NUTRITIONAL BIOCHEMISTRY, BIOENERGETICS, AND NUTRITIVE VALUE OF THE DRY-WOOD TERMITE, MARGINITERMES HUBBARDI (BANKS)

Item Type text; Dissertation-Reproduction (electronic)

Authors La Fage, Jeffery Paul, 1945-

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 05/10/2021 08:17:33

Link to Item http://hdl.handle.net/10150/289496 INFORMATION TO USERS

This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of die original submitted.

The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction.

1. The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity.

2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame.

3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete.

4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced.

5. PLEASE NOTE: Some pages' may have indistinct print. Filmed as received.

University Microfilms International 300 North Zeeb Road Ann Arbor, Michigan 48106 USA St. John's Road, Tyler's Green High Wycombe, Bucks, England HP10 8HR 77-6061

LA FAGE, Jeffery Paul, 1945- NUTRITIONAL BIOCHEMISTRY, BIOENERGETICS, AND NUTRITIVE VALUE OF THE DRY-WOOD TERMITE, MARGINITERMES HUBBARDI (BANKS).

The University of Arizona, Ph.D., 1976 Entomology

Xerox University Microfilms, Ann Arbor, Michigan 48106 NUTRITIONAL BIOCHEMISTRY, BIOENERGETICS, AND NUTRITIVE VALUE

OF THE DRY-WOOD TERMITE, MARGINITERMES HUBBARDI (BANKS)

by

Jeffery Paul La Fage

A Dissertation Submitted to the Faculty of the

DEPARTMENT OF ENTOMOLOGY

In Partial Fulfillment of the Requirements For the Degree of

DOCTOR OF PHILOSOPHY

In the Graduate College

THE UNIVERSITY OF ARIZONA

19 7 6 THE UNIVERSITY OF ARIZONA

GRADUATE COLLEGE

I hereby recommend that this dissertation prepared under my direction by Jeffery Paul La Fage entitled NUTRITIONAL BIOCHEMISTRY, BIOENERGETICS, AND NUTRITIVE VALUE OF THE DRY-WOOD TERMITE, MARGINITERMES HUBBARDI (BANKS) be accepted as fulfilling the dissertation requirement of the degree of DOCTOR OF PHILOSOPHY

&//0 /76 Diss Date

After inspection of the final copy of the dissertation, the follov?ing members of the Final Examination Committee concur in its approval and recommend its acceptance:'"

e)?/7i ^ hi11- >.i Mmu.J/tv,-- ' ( 8/ j/7i

i hln>

This approval and acceptance is contingent on the candidate's adequate performai>ce and defense of this dissertation at the final oral examination. The inclusion of this sheet bound into the library copy of the dissertation is evidence of satisfactory performance at the final examination. STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rule's of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author. ACKNOWLEDGMENTS

My sincere gratitude is extended to Dr. William L,

Nutting for his continued advice and patience throughout the preparation of this dissertation. I am also indebted to Drs.

Larry Crowder, Gordon Waller, Bobby Reid, and William

McCaughey for reviewing this manuscript and for serving on my graduate committee. Appreciation is especially extended to Dr. James Berry for his constant encouragement and technical advice without which much of this study would not have been possible, For their helpful suggestions con­ cerning statistical analyses, I am indebted to Dr. Robert

Kuehl, Dr. David Marx, and Mr, Paul Johnson,

For their assistance in gathering field and labora­ tory data I am grateful to my wife, Wendy, Dr. W. L,

Nutting, Dr. Michael I, Haverty, Mr. Brad Esser, Mr, J. Gary

Eckhardt, Dr. Charles Weber, Mr. Angelo Longo, Dr. Guy

Josens, Dr. Stanley Alcorn, Dr. John A. Rupley, Dr, P-H

Yang, and Dr, Mike McClure. Gratitude is also expressed for secretarial assistance kindly provided by Mrs, Hazel

Tinsley, Mrs. Martha Montgomery, and Mrs. Tina Hutson.

Finally, I am indebted to my wife, Wendy, for her everlasting patience and understanding during the prepara­ tion of this manuscript and throughout my graduate studies, BIOGRAPHICAL SKETCH

Jeffery Paul La Fage was born on June 16, 1945, in

Waterbury, Connecticut. He attended elementary and junior high school in Watertown, Connecticut, and was graduated from the Mount Hermon School, Mount Hermon, Massachusetts.

He received his undergraduate training at the

University of Connecticut, Storrs, Connecticut, and was graduated in January, 1968, with the Bachelor of Arts degree.

During the years 1971 to 1974 he was the recipient of a National Defense Education Act Title IV traineeship,

The subject of his master's thesis was an analysis of en­ vironmental parameters correlated with the foraging behavior of a desert subterranean termite, Gnathamitermes perplexus

(Banks).

He received the Master of Science degree from The

University of Arizona in 1974. During 1975 and 1976 he was supported by the International Biological Program, Analysis of Ecosystems, Desert Biome, for study and research toward the Doctor of Philosophy degree in Entomology with a minor concentration in Nutritional Biochemistry.

He is married to the former Wendy Ellen Lawrence of

Middlebury, Connecticut.

iv TABLE OF CONTENTS

Page

LIST OF TABLES vii

LIST OF ILLUSTRATIONS xi

ABSTRACT xii

INTRODUCTION 1

MATERIALS AND METHODS 6

Biochemical Studies on Dead Saguaro Wood 7 Analytical Determinations 14 Data Anslysis 26 Nutritional Physiology 26 Feeding Trials 27 Respiration Studies 31 Response of Protozoans to Temperature .... 41 Energy Budgets for Laboratory Feeding Groups 42 Trophic Level Interactions 42 Lower Level Interactions 43 Predation 44 Field Estimates on Energy Flow Through an M. hubbardi Population 49

RESULTS 51

Biochemical Studies of Dead Saguaro Wood 51 Analytical Determinations 51 Nutritional Physiology 64 Feeding Trials 64 Respiration Studies 76 Energy Budgets for Laboratory Feeding Programs 8 5 Trophic Level Interactions 97 Lower Level Interactions 97 Predation 98 Field Estimates on Energy Flow Through an M. hubbardi Population 102

v vi

TABLE OF CONTENTS—Continued

Page

DISCUSSION 109

Biochemical Studies on Dead Wood 109 Analytical Determinations 112 Nutritional Physiology 133 Feeding Trials 133 Respiration Studies 142 Energy Budgets for Laboratory Feeding Groups 153 Trophic Level Interactions 154 Lower Level Interactions 154 Predation 154 Field Estimates on Energy Flow Through a Population of M. hubbardi 164

APPENDIX A. MODIFIED METHYL-RED INDICATOR USED FOR KJELDAHL TITRATIONS 167

APPENDIX B. FATTY ACIDS OF SAGUARO WOOD 168

APPENDIX C. SUMMARY OF RESULTS FROM THE INTER­ MEDIATE SIZED FEEDING TRIAL 170

REFERENCES ..... 218 LIST OF TABLES

Table Page

1. The amino acid composition of 2 termite species 47

2. Constituents of diet fed to 7 pairs of weanling mice for 21 days to assess the quality of Reticulitermes flavipes protein . . 48

3. Percentage of saguaro wood (dry wt) removed by successive 6 h extractions in ethanol (.95%), ethanol:benzene (1:2, v/v), and hot water 52

4. Percentage of saguaro wood (dry wt) removed by two 3-min extractions with chloroform: methanol solution (2:1, v/v) 53

5. Fatty acids in total chloroform:methanol (2:1, v/v) extracts of saguaro skeletons of 3 age categories 54

6. Per cent carbon content of saguaro samples representing 3 age categories 56

7. Per cent ash in saguaro samples representing 3 age categories 56

8. Calorific content (kcal/g dry wt) of saguaro samples from the 30 skeletons representing 3 age categories 57

9. Calorific content (kcal/g dry wt) of wood samples from superficial and inner parts of a dead saguaro rib 57

10. Per cent nitrogen content of saguaro samples from skeletons representing 3 age categories 59

11. The amino acid composition of dead saguaro wood determined by the method of Scurfield and Nicholls (1970) 60

vii viii

LIST OF TABLES—Continued

Table Page

*12. Lignin content of extractive-free saguaro wood determined by the acetyl bromide method (Johnson et al. , 1961) 62

13. Delignification of dead saguaro wood using the method of Ritter and Barbour (1935) ... 63

14. Final group weight (fresh), total food intake, and caste composition of 16 Marginitermes hubbardi incipient colonies . . 65

15. Dead termites which remained in feeding groups 67

16. Initial mean weight (mg fresh) of individual termites selected from 3 natural colonies of Marginitermes hubbardi 67

17. Mean total offspring (eggs and larvae) pro­ duced by intermediate feeding groups during 90 d 68

18. Wood consumption and fecal pellet production by intermediate sized feeding groups 69

19. Food consumption and fecal pellet production divided by biomass days for the intermediate sized feeding groups 7 0

20. Apparent assimilation of food by the intermediate sized feeding groups 71

21. Final caste composition of intermediate sized feeding groups of Marginitermes hubbardi 73

22. Percentage of original biomass (fresh) gained or lost by intermediate sized groups during the 90 d feeding trial 75

23. Equations for predicting wood consumption (mg), fecal-pellet production (mg), and apparent assimilation efficiency by groups of 5-50 Marginitermes hubbardi 77 ix

LIST OF TABLES—Continued

Table Page

24. Caste composition of 6 large feeding groups (490-500 individuals) of Marginitermes hubbardi maintained on saguaro wood for 90 d at ca. 100% RH 78

25. Wood consumption, fecal-pellet production, biomass, and carton production by 6 large groups (4H0-500 individuals) of Marginitermes hubbardi fed saguaro wood for 90 d 79

26. Mean oxygen consumption by groups of Marginitermes hubbardi 80

27. Prediction models which estimate oxygen consumption (Y) (yl 02/mg termite [fresh]/ h) for the 4 physiological states tested by manometry 82

28. Comparison of heat production by Marginitermes hubbardi determined by calorimetry and respirometry 84

29. Mean biomass (by caste) of Marginitermes hubbardi colonies collected on or adjacent to the Silverbell site during 1974-75 .... 88

30. Moisture content (by caste) of Marginitermes hubbardi 89

31. Calorific content (kcal/g dry wt) (by caste) of Marginitermes hubbardi, determined by micro- and macro-bomb calorimetry 90

32. Mean energy budget for 16 incipient colonies of Marginitermes hubbardi 91

33. Calculation of respiration (kcal) by intermediate sized Marginitermes hubbardi groups 93

34. Calculation of production (P) for inter­ mediate sized groups of Marginitermes hubbardi 94 X

LIST OF TABLES—Continued

Table Page

35. Mean energy budgets for intermediate sized groups of Marginitermes hubbardi 95

36. Energy budgets for the large laboratory feeding groups of Marginitermes hubbardi ... 96

37. Nitrogen and lignin content of Marginitermes hubbardi fecal pellets produced during intermediate feeding trials 98

38. Ash content of Marginitermes hubbardi 100

3 9. Carbon content of Marginitermes hubbardi determined by the dry-combustion method of Allison et al. (1965) 100

40. Nitrogen content of Marginitermes hubbardi castes 101

41. Termite lipids extracted by ethyl ether in a Goldfisch extractor 103

42. Fatty-acid composition of Reticulitermes flavipes and Marginitermes hubbardi 104

43. Biological assay of Reticulitermes flavipes included as the protein source in a diet fed to mice for 3 weeks 105

44. Chemical elements and organic components in sound wood 113

45. Nitrogen content of stem wood of various gymnosperms and angiosperms 120

46. Nitrogen content of portions of the stem wood of various tree species (after Cowling and Merrill, 1966) ..... 124

47. Summary of literature reports of oxygen consumption by termites (yl/mg termite Jfresh]/h) 144

48. Nitrogen content of termites ..... 157 LIST OF ILLUSTRATIONS

Figure Page

1. Early group A saguaro skeleton 9

2. Late group A saguaro skeleton 10

3. Group B saguaro skeleton 11

4. Group C saguaro skeleton 12

5. Saguaro skeleton in very late stages of decay 13

6. The basal portion of the dead saguaro ICarnegiea gigantea (Engelm.) Britt. & Rose] 15

7. Phase contrast photomicrograph of fluid from the hindgut of the dry-wood termite, Marginitermes hubbardi (Banks) 33

8. Parts of a thermogram produced by a single, young Marginitermes hubbardi (Banks) nymph during a 12-h recording of heat production 86

9. Map which shows the location in 1976 of all live and dead saguaros ICarnegiea gigantea (Engelm.) Britt. & Rose] on plot F2 of the Saguaro National Monument (East) near Tucson, Arizona 107

xi ABSTRACT

The role of a dry-wood termite, Marginitermes hubbardi (Banks), in a Sonoran Desert ecosystem was eval­ uated using biochemical, physiological, and ecological tech­ niques. Samples from dead, but termite-free, and infested saguaro skeletons [Carnegiea qigantea (Engelm.) Britt.&

Rose] were analyzed chemically to assess the importance of wood constituents in nest-site selection and suitability,

As there was little or no difference in the carbon, nitro­ gen, extractive, fatty-acid, ash, lignin, and calorific content of these samples, it was concluded that some other factor must determine nest-site adequacy. Skeletons which had been dead at least 4 years and which retained a partial covering of dry loose cortex for protection of alates during entrance-hole excavation, appeared to be the most attractive to M, hubbardi for colony initiation.

Groups of M. hubbardi fed saguaro wood and main­ tained at 24, 26, 28, 30, and 32°C consumed an average of

56.78 mg/g termite (fresh)/d. Wood consumption which was more closely correlated with termite-group biomass than either temperature or colony origin was predicted by the equation: Y = -94.98 - 0.0623 B + 0,30 T2 + 0.00368 BT; where Y = mg of wood consumed, B = mean daily termite

xii X1X1 biomass multiplied by the duration (in days) of the experi­ ment, and T = temperature (°C),

Oxygen consumption by larvae and nymphs was highly correlated (R 2 =0.93) with temperature from 16 -*-->• 36°C and was predicted by the equation: Log = *-7,736 + 0,424 T -

0.006 T 2; where Y = yl 0^ consumed/mg termite (fresh)/d and

T = temperature (°C). Mean C>2 consumption by actively feeding termites was 0,453 yl/mg/h with an average RQ of 0,97,

Heat production by M. hubbardi measured directly at

20°C in a modified drop-heat capacity calorimeter averaged

0,01895 cal/mg termite (fresh)/12h. Comparative values obtained indirectly by respirometry were considerably lower, averaging 0,0116 cal/mg termite (fresh)/12h, The differ­ ences were attributed to the failure of respirometry to detect heat produced by anaerobic protozoans in the termite hindgut.

Energy budgets calculated for groups of 5, 10, 20, and 50 individuals maintained at 22 -*--»• 32°C revealed that termites dispersed an average of 0,07 cal/mg termite

(fresh)/d.

M. hubbardi is an effective detritivore in the

Sonoran Desert where it releases nitrogen, carbon, and other nutrients which otherwise would remain "locked up" in dead saguaros and other woody detritus. It also facilitates xiv microbial decomposition through, the communition of large decay-resistant woody debris.

In order to evaluate their protein quality, termites were included as the sole source of nitrogen in a diet fed to weanling mice for 3 weeks. Although they did not sustain growth as well as a whole-egg reference protein, termites were deficient only in sulfur amino acids.

The density of M. hubbardi colonies was measured on a 4.55-ha plot on the Saguaro National Monument (East) near

Tucson, Arizona. The estimated termites have a potential to disperse ca. 680 kcal/ha/y, This is much less than the value calculated for a common subterranean species,

Gnathamitermes perplexus (Banks) in an Arizona grassland ecotone (1,5 x 10 £ kcal/ha/y). INTRODUCTION

By 1969, Emerson recognized approximately 2200 species of termites (Wilson, 1971) which comprise the order

Isoptera, found primarily throughout the tropical regions of the world. The United States fauna includes 4 0 species, 17 of which occur in Arizona.

All known termites are eusocial. As such, they exhibit cooperative brood care, a caste system, and over­ lapping generations, Although they arose from widely separated evolutionary bases, the termites and social

Hymenoptera share a number of similar behavioral mechanisms,

Indeed, the social organization of termites rivals that of the most advanced Hymenoptera. Many additional comparisons have been discussed by Wilson (1971). Unlike the Hymen­ optera, however, no known subsocial species are known among the Isoptera (Krishna, 1969),

All termites consume cellulosic materials as their primary energy source. The sole nourishment for many species, notably among the Kalotermitidae, is sound, decay- free dead wood (Hendee, 1933; Pence, 1956), which may con­ tain as little as 2,5-3.0% moisture. Other species that nest in the soil forage for dead wood, grass, plant litter, and humus (Lee and Wood, 1971). Still other species re­ strict their diets to dead roots, animal dung, or, as among

1 2 the Macrotermitinae, to fungi cultured on recycled plant material within their nests.

According to Kirk (1971), cellulose is the most abundant of the continuously cycled organic materials. The two groups of organisms capable of digesting cellulose use different digestive strategies. Among the first group are some protozoans, mollusks, and arthropods, that produce their own cellulolytic enzymes. Members of the second group, which includes ruminants, lack this ability, but possess a rich gut fauna and/or -flora which partially de­ grade cellulose to components that the host can assimilate.

There is mounting evidence that termites belong to both groups (La Fage and Nutting, in press). Among the cellulo­

lytic animals, few have been studied more intensively than the termites and their symbiotes, yet we still know very

little about their digestive physiology. The lower termites

(Mastotermitidae, Kalotermitidae, Hodotermitidae, Rhino- termitidae) have symbiotic protozoans in the hindgut; the higher termites (Termitidae) do not, but contain symbiotic bacteria instead. Honigberg (1970), La Fage and Nutting

(in press), Lee and Wood (1971), Moore (1969), and Noirot and Noirot-Timoth6e (1969) have written general reviews of cellulose digestion by termites. Lasker (1959) presented a valuable discussion of the problem of defining the relative importance of host and microorganism in cellulose digestion.

The ability to utilize cellulose efficiently accounts 3 largely for the widespread success of termites. Emerson

(1955) suggested that, because potential food is so abundant, other factors such as the availability of suitable nesting sites may be more restricting on the growth and dispersal of termite colonies.

Members of the family Kalotermitidae are known as the dry-wood termites because, with few exceptions, their nests are permanently excavated within the wood on which the colony feeds. Mature dry-wood colonies rarely exceed a few thousand individuals (Nutting, 1969). In contrast, several rhinotermitid and termitid species routinely have colonies exceeding a million individuals (Wilson, 1971). Although evolved from the phylogenetically primitive mastotermitid base (Wilson, 1971), present-day kalotermitids display far less behavioral complexity than colonies of the only living mastotermitid, Mastotermes darwiniensis Froggatt,

Owing to the fact that kalotermitids can be main­ tained easily and studied in the laboratory, more is known of their physiology than of the more advanced species, especially the Termitidae. The ecological significance of kalotermitid species remains poorly understood, however. On the other hand, recent investigations including those by

Haverty and Nutting (1975a, 1975b, 1975c); Kaverty, La Fage, and Nutting (1974); Haverty, Nutting, and La Fage (1975);

Johnson and Whitford (1975); La Fage, Nutting, and Haverty

(1973); and Bodine and Ueckert (1975) have done much to clarify the role of subterranean species in desert eco­ systems. These studies suggest that subterranean termites may at times consume a large percentage of the annual dead wood and dead grass production. In western Texas, Bodine and Ueckert (1975) have shown that termites compete sig­ nificantly with large herbivores for grasses. Similar effects on grasslands were noted by Coaton (1951) and Watson and Gay (1970).

Marginitermes hubbardi Banks is the most destructive dry-wood termite in the Tucson area. According to Light

(1934), it ranges throughout central and southern Arizona, south along the coast of western Mexico, at least to the city of Colima. Although the extent of its eastern range in

Mexico is unknown, it is generally replaced at higher eleva­ tions by Incisitermes marginipennis (Latreille). In Cali­ fornia M. hubbardi is restricted to a small area along the southeastern border. It is common all along the eastern coast of Baja California. In Arizona it is replaced entirely by Incisitermes minor (Hagen) above 1000 m, while at intermediate elevations, including Tucson (635 m), the 2 species are sympatric. At Phoenix (27 5 m) and other low elevations 1^. minor is replaced by M. hubbardi, which is considered to be better adapted to high temperature and low moisture,

M. hubbardi consumes a variety of dead woods, in­ cluding cottonwood (Populus Fremontii Wats,), palo-verde [Cercidium microphyllum (Torr,) and C. floridum Benth.], ash

(Fraxinus velutina Torr.), sycamore (Piatanus Wrightii

Wats.), and the ribs of the saguaro cactus [(Carnegiea giqantea (Engelm.) Britt.& Rose],

Its habit of attacking dwellings has earned it the common name "house termite" (Snyder, 1954). Light (1934) considered saguaro and cottonwood to be its most important food sources in Arizona.

The overall objective of this study was to measure the ecological significance of M. hubbardi in a saguaro forest in the Sonoran Desert. Three types of investigations were performed to answer the basic question. These included studies of saguaro wood composition, nutritional physiology,- and trophic interactions, Finally, information on saguaro distribution and termite biomass available from other sources was used to estimate annual energy flow through a population of M. hubbardi on a 10-acre plot (4,55 ha) on the Saguaro National Monument, east of Tucson, Arizona. MATERIALS AND METHODS

Field investigations were conducted at 2 locations near Tucson, Arizona. The first is located on or adjacent to the United States International Biological Program/Desert

Biorae Site on the Silverbell bajada approximately 55 km northwest of Tucson at an elevation of 560 m, M. hubbardi colonies collected in this area provided experimental animals for subsequent laboratory studies. The second site is a 10-acre plot in section 17 of Saguaro National Monument

(East) (SNME) near Tucson at 735 m elevation.

The vegetation of both areas is described by

Niering and Whittaker (1963) as desert scrub and is char­ acteristic of the Lower Sonoran Desert. Dominant trees include palo-verde [Cercidium microphyllum (Torr„) and C, floridum Benth.] and mesquite [Prosopis juliflora var. velutina (Woot.) Sarg.]. The giant cactus or saguaro

[Carnegiea qiqantea (Englem,) Britt, & Rose], though present in both areas, is especially dense on the SNME. Opuntia fulgida Engelm. and 0. spinosior (Engelm. & Bigel) Toumey are common on both sites, as is creosote bush (Larrea divaricata Cav.),

Laboratory studies were carried out at several loca­ tions on The University of Arizona campus including the departments of Entomology, Chemistry, Nutrition and Food

6 7

Science, Plant Pathology, and Renewable Natural Resources,

Additional experiments were done in part at the University

Poultry Research Laboratory, Casa Grande Highway.

Biochemical Studies on Dead Saguaro Wood

The saguaro, largest of the United States cacti, is endemic to the Sonoran Desert (Benson, 1969). A healthy, mature plant may attain formidable dimensions, including a height of 13 m, weight of more than 5000 kg, and life span of 200 years (Berry, Anmin, and Steelink, 1960).

