., geographic variation, and they were not polymorphic in all populations (Table 1)
(Kaiser et al. 1988a). Species A inversions In(3R)a, b and c were never observed in
adults collected from central and south Florida. However, populations in north Florida
contained a few individuals heterozygous for these inversions. Populations of species A
that were north and/or west of Florida were very polymorphic, and in some cases the
homozygous inversion frequency was greater than that of the standard banding arrange-
ment. The frequencies of species B floating inversions also varied between populations.
However, since the species B populations were all above latitude 30N no conclusions
regarding north vs. south geographic variation in inversion polymorphism could be


Adult females from field populations were returned to the laboratory and used to
establish isofemale lines. Cytology was used to determine species A and B lines, and
forced copulation was used to make hybrid crosses (sp. A x sp. B and reciprocal).
Cytology on F1 hybrids demonstrated that some portions of the polytene chromosomes
were asynaptic (Fig. 1), and that others contained inversion loops indicative of fixed
inversions (Fig. lb,c). Complete asynapsis was noted in the X chromosome (Fig. 2c)

Florida Entomologist 71(3)


Population +/+ +/a a/a N Freq. + Freq. a X2b

Goldens Bridge,
Chickamauga Lake,
Guntersville Lake,
Stuttgart, AR
College Station,
Lake Charles, LA
Chatahoochee State
Park, AL
Lake Seminole,
Lake Octahatchee,
Kanapaha Bot.
Garden, FL
Lake Panasoffkeed,
Myakka State Park

0 12 88 100 0.060 0.940 0.407

0 12 88 100 0.060 0.940 0.407

0 10 91 101 0.050 0.950 0.274

1 10 89 100 0.060 0.940
10 46 44 100 0.330 0.670


25 54 21 100 0.520 0.480 0.668
19 48 21 88 0.489 0.511 0.736

28 47 30 105 0.490 0.510 1.145

56 12 6 74 0.838 0.162 12.032

120 4 0 124 0.984 0.016 0.033

0 0 100 1.000 0.000

0 0 115 1.000 0.000

aKaiser et al. 1988a
bSignificant at the 0.05 probability level

and in the left arm of chromosome 3 (Fig. ld,e). Also, the breakpoints for the fixed
inversions, In(2L)c and In(3R)e, were confirmed.
One of the important effects of hybridization is sterility, so to study this we dissected
the ovaries and testes from Fl hybrids and observed them for abnormalities. Some
hybrids had no reproductive organs, although most had some form of atrophied organs
(Figure 4). When the atrophied testes were dissected, the contents revealed small
amounts of deformed sperm (Figure 5). Some of the Fl hybrids were backcrossed and
pre- and post-zygotic mortality were observed in all of the crosses (Table 2).
One of the more interesting aspects of the hybridization studies was the differential
mortality expressed in the Fl hybrids (Kaiser et al. 1988b). When species A females
were mated to sp. B1 males, only F1 females were produced because all the males died
in the pupal stage (Table 3). When sp. B2 males were used in the cross, females and
males were produced in about a 3:2 ratio. The reciprocal cross (B 9 x Ad) resulted in
a normal sex ratio.


We used cytology as a taxonomic tool in determining that a sibling species complex
existed in A. quadrimaculatus (Table 4). This proved necessary because classical


September, 1988

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