After the death of a saguaro, recognized by the total absence of green tissue, the initial stages of decay proceed rapidly. The epidermis and unusually thick, pulpy cortex slough off to expose an inner framework of ligni- ferous ribs, the secondary xylem. I have called this structure the skeleton, for it provides the sole support of the living plant. The woody skeleton decays much more slowly than the other tissues, requiring 30 years or more to disappear from the desert floor. During certain stages of decay, saguaro skeletons are especially attractive as . nesting sites for M, hubbardi. Field observations suggest that neither very young nor very old skeletons are potential hosts for termite colonies, although the reason for this is not clear, One possible explanation might be that the decay process alters the chemical composition such that the wood is attractive only for a limited period, 8

To test chis hypothesis 3 categories of dead saguaro

skeletons were selected: A, B, and C. Group A included

plants which were considered too young to support M,

hubbardi (Figures 1 and 2), Typically these skeletons

remain upright with a firm footing in the soil although the

plant may have been downed by wind or lightning. The basal

third or more is generally covered with intact but very dry

cortex. Also in this region, an abundance of fibrous

material (dried pith) adheres to the ribs. Group B

skeletons (Figure 3) retain less, or at times, no intact

cortex, less fibrous material and, if standing, a loose and

easily disturbed footing in the soil. Quite often they have

become dislodged and lie on the soil surface or propped up

against some other vegetation. There remains in practically

every instance a soil connection with part of the root

system. Those skeletons that rest on the surface often have

been attacked by subterranean termites at points of soil contact. Group B skeletons invariably contain one or more

M. hubbardi colonies. Group C plants (Figures 4 and 5) are older and more decayed than the others. They are always

prone, rarely retain soil connections with the root system,

and at times have lost their roots and basal sections com­

pletely. In advanced decay (Figure 5) little structural integrity remains? ribs are fragmented, and the assaults of

weather and fungi are extensive. Skeletons in this category Figure 1. Early group A saguaro skeleton — The saguaro to the left ICarnegiea gigantea (Engelm.) Britt. & RoseJ died in March, 1976. At this time it was considered too young to contain Marginitermes hubbardi (Banks) colonies and thus was assigned to category A. 10

Figure 2. Late group A saguaro skeleton >— This saguaro iCarnegiea gigantea (Engelm.) Britt. & Rose] died in 1974 and decayed rapidly. Although very little cortex remained in 197 6 the root system retained a solid footing. No Marginitermes hubbardi CBanks) activity was noted. Although it was assigned to category A, it was probably a very late "A.M 11

Figure 3. Group B saguaro skeleton -- The saguaro [Carnegiea gigantea (Engelm.) Britt. & Rose] pictured died in 1973 and by 1976 had the appearance typical of category B. Although no destructive sampling was attempted, this skeleton probably contained active Marginitermes hubbardi (Banks) colonies. 12

Figure 4. Group C saguaro skeleton — Much of the basal portion of the saguaro iCarnegiea gigantea (Engelm.) Britt. & Rose] pictured had decayed during the 24 years since 1952 when the plant died. It was classified as a "C" skeleton because, although it contained Marginitermes hubbardi (Banks) during some period in the past, there was no evidence of active colonies when it was checked in 1976. 13

**•«> a*

Figure 5. Saguaro skeleton in very late stages of decay — Shown are the remains of a saguaro ICarnegiea gigantea (Engelm.) Britt. & Rose] which died in 1945. The decay process was so extensive that only ca. 88" of the original 12-18' plant were found in July, 1976. As there was evidence of past Marginitermes hubbardi (Banks) activity, the skeleton was assigned to category C. 14 never contain active colonies of M, hubbardi, but show evidence that they did at an earlier time.

During late summer, 1975, 30 saguaro skeletons, 10 from each group, were collected on or adjacent to the

Silverbell site. The basal portion (Figure 6) (ca, 0,5 m) of each specimen was removed, labeled, bagged, and returned to the laboratory.

Analytical Determinations

All chemical determinations described in this section were performed on wood removed from the 30 skeletons.

Prior to analysis, 600-g wood samples were collected from each skeleton and cleaned of any dried pith, cortex, or soil, A hatchet was used to split the ribs into slivers small enough to pass through a large Wiley® mill (Arthur H,

Thomas Co., model No. 4) fitted with a 3-mm sieve. The re­ sulting wood dust was further comminuted in a smaller Wiley® mill (intermediate model) fitted with a 60-mesh delivery unit and 4-oz screw-cap glass jar. The resulting wood flour was dried at 105°C for 24 h.

Extractives. Sixty 20-gm wood flour samples, two from each of the 30 skeletons, were subjected to successive solvent extractions in Soxhlet extractors, Individual samples were transferred to dry, single-thickness (123 x

43 mm) pre-weighed thimbles (Whatman). A pre-weighed glass wool plug was placed on top of the wood flour to prevent 15

fJJ j

Figure 6. The basal portion of the dead saguaro Igarnegiea gigantea (Engelm.) Britt. & Rose] — This and other similar skeletons contained active Marginitermes hubbardi (Banks) colonies which provided termites for laboratory experi­ ments. sample loss during extraction and the total preparation was weighed. Samples prepared in this manner were extracted continuously for 3 successive 6-h periods with ethanol

(.95%), ethanol:benzene (1:2 v/v) , and hot water, respec­ tively. After the thimble had drained and cooled, the entire sample was dried for 24 h at 105°C, cooled in a desiccator, and weighed. Weight loss was considered' to represent total extractives and was expressed as a per cent of unextracted dry, wood flour. Filtrates resulting from these extractions were stored in 1-quart glass bottles.

Since it was expected that hot-water filtrates might contain soluble sugars, several samples were tested with Molisch's reagent [a 5% solution of a-napthol in ethyl alcohol (Oser,

1965)]. In this test a reddish-violet zone appearing be­ tween the sample and an equal volume of concentrated sul­ furic acid indicates the presence of carbohydrates in the sample (Oser, 1965).

Wood Lipids, The total lipid content of saguaro wood was measured using a modification of the method described by Folch, Lees, and Stanley (1957), Thirty 4-g wood samples were homogenized individually in 4 0 ml of chloroform:methanol (2:1 v/v) for 3 min in an omni-mixer

(Sorvall Co., model 17150) on speed 4. The mixture was filtered and the solid residue, resuspended in 40 ml of solvent, was homogenized for an additional 3 min. The 17 mixture was filtered again, washed with 40 ml chloroform: methanol, and the filtrates collected in a 500-ml separatory funnel. Thirty ml of 0.8 8% potassium chloride in water was added. The mixture was shaken vigorously and allowed to settle for 24 h before proceeding. During this time, lipids were partitioned into the lower chloroform layer which was subsequently collected in a tared Nalgene® beaker and dried in a vacuum oven for 48 h at 40°C. The resulting lipids were desiccated, weighed, resuspended in a volume of chloro­ form, and stored at 4°C for further analysis.

Preparation of Methyl Esters for Fatty-Acid Analysis.

Methyl esters of the fatty acids contained in the above extracts were prepared using a modification of the I^SO^-

CH^OH procedure described by Rogozinski (1964) as follows:

A ca. 50-mg lipid sample (2-3 drops) was placed in a 15-ml centrifuge tube and ca. 2 ml of benzene added. To this was added 4,5 ml of 0.5% sulfuric acid in absolute ethanol (by volume) and a boiling chip. The preparation was capped, heated for 2.5 h at 90°C on a heating block, and allowed to cool. Two ml of redistilled petroleum ether were added, mixed thoroughly, allowed to settle, and the methyl esters recovered in the ether layer. The fatty-acid methyl esters were stored in redistilled hexane at 4°C.

Purification of Methyl Esters. Fatty-acid methyl esters were separated from contaminates on 1 mm-thick, thin-layer plates (20 x 20 cm) coated with silica gel G-PF

254 (Brinkman Instruments Inc.). Plates were activated at

105°C for 1 h prior to streaking, A 1-dimensional solvent system consisting of redistilled hexane, ethyl ether, and acetic acid (160:40:2 v/v/v) was used to isolate the methyl esters. The plates were sprayed with Rhodamine B (Supelco

Inc.) and read under ultra-violet light. The methyl-ester band was marked, collected in redistilled hexane, and stored at 4°C.

Identification of Fatty-Acid Methyl Esters. Fatty- acid methyl esters were identified and measured using a gas chromatograph (Tracor, model MT220) equipped with a flame- ionization detector. A 6-ft glass column (1/4" ID) was packed with 15% EGSS-Y (Applied Science Laboratories, Inc.), a copolymer of ethylene glycol succinate containing methyl- silicone, on 100/120 mesh Chromosorb AW-DMCS (Supelco, Inc.) as suggested by Christie (197 3). The carrier gas, nitrogen, was delivered with a head pressure of 40 psi and a flow rate of 70 ml per min. The column was held at 17 0°C (isothermal) and the detector temperature was 250°C. Straight-chain, even-numbered, methyl esters of fatty acids from Cg to c^8-3

(Supelco, Inc.) were used as internal standards. Methyl esters prepared from the saguaro flour were injected into the gas chromatograph in hexane with a 1-yl syringe

(Hamilton Co., model 7101N). Peak areas were measured by 19

triangulation (Bartlet and Iverson, 1966) and results

expressed as percentages of total fatty acids in the sample.

Carbon Analysis. Two 50-mg samples from each of the

30 skeletons were analyzed to determine their carbon content with a high-frequency induction furnace (Leco, Inc.), according to the method outlined by Allison, Bollen, and

Moodie (1965). Samples were weighed into tared ceramic crucibles and a catalyst containing iron, copper, and tin was added. The crucible was inserted into the furnace and

burned in a stream of oxygen for 3 min. Carbon dioxide released during combustion was trapped in an Ascarite® bulb which was weighed after each combustion. The weight gained by the bulb represented CC>2 (proportionate to carbon) released from the sample. Blanks consisting of a crucible and catalyst were combusted periodically to determine CC>2 absorbed by the Ascarite® bulb from the atmosphere. After every 7 determinations, a glycine sample was analyzed to determine recovery efficiency.

Ash Analysis. Three 2-g samples from each of the

30 skeletons were analyzed for inorganic mineral content

(ash), Samples weighed into tared ceramic crucibles were burned in a muffle furnace (Thermo Electric Co., model

F-1740) at 550°C for 18 h. The oven was turned off and allowed to cool before the crucibles were transferred to a 20 desiccating jar containing indicating Drierite®, The ash samples were weighed after an additional 24 h.

Calorific Content, Two 1-g samples from 21 of the

30 saguaros were ignited in an oxygen-bomb calorimeter (Parr

Instrument Co., model 1221) to measure heat of combustion, i.e., total calorific content, After loading the sample/ the bomb was charged with at 30 psi and immersed in a water jacket. The system was closed and allowed to equili­ brate before firing. Corrections for heat generated from the fuse wire and formation of E^SO^ and HNO^ in the bomb were applied to the observed heat rise of the water sur­ rounding the bomb.

An additional series of samples was analyzed to determine if the calorific content of weathered saguaro differed from that in the center of a rib. Results were expressed as kcal/g dry weight basis,

Nitrogen Content. Wood contains so little nitrogen that conventional Kjeldahl procedures are inadequate, tending to produce inaccurate and inconsistent results, A modified micro-Kjeldahl procedure described by Rennie (1965) is especially well-suited for analysis of woody tissue and therefore was used in this study.

Two 200-300 mg wood-flour samples from each of the

30 skeletons were analyzed as follows: The samples were weighed into 100-ml Kjeldahl flasks to which were added ca. 21

40 mg of mercuric-oxide catalyst, 4 g of potassium sulfate, and 5 ml of concentrated sulfuric acid. To minimize foaming,

2-3 drops of Dow-Corning antifoam solution were added to each flask. Two blanks were prepared for every 10 unknowns.

The flasks were swirled gently and allowed to stand over­ night before the Kjeldahl digestion was effected on electric heating racks in a fume hood, The thermostats were set initially as low as possible to minimize frothing and loss of samples from the flask necks. After clearing (ca. 2 hr) the settings were gradually raised to a final position of

No, 6. Total digestion required 6 h. After cooling, but while still warm, ca, 10 ml of water was added to each flask. The addition of water before the preparations reach room temperature is critical because, when completely cooled, the sample becomes solid and extremely difficult to remove from the reaction flask.

After further cooling, the digests were removed and brought to volume in 25-ml volumetric flasks. At this point the samples were stored in a refrigerator at 4°C until they could be distilled. The distillation of ammonia from a 10-ml aliquot of the digest was carried out in a micro-Kjeldahl distilling unit (Arthur H. Thomas Co,, Kirk model) as follows: A 50-ml Erlenmeyer flask containing 10 ml of a 2% boric acid solution and 2-3 drops of modified methyl-red indicator (Appendix A) was attached to the delivery tube of the distilling apparatus, Ten ml of base 22 solution [containing 60% (w/v) sodium hydroxide (pellet form) and 5% (w/v) sodium thiosulfate] were added slowly to the apparatus which held the sample. Heating commenced and continued for 7 min after the first drop of condensate appeared on the cold finger. The boric acid solution, now green, was titrated to its blue-purple end point with standardized hydrochloric acid dispensed from a 10—nil buret.

Per cent nitrogen in the sample was calculated by the fol­ lowing formulae:

Nitrogen =

[titration (ml) - blank(ml) ] x standard factor (mc^/ml) sample weight(mg)

Standard Factor =

normality of HCL x 0.014 g nitrogen x 1000 mg 1 meq wt x 1 g nitrogen

Amino Acid Profile. Contrary to most other opera­ tions, only 1 amino acid determination was carried out on the saguaro wood. Amino acid determinations are extremely difficult to perform on woody tissues (La Fage and Nutting, in press). The procedure selected to analyze the sample

(B7) was proposed by Scurfield and Nicholls (1970), Seven g of extractive-free saguaro flour were combined with 700 ml

6N HCL in a 1-liter round-bottom flask. After sitting over­ night, the sample was refluxed for 16 h at 115°C and 15 psi.

The hydrolyzate was dried in a rotavapor-flask evaporator 23

(Brinkman Instrument Co., model R/A) at room temperature,

The resulting residue was resuspended in water, filtered, and evaporated once more before bringing to volume with citrate buffer, pH 2.2. Amino acids were measured with an automatic amino acid analyzer (Beckman Instrument Co., model 121). Resulting peaks were identified by comparison with the Beckman amino acid calibration mixture, type 1, and quantities of individual amino acids calculated by a computer program. Results were expressed as percentages of total amino acids present in the hydrolyzate,

Lignin Analysis. Classical lignin analysis (Ritter and Barbour, 1935) suffers from 2 major drawbacks; large samples (2 g) and long reaction times (ca. 15 h) are re­ quired. A rapid (<1 h) spectrophotometry method for the analysis of samples containing 3-6 mg of lignin was proposed by Johnson, Moore, and Zank (1961). A total of 50, 10-20 mg extractive-free samples from the 3 saguaro groups was examined using this method. The saguaro flour was added to

15 x 150 mm test tubes containing 10 ml of a 25% acetyl bromide solution (in acetic acid). A glass marble was placed over the tube opening and the preparation heated for

30 min in a waterbath at 70°C. The tubes and their contents were swirled gently at 10-min intervals to insure that the reagents were mixed well and to promote dissolution of the sample. After cooling in water at 15°C, the reaction mixture was transferred quantitatively to a 100-ml vol­

umetric flask containing 9 ml of 2M sodium hydroxide and

ca. 50 ml of acetic acid. One ml of 7.5M hydroxylamine

hydrochloride was added, the contents mixed thoroughly, cooled, and brought to volume with acetic acid. The ab­

sorbency of this solution was read at 28 0 my on a grating spectrophotometer equipped with a recorder (Beckman 'Instru ment Co., model DB-GT) and per cent lignin calculated by comparison with the absorbency of a standard prepared from

pure lignin. The following equations were used:

A st. - A,b ast C

„ , . • 100V (A - A. ) % lignin __ s b_ in unknown wa st. where

a £ = absorptivity of lignin standard

A £ = absorbency of lignin standard

A^ = absorbency of blank

C = concentration of lignin in g per 1

V = volume (1) of solution used

Ag = absorbency of sample

w = weight of sample in g

To check the accuracy of this method, 3 replicates from skeleton A8 were analyzed by the highly reliable I^SO method (Ritter and Barbour, 1935). 25

Delignification. The quantitative determination of

holo-cellulose in wood is possible only after lignin has

been removed from the wood (Browning, 1969). The chlorite

delignification technique described by Ritter and Barbour

(1935) was employed to reduce the lignin content of saguaro

flour to 2-3% of the original sample weight. Skeleton A8

was selected for a preliminary investigation. Three repli­

cates of extractive-free flour from this sample had an

average lignin content of 25,5% determined by the H^SO^

method and 28.4% by the acetyl bromide procedure. One g of

extractive-free flour was transferred to a 125-ml Erlenmeyer

flask and 32 ml H^0, 0.1 ml acetic acid, and 0.3 g sodium

chlorite added. Additions of sodium chlorite and acetic

acid in the amounts specified above were made at regular

intervals for the duration of the experiment. The reaction

was carried out in a waterbath at 72°C during periods of

2.25 to 4 h. The reaction flask was fitted with an externally-driven stirrer. After the reaction was com­

pleted, the flask was cooled in ice water and the reaction mixture was filtered and washed successively with cold water, acetone, and ethyl ether. The preparation was dried

at 30°C under vacuum for 12 h and weighed. The filtrate was tested with Molisch's reagent to determine if the washing had removed any carbohydrates. No attempt was made to

analyze cellulosc components because delignification had not been entirely successful. 26 Data Analysis

Wood-composition data were examined where appro­

priate with a one-way analysis of variance. Treatment means

were tested for significance at the a = 0.05 level,

Nutritional Physiology

Natural ecosystems are extremely complex associa­

tions of living organisms with their non-living habitat

(Evans, 1956). Very few investigators have estimated energy

flow at the community level. Notable exceptions include the

studies by Lindeman (1942), Odum and Odum (1955), Odum

(1957), and Teal (1957). Darnell (1968) noted that an

understanding of energy flow through a community depends to

a large extent on an understanding of the energy dynamics

of the individuals comprising that community. Indeed,

certain physiological parameters, e.g., total heat produc­

tion, cannot be measured in the field. Others, e.g., assimilation, are much easier to assess in the laboratory,

Engelmann (1961, p. 221) suggests that, "There are two aspects to any study of energetics: the field survey and the laboratory experiment," The same author (Engelmann,

1966, p. 80) declared, "Paradoxically, a good portion of the data necessary for field estimates must come from laboratory studies," The situation has not changed dramatically during the last 10 years, Although radio­ tracers appear to offer a promising approach to identifying 27 trophic exchanges in nature (Paris and Sikora, 1967), the techniques developed so far are not well suited to dry-wood termites living entirely within wood, It was decided, therefore, that population level estimates on energy flow in M. hubbardi would necessarily be projected from labora­ tory data. The studies described below have provided such data.

Feeding Trials

A complete energetics analysis must include esti­ mates on rates of ingestion, egestion, assimilation, growth, death, numbers, biomass, and the calories represented by these figures. Laboratory feeding studies under controlled conditions can provide data on many of these parameters.

Brian (1971) suggests that accurate estimates of feeding, assimilation, and respiration rates on social insects require that the experimental unit be representative in age and caste structure of the parent society. Wilson (1971) discusses another facet of social insect biology which may affect physiological measurements, As defined by Wilson, the "group effect is an alteration in behavior of physiology within a species brought about by signals that are directed in neither space nor time" (p. 297). The density of indi­ viduals in a social insect colony is known to have an effect on physiological processes (Wilson, 1971), It was decided, 28 therefore, to measure food intake for different sized groups of M. hubbardi.

Incipient Colonies. M. hubbardi alates were collected at lights during dispersal flights during July,

1972. Within 2 days they were sexed, paired (1

1-3 pre-weighed and slighly moistened saguaro discs as food.

The incipient colonies were held at 32°C and ca. 100% R.H. until September 3, 1975, when the 16 surviving colonies were dismantled, the termites counted, and remaining wood weighed. In addition, the caste of each surviving termite was recorded.

Intermediate Groups, Numerous saguaro skeletons collected from an area adjacent to the Silverbell site during June and July, 1975, contained large M. hubbardi colonies (>100 0 individuals) which provided animals for the following experiment. Five groups of 5, 10, 20, and 50 termites (larvae, nymphs, and 1 soldier) were selected from each of 3 large colonies. Since termites from each colony were represented in every size category experimental unit, variation among the 3 original termite sources could be ascertained. Termites were allowed to feed on saguaro discs (dried at 105°C for 24 h) for 90 d in 8-oz wide-mouth glass jars. The following data were recorded for each of the 60 experimental units: initial and final termite bio- mass; number and caste of survivors after 90 d; ingestion, egestion, and the weights of exuviae, carton (a mixture of feces and undigested wood used for construction); dead termites? and unused material. These data and several additional variables calculated from them were examined by an analysis of variance utilizing a randomized complete- block design with a factorial selection of treatments. The survival percentages were transformed to probits before analysis. Treatment means were tested for significance at the a = 0.05 level and separated using Student-Newman-

Keul's test (Steel and Torrie, 1960) or the least signifi­ cant difference. Where applicable, multiple linear regres­ sion equations were developed by the forward stepwise method described by Kim and Kohout (197 5).

Daily inspection of the rearing units revealed that mortality occurred predominantly during the early days of the experiment. A biomass-time estimate was therefore required to calculate food consumption and other variables on a basis of response per termite wt per unit time, To be realistic, this estimate should represent average standing- crop biomass and the total time of the study. Petrusewicz

(1967) and Petrusewicz and Macfadyen (1970) have discussed such a term, "biomass-days" (BT) which they define by the following expression; 30

BT = I B. i=l 1 or

BT = 5 • T where

B^ = daily standing-crop biomass

T = time of the study

B = average standing-crop biomass.

5 is further defined by the expression:

- 1 K B K .S, Bi 1=1 where K = the number of records (days in the study).

Three assumptions were made regarding the behavior of the intermediate sized feeding groups in order to cal­ culate BT.

1. Mortality occurred only during the first 10 days of

the experiment and was linear,

2. Although the group weight changed during this 10-

day period due to removal of individuals through

mortality, surviving individuals retained their

initial weights until the beginning of day 11.

3. Individual weight change over the last 80 days was

linear.

The first assumption, though not entirely valid, is largely substantiated by direct observations on the experimental units. Assumption 2", on the other hand, is based on speculation that groups did not settle into a uniform feeding regime for some time after establishment.

It is possible that a small weight loss occurred during this period. The third assumption is sound when, as in this experiment, weight change is less than two-fold (Gordon,

1968). Based on these assumptions, a computer program re­ quiring input of only initial and final termite biomass and the number of survivors in each group was used to calculate

BT. A number of variables including ingestion, egestion, and egg production were subsequently expressed on the basis of response/biomass-days.

Large Groups. Seven M. hubbardi groups containing

490-500 individuals and prepared in the same manner as the intermediate groups were maintained for 90 d at 20, 22,

24, 26, 28, 30, and 32°C, Termites at each temperature were from the same colony and were representative of original caste composition of the colony. The 20°C group suffered excessive mortality from a bacterial infection (probably

Serratia marcescens Bizio) and was not included in the data analysis.

Respiration Studies

Fundamental to the understanding of animal ener­ getics is the knowledge that animals dissipate heat as a result of their incomplete utilization of food (McNeill, and Lawton, 1970). Both direct and indirect methods have been developed to measure heat loss. Direct methods always require total enclosure of the animal and involve measure­ ment of temperature rise in some type of surrounding medium.

Indirect methods are more widely used owing to their simplicity and reliability. If the protein'.carbohydrate:fat ratio of the food being metabolized is known, heat produc­ tion can be calculated by measuring the liters of C>2 con­ sumed or C02 produced (Brody, 1945). An animal consuming a

100% carbohydrate diet emits 5.047 kcal for each liter of

C>2 respired. Because M. hubbardi normally consumes a diet which is close to 100% carbohydrate, it would appear that indirect calorimetry would be a valid method to determine heat production. A constant-pressure differential respi- rometer (Gilson Medical Electronics, Inc., model GRP20) was used to measure O^ consumption directly and CC>2 production indirectly for groups of 5 or 10 M. hubbardi. Each group included 1 soldier in addition to 4 or 9 nymphs and/or large larvae. Four physiological states were examined: normally faunated/starving, defaunated/starving, normally faunated/feeding, and defaunated/feeding.

Normally faunated M. hubbardi contain a protozoan species; Tricercomitus divergens Kirby, Staurojoenina sp.,

Oxymonas hubbardi Zeliff, and Ketadevescovina debilis Kirby

(Figure 7). In addition, a variety of bacterial forms are present in awesome numbers. Cleveland (1925a) demonstrated 33

Lai 210Xf B = Metadevescoyina debills Kirby

Figure 7. Phase contrast photomicrograph of fluid from the hindgut of the dry-wood termite, Marginitermes hubbardi (Banks). (b) 570X. A = Staurojoenina sp.; S = Spirochetes

Figure 7.--Continued 35

that termites held for 1.5 h at 45 psi lose their proto­

zoans but they themselves suffer no adverse physiological

consequences. When M. hubbardi is so treated, the entire

protozoan faunule as well as a substantial proportion of the

bacteria are killed. Defaunated termites may be completely

refaunated and reflorated within 24 h after being combined

with normally faunated individuals. Refaunated termites

behave normally and appear to suffer no ill-effects from the

defaunation-refaunation process.

Each group size-temperature treatment combination

was tested for 8-12 h at 16, 20, 24, 26, 28, 30, and 32°C.

The respirometer had 18 functioning manometers, which pro­

vided space for 14 experimental and 4 control flasks. All

measurements were made on termites held in 15-ml reaction

flasks containing an internal reservoir and a single side

bulb. Flasks containing groups on which consumption was

to be measured contained a few ml of 10% KOH solution (for

absorbing CC^) and a small piece of fluted filter paper in the central reservoir. Those used in conjunction with the

indirect method of measuring CO 2 production contained no

KOH (Umbreit, Burris, and Stauffer, 1972). Several small saguaro chips were added to the flasks containing feeding termites, while starving groups were provided with small

pieces of rubber tubing for footing.

Barometric pressure and room temperature were re­

corded at the beginning of each 8-12 h experiment. Gns J

36 uptake measurements were recorded hourly and the readings corrected to STP using the following formula:

273(P-P )AV True gas change = —2. where

P = barometric pressure in mm Hg corrected for tempera­

ture and latitude,

P^ = vapor pressure of water at temperature, T,

T = temperature in K° at the level of the manometer, and

Vg = vi 1 of gas change measured.

A theoretical argument for selecting this equation is given by Gregory and Winter (1965). A total of 672 replicates were tested. A computer program was written to reduce the resulting hourly readings to STP and check linearity of gas-consumption rates over time. Final results were expressed as average ul of 0^ consumed/mg termite

(fresh wt)/h. Respiratory Quotients (RQ's) were calulated according to the method described by Umbreit et al. (1972).

Gas exchange data were examined by an analysis of variance utilizing a completely randomized design with a factorial selection of treatments, Means were tested for significance at the a = 0.05 level and separated using Student-Newman-

Keul's test. Regression equations were developed by the method of orthogonal polynomials (Little and Hills, 1972). 37

Manometry (I^). One of the basic assumptions of indirect calorimetry states that measurements on 02 and CC>2 will be accurate only when these are the only gases ex­ changed in the flask. Hungate (1938, 1939, 1943) used manometric techniques to demonstrate that Zootermopsis sp. produced both CC^ and hydrogen during respiration. On the other hand, ruminants, the cockroach Cryptocercus pu'nctulatus

Scudder, and the termites Reticulitermes flavipes (Kollar) and Cryptotermes brevis (Walker) produce methane (Breznak et al., 1973). It was therefore necessary to determine if

M. hubbardi also produced these gases,

Hungate's (1943) method which used palladium black to absorb during manometric measurements was employed.

Termites were allowed to respire in a specially constructed reaction flask containing KOH to absorb CO2 until the 500 yl capacity of the Gilson micrometer unit had been exhausted.

At this point a valve was turned and the gas path directed from the respiration flask into a side bulb containing ca,

0.5 g of palladium black. If were present, it should have been detected by a lowering of the manometer fluid level. This was not observed, however, and a more sensitive technique was attempted.

Gas Chromatography of Respiratory Gases. Whitney and Ortman (1962) and Carlson (1966) are among those who have adapted gas chromatography to the measurement of insect 38 respiratory gases. Their methods were applicable only to

O2, and No one had at this time (1975) described a method for H2 or CHThe assistance of the University

Analytical Chemistry Laboratory was enlisted to develop a method to measure these gases.

Normally faunated M. hubbardi groups weighing ca.

1.5 g (fresh wt) were allowed to feed in a sealed container

(182 cc cap) for 2 h before 2 cc gas samples were withdrawn and injected with a gas syringe (Hamilton Co., model 1005) into a gas chromatograph (Hewlett Packard, Model 762OA).

The carrier gas, argon, was delivered with a head pressure of 40 psi and a flow rate of 60 ml per minute to a thermal conductivity detector at 250°C and a bridge current of 100 mamp. Fifteen ft x 1/4 in (O.D.) molecular sieve (5&)

60/80 mesh columns, held at 225°C (isothermal) were used to measure H^r C^, N£, and CH^. Six ft x 1/8 in (O.D.) molecular sieve (5A) 80/100 mesh columns held at 225°C

(isothermal) were used to measure CC>2 • Two-cc gas samples were injected at a port temperature of 250°C. Peaks were recorded on a 1-mv full-scale recorder (Hewlett Packard, model 7101B ) and compared with those produced by the in­ jection of standards. Although these procedures were excellent for detecting CC^, N^, and C^, a more sensi­ tive method was required for CH^. The same chromatograph and recorder were used in conjunction with a flame- ionization detector at 250°C and 15 ft x 1/4 in (O.D.) 39 molecular sieve (5&) 60/80 mesh at 225°C (isothermal) to measure CH^, The carrier gas, nitrogen, was delivered at a head pressure of 40 psi and a flow rate of 60 ml per min.

All columns were made of stainless steel.

Termite Thermogenesis. The gradient-type calori­ meter has been used for many years for direct measurements of heat production under equilibrium conditions (Hammel and

Hardy, 1963). The early instruments constructed by Atwater and Benedict (1905); Lusk (1915); DuBois (1936); and more recently Prouty, Barrett, and Hardy (1949) were relatively insensitive to rapid fluctuations in heat formation owing to low thermal conductivity of the air-space gradient layer.

As such, these instruments were well-suited only to studies on rather large animals over long periods.

Prat (1954) described a modified Calvet-type micro- calorimeter suitable for measuring the heat production of very small animals, bacteria, germinating seed, and other tissues, A commercially produced version of this instru­ ment is currently available from Setaram, Lyon, France.

More recently, Peakin (1973) used a modified LKB® flow micro-calorimeter to study the costs of maintenance in terrestrial poikilotherms. A precise drop-heat capacity calorimeter similar to the one described by Konicek,

Suurkuusk, and Wadso (1971). has been constructed by Dr.

J. A. Rupley in The University of Arizona Chemistry Department. This instrument was made available for a study of M. hubbardi. Heat production from individuals and groups of 5 termites was measured at 20°C for 18-24 h. The sample ampoule in which the termite(s) were tested had a total gas capacity of 5 cc. It was fabricated from stainless steel tubing (0.3 mm wall thickness) and was fitted with a teflon seal. A simple mechanical lift was used to lower the ampoule into the calorimeter which was allowed to equili­ brate for 4 h or longer to obtain a stable baseline on a

1-mv recorder. The voltage signal output from the calori­ meter was amplified with a micro-ammeter (Keithley, model

150B). The amplified signal was sensed by a digital volt­ meter (SE Laboratories, model SM213, DVM) , integrated every

10 seconds, and recorded on paper tape. The tape was read at The University of Arizona Computer Center and the data reduced with a computer program to calories produced. A continuous record of the amplified signal was produced on the 1-mv recorder. Seventeen experiments, each lasting 18-

24 h, were executed with groups consisting of 4 large larvae or young nymphs and a single soldier. Because normally faunated and defaunated termites were tested while feeding or starving, the relative contributions of host and symbiote to total heat production and the heat increment (Harris,

1966) could be estimated. Six additional experiments were performed on single nymphs or soldiers. Where appropriate, the heat production data were subjected to a one-way 41 analysis of variance. Treatment means were separated at the a = 0.05 level, using the least significant difference (lsd)

(Steel and Torrie, 1960).

Response of Protozoans to Temperature

Variations in the constitution of protozoan faunules present in the termite gut may account for differing physio­

logical responses from the host under certain environmental regimes (Honigberg, 1970). Becker (1969) found that optimum brood-rearing temperatures were slightly lower than those required for high wood-consumption rates. Moreover,

Mannesmann (197 2a) reported that there is a selective de- faunation of cold-sensitive hindgut flagellates in

Copotermes formosanus Skiraki. It was necessary, therefore, to assess the viability of M. hubbardi protozoans over the temperature range in which experiments were to be performed.

Groups of 20 individuals including larvae, young and old nymphs, and secondary reproductives were fed saguaro wood at temperatures of 16, 20, 24, 28, 30, and 32°C. After

14 d, the gut fluid from 3-5 individuals in each group was inspected under a phase-contrast microscope. The general appearance and relative number of the protozoans present were estimated on the basis of what is felt to be normal for field populations. In no case did any experimental individual differ in any observable manner from either field populations or groups held at 30-32°C, a temperature' range particularly well suited to rearing M. hubbardi colonies.

Energy Budgets for Laboratory Feeding Groups

Total energy budgets were constructed where possible for all laboratory feeding groups. When more than 1 repli­ cate was run on a treatment combination involving group size and temperature, mean values were used to calculate budgets.

Respiration figures were estimated from indirect calorimetry unless otherwise noted. The following general formula adopts the notation presented by Petrusewicz (1967):

C = R + P + FU, where

C = ingestion

R = respiration

P = production, and

FU = egestion (rejecta).

Trophic Level Interactions

The a priori observation that organisms interact with other organisms in both higher and lower trophic levels leads to the inevitable conclusion that energy must flow through ecosystems. Efforts were made to assess the energy flow from M. hubbardi populations to higher trophic levels through predalion and to lower trophic levels as a result of death and ecjesbion. 43

Lower Level Interactions

Microorganisms are responsible for a large propor­ tion of the energy flow through any ecosystem (Phillipson,

1966). Detritivores represent an intermediate trophic grouping between the decomposers and the autotrophs. As such they serve to facilitate the activity of micro­ organisms by comminuting litter. It is not clear how, or even if, they alter litter chemically.

A certain amount of material removed by a population from a higher trophic level may be wasted. M. hubbardi . appears to produce a fine sawdust-like residue while feeding on saguaro. This material was collected and weighed for each intermediate and large feeding group, A total of 123,5 mg collected from the 60 intermediate groups tested repre­ sented only 0.27% of the material removed (MR) from the saguaro discs and was thus ignored in subsequent calcula­ tions. The complete results are given in Appendix C,

Egestion was another variable measured during the intermediate and large laboratory feeding experiments.

Nitrogen, lignin, and calorific content, important indi­ cators of nutritional quality, were determined for several fecal-pellet samples obtained from the intermediate feeding groups. Nitrogen was determined by the method described by

Rennie (1965), lignin by the acetyl bromide method (Johnson et al., 1961), and calorific content by oxygen-bomb calori- metry. In addition several very small (2-8 mg) fecal-pellet 44 samples were analyzed for calorific content using a com­ mercial version of the Phillipson micro-bomb calorimeter

(Gentry Instruments, Inc.) according to the manufacturer's recommended procedures.

Although the resulting data are not easily projected for field estimates, biomass records were kept of all dead termite bodies, exuviae, and carton produced by the inter­ mediate and large feeding groups. The calorific content of carton was measured.

Predation

No direct estimates were attempted of annual alate production by M. hubbardi or of the loss to predators. How­ ever, a series of chemical and biological studies were carried out to assess the nutritional quality of termites.

Chemical Assays. Owing to their cryptic behavior, termites remain well-protected from potential predators except during dispersal flights. Ants occasionally penetrate termite galleries (Beard, 1972) and feed upon the wingless castes. It was necessary, therefore, to determine the nutritional characteristics of larvae, young and old nymphs, and soldiers.

There is no general agreement on a standard method for determining the moisture content of biological materials.

High drying temperatures (>100°C) may cause a loss in the calorific content of tissues containing large amounts of fat 45

(Wiegert, 1968). To avoid this, samples were dried for 7 days at 60°C. Unless stated otherwise, all chemical analyses were carried out on termites dried in this manner.

The methods used for determinations of ash, carbon, and calorific content determinations have been discussed previously. Specific techniques will be referred to where appropriate.

Micro-Kjeldahl digestions, were performed on 20-50 mg termite samples in 30-mm Kjeldahl flasks. Two ml of di­ gesting solution [I^SO^ and H^PO^ (1:3 v/v) and ca. 20 mg of catalyst (Appendix A)] were added and slow heating commenced on electric digesting racks. Digestions were complete after 2 h.

Total amino acids were quantified using the same procedure for the woody tissues (Scurfield and Nicholls,

1970), except that smaller samples (100 mg) were analyzed.

An alternate procedure (La Fage, Crowder, and Watson, 1974) which employed sealed hydrolysis tubes was used for early determinations. Total lipids were measured using the

Goldfisch apparatus (Arthur H. Thomas Co., 6-unit model).

Termite samples initially weighing 0.4-3,2 g were extracted continuously for 6 h with ca-40 ml of ethyl ether, dried, and re-weighed. The fatty acids of these lipids were identified and measured using the procedure described pre­ viously under "Identification of Fatty-Acid Methyl Esters." 46

Biological Assay. Although the analyses outlined in the previous section can rapidly provide useful information about the nutritional characteristics of a particular food, they fail to answer a number of questions, especially re­ garding digestibility. The capacity of a food source to support animal growth is often considered the most important measure of nutritional quality. Generally this capacity is directly dependent on the amino-acid composition of the protein in food (Eggum, 1970). It was not possible to collect sufficient M. hubbardi to carry out practical feeding trials nor to find a method for performing such trials with any of its known predators. Munck (1970) recog­ nized several advantages in using mice as test animals for screening protein quality: well-defined breeding stocks are available, they require much less food than rats or chickens more commonly used in such experiments, and their protein requirements correlate well with those of the rat,

Since the amino acid composition of M. hubbardi is re­ markably similar to that of R. flavipes (Table 1), the latter species was considered suitable for the determination of termite protein quality. Fourteen mice (Charles River

CD-I) weaned at an age of ca 2 0 d were paired (Ij", ly) and fed a diet containing R. f.lavipes as the sole nitrogen source ad libitum. The complete diet is doscribcd in Table

2. Food intake was monitored and the supply replenished twice weekly. The mice were weighed weekly. The results Table 1. The amino acid composition of 2 termite species — Values represent the percentage of total amino acids in the sample.

Marginitermes hubbardi

Old Young Reticulitermes flavipes Alate Alate Nymph Nymph Larva Soldier Amino Acid (St)a Cot) Cot) Cot) Cot) Cot) Mixed Castes Cot)

Lysine 6.1b 5.3 5.7 5.9 5.5 4.2 5.6 Histidine 2.9 2.4 2.3 2,4 2.2 2.2 2.3 Arginine 5.9 4.2 3.9 4.3 4.1 3.8 4.1 Aspartic Acid 8.2 8.6 8.5 9.0 8.5 7.4 8.3 Threonine 3.9 4.9 5.1 4.8 5.3 5.3 5.2 Serine 3.4 5.2 5.2 5.0 4,6 4,4 5.1 Glutamic Acid 12.6 11.7 10. 6 10.5 10.6 9,4 11.3 Proline 6.6 7.4 6,9 6.4 7.5 7.0 6.7 Glycine 12.0 10.8 9.3 13.3 13.6 13.0 10.4 Alanine 8.7 11.9 10.8 10. 7 11.5 16.3 11.9 Cystine 0.Q 0.0 0.0 0.4 0.4 0.0 0.6 Valine 6.4 7.1 7.0 6.4 6.4 6.9 7.4 Methionine 1.6 1.8 2.1 1.8 1.5 1.2 1.7 Isoleucine 4,7 4.6 4,7 4.5 4,7 4,9 4.8 Leucine 7.6 7.6 7.6 7.2 7.0 7.4 7.6 Tyrosine 5.7 4.0 5,9 4,7 4.6 4.3 4.4 Phenylalanine 3.8 3.3 3,9 3.5 3.0 2.3 3.1 % AAC 40,17 29.84 38.89 39.93 49,90 51.60 % EAA 42.76 40.67 44.39 46.21 . 35,84 53,76

aSt = Sealed hydrolysis tube; Ot = open hydrolysis tube. b Values are means for 2-10 replicates/sample. % AA = Total amino acids as a percentage of sample weight Cdwb]. % EAA = Percentage of total amino acids that are dietary essentials. 48

Table 2. Constituents of diet fed to 7 pairs of weanling mice for 21 days to assess the quality of Reticulitermes flavipes protein.

Ingredient Amount Per Cent

R. flavipes 233. 3 10.9

Corn oil 57.0 2.7

Cerelose 1396.7 65.3

Cellulose 297.2 13. 9

NaCl 7.6 0.4

Purified Vitamin Mix 76.0 3.6

Dicalcium Phosphate (USP) 57.0 2.7

Trace Mineral Mix 3.8 0.2

Potassium chloride 3.8 C. 2

Choline chloride (50%) 3.8 0.2

Cr2°3 3.8 0.2

Total 2140.0 100. 0

Per cent Crude Proteins Termite = 64.20 Diet = 7.00 Per cent Crude Fat: Termite =12.37 Diet = 4.00 Metabol.i.zable energy supplied by diet = 3.086 kcal/g dry wt. % Calcium = 1,00 % Available phosphorus = 0.5 49 were compared with values obtained from feeding a whole-egg control diet to similar mice. Several indices of protein quality were calculated including protein efficiency ratio

(PER) (Sebrell, 1963), net protein retention (NPR) (Bender and Doell, 1957), protein score (Frost, 1959), and essential amino acid index (EAAI) (Oser, 1959).

Field Estimates on Energy Flow Through an M. hubbardi Population

The possibility of using laboratory measurements to estimate energy flow through a field population of M. hubbardi was not foreseen during the early stages of this study. However, when it was learned that long-term records on the demography of a saguaro forest were available

(Alcorn, 1976), it appeared that energy flow estimates might indeed be possible.

During 1941/42, six 10-acre plots on SNME were selected to study population dynamics of the saguaro cactus

(Alcorn and May, 1962). At that time, maps were prepared which identified the location of every saguaro on the 60- acre study area. The plots were inspected annually and records made of the location of saguaros which had died during the previous year. It was possible to inspect the total population of dead saguaros on plot F-2 during a single morning in June, 1976. On the basis of previous experience cach skeleton was assigned to 1 of 3 categories

(A, B, C) according to the criteria discussod earlier. Due to the potential for future study it would not have been prudent to dismantle the skeletons on this study plot to verify the presence or absence of M. hubbardi. Every skeleton assigned to category B is assumed to contain an active M, hubbardi colony.

Nutting (1969) reported actual counts of individuals present in 10 M. hubbardi colonies. The mean of these counts was used as an estimate of individuals per colony in this study. The caste composition of a large number of groups from field colonies was recorded during the present study. In addition, calorific values were determined for each caste. Sellers (1960) presented mean annual air temperatures for Tucson between 1895 and 1960. The con­ sumption rate at 22°C for a large group (500 individuals) of M. hubbardi was judged to be the closest estimate of a mean annual field consumption rate. Collectively, the preceding information permitted an estimate of total calories dispersed and stored by M. hubbardi on the F-2 plot during a single year. RESULTS

Biochemical Studies of Dead Saguaro Wood

Analytical Determinations

After the initial comminution of 600 g of wood col­ lected originally from each saguaro skeleton, only 500 g remained. The second grinding (in the smaller Wiley mill) to produce 60/80 mesh wood flour was very time consuming and therefore continued to produce only ca. 100 g of each sample,

Extractives. Due to the strong retention of ethanol by wood, it was not possible to measure accurately the specific amount of extract removed by individual solvents.

The results of 60 saguaro extractions are summarized in

Table 3. Three separate samples from the hot-water fil­ trates showed positive reactions with Molisch's reagent, As cellulose is soluble in neither hot nor cold water (Oser,

1965), it is probable that starch, gums, mucilages, and monosaccharides were present in these filtrates,

Wood Lipids. Chloroform and methanol in the ratio of 2:1 (v/v) is generally considered the most effective, simple solvent system for the extraction of lipids from plant and animal tissues (Christie, 1973). However, it was

51 52

Table 3. Percentage of saguaro wood (dry wt) removed by successive 6 h extractions in ethanol (95%), ethanol:benzene (1:2, v/v), and hot water.a

Category No. of Samples Mean % SE Range

A 10 4.50? 0.35 3.12-6.15

B 10 5. 25a 0. 52 2.44-7.73

C 10 4.09a 0.31 Z.97-6.13

Each sample value is the mean of 2 replicates.

Means followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by the least significant difference.

difficult to use this system with samples of less than 4 g, especially when lipid content was low (<10%). In most cases the 4-g saguaro flour samples contained little more than 8 0 mg of lipid (Table 4),

Fatty Acids in Saguaro Wood, Fatty acids in saguaro extracts were expressed as percentages of total peak area.

The results reported in Table 5 are mean values from duplicate chromatographic analyses on 21 wood samples.

Actual values for each of 21 samples are presented in

Appendix B, Methyl esters with carbon chains less than 12 were unresolved. A large percentage of the fatty-acid methyl esters present in the wood samples consisted of palmitic (C16:0), palmitoleic (C16:l), stearic (C18:0), 53

Table 4. Percentage of saguaro wood (dry wt) removed by two 3-min extractions with chloroform:methanol solution (2:1, v/v).a

Category No. of Samples heV an %g.b SE Range

a • o 00 A 11 2.41 M 2,10-3.08

B 10 2.17a 0.29 1.32-3.95

C 9 2.95a 0.41 1.79-4.99

£Each sample value is the mean of 2 replicates.

Means followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by Student*-Newman-Keul' s Multiple Range Test. Table 5. Fatty acids in total chloroform:methanol (2:1, v/v) extracts of saguaro skeletons of 3 age categories — Values are mean percentages of total peak area.

Summary of Means

No. of Group Group Group Category samples 1 C16: 0 C16:1 2 C18 :0 C18:1 C18 :2 3

A 5 5,01 a 31.58a 2,56 a 2.54a 27.47a 25.79a 5 .31 a 0.00a

B 8 4,91 a 30.79a 1.7la 3.13a 23.41a 29.31ab 6. 8la 0.00a

C 7 4. 40a 25.00b 3 . 04a 2,75a 19.48b 33.09b 11,02b 0.14a

aEach sample value is the mean of 2 replicates.

Means in the same column followed by the same letter are not signifi­ cantly different at the a = 0,05 level as determined by the least significant difference, 55 oleic (C18:1), and linoleic (C18:2). Three groups of minor

components were recognized including: (1) methyl esters

with retention times less than palmitate, (2) those between

palmitoleate and stearate, and (3) those higher than

linoleate.

Carbon Analyses. The carbon content of 30 saguaro

samples, 10 from each of the 3 age categories, are pre­

sented in Table 6.

Ash Analyses. The ash residue from the combustion

of saguaro samples from the 30 skeletons, 10 from each category, are shown in Table 7.

Calorific Content. The calorific content (kcal/g dry wt) was determined for 21 of the 3 0 saguaro skeletons

by macro-bomb calorimetry and the data summarized in Table

8. Calorific differences between inner and superficial surface wood from the saguaro ribs were of potential interest with regard to both dry-wood and subterranean termite nutrition, as Gnathamitermes perplexus (Banks) attacks the weathered surfaces of dead saguaro skeletons.

The data in Table 9 show that there was a statistically sig­ nificant difference between the samples from the 2 regions of the rib.

Nitrogen Content. Rennie (1965) suggested the use of 30-rnl Kjeldahl flasks for the digestion of wood flour 56

Table 6. Per cent carbon content of saguaro samples repre­ senting 3 age categories.3

Category No. of Samples Mean % SE Range

A 10 45.45a 0.33 43.35-46.57

B 10 44,96a 0.17 44.16-45.58

C 10 46,02a 0.17 45.09-46.83

aEach sample value is the mean of 2 replicates.

bMeans followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by the least significant difference.

Table 7. Per cent ash in saguaro samples representing 3 age categories.3

Category No. of Samples Mean %b SE Range

A 10 3. 53a 0.40 1.94-6.24

B 10 3.3 9ab 0.39 1,70-5.72

C 10 2.62 b 0.21 1.80-4.07

Each sample value is the mean of 3 replicates.

Means followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by the least significant difference. 57

Table 8. Calorific content (kcal/g dry wt) of saguaro samples from the 30 skeletons representing 3 age categories •— Data obtained by macro-bomb calorimetry.a

1 • 1 • • 1 • .i b Category No. of Samples Mean kcal/g SE Range

A 7 4,43a 0.04 4.28-4.56

B 0 4.44a 0.03 4.37-4.55

C 6 4.49a 0,03 4.37-4.56

aEach sample value is the mean of 2 replicates.

Means followed by the same letter are not signifi­ cantly different at the a = 0,05 level as determined by the least significant difference.

Table 9. Calorific content (kcal/g dry wt) of wood samples from superficial and inner parts of a dead saguaro rib.a

Sample No. of Samples Mean kcal/g*3 SE Range

Surface 5 4. 43a 0.03180 4.328-4.509

Inner 4 4. 63b 0.01600 4.604-4.669

Each sample value is Lhe mean of 2 replicates.

Means followed by the same letter are not signifi­ cantly different at the a = 0.05 level as tested by the t distribution. samples. However, trials with this size flask were un­ successful due to excess frothing during heating. The addition of anti'-foaming solution and the use of the 100-ml flasks eliminated much of this problem. The results of the nitrogen analyses are presented in Table 10.

Amino Acid Profile, Total amino acid analyses are often inaccur.ite measurements even under the best conditions.

During hydrolysis with 6N HCL, tryptophan is so labile that it is lost completely. Other amino acids including cystine, threonine^ and serine, though less labile, are partially destroyed. The addition of sodium thioglycolate to the hydrolysis mixture protects methionine but destroys cystine.

It was not added to the single wood sample hydrolyzed.

Scurfield and Nicholls (1970) found that many factors affected the outcome of their amino acid measure­ ments on wood. Among these were the ratio of acid to wood during hydrolysis, humin formation, and the loss of amino acids through interaction with wood constituents such as lignin. Also, certain non-amino acid wood components may appear a; peaks on the chromatogram. Thirty-three peaks were resolved from the saguaro wood sample. Of the total area resolved, 6.63% was associated with 15 unknowns. The compounds for which authentic standards were available are listed in Table 11, 59

Table 10. Per cent nitrogen content of saguaro samples from skeletons representing 3 age categories.

•Li Category No. of Samples Mean % SE Range

A 9 0.34a 0. 02 0.29-0.44

B 10 0.39a 0. 02 0.27-0.55

C 9 0. 38a 0.02 0.27-0.53

Each sample value is the mean of 2 replicates.

Means followed fay the same letter are not signifi­ cantly different at the a = 0.05 level as determined fay the least significant difference. 60

Table 11. The amino acid composition of dead saguaro wood determined by the method of Scurfield and Nicholls (1970) •— Values represent the per­ centage of total amino acids in the sample hydrolyzed in 6N HCL in an open tube.

Percentage of Total Amino Amino Acid Acids in the Sample

Lysine 3. 13

Histidine 1.19

Arginine 2. 81

Aspartic Acid 11. 83

Threonine 7. 81

Serine 5. 36

Glutamic Acid 10. 47

Proline 8. 42

Glycine 12.17

Alanine 11. 26

Cystine 0.15

Valine 7. 45

Methionine 0. 53

Isoleucine 5.18

Leucine 7. 81

Tyrosine 1.33

Phenylalanine 3.11 61

Lignin Analysis. The results of 50 lignin measure­ ments on saguaro samples from the 3 groups of skeletons are presented in Table 12, All determinations were carried out on extractive-free samples to avoid contamination by catechol and tannins. The sulfuric acid delignification procedure (Ritter and Barbour, 1935) was used as a check on the accuracy of the acetyl bromide determinations. Three

treated samples from skeleton A8 had average lignin content of 25.00%, a value slightly lower than the mean for any 4 acetyl bromide replicates from group A skeletons. It may be concluded that the lignin values in Table 12 are slightly high.

Delignif icat ion. Saguaro wood is particularly re­ sistant to delignification by the chlorite method described by Ritter and Barbour (1935). This procedure calls for the addition of delignifying reagents to wood samples at 1-h intervals until lignin content is reduced to ca 2-3% of the initial sample weight (dry wt). In the first delignifica- tion experiment (Table 13) four 1-g extractive-free saguaro samples were subjected to successive additions of solvents at 1, 2, 3, and 4 h. The sample that received 4 additions lost the greatest amount of lignin, ca. 50%. In a second experiment (Table 13), solvents were added to the reaction mixture at 15-min intervals. In one trial larger amounts of solvents (+50%) were added. After 9 solvent additions the 62

Table 12. Lignin content of extractive-free saguaro wood determined by the acetyl bromide method (Johnson et al., 1961).a

Category No. of Samples Mean %b SE Range

A 10 29.57a 0.76 27.9-33. 7

B 10 26.78b 0,44 2'3. 9-29. 8

C 10 30.97a 0.63 28.1-33. 3

aEach sample value is the mean of 4 replicates.

Means followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by the least significant difference. Table 13. Delignification of dead saguaro wood using the method of Ritter and Barbour (1935) — All determinations were performed on extractive- free, 60/80 mesh, wood flour dried at 105°C for 24 h.

Per Cent Lignin Content

Weight, g Duration, No. Solvent Wt Loss, Calculated from Analysis by Sample Cdry wt) h Additions g Wt Loss H2S04 Method

Experiment No. la A 1.00 1 1 0. 039 24.5 21.6

B 1.00 2 2 0.102 18.2 18.5

C 1.00 3 3 0.123 16.1 15.5

D 1.00 4 4 0.152 13. 2 12.3

Experiment No, 2b AC 1. 00 2-1/4 9 0.124 16.0 13.8

B 1,00 2-1/4 9 0.163 12.1 12.5

£Solvents added at 1-h intervals.

Solvents added at 15-min intervals. Q Sample B, experiment No. 2 received solvent additions which were 50% larger than those added to Sample A. greatest drop in lignin content was only slightly better

(52.5%) than in the first experiment. Since these data were not essential to the overall purpose of the study, no further efforts to determine them were attempted.

Nutritional Physiology

Feeding Trials

Incipient Colonies. The incipient colonies were inspected after 3 y and 1 mo. The resulting data have been reduced in Table 14, Fecal-pellet production was not measured in this experiment. When the experiment was terminated, the mean number of termites per group was 20,13.

It is interesting to find that young and old nymphs to­ gether comprise 35.71% of the survivors. Although there were slightly fewer than 2 reproductives per colony (1,82),

1 group contained 4 alates, which demonstrated that dis­ persal flights can occur in very young colonies. Although

3 colonies produced supplementary reproductives (.1 each) , all but 1 colony retained at least 1 primary reproductive.

The amount of wood consumed by each of the 16 young colonies was remarkably similar (3( + SE = 6.21 + 0.42 g). This corresponds to a daily consumption of 27.17 mg wood/g termite (fresh).

Intermediate Groups. This experiment was the most comprehensive of the feeding trials, The 60 experimental Table 14. Final group weight Cfresh), total food intake, and caste composition of 16 Marginitermes hubbardi incipient colonies — Termites were main­ tained for 3 years and 1 month at 32°C and ca. 100% R. H. on saguaro wood.

Caste Composition (Mean No. per Group)

Wood Primary Secondary Group Wt Consumed Reproduc- Reproduc- Young Old (mg fresh) (mg dry wt) tives tives Larvae Nymphs Nymphs Soldiers Total

Mean 202,99 6,21 1.53 0.19 9,31 3.81 3.38 1.75 20.13

SE 18.99 0,42 0.22 0,10 1,70 0,77 0.77 0,27 2.04

Range 114,8-365,9 3.42-9.10 0-4 0-1 1-23 Ot-IO 0-13 0-3 9-35 66

feeding groups provided a vast amount of data which are

reported in subsequent tables. While space does not permit

the inclusion in text of all data for each feeding unit,

several additional tables have been made available for study

in Appendix C, Termite response was measured as a function

of colony origin, rearing temperature, and group size.

Biomass days (BT) or survivors were sometimes used in

preference to initial group size as independent variables

to assess termite response.

Where appropriate, one-way analyses of variance were

performed to identify differences in response at the a =

0.05 level, The only significant differences noted among

colonies were initial mean weight of individual termites and

biomass of dead bodies. These data are summarized in Tables

15 and 16. It was concluded that the three colonies were

similar enough so that this variable need not be considered

in further analysis.

The only response variable which differed signifi­

cantly with changes in temperature was the quantity of

carton produced (Appendix C). Groups at 24°C produced the

least (2.17 mg) and those at 28°C, the most (16.13 mg).

The initial number of termites in a group appeared to affect termite response far more than either colony

origin or temperature. One measure of colony success is the

J total number of offspring (L r) produced by a feeding group

Table 17). As one might expect, large groups consumed more 67

Table 15. Dead termites which remained in feeding groups — Data are grouped by colony.

Mean wt of No. of bodies/group Colony Groups (mg dry wt)a SE Range

A 20 0. 94a 0.43 0-7.4

B 20 4. 71b 1,61 0-27.1

C 20 2.2lab 0.78 0-12,0

All 60 2.62 0,64 0-27.1

cl Means followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by the least significant difference.

Table 16, Initial mean weight (mg fresh) of individual termites selected from 3 natural colonies of Marginitermes hubbardi.

No. of Feeding Mean wt Colony Groups (mg fresh)a SE Range

A 20 10.7a 0.20 9.16-12. 34

B 20 9.33b 0. 20 8. 08-11.12

C 20 9.64 b 0.19 8. 19-11.58

All 60 9. 90 0.14 8. 08-12.34

Means followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by the least significant difference. 68

Table 17. Mean total offspring (eggs and larvae) produced by intermediate feeding groups during 90 d.

Group No. of Feeding No. of Size Groups Offspring SE Range

5 15 0.00a 0.00 0-0

10 15 0.27a 0.15 0-2

20 15 2.13a 0.76 ' 0-9

50 15 7,27b 2.18 0-23

All 60 2.42 0.68 0-23

aMeans followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by the least significant difference.

wood and produced more feces than smaller groups, Table 18 summarizes consumption and fecal-pellet production data for the 60 experimental groups.

When wood consumption and fecal-pellet production are divided by BT, the resulting data (Table 19) look quite different from those in Table 18. On a basis of mg termite/d,fecal production is quite constant among the 4 groups. Food consumption, however, is higher for indi­ viduals in the smallest sized groups (5),

The data in Table 19 suggest that, although indi­ viduals in smaller groups consume more food than those in larger groups, they excrete about the same amount of feces; Table 18. Wood consumption and fecal pellet production by intermediate sized feeding groups.

Mean Fecal No. of Mean Wood Pellet Jroup Feeding Consumption^ Productiona Size Groups (mg dry wt) SE Range (mg dry wt) SE Range

5 15 236,33a 23, 28 87-382 31.61a 6.19 2.5-73.9

10 15 388.33a 34. 04 157-597 87.54a 10.17 13.6-145.2

20 15 724,60b 61,86 277-1037 209,53b 17. 33 79.5-304.3

50 15 1680.67° 141.77 817-2847 611,35° 60. 27 179.0-1070.9

All 60 756.48 82. 85 87-2847 237,51 33. 81 2.5-1070.9

aMeans followed by the same letter are not significantly different at the i = 0.05 level as determined by the least significant difference. Table 19. Food consumption and fecal pellet production divided by biomass days for the intermediate sized feeding groups.

Mean Wood Mean Fecal- Consumption Pellet No, of (mg/g (mg/g Group Feeding termite, termite, Size Groups fresh/d)a SE Range fresh/da SE Range

5 15 294b 39 0,05-1. 27 0.016a 0.003 0.006-0. 055

10 15 84a 10 0.03-0. 43 0.015a 0. 002 0.008-0.037

20 15 55a 2 0,02-0. 10 0.016a 0.001 0.006-0. 026

50 15 45a 3 0.02-0,07 0,016a 0,001 0,004-0,023

All 60 119 22 0,02-1, 27 Q.016 0.001 0.004-0.055

^eans followed by the same letter are not significantly different at the a = Q.Q5 level as determined by the least significant difference. i.e., they assimilate a greater percentage of their food.

Apparent assimilation, defined as I(consumption - feces) * consumption], was examined by a one-way analysis of variance and the results summarized in Table 20,

Table 20. Apparent assimilation of food by the intermediate sized feeding groups.

Group No. of Feeding Mean Apparent Size Group Assimilation3 SE Range

5 15 0.876a 0. 018 0. 78-0.99

10 15 0.775b 0.018 0.66-0.91

20 15 0.7 01° 0.012 0.62-0.84

50 15 0.641d 0.015 0.58-0.84

All 60 0. 749 0.014 0. 58-0.99

Means followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by the least significant difference. 72

The 6 0 feeding groups initially contained a total of

1275 termites. Eight hundred ninety-six (70.3%) of these survived the experiment, during which time 90 new larvae and 55 eggs were added to the population. Survival per­ centages within the temperature group-size categories

(Appendix C) were analyzed after being transfored to probits.

No significant differences were noted among temperature or group sizes.

The caste composition of each of the 6 0 feeding groups was recorded (Table 21) at the end of the feeding trials, Orie-way analysis of variance was used to determine if the percentage of individual castes varied among the feeding groups according to temperature or group size.

While no differences in caste composition were noted among the 5 temperature treatments, group size did influence caste ratios. Larger groups tended to allocate more biomass to nymphal castes and thus maintained smaller percentages of soldiers, supplementary reproductives, and larvae than did smaller groups,

Per cent of biomass gained by a feeding group is one estimate of productivity. These data are given in Table 22,

Productivity was noted only in groups which initially con­ tained 50 termites.

Multiple linear regression equations were developed which predict termite food consumption, fecal-pellet pro­ duction, and assimilation rates as functions of biomass days Table 21, Final caste composition of intermediate sized feeding groups of Marginitermes hubbardi — Data are presented for temperature and group size categories.3

Secondary No, of New Reproduc- Young Old Row Groups Eggs Larvae tives Larvae Nymph Nymph Soldiers Total

Temperaturs, °C: Total Number of Individuals

24 12 5 0 20 5 138 17 5 . 190 26 12 12 1 24 9 100 63 9 218 28 12 33 0 22 9 82 31 10 187 30 12 7 28 24 7 100 37 9 212 32 12 33 26 23 14 95 30 13 234 Column Total 60 90 55 113 44 515 178 46 1041

% of Grand Total 8.65 5. 28 10.35 4.23 49.47 17 .10 4. 42

Mean No. of Individuals

24 12 0.423 0. ooa 1.67a 0.42a 11.50a 1 .42a 0. 42' 15.83 a 'ab 26 12 i.ooa 0. 08a 2,00a 0.75a 8.33a 5 .25a 0. 75' 18.17 a * a ab 28 12 2.75a 0. ooa 1.75 0.75a 6.83a 2 .58a 0. 75' 15 • 58a ab 30 12 0.58a 2. 3 3a 2.00a 0.58a 8.33a 3 .08a 0. 17.67 a 75ib 32 12 2.75a 2. 17a 1.92a 1.17a 7.92a 2 .50a 1. 08 19.50 All 60 1.50 0. 92 1.87 0,73 8.58 2 .97 0. 75 17.35 Table 21.—Continued

Secondary No. of New Reproduc- Young Old Row Groups Eggs Larvae tives Larvae Nymph Nymph Soldiers Total

Group Size: Total Number of Individuals

5 15 0 0 15 2 15 1 4 37 10 15 4 0 27 5 58 4 10 108 20 15 21 11 34 16 136 14 14 246 50 15 65 44 37 21 306 159 18 650 Column Total 60 90 55 113 44 515 178 46 1041 % of Grand Total 8. 65 5. 28 10.85 4.23 49.47 17 .10 4. 42

Mean No, of Individuals a ooa 1 a 0.13d i.ooa a 5 15 0. oo 0. . oo 0 -67 0. 27 2 47 a a 73 a a b • b 10 15 0. 27 0. oo i- ^ °"33b 3.87 0 .27 0. 60 7 .20 20 15 1. 0. 2.27 1.07^ 9.07 0 0. 93° 16 40b 73b •93h •40d 50 15 4. 33 2. 93 2.47C 1.40 20.40C 10 .60 1. 20C 43 ,33d All ' 60 1. 50 0. 92 1.87 0.73 8.58 2 .97 0. 75 17 .35

Mean No. t Group Size 5 15 0. ooo a, 0. oooa 0.027a 0.200a, 0.200a 0 .013a 0.053 a 0 .493 027?b 0.000 , 0.033 0,173*° 0.387° .027a 0.060 a 0 .720 10 15 0, be 0 ab a a 20 15 0. 070 0.037 0.053 0.113 °-453b 0 .047a 0. 047 0 .820 50 15 0.08 7C 0.059 0.028 0.049° 0.408 0 .212 0. 240a 0 .867 All 60 0- 046 0, 024 0.035 0.134 0,362 0 .075 0.046 0 ,818

a Means in the same column followed by the same letter are not significantly different at the a = 0.05 level as determined by the least significant difference. 75

Table 22. Percentage of original biomass ffresh) gained or lost by intermediate sized groups during the 90 d feeding trial — Data are categorized by temperature and group size.

No, of Feeding Treatment Groups % Wt Change3 SE Range

Temperature, °C: 24 12 -30.64a 10.23 -4.12 to -100,00 26 12 "8-00a 10. 46 -42. 40 to -100. 00 28 12 -24.02 9. 05 -7. 30 to -100. 00 30 12 -19.65® 12.64 -28. 38 to -100. 00 32 12 -25.67 5.17 -6. 81 to -59,79 All 60 -21.60a 4.36 -42. 40 to -100,00

Group Sizet 5 15 -51.95? 9.91 -6. 81 to -100,00 10 15 -25-73bc 7. 42 -10.14 to -100,00 20 15 -11.05®° 3.94 -7. 30 to -49. 20 50 15 +2.34 5.93 + 42. 40 to -32. 53 All 60 -21.60 4.36 +42. 40 to -100.00

g Means followed by the same letter are not signifi­ cantly different at the a = 0,05 level as determined by the least significant difference. and temperature (Table 23), They explain more than 89% of the observed variation.

Large Feeding Groups. Although 7 temperatures were

used to test termite response in large feeding groups, there

was only 1 replicate per treatment and, therefore, no

statistical analyses were performed. Tables 24 and 25

summarize the results of this experiment.

Respiration Studies

Manometry (02, C02)• The results of 33 6 oxygen-

consumption tests, each lasting 8-12 h, are presented in

Table 26. These values are means for the total duration of the experiments. There appears to be no standard for expressing metabolic activity, although dry weight, fat-free dry weight, unit-protein nitrogen content, and fresh weight

are measures often encountered in the literature. Since

"dry" tissue is almost always partially hydrated, Keister

and Buck (1974) suggest that fresh weight is as good a measure as any. As it was not possible to obtain initial dry weights for the groups, oxygen consumption is expressed in terms of fresh weight.

To estimate an oxy-caloric equivalent (Brody, 1945),

it is necessary to know the chemical nature of nutrients

being metabolized by the termites. A respiratory quotient

(CC>2 evolved v 0^ consumed) near 1.0 suggests that Table 23. Equations for predicting wood consumption (mg), fecal-pellet production (mg), and apparent assimilation efficiency by groups of 5-50 Marginitermes hubbardi — B = biomass days, T = °C.

Response Variable Equation R^a

Wood Consumption (mg) Y = -94.98 - 0.0623 B + 0.30 + 0.00368 B'T 0.961

Fecal Production (mg) Y = 53.06 - 0.03 B - 0.0842 + 0.00175 B*T 0.944

Apparent Assimilation n ~ Efficiency Y = 0.914 - 1.99 x 10 B + 3,166 x 10 B 0.894

aR^ = coefficient of determination (Steel and Torrie, 1960). b^ . j » ,. Consumption - Feces Definedc by the equation: •- t J Consumptionr Table 24. Caste composition of 6 large feeding groups (490-500 individuals) of Marginitermes hubbardi maintained on saguaro wood for 90 d at ca. 100% RH,

Production Caste Composition of Survivors Tempera­ ture Initial No, Final No. New Supplementary Young Old (°C) in Group in Group Eggs Larvae Reproductives Larvae Nymphs Nymphs Soldiers

22 500 444 0 0 4 20 156 250 14

24 500 348 0 0 6 40 126 170 6

26 500 387 0 0 1 1 83 291 11

28 490 334 0 1 9 2 38 268 16

30 500 324 0 0 3 1 87 231 2

32 500 355 13 0 3 •2 104 220 13

Column Total 2990 2192 13 1 26 66 594 1430 62 % % of Grand Total 0,59 0.05 1.19 3.01 27.10 65.24 2.83 Table 25. Wood consumption, fecal-pellet production, biomass, and carton produc­ tion by 6 large groups (490-500 individuals) of Marginitermes hubbardi fed saguaro wood for 90 d — Experimental units were maintained at ca. 100% RH.

Rearing Tempera­ Initial Final Wt Change Wood Consumption Fecal-Pellet Carton ture No. of Group Group of Group (mg/g termite Production Production (°C) Groups No, No. (mg fresh) fresh/d) (g) (mg dry wt)

22 1 500 444 + 26.8 16.53 5.2170 0.00

24 1 500 348 -613.5 19.44 4.2010 0.00

26 1 500 387 +1251.7 24.48 6.2207 8.10

28 1 490 334 -280.1 23.97 5.3082 201.80

30 1 500 324 +284.9 30.73 8.2760 21.60

32 1 500 355 -808.7 29.93 6.5193 157.50

Mean 498.33 365.33 -23.15 24.18 5.9570 64.83 Table 26. Mean oxygen consumption by groups of Marginitermes hubbardi — Values represent means of 14 replicate groups, each containing 5 or 10 indi­ viduals, Data were obtained by manometry and expressed as pi O2 consumed per mg termite (fresh wfc) per h.a

Feeding Starving

ture Normally Normally C°C) Faunated R.• Q. Defaunated R,, Q • Faunated R.• Q. Defaunated R. Q.

16 0, 083 1.,07 0.,041 0., 93 0,,072 0.,63 0., 060 1.,28

20 0. 225 0., 97 0.,176 0., 76 0.,145 0., 75 0.,114 0.,76

2 A 0., 398 0,, 96 0,, 219 0,,84 0,,195 0.,64 0,,152 0.,75

28 0,, 578 0,,96 0.,378 0,, 85 0., 202 0,,69 0.,224 0..79

32 0., 597 0.,98 0..482 0.,81 0.,326 0..71 0.. 283 0.,78

36 0., 839 0., 85 0.,542 0.,75 0.,399 0,,69 0..456 0.,71

X 0,,453 a 0,, 306b 0., 223C .0,, 215d

Means in the same row which are followed by the same letter are not significantly different at the a = 0.05 level as determined by the Student- Newman-Keul's Multiple Range Test. 81 carbohydrate is being utilized. Starving animals, deriving their energy exclusively from stored fat, show RQ1s near

0.7. The consumption of mixed protein produces RQ's around

0.83. The mean RQ for normally faunated M. hubbardi (Table

26) was 0.97. This is sufficiently close to 1.00 to select

5,05 kcal/L C>2 as the oxy-caloric equivalent. As the C>2 consumption rates reported in Table 26 suggested a pre­ dictable relationship with temperature, polynomial tests for trends were attempted. The dependent variable [yl C^/mg termite (fresh)/h] was transformed to logg and regressed, step-wise, on the linear, quadratic, and cubic components of temperature. The resulting prediction models and their coefficients of determination (R 2's) are listed in Table 27.

Gas Chromatography of Respiratory Gases. The main objective of 2 gas chromatographic procedures was to identify hydrogen and/or methane in termite respiratory gases. Should these gases be found in large amounts, their specific calorific equivalents would necessarily have to be considered in energy budgeting. Although no serious attempt was made to assess gas production quantitatively, a few statements can bo made regarding the relative rates of l-^ and CH^ evolution. While CO^ concentrations reached as high as 5.0% of the total gas volume of the respiratory chambers, in no case did concentrations exceed 1.0%. The highest concentrations of CC>2 and I-I^ were reached after termites had Table 27. Prediction models which estimate oxygen consumption (Y) (yl (^/mg termite tfresh]/h) for the 4 physiological states tested by manometry — The dependent variable (Y) is regressed on (T) temperature C°C).

Treatment Equation

Normally faunate

Feeding loge Y = -7.737 + 0.424 T - 0.006 T' 0.927

log Y = -3.815 + 0.0830 T 0,891 Starving e

Defaunated

Feeding log Y = -9.431 + 0.513 T - 0.008 T 0.895 3e Starving log Y = -4.315 + 0.0976 T ' 0.935 e 83 respired for 12 h in closed chambers. At this time C02 levels began to have a narcotic effect and after 24 h the termites died. For this reason, chambers were flushed with fresh room air after each 12 h. CO^ and H2 standards were used to estimate the concentration of gases. Methane pro­ duction determined by the second procedure was equally un­ impressive. The highest CH^ concentrations measured (22 ppm) were achieved after 4 hours in a closed chamber by normally faunated/feeding groups on filter paper (Whatman

No. 2). Termites which fed on saguaro wood produced less than 11 ppm CH4. The combined H2 and CH^ production was less than 1% of the respiratory gas output from M, hubbardi and thus was not considered in energy budgeting.

Termite Thermogenesis. The results of M, hubbardi thermogenesis experiments are reported in Table 28 as calories produced per mg termite (fresh) per 12 h. The period for which heat production was recorded began after a stable baseline response had been achieved by the empty calorimeter. The ampoule which housed termites during heat measurements had a gas volume of ca, 5 cc and contained enough 02 for a 75-mg M. hubbardi group to survive almost

47 h at 20°C. Since C02 levels would have become fatal after ca,24 hf heat measurements were made only during the initial 16. Although termites removed from the ampoule after this period showed no immediate ill effects, they were 84

Table 28. Comparison of heat production by Marginitermes hubbardi determined by calorimetry and respirometry.

Treatment Mean Heat Production3 cal/mg Termite (Fresh)/12 h Groups of 5 No. Termites Replicates (Calorimetry) (Respirometry)

Normally Faunated

Feeding 5 0.01895a 0.0116

Starving 4 0.00770° 0.0070

Defaunated

Feeding 4 0.01419b 0.0057

Starving 4 0.00903c 0.0057

Single Individuals

Large Nymph 3 0.01513

Soldier 3 0.00768

3 Means followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by the least significant differencej b Oxy-caloric equivalent = 5.050 kcal/L O2. 85 held for observation under normal rearing conditions for

several days.

Figures 8a and 8b show parts of thermograms produced by a single young nymph. The event pictured, a sudden fall and rise in heat production, is remarkably similar to a thermogram produced by Galleria mellonella (L.) during adult emergence (Prat, 1954).

Energy Budgets for Laboratory Feeding Programs

The following equations were used to calculate energy budgets for incipient colonies, and the intermediate and large sized laboratory feeding groups, The notation is that of Petrusewicz and Macfadyen (1970) and Petrusewicz

(1967). MR = C + NU

C = A + FU

A = P + R

P = G - L where

MR = material removed;

C = total intake of food during a defined time period;

A = assimilation, the sum of production and respiration;

FU = rejecta, that part of consumption not used for

production and respiration; 15 min -I (a) Isolated event which, shows rnaximum amplitude

20 min I —i (b) Three successive events of lesser intensity

Figure 8. Parts of a thermogram produced by a single, young Marginitermes hubbardl (Banks) nymph during a 12-h recording of heat production. 87

P = production, the total amount of body tissue

generated (G) by a population less weight losses

(L) during a defined period of time; and

R = respiration, that part of consumption converted to

heat and dissipated in life processes (metabolism)

during a defined period of time.

Direct measurements were available for the following

variables: MR, NU, C, FU, and P. Respiration (R) was estimated on the basis of observed and predicted 0^ con­ sumption (Tables 26 and 27). For incipient colonies, caste

biomass was partitioned according to the percentages re­ ported in Table 29. Estimates for dry weights and calorific content of termite castes were made using the values pre­ sented in Tables 30 and 31, respectively. With these data it was possible to calculate energy budgets for all labora­ tory feeding groups.

Incipient Colonies. An energy budget for 16 in­ cipient colonies was constructed. Although fecal-pellet production data were not recorded for these colonies, a regression equation had been developed earlier which pre­ dicts wood consumption on the basis of fecal-pellet produc­ tion by M. hubbard1 at 32°C. Consumption data for the in­ cipient. colonies were substituted into tho following equation to predict fecal-pellet production: 88

Table 29. Mean biomass (by caste) of Marginitermes hubbardi colonies collected on or adjacent to the Silver- bell site during 1974-75.a

Per Cent Total Mean Biomass of Sample Wt Termite Individual Caste (g dry wt) Biomass (mg dry wt)

Nymph (Old) , 10.1638 GO. 73 6. 31

Nymph (Young) 5.8604 35.02 31 64

Larva 0.3671 2.19 1. 57

Soldier 0.2925 1.75 4.22

Supplementary Reproductive 0,0527 0,31 4.98

aDried at 60°C for 7 days. 89 Table 30. Moisture content (by caste) of Marginitermes hubbardi.a

No. of Mean Sample Wt Mean % ^ Caste Samples (mg dry wt) Moisture SE

Alate

Newly Emerged 12 4.5 72.46a 0.64

Pre-dispersal 5 48.5 66.17b- 0.59

Post-dispersal 6 369.4 61.34C 0.26

Nymph (Old) 7 34.3 67. 41 1.21

Nymph (Young) 5 32. 8 76. 33 1.03

Larva 8 25.8 78.49 0. 54

Soldier 4 38.0 67. 33 0.40

aDetermined as difference between fresh weight and dry weight after 7 days at 60°C.

Means followed by the same letter are not signifi­ cantly different at the a = 0.05 level as determined by the least significant difference. Table 31. Calorific content (kcal/g dry wt) (by caste) of Marginitermes hubbardi, determined by micro- and macro-bomb calorimetry.

No, of Mean Calorific Caste Samples Content SE Range Method

Alate 10 6. 41 0.31 5.87-7.02 Micro

Nymph, Old 11 6,78 0.15 5.76-7.39 Micro

Nymph, Old 3 6.84 0. 04 6.77-6.90 Macro

Nymph, Young 10 6.26 0.21 4.89-7.12 Micro

Nymph, Young 3 6.35 0. 35 5.89-7.02 Macro

Larva 5 5. 82 0.20 5.31-6.29 Micro

Soldier 5 4.43 0,14 3.94-4.72 Micro Wood consumption =

70.77 + 2.4s (fecal-pellet production).

The data presented in Table 32 suggest that the incipient colonies dispersed slightly more energy (9.9%) than they consumed. However, considering the large number of estimates involved in computing the budget, this error is small.

Table 32. Mean energy budget for 16 incipient colonies of Margin iterates hubbard i. — The colonies were maintained at 32°C and ca. 100% RH for 3 years and 1 month.

Energy of Biomass Source (kcal/g dry wt)

Input

MR 27.57

NU 0. 00

C 27.57

Output

P 0.3166

Ra 16. 52

FU 13.47

Total 30, 31

aBased on estimate i com Table ;6. 92

Intermediate Sized Groups. Mean energy budgets were

calculated for 15 feeding units in each of the 4 group-size

categories (5, 10, 20, and 50 termites). All energy terms

were measured directly except respiration (R), 02 consump­

tion was estimated for energy budgeting from Table 27. An

oxy-caloric equivalent of 5.05 kcal/L O^ was used because

normally faunaied/feeding termites had RQ's near 1.00 (Table

26). One additional variable, biomass days, was used to

calculate R estimates which appear in Table 33. The cal­

culation of production by these groups is sufficiently

complex to warrant further explanation. Production (P), as

defined in this study, is the sum of weight gained minus

weight lost during a defined period of time. However, the

concept of ecological productivity has several additional

definitions, one of which measures the total generation of

organic matter during time, regardless of whether this

matter remains in the population to the end of the observa­

tion period (Petrusewicz and Macrfadyen, 1970). Two types of

growth are recognized in ecolog..cal studies. The first, P ,

is a measure of biomass of new individuals, i.e., natality.

The second, P , is production resulting from weight gained

by individuals present at the start of the observation

period. Several forms of weight loss are also recognized.

Among these are losses through death and ecdysis, glandular secretions, and endogenous (body) and metabolic fractions of the urine and feces respectively. Since they are recycled 93

Table 33, Calculation of respiration (kcal) by inter­ mediate sized Marginitermes hubbardi groups — O2 consumption rates are based on direct manometric measurements at 2 8°C for groups of 5-10 termites. Mean biomass days (Appendix C) were used for a wt-time factor. Oxy-caloric equivalent = 5.05 kcal/L O2 consumed with R.Q. = 1.00.

Group Total 0O Consumed Calorific Equivalent Size Biomass Days ID (kcal.)

5 2471. 93 0. 034 0.173 10 6742. 67 0. 094 0. 472 20 13551. 40 0.188 0.949 50 38181. 80 0.530 2. 675

quickly by the termite, very few exuviae were noted among any feeding groups. Although they are generally recycled, a measurable quantity of dead termites or body parts were present when the feeding trials were terminated. Produc­ tivity for the intermediate sized groups was partitioned into 3 components: P , P , and weight losses (L) (Table 34) , r g Only 1 of the 4 groups showed positive production values.

Since negative production has a little ecological meaning, all negative eases were assigned 0 values in the construc­ tion of energy budgets (Tables 35 and 36) .

The othei: variable, carton, affects the value of both input and output energy. The carton and calorific 94

Table 34. Calculation of production (P) for intermediate sized groups of Marginitermes hubbardi.

kcal

L p P Group r g Dead " Id 0 Size Larvae + Eggsa Biomass Change (AB) Bodies P

5 +0.00000 -0.048 -0. 023' -0.071

10 +0.00124 -0.042 -0.018 -0.059

20 +0.00122 -0.037 -0.005 -0.041

50 +0.00415 +0.022 -0.017 +0.009 \

a Pr = Production from birth of new individuals or eggs.

Pf = Changes in biomass of individuals present at the beginning of the experiment.

L = Loss of biomass due to death and ecdysis. 95

Table 35. Mean energy budgets for intermediate sized groups of Marginitermes hubbardi.

Energy (kcal)

Group Size (No. of Termites)

Source 5 10 20 50

Input

MR 1. 060 1.740 3,250 7.530

NU 0. 010 0,040 0. 040 0. 050

C 1.050 1.700 3.210 7. 480

Output

Pa 0. 000 0. 000 0. 000 0. 009

R 0.173 0. 472 0. 949 2.675

FU 0.180 0. 510 1.170 3. 330

Total 0. 353 0. 982 2.119 6. 041

aProduction estimates <_ 0.00 have been assigned 0 values in the calculation of total energy flow. 96

Table 36. Energy budgets for the large laboratory feeding groups of Marginitermes hubbardi — A single replicate was used for each temperature. Groups were allowed to feed for 92-97 d at ca. 100% RH (kcal).

Energy (kcal)

Temperature

Source 22 24 26 28 30 32

Input

MR 48.115 41.754 61.421 57.926 78.400 70.918

NU 0. 000 0. 000 0. 020 0.505 0. 054 0.394

C 48.115 41,754 61.401 57.421 78.346 70.524

Output

P 0. 047 0. 000 2.162 0. 001 0. 492 0. 007

R 21.189 21.034 31.430 36.769 45.485 46.827

FU 28.015 22.559 33.425 29.010 44.496 35.403

Total 49.211 43.593 67.017 65.779 89.981 82.230 contents were determined respectively as follows: carton,

50% and 5.0 kcal/g; saguaro wood, 45.5% and 4.48 kcal/g; and feces, 51.8% and 5.37 kcal/g. These values suggest that carton is a mixture of feces and undigested wood. For energy budgeting, it was assumed that the mixture was half wood and half feces. Half of this energy was added to FU and the other value to NU. The final energy budgets for the intermediate sized group are given in Table 35.

Trophic Level Interactions

Lower Level Interactions

The nitrogen and lignin content of M. hubbardi fecal-pellet material was determined for samples from each of the 5 temperatures used to test intermediate sized groups. Pellets from all groups sizes had to be combined to provide adequate samples. The results appear in Table 37.

The mean carbon content of 4 carton samples was

50.03% and for 10 fecal-pellet samples, 51,84%. Calorific content was determined on fecal-pellet sub-samples from the total pellets of all 60 intermediate sized groups. The micro- and macro-bomb determinations on carton had a mean and standard error of 5.4 4 j. 0.24 kcal/g (dry wt) . Very little carton was collected from the intermediate and large groups (Appendix C and Table 25). Carton samples had an average calorific content of 5.00 kca.L/g dry wt of 0.295.

So few dead bodies or body parts remained in the feeding 98

Table 37. Nitrogen and lignin content of Marginitermes hubbardi fecal pellets produced during inter- mediate feeding trials — Nitrogen was determined by the Micro-Kjeldahl method (Rennie, 1965), and lignin by the acetyl bromide method (Johnson et al., 1961).

Temperature, No. of Lignin No. of Nitrogen °C Samples (%) Samples (%)

24 2 60. 52 3 0.225

26 1 66. 56 3 0.175

28 2 61.12 3 0. 225

30 2 60.40 3 0.195

32 2 59. 55 3 0.195

Mean 9 61. 63 15 0. 203

units after 90 d (Appendix C) that noi measurements other than biomass were recorded. For the purposes of energy

budgeting, dead bodies were assigned a mean calorific value

for all casts of 6.02 kcal/g dry wt.

Predation

A large number of field-collected colonies of M.

hubbardi provided the termites for the chemical analyses described below. Although time did not permit a complete census of each colony, a large number of individuals were

collectcd, separated into casts, and weighed. Thus, it was

possible t:o est iinnto how t: r rm i I: e. hi. omn r; r, is pnrl ilrioncd 99 into castes in natural colonies, These data were summarized in Table 29.

Chemical Assays, The moisture content of M, hubbardi is given in Table 30. Animal tissues generally contain less than 10% mineral ash (dry wt). In order to minimize weighing error, a few large samples were analyzed instead of many smaller ones. The ash content of M, hubbardi castes appears in Table 38. Differences in the calorific content of M. hubbardi apparently reflect a gradual deposi­ tion of energy-rich fat as larvae develop into nymphal and adult stages. The results of numerous micro- and macro-bomb calorimetric determinations on M. hubbardi castes appear in

Table 31. Table 39 contains the results of 32 carbon determinations performed on M. hubbardi samples. Too few soldiers were available in stock cultures to permit their inclusion here. The recovery efficiency of the carbon analyses was very high (96-98%) as determined by the combustion of glycine samples. Carbon determination by combustion in oxygen is much more accurate than wet oxida­ tion with dichromate, a procedure often used for biological material. Strickland and Parsons (1968) state that the latter method often causes the oxidation of compounds other than carbon.

Kjeldahl-nitrogen content was determined for several, castes (Table 40). Since protein contains about 16% 100

Table 38, Ash content of Marginitermes hubbardi — Values represent residues of termites combusted in a muffle furnace at 550°C for 18 h.

Mean Ash Content NO. Of Mean wt of Sample (% of Sample Caste Samples (mg dry wt) wt dry wt)

Alate 5 206,2 4.29

Nymph (Old) 4 509. 8 4:37

Nymph (Young) 2 152.4 5.20

Larva 1 363.3 6.19

Soldier 1 84.3 4.27

All 11 263.2 X 4.86

Table 39, Carbon content of Marginitermes hubbardi deter­ mined by the dry-combustion method of Allison et al. (1965).

No. of Mean % Caste Samples Carbon SE Range V

Alate 6 52. 90 0.54 52.14-55. 52

Nymph (Old) 10 53.31 0.36 51. 17-54. 79

Nymph (Young) 10 51. 62 0.41 49. 03-53. 45

Larva 6 48.06 0.08 47. 85-48. 36 101

Table 40. Nitrogen content of Marginitermes hubbardi castes.

No, of % Nitrogen Caste Samples (dry wt) SE Range

Alate 5 8.65 0.19 8. 18-9.24

Nymph (Old) 7 6. 95 0.33 5. 46-8.18

Nymph (Young) 16 8.49 0.51 6.•18-11. 69

Larva 15 9.45 0. 40 6. 62-11.63

Soldier 12 12. 88 0.27 11. 32-14.21

£Each sample represents 2 replicates.

nitrogen, many authors predict a quantity called "crude protein" which is equal to nitrogen content x 6,25,

Nitrogen-to-protein conversion factors, based on measure­ ments of the true protein content of animal tissues, vary widely between 5.18-7.46 (Merrill and Watt, 1955).' Tassoni

(1951) determined that the nitrogen-to-protein conversion factors for the true protein in the pupa of the moth Telea polyphemus Cramer was 6.19, He also determined that the factor for the whole insect, including exoskeleton, was somewhat higher than 6.41. Since a true conversion factor was not determined for the termites in this study, a value of 6.25 was used when it became necessary to estimate protein. 102

The amino acid composition of M. hubbardi and R,

flavipes appears in Table 1. The two hydrolytic methods

used were designated by ST for sealed tubes and OT for open

tubes.

The lipid content of M, hubbardi and R. flavipes

was determined for several castes. Measurements using a

modified "Folch" technique were largely unsuccessful due to

weighing error and large numbers of quantitative transfers

required by the procedure. The Goldfisch extractor, how­

ever, provided rapid and reproducible means to determine

total lipid content. Termite samples were initially

extracted with ethyl ether for 6 h, dried, and weighed.

They were then extracted for an additional 2 h. In no case

was additional weight loss greater than 1.0% of the already

extracted lipid weight. The results of this experiment are

summarized in Table 41, Lipid extracts prepared in this

manner were esterified and analyzed with the gas chromato- graph. The results are presented in Table 42,

Biological Assay. The results of the biological

assay in which termite protein was fed to mice are sum­

marized in Table 43.

Field Estimates on Energy Flow Through an M. hubbardi Population

Saguaro density is highly vari.nblo From onn location

to another. In 1941, section 17 of SNME had ca.15 live 103

Table 41. Termite lipids extracted by ethyl ether in a Goldfisch extractor.3

No. of Mean % of Species and Original Weight Caste Samples (dry wt) SE Range

Marginitermes hubbardi

— Alate 1 46.43 • —

Nymph (Old) 3 42.38 0. 54 41.60-43.42

Nymph (Young) 3 22.07 0.25 21.70-22.56

Larva 1 17.05 — —

Soldier 1 6.24 — —

Reticulitermes flavipes

Mixed Castes 3 12.37 0.15 12.22-12.66

aSamples were extracted during 2 successive periods totaling 8 h, Table 42. Fatty-acid composition of Reticulitermes flavipes and Marginitermes hubbardi.a

Mean Per Cent Fatty Acids Species and No. of Castes Samples Group 1 C16 :0 C16:1 Group 2 C18:0 C18:1 C18:2

M. hubbardi Alate 1 5. 03 19. 84 2. 44 1.60 5.15 55.14 10.54

Nymph, Old 2 4. 02 21.06 1.98 1.31 3,92 60.64 7.06

Nymph, Young 3 3.91 16.35 1. 81 1,53 4.19 63.31 8.84

Larva 1 4.91 14.19 1.98 1. 69 6. 24 63.61 7.37

Soldier 1 3 . 67 8.69 2. 27 3.20 8.43 63.11 10.77

R. flavipes Mixed castes 3 4. 94 2.31 11. 24 2.79 8.75 61.19 9. 86

aEach sample value is the mean of 2-4 replicates.

Group 1 fatty acids are the combined percentages for compounds with retention times less than C16:0. Group 2 fatty acids had retention times greater than C16:l and less than C18:0. There were only traces of compounds beyond C18:2, and these were not included in the table. 105

Table 43. Biological assay of Reticulitermes flavipes included as the protein source in a diet fed to mice for 3 weeks.

Dietary Protein Sources

Whole Egg Evaluation Technique (Control) R. flavipes

Third week body weight (g) 24.1 13.1

Gain in body weight (g) 15.4 4.4

Mean feed intake. „(g/d/mouse) 4.2 3.3

Apparent protein absorption 81.0 73.0

Protein efficiency ratio 2.50 1.1 Q Net protein retention 1.78 u.'7

First limiting essential amino acid none sulfur amino acids

Protein score 100 40

Essential Amino Acid Index 100 78 s Biological value 97. 3 73.3

aDetermined according to the Cr_0_ method of Schiirch, Lloyd, and Crampton (1950).

Sebrell (1963). P.E.R.'s adjusted to 2.50 for egg standard.

Determined by the method described by Bender and Doell (1957).

^Determined by the method described by User (195 9).

eCalculated by the formula: BV = 1.09 (EAAT) - 11.73 (Oser, 1959), 106 saguaros per acre. Since that time, however, mortality has greatly exceeded natality (principally because of bacterial necrosis), On plot F2 (Figure 9) more than 80% of the original plants have succumbed. Of the 44 dead skeletons inspected there during early July, 1976, only 7, all dead for 4 years or less, had not been attacked by M. hubbardi,

I have estimated that 13 skeletons contained active colonies. Nutting (1970) reported actual counts for 10 colonies. The mean number of termites per colony (.117 3.7) has been partitioned into castes and assigned energy equivalents according to the values presented in Tables 29-

31. Based on Nutting's data 14% of the termites should be alates awaiting dispersal flights. Thus, a single M. hubbardi colony would have a mean weight of ca. 6.00 g dry wt, and contain an equivalent of 39.51 kcal. The 13 colonies on plot F2 (10 acres) SNME would weigh 78 g and contain an energy equivalent of 513.63 kcal. The mean annual temperature for Tucson during the period 18 95 to

1960 (61 of the 65 years were recorded) was 19.65°C.

Daily wood consumption was calculated using estimates for large feeding groups at 22°C (Table 25). The 13 colonies on plot F2 would then consume ca. 1,557 kg dead wood annually (0.34 kg/ha/y; 153.23 kcal) and, through respiration, would dissipate 3094,14 kcal (680,7 kcal/ha),

Insufficient data were available to compute either pro­ duction or rejecta. These figures would suggest that more Figure 9. Map which shows the location in 1976 of all live and dead saguaros [Carnegiea gigantea (Engelm.) Britt, & Rose] on plot F2 of the Saguaro National Monument (East) near Tucson, Arizona — Original identifica­ tion numbers were assigned to each live plant on this 10-acre plot during 1941/42 (Alcorn and May, 1962). NW CORNER 4- 14 HE CORNER 13 19 137 13 '«

134 140

130 9 8• 10 129 6 23 '5« 7 m2, »9 12 107 108 2 "" |oo100 «« 3 . 26 103 S m 23 2* 102103(04 20 *27 123 99oo 133 » 118 97 29 ^93 327 94 122,21 3? 3435 »2 ^a7 135 48 36 90 B6 ..S3 49 <7 30 46 w 3.37

^ r)72 «> "£2 116 37 . «S M 470 „ « -"4 127 69 73 re 96 117 eo ^ 787 132 ft H + SW CORNER + SE CORNER Figure 9, Map which shows the location in 1976 of all live and dead saguaros jCarnegiea gigantea CEngelm.) Britt. & Rose] on plot F2 of the Saguaro National Monument (East) near Tucson, Arizona. o -j 108 than 4 times as much energy would be dissipated by respira­ tion than would be consumed (.680,71 vs. 153,23), Respiration energy appears to have been underestimated at low tempera­ tures and overestimated at high temperatures for the large feeding groups (Table 36), If this pattern holds true for the present situation, the gap between intake and respira­ tion energy for termites on SNME would be even greater,

Thus, it appears that wood consumption must be under­ estimated. DISCUSSION

Biochemical Studies on Dead Wood

The preceding laboratory and field observations strongly suggest that M. hubbardi can maintain itself and grow on saguaro wood that macroscopically appears free of decay. Since approximately a third of this study is con­ cerned with differences in the chemical constituents of aging wood, it is necessary to include some preliminary remarks about its structure and chemistry.

The term "wood" has been used rather loosely by many authorities. The reader should be aware that technically it is only the xylary portion of the vascular tissue of trees, shrubs, and, to a limited extent, herbaceous plants.

It is well known that cellulose, the basic energy source for all termites, forms a large part of sound woody tissues.

Curiously enough there is no general agreement as to its exact chemical nature. Green's (1963) definition has gained wide acceptance: "Cellulose is a naturally occurring poly­ meric fiber found in woody tissues composed primarily of

B-(l->-4)-linked D-glycopyranose units" (p. 127), Al­ though cotton fibers are almost "pure" cellulose (Cot6,

1968), a small number of non-glucose residues occur regularly in this native (naturally occurring) form. The

109 110 existence of a naturally occurring "true," or chemically pure cellulose composed of only glucose units is thus doubtful.

Unfortunately, chemically pure cellulose is rarely produced in the laboratory fractionation of woody tissues, and those cellulosic compounds which are obtained vary considerably according to the methodology and raw materials employed, Jermyn's (1955) admonition is worth repeating:

"It cannot be too strongly emphasized that cellulose and hemicellulose are normally determined as the result of certain sets of operations, rather than as chemically de­ fined species" (p. 283). A brief survey of the "operations" employed in analyzing woody tissue carbohydrates is pre­ sented here to clarify certain commonly encountered terms and to emphasize the necessity for accurately describing methods employed. As Jermyn notes, no single set of pro­ cedures will suffice for every situation and the analyst will often be required to develop or modify existing techniques.

The preparation of wood tissues for analysis in­ volves a careful selection of samples which will reflect the overall composition of the material under study. The optimum particle size to begin with has been debated

(Browning, 1952). While some investigators prefer rather large shavings, others utilize only those particles which will pass an 80/100 mesh sieve. The first chemical step Ill normally involves extraction of the raw material in organic solvents such as ether, alcohol, or alcohol'benzene.

Materials soluble in these solvents include fats, fatty acids, resins, resin acids, phytosterols, waxes, and hydro­ carbons (Browning, 1952). Succeeding extractions in hot and cold water remove inorganic salts, sugars, cyclitols, and various polysaccharides including gums, mucilages, starch, pectin, and galactans (Browning, 1952). The insoluble resi­ dues are subsequently de-lignified by the chlorite method

(Jermyn, 1955; Ritter and Barbour, 1935). The resultant material, "holocellulose," contains the bulk of the re­ maining cell-wall polysaccharides, with small amounts of lignin invariably included. Even though I was unable to carry my wood carbohydrate investigations beyond this point, it would benefit the reader to continue this discussion to its logical end, Holocellulose is separated by mild alkali extraction into insoluble alpha-cellulose, and soluble hemi- cellulose components. As in the previous step this separa­ tion is often not distinct and is greatly influenced by alkali concentration. The alpha-cellulose may contain other sugar residues such as xylose and mannose. Both the alpha- cellulose and hemiccllulose fractions may then be hydrolyzed and subjected to qualitative and quantitative analyses.

Finally, Jermyn (1955) cautions that although the terms

"hemicellulose," "xylan," and "pentosan" are sometimes used almost interchangeably, they are indeed different entities. 112

The xylans are uimply those polysaccharides which consist

primarily of xylose; the pentosans are likewise those made

up of 5-carbon pentose sugars. The hemicelluloses, as we

have seen, are defined by a set of extraction procedures

rather than the characterization of chemical species.

The basic approach to woody tissue analysis has pro­

gressed little during the last 20 years and, although much

of the methodology has become somewhat standardized, con­

fusion remains because some investigators have been

negligent in reporting their laboratory procedures and

careless with terminology. The necessity for paying strict

attention to details should be especially obvious to those

who study the organisms which habitually consume woody tissues. Although the composition of wood varies widely

according to species, age, tissue, and growing conditions, the data in Table 44 provide general values for the major

elements and organic compounds as they occur in decay^free

wood.

Analytical Determinations

Extractives, Wood Lipids, and Fatty Acids, Quanti­ tatively neither the ethanol/ethanol:benzene/hot water nor chloroform:methanol extracts were statistically different for the 3 saguaro groups (Tables 3 and 4). As one would expect, the compounds extracted by the first solvent system were almost double those removed by the second. It should 113

Table 44. Chemical elements and organic components in sound wood. s % Oven-dry Element or Component Weight Reference

C 49-50 Tsoumis, 1968, modified

H 6 Tsoumis, 1968, modified

0 44-45 Tsoumis, 1968, -modified

N 0.03-0.10 Cowling & Merrill, 1966

Ash 0.2-1.0 Tsoumis, 1968

Polysaccharides

a-Cellulose 40-50 C6te, 1968; Tsoumis, 1968

Hemicellulose (= xylans, mannans, etc. ) 15-30 Tsoumis, 1968, p. 61

Pectic substances <1. 0 Browning, 1952, p. 1189

Starch 0.5-5.0 Wise, 1952, p, 644

Gums and mucilages -

Lignin 15-35 Stecher, 1968, p. 619; C6t6, 1968

Protein 0.01-0,20 Cowling & Merrill, 1966

Extractives 1.0-10.0 Tsoumis, 1968 114

also be noted that the chloroform:methanol treatments were

much shorter (6 min) than those in ethanol/ethanol:benzene/

hot water (18 h). The latter solvents probably removed a

number of constituents in addition to lipids. The fact that

these compounds were not identified should not suggest that

they are unimportant, but rather than their relevance to

termite nutrition is as yet unknown or only suspected. Many

of these compounds are probably involved in modifying the

attractiveness or resistance of wood to termites and thus

its suitability as food. Although specific termite attrac-

tants such as vanillic acid, p-hydroxybenzoic acid, p-

coumaric acid, and protocatechuic acid have been isolated

from fungi (Becker, 1964), none has yet been reported from

the higher plants. Becker (1971) has reviewed the studies .

on substances found in wood which attract xylophagous

insects. For example, females of Hylotrupes bajulus (L.)

and another large cerambycid, Ergates faber (L.), are

attracted by the c^qH16 hydrocarbons (a~ an<^ 6-pinenes) ,

carene, and sabinene (Becker, 1971). It seems reasonable

to expect that similar examples will be found, particularly

of the sort which would lead to the establishment of dry-

wood termites in suitable dead-wood nesting sites.

Becker's (1971) review provided an exeollont litera­ ture base for further reading on the susceptibility of wood

to termite attack. In general natural termite resistance

is a function of many variables including toxic or repellent 115 chemicals, nutrient imbalance or lack of specific growth

factors, adverse physical characteristics, and pre­

conditioning. Termites are repelled or poisoned by a number of plant compounds, among them a steam-volatile oil isolated from Chrysocoma tenuifolia (.Berg) (Compositae) which possesses both contact and vapor toxicity for the harvester,

Hodotermes mossambicus (Hagen) (Hewitt and Nel, 1969a).

Saeki, Sumimoto, and Kondo (1973) have isolated chamae-

cynone, a termiticidal agent from the essential oil of

Japanese sawara wood, Chamaecyparis pisifera Don, Becker

(.1971) reports that stilbenes, quinones, pyran derivatives,

and furfural are among the many other compounds found in

plant tissues which are toxic or repellent to termites.

In addition to the B vitamins, plants contain certain water-soluble growth factors, which are now dis­ tinguished as lipogenic factors because they function pri­ marily in fatty acid metabolism rather than as enzyme co-

factors (Dadd, 1973). As such the exogenous requirements

for choline and inositol are far greater than for the B

vitamins. While only phytophagous insects have a demon­ strated need for inositol, Dadd speculates that dietary choline is probably essential for all insects. Fortunately for termites both myo-inositol (=mesoinositol) and its derivatives and choline are common in many types of plants.

Choline occurs in many phospholipids and functions as a phosphate carrier in plant sap (Robinson, 1963). It is 116 possible that Goetsch's "vitamin T," found in termites and other organisms, is actually a mixture of these growth sub­ stances with additional known vitamins (Stecher, 1968).

Buchner (1965, p. 819) has remarked that no attempt has been made to compare the vitamin content of different species of wood or areas within wood, and that such studies are critical to understanding the symbioses of certain wood- boring beetle larvae. Many other minor and trace compounds have been isolated from extracts of wood. Among those of probable significance in termite nutrition are vitamin A precursors such as B-carotene. Tsoumis (1968) states that total wood extractives normally vary between 1 and 10% of the oven-dry weight of wood. Many tropical species, how­ ever, contain extractives in excess of 30%,

The lipid content of wood (ether or chloroform: methanol extracts) is complex as well as variable. It con­ tains numerous fats, fatty acids, phytosterols, and wartes, all of which may influence wood-termite relationships, The fatty acids in tall oil, a resinous by-product from the manufacture of wood pulp, include: oleic, linoleic, lino- lenic, and palmitic (Harris, 1952, p. 611). Carter, Dinus, and Smythe (1972a) reported the fatty acid (FA) contents of

2:1 chloroform:methanol extractions of sound Pinus taeda L.

(loblolly pine), Pinus elliottii Engelm. var. elliottii

(slash) and Acer saccharum (Marsh.) (sugar maple). Oleic acid was the most abundant fatty acid from all 3 types of 117 wood, with concentrations of total FA present varying from

12.5 in maple to 45.4% in slash pine. The other fatty acids found in descending quantitative order were: lino- leic, palmitic, palmitoleic, stearic, and myristic.

Three fatty acids (palmitic, stearic, and oleic) in similar, but not quite equal, ratios made up practically all of the FA (78.38-84.84%) in the saguaro extracts (Table 5),

The 2 saturated compounds decreased in relative abundance in group C plants. Two unsaturated.fatty acids (oleic and linoleic) showed a commensurate increase in the C skeletons.

Although the reason for this is not entirely clear, it might have been caused by the presence of fungal tissues high in unsaturated FA (Carter et al., 1972a) in the older skeletons.

It is difficult to relate such slight lipid differences to nutrition. Indeed, an increase in polyunsaturated FA as in group C saguaros should enhance rather than retard termite growth.

Carbon Analysis. Wood contains about 49-50% carbon

(Table 44). If, however, other constituents such as nitro­ gen and ash are abundant, the percentage of carbon may appear lower. The carbon contents of the 3 saguaro groups were statistically similar between ca 45-46%.

Ash Analysis. The ash content of the 3 saguaro groups (Table 7) was not significantly different. It is doubtful that they have any effect on the presence of 118

colonies in the 3 groups, Inorganic minerals occur- in

varying amounts in plant tissues with quantitative and qualitative differences most closely related to soil type.

Wise (1952, p. 658) reports that 27 different minerals'have

been isolated from Pinus strobus L. Although the mineral requirements of insects are poorly known, it would appear that M. hubbardi has an ample supply of these nutrients at

its disposal.

Calorific Content. No difference was noted in the

calorific content of the saguaro groups (Table 8). Although the data in Table 9 indicate a significant difference be­ tween the rib interior and the outer weathered surface, the

difference, 0.2 kcal/g dry wt, seems too small to reflect a true biological difference. It is possible that energy-rich

compounds have been leached from the surface wood. Since termites generally nest in inner areas of the wood, this difference should not have been a factor in determining the

presence of M. hubbardi in the 3 saguaro groups, I know of

no other calorific values for saguaro wood. For comparison, determinations of gross energy (kcal/g dry wt) were per­

formed on additional desert woods consumed by M, hubbardi

and other termites found in the IBP sites, These included

O, spinosior (Englem, & Bigel.) (4.33), C. floridum Benth,

(4.30) , and Acacia Greggii Gray (4,55) , Since a number of

M. hubbardi incipient colonies were reared on birch (Betula 119 sp.) calorific content was measured and found to be 4.50 kcal/g dry wt.

McBrayer, Reichle, and Witkamp (1974) determined the ash-free calorific content of forest floor litter 5 times during the year. They found that winter litter held substantially less energy (3.9 vs. 4.3 kcal/g dry wt) than spring or summer detritus and concluded that leaching was responsible. Golley (1961) reported a mean calorific con­ tent for litter (4.3 kcal/g dry wt) and for standing-dead vegetation (4.1 kcal/g dry wt).

Nitrogen Content, Quantity/ and Distribution in

Wood. All living organisms require nitrogen for protein and nucleic acid synthesis. Since the nitrogenous compounds in wood are quantitatively minimal (Table 44), they are of more than casual interest in any consideration of the nutrition of xylophagous organisms.

The nitrogen content of plant materials varies tremendously. Herbaceous tissues normally contain 1-5% nitrogen by weight whereas woody tissues contain only 0.03-

0.10% (Table 45) (Cowling and Merrill, 1966). Levi,

Merrill, and Cowling (1968) reported carbon to nitrogen ratios for several plant materials: tomato foliage (10:1), tobacco stems (55:1), cotton seed hairs (200:1), micro­ biological agar medium (200:1). In the present study the ratio for saguaro was low, ca,120:1. 120

Table 45. Nitrogen content of stem wood of various gymnosperms and angiosperms — The data are for individual samples from a given tree (Cowling and Merrill, 1966),

N Content Tree Species (% by wt)

Gymnosperms Abies concolor 0. 045 Abies magnifica 0. 227 Juniperus virginiana 0.139 Larix occidentalis 0.180 Libocedrus decurrens 0. 097 Picea engelmanni 0.118 Pinus contorta 0. 071 Pinus echinata 0.130 Pinus elliottii 0.050 Pinus lambertiana 0.124 Pinus mSnticola 0,113 Pinus palustris 0. 038 Pinus ponderosa 0. 052 Pinus strobus 0. 087 Pinus taeda 0. 068 Pseudotsuga taxifolia 0. 051 Sequoia sempervirens 0.067 Taxodium d'istichum 0.057 Tsuga canadensis 0.106

Angiosperms Carya ovata 0.100 Castanea dentata 0.072 Juglans nigra 0.100 Liquidambar styraciflua 0.057 Liriodendron tu'lipifera 0. 088 Quercus alba ~~ 0.104 Quercus rubra 0.099 Quercus stfeilata 0. 096 Quercus vfelutina 0.070 121

Three factors contribute to the highly variable determinations which have been published on total nitrogen in woody tissues, These are the low nitrogen content, high carbon content, and extreme heterogeneity of wood materials.

Additional variation stems from the many modifications of the basic Kjeldahl process which have been used almost exclusively for these determinations. In general all

Kjeldahl procedures suffer the similar shortcoming of not

fully recovering nitrogen bound in N-N or N-0 linkages

(Rennie, 1965). Much of the nitrogen budgeting among termite societies should be re-examined due to the general

lack of uniform methodology and the current knowledge that

the N-0 linkage is common in several woody tissues, notably

among the Leguminosae,

Early biologists often speculated about the ability of termites to survive and reproduce on diets apparently containing very little or no nitrogen (Cleveland, 1925b).

The specific quantity of nitrogen in wood is most critical for the dry-wood species, After swarming, the reproductives establish their nest in a dead branch or log of limited size which serves not only for shelter but also as the sole source of nutrients. One would think that the composition of the nest material would change qualitatively as well as quantitatively, and eventually limit colony growth (Nagin,

1972). Although the relationship between available nitrogen and termite growth is not well understood, we do have data 122

on the distribution of this element in wood (Cowling and

Merrill, 1966) and some knowledge about the effects of

varying nitrogen levels on the development of two xylo-

phagous beetles, Hylotrupes bajulus (L.) (Becker, 1971) and

Anobium punctatum De Geer (Bletchly, 1969).

Nitrogen compounds are synthesized in the cambial

cells as structural proteins, enzymes, peptides, amino

acids, lipoprotein membranes (Cowling and Merrill, 1966),

and certain special hydroxypyroline-rich proteins deposited

in the primary walls of differentiating cambial cells

(Scurfield and Nicholls, 1970). These cells contain 1-5%

nitrogen dry wt (Cowling and Merrill, 1966). Protein found

in old or dead wood has often been attributed to dried

cytoplasmic remnants derived from differentiating cambial

cells (Wise, 1952), but Cowling and Merrill (1966) suggest

that little or none of the cytoplasmic nitrogen of these

cambial cells remains after lignification. Tissue nitrogen

drops rapidly as wood matures. Shortly after cambial cell

formation the gradual deposition of cellulose, lignin, and

other secondary cell-wall substances dilutes the nitrogen

content of the immature xylem from 5.0 to 0.5%. Lignifica­ tion is normally completed by the beginning of the second

stage of maturation (elution phase) 10 to 30 days after

cell division at the cambium. It is during this period

that the vascular elements (wood-fiber cells) die and lose

most of their nitrogen-rich cytoplasmic constituents to more 123

active tissues. Although experimental evidence is lacking,

Cowling and Merrill (1966) have theorized that the xylary cell contents undergo an autolytic depolymerization and are subsequently eluted upward in the transpiration stream.

This process presumably functions as an internal recycling mechanism whereby limited supplies of nitrogen and other nutrients are removed from essentially dead tissues and re­ distributed among the living tissues of the stems and crown.

By the end of the first growing season, and following cell division at the cambium, the nitrogen content of most woody tissues has dropped to 0.06-0.05%.

The longitudinal and ray-parenchyma cells of the sapwood live considerably longer and maintain a somewhat higher nitrogen concentration than other vascular elements of the xylem. The loss of nitrogen from these cells to the transpiration stream proceeds slowly through what is known as the "parenchyma death phase." The nitrogen content of most wood species has stabilized near 0.03% by the time the sapwood has transformed into heartwood (stable phase). In conifers a slightly higher nitrogen concentration is often found near the pith. Table 46 lists the nitrogen content of various parts of several species of stem wood, Despite considerable variation in the literature, Cowling and

Merrill (1966) concluded that a higher percentage of nitrogen exists: in angiosperms than in gymnosperms, within the crown of the tree than below it, in sapwood than in 124

Table 46. Nitrogen content of portions of the stem wood of various tree species (after Cowling and Merrill, 1966).

Nitrogen Content (% Dry Weight)

Immature Tree Species Cambium Sapwood Sapwood Heartwood

Gymnosperms Picea mariana 1.10 0. 056 0.059

Picea mariana 1,11 0. 27 0. 047 0. 062

Pinus sylvestris 3, 25 0. 012

Pinus 3. 33 0. 130

Corsican pine, butt 0. 108 0. 067

Corsican pine, crown 0. 085 0. 063

Pinus sylvestris, butt 0. 086 0. 079

Pinus sylvestris, crown 0. 052

Pinus sylvestris 0,115- 0. 047- 0. 031 0,078 0. 03 9 0.047 igiosperms

Eucalyptus regnans 2. 03 0. 52 0. 062. 0. 031

Fraxinus elatior 4, 59 0. 88 0. 22

Fraxinus sp. 4. 70 0. 89 0. 22

Ulmus sativa 4. 69 0. 81 0. 27

Ulmus sp. 4. 80 0. 83 0. 28 125 heartwood, nearer the cambium than close to the sapwood- heartwood interface, and nearer the pith (in conifers) than in the more recently formed heartwood.

The 3 saguaro groups had a mean nitrogen content

(0.37) more than 3.5 times that of the high value for wood shown in Table 44. Wood species which contain unusually high amounts of nitrogen may immediately be suspected of possessing large quantities of alkaloids (Wise, 1952). The i alkaloids comprise a fairly large heterogeneous group of compounds, all of which contain nitrogen, usually in a heterocyclic ring. Some of the more common alkaloids are: strychnine, quinine, morphine, cocaine, mescaline/ ephedrine, and nicotine. They are usually,.found primarily outside the xylem.

Although he did not determine their structures,

Caldwell (1966) confirmed the presence of 2 alkaloids in

3-year-old saguaro ribs. It seems probable that the high nitrogen content of saguaro is at least partially explained by the presence of these compounds. The question of whether or not they are available to termites or their symbiotes remains unanswered» Although they have received intensive chemical study because of their toxic properties, many of the alkaloids, poisonous to man, appear to have little effect on xylophagous organisms (Robinson, 1963).

Since nitrogen is an indispensible building block of biological tissues, it tends to remain in a constant state 126 of flux passing from producer to consumer to decomposer.

It is surprising that no quantitative differences were noted among the 3 groups. The high nitrogen level in old

(C) saguaro wood suggests that this valuable nutrient is

"locked up" for many years in a desert ecosystem where nitrogen may well limit production. Should this be true, the activities of M. hubbardi, which enhance nitrogen avail­ ability through their feeding habits, may be of great value to the community.

Amino Acid Profile. The quality of nitrogen is also an important factor in nutrition. Scurfield and

Nicholls (1970) have isolated and characterized the protein- bound amino acids (PBAA) in the wood of Eucalyptus sp. and

Pinus radiata D. Don, Results of these analyses showed that approximately 50% of the total nitrogen in P, radiata was recovered as amino or ammonia nitrogen and included all of the commonly encountered amino acids, Lincoln and Mulay

(.1929) found a similar situation in pear wood. There were, however, both quantitative and qualitative differences between the wood species tested. Many non-protein amino acids have been isolated from plant tissues, including gamma-amino butyric acid and beta-alanine (Robinson, 1963),

Similar data were obtained for saguaro wood.

Lignin Analysis. After cellulose, lignin is the most abundant compound in wood (Kirk, 1971). It comprises an 127 entire family of polymeric phenolic compounds which account for most of the methoxyl content of wood. They are resist­ ant to acids, readily oxidized, soluble in hot alkali and bisulfate, and condense with phenols and their compounds

(Schubert, 1965). Kirk and Harkin (1972) describe lignin as "an amorphous, three-dimensional, highly branched aromatic polymer" (p. 1) and suggest that one can'best understand its structure from a "biosynthetic viewpoint,"

As such it is composed of many phenylpropane (guaiacyl) subunits derived from hydroxycinnamyl alcohols. These undergo enzymatic oxidation to form free radicals which sub­ sequently couple with one another at the site of lignifica- tion (Kirk, 1971).

The resultant lignin is deposited in close associa­ tion with the cellulose of the' plant cell wall to form an intractable complex which has generally been thought to consist of mutually interpenetrating polymers held together only by physical attraction (Cowling, 1961) , Kirk and

Harkin (197 2) have suggested, however, that the association also involves a degree of chemical bonding between lignin and the wood polysaccharides of the cell wall. Regardless of the true nature of this association, it is clear that lignin strengthens the cell wall and protects the cellulose from attack by cellulolytic enzymes, For example, there are microorganisms such as Bacillus polymyxa (Praxmowski) Mac6 128 that readily attack isolated cellulose, but cannot utilize it in wood at all. Owing to the many possible linkages between the guaiacyl monomers and the variability among species, a definite structural formula for lignin cannot be given.

Approximately 30% of the guaiacyl monomers of softwood lignin appear to contain a free phenolic group while' 70% of the monomers possess carbonyl groups on one of the carbons of the propyl group. Benzyl ether or benzyl alcohol groups are found on 4 3% of the units. Linkages commonly found between lignin monomers include C-O-C and C-C. The texts by Schubert (19 65) and Marton (1966) provide a more complete discussion of the chemistry of lignin. As a group, the lignins are digested less efficiently by the termites and their symbiotes than the hemicelluloses or cellulose (Lee and Wood, 1971; Wolcott, 1946), Although the lignin content of group B saguaros was statistically lower than that of A and C (Table 12), the difference was small enough that it was not considered biologically significant. The overall percentage of lignin in saguaro, a hardwood, falls well within the range given in Table 44.

Summary of Wood Analyses. With minor exceptions/ the chemical composition of the 3 saguaro groups was re­ markably similar. On the basis of nutritional constituents assayed, I have concluded that they should maintain M, 129 hubbardi equally well. As they do not, however, undeter­ mined nutritional components such as those discussed under extractives, environmental parameters, or perhaps behavioral factors must determine the nutritional suitability of saguaro wood. The possible effects of wood malnutrition were emphasized in the following experiment. In an attempt to attain a homogeneous diet, birch dowels were provided as food for incipient colonies of M. hubbardi. Although this wood was readily consumed and initial colony growth appeared adequate, birch was not as good a nutrient source as saguaro. After 3 years and 1 month the young colonies fed birch had consumed an average of 1.59 g dry wt while those fed saguaro consumed 6.21 g. On birch only 24 of 500 sur­ vived (6.8%) as opposed to 16 of 41 (39%) on saguaro. The mean weight of birch-fed colonies was 47,62 mg versus 202,99 for the saguaro-fed termites. The former groups also had fewer individuals (X = 4.7) than the latter (X = 20,13).

Most interesting, however, was the observation that while the coloniep on birch produced no nymphs whatsoever, those on saguaro contained 36.7% of this pre-alate caste. In addition, one of the colonies fed saguaro produced 2 alates during the relative short growth period, The only chemical analyses done on birch wood were per cent nitrogen (0,07) and calorific content (4,55 kcal/g dry wt). The calorific values of birch and saguaro were essentially similar al­ though birch contained much less nitrogen. The nitrogen 130 levels among the saguaro groups were not different, Al­ though it has been known for some time that both the quantity and quality of nitrogen compounds affect insect growth rates, I did not assess possible qualitative dif­ ferences, Xylophagous organisms consume cellulosic materials which are invariably low in nitrogen. As a result, many of them have evolved special mechanisms to enhance their nitrogen-accumulating capabilities, Cowling and Merrill

(1966) presented data and speculation on the role of nitro­ gen in wood decay. They concluded that wood-destroying fungi must conserve nitrogen through 3 possible mechanisms: re-utilization of tissue nitrogen by autolysis of less active fungal tissues, atmospheric nitrogen fixation, and physiological adaptations that result in preferential allocation of available nitrogen to metabolically active substances and pathways that are high efficient in the utilization of wood constituents, It seems likely that one or more of these mechanisms must be used by termites. We know that nitrogen fixation occurs in the termite gut due to the presence of certain bacterial species (Breznak et al., 1973). We also know that the growth rates of termites and beetles are retarded by low nitrogen levels in wood

(Hungate, 1941; Bletchly, 1969). Becker (1971) discussed the role of amino acid ratios relative to the growth of xylophagous organisms. He noted that change in the amino acids during normal aging of wood may lead to natural 131 resistance against H. bajulus, The addition of suitable amino acids to old wood made them once more susceptible to attack.

Although environmental parameters may have in­ fluenced the suitability of saguaro skeletons for M. hubbardi, it is probable that they would be important only in very old C skeletons which have been reduced in size so as to afford no further protection from temperature and moisture extremes. No data were gathered to substantiate this speculation, however.

There is some reason to implicate the nest site exploration strategy of M, hubbardi with its avoidance of

A skeletons. Earlier studies (Wilkinson, 1962; Usher, 1974) suggest that some kalotermitids and termitids preferentially select cracks for entry into wood. Other species, C, brevis

(Coaton, 1948) and I_. minor (Harvey, 1934) , tend to excavate entrance holes on smooth surfaces.

Based on several casual and one documented observa­ tion, I submit that M. hubbardi alates employ a combination of the above strategies for selecting their nesting sites.

It has already been stated that group A saguaros are generally covered by intact but dry cortex over much of their surfaces. Saguaro No. 90, plot F2, SNME, died from bacterial necrosis during 1971, On close inspection 3 M. hubbardi entrance holes were found ca. 15 cm below the upper margin of the dried cortex and ca. 46 cm above ground level. 132

It appeared that alates entered the space between the cortex

and skeleton, worked their way downward and under the pro­ tection of the cortex began the process of excavating the entrance hole. Such a strategy would at least partially

explain the absence of termites from very young skeletons which are more fully covered by cortex. Moreover, very old

skeletons which lack cortex completely would not provide any

protection during the critical excavation period, M,

hubbardi flights occur during the. early evening hours from

late June through early September. During this period a

vast number of insect and vertebrate predators are abundant

and active.

Although the maximum longevity of kalotermitid colonies is not known with certainty (Wilson, 1971), Grasse

(1949) speculated that dry-wood colonies generally die after

10-15 years as a result of declining egg production. Field observations tend to substantiate this theory. Moreover, it

appears that M. hubbardi colonies are in fact rather short­

lived. No skeleton on plot F2 of SNME had been attacked

until it had been dead at least 4 years. These data, though

not conclusive, suggest that M, hubbard1 colonies may live as few as 2-5 years. Under favorable laboratory conditions

M. hubbardi colonies matured, i.e., produced new alates, in ca,3 y. Regardless of age, field-collcctcd M. hubbardi colonies contained more than 95% nymphal stages. This

suggests a tendency to produce reproductive forms very 133 quickly which are capable of dispersing and forming new colonies. Nagin (1972) speculated that kalotermitid colonies might die as a result of failing to produce supple­ mentary reproductives and declining egg production potential when colonies enter a physiological state associated with maturity, Although it was not entirely clear whether these physiological changes were endogenous or exogenous,'the net result was the appearance of large numbers of pre-alate forms, i.e., nymphs. A parallel phenomenon is known in certain flowering plants which, when faced with potentially lethal environmental stress, allocate much of their metabolic activity to seed production.

Nutritional Physiology

Feeding Trials

Becker (1969), Gay et al. (.1955), and Haverty (1974) are but a few who have reviewed the methods for testing termites in the laboratory. It is particularly disturbing that in spite of their recommendations, no uniform method­ ology or standards for expressing results have been adhered to. Haverty (1974) reported that a large number of variables can affect wood-consumption rates. Among these he noted wood species and hardness, toxic substances and other feeding inhibitors or deterrents, presence or absence of fungi and degree of decay, moisture content, and tempera­ ture. In the present study, group size, temperature, and 134 colony origin were the variables used to test termite response.

Few authors have recognized the effect of group size on termite response. Nel, Hewitt, and Joubert (1971) found that the relationship between brood production (P ) and colony size (50-600 individuals) of H. mossambicus (Hagen) was related in a linear fashion during the first 116' d.

After 168 d, however, groups containing more than 200 termites showed a negative curvilinear response due, in the author's opinion, to competition for space. McMahan (1962) and others have noted changes in caste ratios and oviposi- tion rates as incipient kalotermitid colonies increased in size. Gay et al. (1955) reported no significant differences in survival among laboratory test groups of 2500, 5000, and

7500 workers of Nasutitermes exitiosus (Hill). None of the above, however, evaluated wood-consumption rate as a function of colony size.

The observed patterns of growth (A biomass) for the

hubbardi feeding groups were consistent with Nagin's

(1972) concept of differing physiological states in developing colonies, According to his theory, very young colonies allocate energy reserves to the production of large larval populations which tend to accumulate biomass very quickly. Mature colonies, on the other hand, show Icbb Ox no growth (P or P ) as they mark time awaiting the proper environmental cues for dispersal flights. The observation 135

that the larger intermediate and large sized feeding groups

showed little or no production while maintaining large

nymphal populations is consistent with Nagin's reasoning.

McNeill and Lawton (.1970) examined the relationship between

production (P) and annual respiration (R) (kcal/m 2/y) for

53 animal species. Their regression analyses demonstrated

convincingly that long-lived poikiloterms that maintain

large populations of old individuals (>2y) generally expe­

rience one or more high respiratory-cost, non-productive

periods. Consequently, they display low annual production

efficiency, i.e., P/R ratios.

According to Nutting (1969) the initial rate of

population growth among the kalotermitids is invariably

slow. At the end of the first year such colonies generally

contain fewer than 60 larvae and nymphs and 0-3 soldiers.

To my knowledge comparable data are not available for 3- year-old colonies.

McMahan (1966) recognized that past feeding expe­ rience might influence the selection of habitat, survival,

and offspring production by dispersing alates of C. brevis

(Walker). More recently Honigberg (1970) reported that gut

faunules of the same termite species may differ in composi­ tion and metabolic patterns. For these reasons I examined

colony origin as a factor in the feeding trials of the

intermediate sized groups. No significant differences were

noted other than the initial mean wt (Table 16) and the 136

accumulation of dead bodies (Table 14). The absence of additional differences should be acceptable with caution

in that the colonies from which the termites were obtained were collected within a mile of each other and thus probably represent a single breeding population. As noted in the introduction, M. hubbardi enjoys an extensive range from

California and Arizona south to Colima, Mexico.

Surprisingly, I was not able to detect a significant temperature effect on M. hubbardi wood-consumption rates.

This was especially disturbing as very strong relationships between temperature and oxygen consumption were noted (Table

(Tables 26 and 27). Becker (1969) noted a similar feeding response by Heterotermes indicola (Wasmann) which maintained

- \j . essentially constant food intake over the temperature range

20-32°C. Nevertheless, additional experiments encompassing wider temperature ranges and more replicates must be per­ formed before it can be concluded that M. hubbardi is in­ sensitive to changing temperatures. Although there were too few replicates for a statistical analysis, the rate of wood consumption did appear to respond to temperature for the 6 large groups (Table 25), Temperature did not, how­ ever, affect survival, productivity (A biomass), fecal- pellet production, assimilation efficiency, carton forma­ tion, or caste composition. Egg production was restricted to the 28 and 32°C groups. 137

The most pronounced differences in response were due to the groups' sizes, The intermediate sized groups (5, 10,

20, 50 individuals) were influenced by this variable in practically every response category (Tables 17-23). The following generalizations apply to the 3 feeding experiments collectively. Although similar in number, the incipient and smaller intermediate sized groups (5, 10) differed greatly in ultimate caste composition. Whereas the former began as pairs of primary reproductives capable of immediate egg production (82% were alive after 3 years), the latter were established only with larvae, nymphs, and soldiers. Any reproductives which appeared arose gradually as replace­ ments (SR) from larvae or nymphs. During the relatively short duration of these trials (90 d) it is doubtful that the smaller intermediate groups (5, 10) had enough time to produce both SR and brood. On the other hand, groups of 20 and 50 were in a better position to do this, as their larger foraging populations allowed newly formed SR to devote more of their energy to production. This speculation is at least partially substantiated by the data in Table 21 which show that although the mean number of SR per group divided by the initial group size was statistically larger for groups of 5 and 10 than for the 20 and 50, the smaller

p groups produced significantly fewer offspring ( r)» In terms of colony maturity, I speculate that the 20- and 50- termite groups correspond most closely to natural colonies 138

2 to 5 years of age in which one or both of the primary reproductives have been replaced by SR,

The ratio of soldiers to non-soldiers was lowest

(2,83%) in the large groups and highest (8,69%) in the

incipient colonies. The mean percentage for the inter­ mediate sized groups was 4,42, The groups of 20 and 50 did

not differ statistically with respect to soldier numbers

which, as one might expect, were higher than in the smaller

groups (Table 21). Although the proportion and occurrence

of a particular caste may achieve a reasonably constant,

species-specific ratio in mature colonies (Haverty, Nutting,

and La Fage, 1974), it is doubtful that the same ratio would

apply to incipient or developing colonies. The 1 or 2

soldiers which are produced in the first M. hubbardi brood

appear to meet the colony's needs for the first 3-5 years or

until the population exceeds 50-60 individuals. Nutting

(1969) found that ratios of soldiers in 10 M. hubbardi

colonies generally varied from 1:13 in a colony of 68 indi­

viduals to 1:68 in one of 1477. There were inconsistencies

in this trend, however, as one group containing 312 had a soldier to non-soldier ratio of 1:20 and a larger group

(450) a ratio of 1:7, In his largest group (3119) the

soldier percentage was about one-half that of the 500- termite groups in this experiment (1.50% vs. 2.83%). Large

groups appear to require fewer soldiers than small groups

do. 139

With the possible exception of predation by ants, M.

hubbardi is well protected within its woody nest and thus

probably requires the services of a rather small soldier

population. This is in contrast with the surface feeding

termite Tenuirostritermes tenuirostris (Desneux) in which

foragers are accompanied by a great number of nasutes.

As expected, larvae constituted the largest 'per­

centage (46.24) of the survivors in the incipient colonies

and the lowest (3.10) in the large groups. This is also

consistent with Nagin's (1972) contention that mature

colonies develop and maintain a large proportion of sub-

imaginal forms at the expense of the younger, undiffer­

entiated castes. A relatively constant proportion of larvae

(8.58%) was noted among the intermediate sized groups.

Nymphs made up 36% of the incipient groups, 67% of

the intermediate groups, and 92% of the large groups, Again

this substantiates my contention that M, hubbardi colonies

rapidly attain a "waiting" posture in anticipation of

dispersal flights.

The incipient colonies, as expected, and again

consistent with Nagin's (197 2) reasoning, had the highest

production-to-respiration ratio (0.0192). The intermediate

and larger groups did not respond as expected in that the

mean P/R value was lower for the former (0.0034) than for the latter (0.0134). Due to the many differences among the

3 experimental groups, P/R ratio3 were not examined 140

statistically. However, a general productivity pattern was

recognized. Incipient colonies showed both reproductive

(P ) and growth (P ) productivity. Intermediate groups

produced more eggs but less overall biomass than the others.

While the large groups laid very few eggs, they accumulated

the greatest biomass.

Wood-consumption rates by the different groups were

of considerable predictive interest as they were strongly

correlated with biomass days (BT) and, to a much lesser

extent, temperature. In the development of multiple linear regression equations (Table 23), the only variables included

in the final prediction model were those which were shown by

orthogonal contrasts to increase R 2's at least 5%. The

first equation in this table explains 96% of the observed

variation in wood consumption. While more than 80% of that

variation was explained by the linear component of biomass days, the temperature component explained only 5.5%, It was

included only because it was easy to measure and met the

minimum requirement for acceptance.

Because they produce little or no cellulase,' most

metazoan soil organisms are generally inefficient processors

of woody detritus. This is also true of many herbivorous insects such as the larvae of Lepidoptera. In contrast a few animals, including the silverfish, Ctenolepisma lineata,

land snails of the genus Helix, ruminants, and termites are

surprisingly efficient cellulose assimilators. Lasker and 141

Giese (1956) reported that C. lineata absorbed 71.7-87,0% of a pure cellulose diet. The literature reports on assimila­ tion efficiency by laboratory colonies of termites have been summarized by Wood (in press), For the termite species and four species of wood listed efficiencies vary from 54-61%

(dry wt). This may be compared with a dairy cow (a ruminant) which digested 72% of a ration of dry grass (Maynard, 1937).

M, hubbardi apparent assimilation efficiency was highest among the intermediate groups (64-87%) and lowest among the groups of 500 (53%). C-FU Assimilation efficiency (—-—) was best estimated by the linear (54.7%) and quadratic (25.2%) components of biomass days. Temperature was not a significant factor based on the above criterion. Prediction models were developed only for the intermediate sized groups. Based on the equation for wood consumption (Table 23) an M. hubbardi colony containing 1000 individuals, each weighing ca,12 mg, would consume only 17 8.58 g of wood annually if maintained at a constant 28°C (Table 25) •

As it is unlikely that M. hubbardi colonies reach these sizes in fewer than 5 years [Nutting (1969) found a large M. hubbardi colony of 3119 individuals in a cottonwood fence-post set in the soil no earlier than

1960] it would appear that M. hubbardi poses little real threat to wood in dwellings unless many colonies 142 are present. Chances are that it can be dealt with best by replacing infested structures if accessible.

Respiration Studies

In contrast with the response variables just dis­ cussed, the respiratory rate of M. hubbardi was highly correlated with temperature, Oxygen-consumption rates for old larvae, nymphs, and soldiers were statistically similar for groups of 5 and 10 individuals tested in the Gilson® differential respirometer. Consequently, group size (and biomass) was eliminated as a factor in statistical analyses.

Oxygen consumption and CO2 evolution were measured at 4°C intervals from 16-36°C. At 16°C M, hubbardi appears to assume a state of respiratory torpidation (an average of only 0,083 yl/ml termite (fresh)/h). The lowest 02 con­ sumption (0.060 vil/mg/h) for any of the 4 physiological states was recorded at 16°C for the defaunated/starving groups, The 16°C replicates were at times greatly in­ fluenced by minor changes in the room environment especially when the air-conditioning was shut off daily at 2100 h.

Even though reaction vessels were submerged in a constant temperature water bath, gas exchange measurements, monitored by a micrometer device at a level several cm above the water bath, were affected by changes in room temperature.

Measurements of very low respiratory rates were inaccurate.

This is substantiated by the rather erratic RQ's reported 143 for 16°C groups in Table 26, especially for the defaunated/ starving groups which could not possibly have attained an

RQ of 1.28. The remaining treatment combinations followed the expected patterns. The treatment with the highest rates of C>2 consumption (normally faunated/feeding) had RQ's near

1,00, suggesting that pure carbohydrate was being metab­ olized. Keister and Buck (1974) compiled a large table of

C>2 uptake for insects. These values, expressed on a fresh wt basis, ranged from a low of 0.037 yl/mg/h for Crypto- cercus punctulatus Scudder to a high of 22,1 for actively feeding G. mellonella larvae. At 16°C the defaunated/ starving groups actually consumed less C^/h than many diapausing insects listed by the same authors. Several literature reports of (^-consumption rates by termites are presented in Table 47.

The (^-consumption data for normally faunated/ feeding M. hubbardi are comparable to those in Table 46.

They are, however, low by comparison with other insects.

Wiegert and Coleman (197 0) concluded that termites have low rates as physiological adaptations to conditions of crowding.

Also if alates did not feed during the initial stages of colony development, a low, fasting metabolic rate would be advantageous to conserve energy. The normally faunated/ feeding tormifcia consumed more (X = 0.4 53 lil/mg/h) nnd had a higher RQ (ca. 1.0) than any other treatment. The mean rates for the starving treatments wore 0.223 iil/mg/h Table 47. Summary of literature reports of oxygen consumption by termites (yl/mg termite [£resh]/h) — Modified from Hebrant (1970).

Temp, °2 Species Caste (°C) Consumption Reference

Kalotermes flavicollis — 26 0.54 Luscher (1956)

Zootermopsis — 32 0. 81 Cook (1932) nevadensis 20 0.42 Cook (1932) 19 0.51 Luscher (1956) 23-25 1,01 Hungate (1938) 3 0.037 Gilmour (1940)

Z. angusticollis — 12-32 .140-.655 Cook and Smith (1942)

Reticulitermes worker 20 0,182 Ghidini (1939) lucifugus unpigmented imago 20 0.173 Ghidini (1939) soldier 20 0.057 Ghidini (1939) 1st stage nymph 20 0.086 Ghidini (1939)

R, santonensis 24-30 .62-.99 Damaschke and Becker (1964)

R. flavipes 24-32 .53-1.24 Damaschke and Becker (1964)

Heterotermes indicola 24-32 .64-.89 Damaschke and Becker (1964)

Nasutitermes 1 0 • • 00 to ephratae 22-30 tn Damaschke and Becker (1964)

N. costalis 21-25 0.6 Wiegert and Coleman (1970) Table 47.—Continued

Temp. °2 . Species Caste (°C) Consumption Reference

Macrotermes bellicosus large worker 30 0.29 Ruelie (1962) small worker 30 0.28 Ruelie (1962) large soldier 30 0.17 Ruelie (1962)

M. ukuzii 28 .11-.14 Rohrmann (1975) 146

(faunated) and 0.215 yl/mg/h (defaunated) (Table 26), While these values differed statistically, it is doubtful if the difference was biologically significant. Since starving termites (and most other animals) rely on stored fats for energy, it is unlikely that the presence of starving proto­ zoans in the gut would affect overall oxygen consumption.

This conclusion is especially warranted in view of the evidence which suggests that these protozoans are anaerobic

(Honigberg, 1970).

The most interesting result of this experiment con­ cerns the defaunated/feeding groups which, though lacking their protozoans, consumed oxygen at a rate approximately intermediate between the normally faunated/feeding and starving individuals. The observed 02 consumption dif­ ferences between the normally faunated/feeding and de­ faunated/feeding treatments could thus not be directly due to the loss of protozoans, The difference resulted from lower levels of available nutrients in the gut as a result of losing the protozoans. The important observation here, however, is that termites assimilate a respectable quantity of wood constituents on their own without the aid of their symbiotes. It should be noted that although most of the gut microfauna was removed by the 45 psi O^ defaunation treat­ ments, surviving species may have digested some cellulose.

Pew efforts have been made to assess the cellulo- lytic capability of the microflora of the hindgut, or of 147 completely defaunated and deflorated termites. Brief reviews of these attempts have been given by McBee (1959) and Mannesmann (1972b). Work which suggests that gut bacteria contribute to their host's digestive capacity in­ cludes that of Beckwith and Rose (1929), Mannesmann (1969),

Ritter (1955), and Tetrault and Weis (1937). Cleveland

(1928), Dickman (1931), Honigberg (1970), and Hungate (1936,

1955) all argued that, although cellulolytic gut bacteria may in fact be present, they are insignificant in the overall process of providing energy for their host,

Mannesmann (1972b) recently demonstrated cellulolytic bacteria in the guts of two rhinotermitids, Reticulitermes virginicus (Banks) and Coptotermes formosanus Shiraki. The slow rate at which these bacteria attacked filter paper, however, suggested that their cellulolytic importance was minor. Tetrault and Weis (1937) demonstrated that bacteria from the gut of R. flavipes also are capable of degrading cellulose.

There are, however, some data which suggest termites have the ability to produce their own digestive enzymes,

Cleveland (1924) noted that, following defaunation, workers of R, flavipes died within 10-20 d when fed sound wood. On a diet of somewhat decayed wood, however, their survival was prolonged and, when fed dextrose (glucose), peptone, and starch, either separately or together, defaunated indi­ viduals survived considerably longer. These observations 148 suggested that the defaunated termites were indeed capable of digesting and absorbing certain nutrients and, thus, must possess digestive enzymes of their own,

Hungate (1938) quantified the relative importance of protozoans and host in digestion and utilization of wood by

Zootermopsis angusticollis, Extracts from fore-, mid-, and hindguts were incubated with a substrate prepared from pre­ cipitated filter paper which contained cellulose and small amounts of other carbohydrates. Reducing substances were found in these preparations after 16 h of incubation at room temperature: slight accumulation in the fore- and midgut, relatively large amounts in the hindgut preparations.

Similar experiments uping more highly purified cellulose, however, revealed no reducing substances in either the fore- or midgut regions, and considerably less in the hindgut. This suggested that, although cellulose was attacked only by microbes, enzymes other than cellulase were present throughout the gut which could hydrolyze soluble carbohydrates. From additional experiments in­ volving defaunated termites Hungate concluded that enzymes produced in the wall of the gut itself could account for about one third of the digestion of normal termites. Al­ though Zootermopsis might be able to oxidize about one third of its total wood intake independently of its symbiotes,

Hungate (1938) showed that it would have to oxidize ca. seven eighths of its intake to satisfy its energy 149

requirement. These data were based on sawdust of "fairly

sound" Monterey pine as the nutrient source and it is

probable that assimilation efficiency would vary according

to wood species and level of decay.

Cook (1943) fed several different carbohydrates to

normally faunated and defaunated groups of Z_. angusticollis.

Those fed wood or cellulose showed RQ1s between 0.9'and 1,0,

typical of carbohydrate utilization; however, post-

absorptive RQ's fell to between 0,70-0.75, suggesting that

lipid reserves were being metabolized. The participation of

protozoans in metabolism, signaled by the evolution of

hydrogen, was noted on all diets except starch, Defaunated termites showed RQ's near 1.0 when fed starch, sucrose,

maltose, lactose, glucose, fructose, or galactose. Since no

hydrogen was evolved, Cook concluded that this termite must

possess enzymes to digest, absorb, and oxidize several

carbohydrates independently of its symbiotic microfauna.

Among other attempts to feed termites non-cellulosic

diets are those of Montalenti (1927) and Lund (1930).

Montalenti was able to maintain K. flavicollis for several

months on soluble starch, either alone or mixed with glucose. During this period the number of gut symbiotes declined rapidly, Lund on the other hand found that, when carbohydrates other than cellulose were fed to Zootermopsis

(as TurmopB.I.B) sp. , the termites generally died shortly after the disappearance of the last protozoans. In general 150

his experimental diets enabled termites to live no longer

than starved controls. Dextrose prolonged the life of the

defaunated termites for approximately 40 days although

bacteria became numerous during this period. Although early

investigators had recognized the presence of protozoans in

the termite gut, Cleveland (1923, 1926) was the first to

present experimental evidence that the protozoans actually

contributed to the digestion of wood. He observed that,

when the protozoans were killed, either by starvation, in­

cubation at elevated temperatures, or by increased oxygen,

tension, the termites lost their ability to utilize wood as

a nutrient. When, however, the same termites were re-

faunated by proctodeal feeding they regained their former

capacity. Trager (1932) was the first to demonstrate a

cellulase from the intestinal flagellates of the wood-

feeding cockroach, Cryptocercus punctulatus Scudder,

flavipes, and angusticollis. He maintained one of the

cellulolytic flagellates, Trichomonas termopsidis Cleveland,

for 3 years, a feat yet to be equalled. Honigberg (1970)

discusses Trager's work in detail along with several other

less successful culture attempts,

Laskcr and Giese (1956) point out that the mere

presence of cellulolytic organisms in the gut of cellulose- consuming animals does not necessarily signify that they

benefit the host. It is possible that other microbes are contributing to the host's well-being and/or that the host 151

is partially self sufficient. Other data which support the

independent elaboration of cellulolytic enzymes by termites

come from the family Termitidae which do not possess

symbiotic protozoa. An adequate discussion of the mechanisms

involved here would require more space than can be reason­

ably allocated. The reader is directed to the review by

La Fage and Nutting (in press) for further information.

As a result of the present study I have concluded

that M. hubbardi can assimilate a significant amount of saguaro constituents without its gut fauna. It remains un­ known, however, whether a termite-produced cellulase is elaborated.

The observed 02 consumption data were transformed to l°ge and regressed on temperature (°C). The resulting

prediction models (Table 27) were useful in energy

budgeting. Over a similar temperature range both feeding groups responded in a curvilinear fashion while the starving groups followed a simple linear pattern. No biological significance is assigned to these differing responses, however.

Termite Thermocfenes is. Table 28 contains estimates of M. hubbardi heat production obtained by calorimetry and by calculations based on 02 consumption predicted from the equations in Table 27 and the oxy-caloric equivalent of

5.05 kcal/L • Discrepancies between the calorimetric 152

and respiration estimates can be explained only partially.

Although they are anaerobic, the protozoans must, as a

result of their metabolic processes, give off heat which

cannot be estimated on the basis of oxygen consumption. One

would expect and, indeed, does observe, that heat production

estimates by calorimetry are higher than those calculated

from respirometry, One other study has used both of these

methods to estimate heat production by insects (Peakin,

1973). Although he found that the 2 methods produced

similar results, his experimental insect (Tenebrio molitor

L.) contains few if any anaerobic symbiotes. The termite by

contrast may contain as much as one third of its total fresh

biomass in protozoans (Katzin and Kirby, 1939). Only one

person has used calorimetry extensively in the study of

insect thermogenesis. The now retired Professor H. Prat of

Marseille, France, has, through personal communication, re­

affirmed by belief that calorimetry should discern heat

production by termites more efficiently than classical

respirometry (Prat, 1975).

The thermogram in Figures 8a and 8b are parts of a

12-h record produced by a single M. hubbardi nymph. The

thermogenetic event pictured occurred repeatedly, though

irregularly during several but not all runs on single

termites. It is possible and probable that the same event

occurred in the group runs but, because these thermograms

were composites of 5 individuals, individual events were

i I 153 difficult to resolve, A possible explanation of the phenomenon pictured might involve a mode of discontinuous respiration in which a sudden burst of moisture-laden respiratory gases escape and cool the ampoule wall. The sudden increase and subsequent leveling off of the heat record remain unexplained.

Energy Budgets for Laboratory Feeding Groups

Energy budgets developed for the different sized feeding groups will be discussed collectively.

Since heat production was measured only at 20°C, C^- consumption measurements (Tables 26 and 27) were used to estimate R for all energy budgets. As R was the only variable not measured directly, it is probable that it was responsible for unbalanced budgets (Tables 32, 35, and 36).

On the positive side, none of the budgets was off by more than a factor of 2. R was overestimated for the incipient and large groups while underestimated from the intermediate category. As biomass days were calculated with higher accuracy for the inter mediate sized groups than the others,

R estimates for the mid-sized groups were probably the most realistic in spite of the fact that energy budgets for these groups do not balance as well as those for the others.

The components of productivity (P) used in energy budgeting have been noted for the intermediate sized groups in

Table 24. 154

Trophic Level Interactions

Lower Level Interactions

M. hubbardi fecal pellets represent a rather poor quality nutrient for lower trophic levels in that they are

low in nitrogen (0.203%) and high in lignin (61.63%), Al­ though it was not confirmed analytically, the remainder is

probably cellulosic in nature. Fecal pellets are probably nutritionally suitable only for a few wood-destroying fungi

(white rots) and specialized bacteria. Nevertheless, pellets are small enough that abiotic degradation may pro­ ceed rapidly as a result of the large ratio of surface area to volume. My observations indicate that M. hubbardi does

not practice coprophagy as do some millipeds to obtain

adequate energy from their feed (McBrayer, 197 3).

Predation

Many animals consume termites as the opportunity arises. The data presented in the results section under the heading "Predation" indicate that termites are indeed

nutritious prey. Nutting (1969) reviewed the available literature on the kinds of animals which eat termites. M.

hubbardi is high in nitrogen ana, depending on caste, fat and thus energy. The larvae and young nymphs contain more than 75% moisture and therefore could provide occasional water for desert animals. Although M, hubbardi larvae and nymphs are generally unavailable as prey as a result of 155

their protective nest other termits such asv Gnathamitermes

perplexus (Banks) which contain more water and forage near

the surface are habitually eaten by birds and reptiles.

Chemical Analyses. Moisture content tends to be

highest in non-reproductive castes, i.e., larvae, workers,

and soldiers. In M. hubbardi moisture content is a function

of fat deposition with the possible exception that

different-aged alates possess differing quantities of water

while maintaining approximately the same fat content. Newly

emerged, unflown alates (pre-dispersal) (Table 30) contained

more moisture (72.46%) than older unflown (65.17%) or flown

individuals (61.34%). Water lost during the preflight

maturation period would tend to lighten subsequent flight

loads and consequently aid in the conservation of energy

required later for brood production.

The ash content of all M. hubbardi castes averaged

4.86% (Table 38). Although no statistical analysis was done to determine differences among the castes, the range (4,27-

6.19) was considered narrow. Humus-feeding or soil-

ingesting termites are the only species which have a high

ash content. This results from soil in the gut. Matsumoto

(1976) found that workers of Dicuspiditermes nemorosus

(Haviland), a humus-feeding species, contained as much as

65.6% ash, dry wt. 156

The calorific content of the individual castes generally reflects their lipid content. In M. hubbardi the alates and nymphs have the highest energy content (6.41 and

6.56 kcal/g dry wt). Although high for insects, these values are much lower than the mean (7,79 kcal/g dry wt) for

Very fat .fall migrant birds reported by Odum, Marshall, and

Marples (1965). Nevertheless M. hubbardi alates, the caste most often preyed upon, are definitely energy-rich. The carbon content of M. hubbardi castes follows the same relative pattern as calorific content and total lipid con­ tent and is, in my opinion, a reflection of the higher carbon content of lipid-rich compounds.

The nitrogen content of M. hubbardi (Table 40) is close to that reported for other termite species (Table 48).

The amino acids reported for the various castes of M. hubbardi (Table 1) are present in essentially similar ratios among all castes tested. The results of determinations on

R. flavipes also given in this table suggest an extra­ ordinary resemblance in the amino acid profiles of the two species.

The whole-body lipid content and constituent fatty acids of M. hubbardi and R. flavipes presented in Tables 41 and 42 represent termite body lipids plus the small quanti­ ties held in the gut contents. Lipids tend to accumulate in those individuals destined to participate in dispersal flights. Owing to their very high energy-to-mass ratio, but Table 48. Nitrogen content of termites.

% N Species Sexa Caste (dwb) Author

Pteroterxues occidentis M Alate 9.94 La Fage (1975)

Pterotermss occidentis F Alate 10.32 La Fage (1975)

Zootermopsis angusticollis M+F Larva & nymph 9.07 Hendee (1935)

Zootermopsis nevadensis M+F Alate 14,7 Hungate (1941)

Zootermopsis nevadensis - Nymph 9.1 Hungate (1941)

Undetermined — Alate 5.71 Leung (1968)

^ = male, F = female. 158 unlike carbohydrates, lipids can be stored in nearly an­ hydrous form. They are the most commonly stored metabolic fuel and, for fasting and hibernating animals, the only energy source (Lehniriger, 1970). When oxidized they generally yield twice as much energy per unit mass as either carbohydrates or proteins. As with most other animals, it appears that lipids, probably as triglycerides (TGL), con­ stitute the most important long-term energy reserve for M. hubbardi. Because of their small size and the diffuse nature of the fat body, few analyses have measured the lipid content of specific insect crgans or tissues. Cmelik

(1969a) used "semi-quantitative" methods to estimate per cent composition of total lipid found in various organs of unflown Macrotermes falciqer (as M. goliath) alates. He

He found approximately 60% in fat body, 6% in thoracic muscles, 5% in gut, 5% in reproductive organsf 2% in heads, and 22% in other body parts. Forty-seven per cent of the total dry body weight was fat. In physogastric queens of

Macrotermes natalensis (Haviland) the same author (Cmelik,

196 9b) found that fat body contained only 22% of the total body fat, whereas the reproductive organs (including eggs) contained 72%, Furthermore, in M. falciger queens, only

11% of the total fat was located in the fat body while 83% was found in the reproductive organs. Combined, these two organ systems accounted for 94% of the total fat in both species. The fat content as a percentage of total dry body 159

weight was not reported for either species; however, it

would likely be considerable.

The lipid content of the worker caste has been

determined for several species as follows: Reticulitermes I flavipes, 5.0% (Carter, Dir.us, and Smythe, 1972b); Copto-

termes formosanus, 4.2% (Mauldin, Smythe, and Baxter, 1972);

C. lacteus, 0.4-0.6% (Moore, 1969); Macrotermes falciger,

2,2% (avg) (Cmelik, 1972); and Nasutitermes exitiosus,

1-3% (Moore, 1969). Major soldiers of M. falciger contained

1.8% lipids (Cmelik, 1972). The value of these figures,

especially for comparison, is questionable. It might be

assumed that they were figured on a dry weight basis, but

this was not clearly stated in any case. Colonies of C.

lacteus maintain larger cellulose stores than N. exitiosus

and are thus apparently less dependent on body lipid re­

serves (Moore, 1969). No attempts have been made to extend

any of these analyses to the organ or tissue level; how­

ever, Carter et al. (1972a) claim that total body extracts

of lipids tend to approximate fat body composition. While

TGL is the primary lipid stored in insect fat body (Gilbert,

1967), Cmelik (1972) found phospholipid to be the pre­

dominant class in workers of M. falciger (56% of total

lipid). It seems likely that this high phospholipid content

represents membrane components rather than fat body lipids

since the fat body is relatively small in workers.

i 160

Basalingappa (1970) studied the quantitative varia­

tion in lipids among the castes of Odontotermes assmuthi

Holmgren. On a dry-weight basis lipids constituted 9.7% of

workers, 17.8% of soldiers, 26,4% of royal pairs, 17.8% of

the undifferentiated instars, and 48.4% of alates. In

Paraneotermes simplicicornis (Banks), a North American

subterranean kalotermitid, older nymphs contained 45,5%

lipid; younger nymphs, 39.3%; larvae, 22.0%; pseudergates,

28.4%; and soldiers, 16.6% (dry wt) (La Fage, unpublished

data). In this species, as in M. hubbardi, there appears to

be a progressive lipid buildup in the reproductive line.

Collins (1951) also recognized the variation in lipid con­

tent amon.g the castes in R. f lavipes. She found that lipids

from nymphs and young imagos (up to 6 days after meta­

morphosis) were histochemically much different from those in

workers and imagos 11 .days old or more. She concluded that

products stored during nymphal development were being used

during alate development. From data given by Hewitt, Nel,

and Schoeman (1972) it can be calculated that unflown

alates of a termitid, T. trinervoides, contain approximately

42% dry wt lipid in the form of neutral and phospholipids,

with the former predominating. Alates of another termitid,

G. perplexus, caught in flight, are also lipid-rich, con­ taining 56.3?. dry wt (M. D.immitt, 1974). Unspecified winyud termites offered for sale in a Leopoldville (Kinshasa)

market contained 44.4% fat (Tihon, 1946), 161

Among these admittedly random observations only one trend seems to surface: the subimaginal instars and alates contain substantially more lipids than other castes. These stores are undoubtedly used both for energy and for egg production during the critical period of colony foundation and early development by many of the Termitidae and perhaps, on occasion, by the lower termites. The belief that thoracic flight musculature is rapidly mobilized for nutrient requirements during this period, as in the ants, requires further investigation, since Noirot (1969) has noted that this process continues over a period of years and probably contributes very little to the nourishment of the royal pair, Hewitt and Nel (1969b) also noted that primary reproductives of H. mossambicus had lost only one third to one half the mass of their dorso-ventral thoracic muscula­ ture after 18 months, From the data available. (Carter et al,, 1972a; Cmelik, 1969a; Mauldin et al., 1972) it appears that TGL is quantitatively the most important lipid class stored by termites and and C^g FA the most important

TGL constituents (Table 42).

Biological Assay. Although a number of authors have extolled the virtues of eating insects (Bodenheimer, 1951;

DeFoliart, 1975; Ruddle, 1973; Tihon, 1946), only one study

(Teotia and Miller, 1973) has attempted to analyze insect protein by actual feeding trials. These authors used 162 housefly pupae as the protein source for 135 day-old broiler chicks. The weight gained by the birds during 4 weeks was statistically similar to that gained by birds on a control diet containing corn, milo, soybean meal (44% protein), fish and bone meal, and alfalfa. It is doubtful, however, that much information regarding protein quality can be gained from this type of experiment as both diets contained more than 25% crude protein. Protein quality should be assessed when supplied at a rate of 7-10% in the diet, i.e., sub-optimum levels.

Several conclusions were drawn about the quality of termite (R. flavipes) protein as a result of the feeding trials (Table 2). Perhaps the most important is the ob­ servation that the protein diet was both non-toxic and palatable as it was consumed at a rate (3.3 g/mouse/d) which was close to that for the whole-egg control diet (4.2 g/mouse/d). The mice which were fed termites did not gain as much weight during the 3-week feeding period as those on the standard diet, suggesting that 1 or more essential amino acids were in short supply. The protein score of R. flavipes protein (40) indicates that the most limiting amino acid was present at 40% of its optimum concentration. The amino acid profile of R, flavipes protein reported .in Table

1 indicates that methionine and cystine (considered col­ lectively in calculating protein score) are the moBt limiting amino acids. 163

Although protein scores abound in the literature, techniques and animals used in various laboratories differ to the extent that comparisons with R. flavipes would be suspect. Many additional analyses on amino acid composition and protein quality have been performed by the laboratory which tested R. flavipes (Department of Nutrition and Food

Science, The University of Arizona), The following protein scores provided by Dr. C. Weber (1976) are for conventional proteins: casein, 47; isolated soybean, 21; lactalbumin,

52; blood fibrin, 53; wheat (cajeme-71), 40; cottonseed meal, 36; brewer's yeast, 35; pinto bean, 21, Among the less conventional proteins, but all of which have been tested as possible food additives, are the following: palo- verde bean, 22; mesquite bean, 22; mesquite pods, 28; buffalo-gourd seed, 28; papago pea, 14. A quick examina­ tion of these protein scores shows that R. flavipes, with a score of 40, is a better protein than all of the non- conventional sources and a few of the conventional ones.

On the basis of Oser's (1959) essential amino acid index

(EAAI) which, unlike protein score, evaluates all the amino acids in a protein, R. £lavipes also compares well with other proteins evaluated in Dr. Weber's laboratory.

Although some proteins appear quite adequate from their protein scores or EAAI's, they are not readily ab­ sorbed by the animals and are thus of little nutritional value. Chromic oxide (C^O^) was added to the diet as a 164 marker to determine the digestibility of R. flavipes protein (SchUrch et al., 1950), Ca. 73% of the termite protein was absorbed. The calculated biological value

(Oser, 1959) suggests that 73,3% of the absorbed protein was retained by mice; i.e., not excreted in the urine. On the basis of the above data I have concluded that R. flavipes by itself is a good, but not wholly adequate, protein. However, it is unlikely that any predator would depend entirely on a single protein source. Termites could be considered an excellent protein supplement. The addi­ tion of sulfur-containing amino acids would undoubtedly raise their protein quality to a level adequate for normal growth in weanling mice and a variety of other animals. It is little wonder that M. hubbardi and other desert termites are routinely eaten.

Field Estimates on Energy Flow Through a Population of M. hubbardi

The density and distribution of dry-wood termite colonies depends to a large extent on the availability of suitable nesting sites which, for M. hubbardi, are few and scattered. As a result this species and many other dry- wood termite populations are far less impressive numerically than subterranean termites. The foraging populations alone of G. perpicxub can under favorable environments1 condi- 5 tions, exceed 7 x 10 individuals/ha (La Fage et al.,

1973). This is compared with 1.5 x 10 3 individuals/ha for 165

M, hubbardi on plot F2 of SNME, an area which probably has

a very high density for this species. The total annual

respiration (kcal/ha) for M. hubbardi was estimated at

680,71 kcal. Based on Haverty and Nutting's (1974)

estimate that the subterranean termites in a desert grass­

land annually consume 413,9 kg of dead wood (4.116 kcal/g

dry wt) I have calculated that more than 1,5 x 10 kcal/ha

would be dispersed in a single year. This amounts to 2,5 x 3 10 times the energy released by M. hubbardi on SNME. The

above-mentioned subterranean termites display approximately 2 2 the same metabolic activity/m (150 kcal/m /y) as other

small decomposers and herbivores such as mites, Collembola, 2 and nematodes (175,7 kcal/m /y) and large decomposers such 2 as earthworms and arthropods (128,5 kcal/m /y) (Lee and

Wood, 1971). Assuming an oxy-caloric equivalent of 4.8 kcal/L C>2 Engelmann (1966) reported that 46 species of oribatid mites with a biomass of 5,377 g/m 2 released 21,54 kcal/m 2/y in forest soil in Belgium, He also summarized respiration data for "mice" (Reithrodontomys, Peromyscus, and Microtus) which released ca, 743,6 kcal/ha/d during the summer.

The ecological importance of M. hubbardi need not o be diminished on the basis of its low respiration/m '/y, since organisms interact with each other and the environ­ ment in many and unsuspected ways. For example, it is

possible that some relationship exists between swarming 166 termites and seed-harvesting ants of the genus Pogonomyrmex which favors brood production by the ants. Throughout much of the year, harvester ants in New Mexico rely solely on seeds for food. During the summer months, however, these ants supplement their seed diet with termite alates, It so happens that termite dispersal flights occur during the period of greatest brood-rearing activity, Whitford (1976) has suggested that the protein supplementation by the termite prey may indeed by necessary to meet the increased nitrogen requirements of tissue production in the young ant larvae.

The study of dry-wood termites has been restricted largely to taxonomic, control, and a few physiological con­ siderations. I hope that the above experiments and speculations stimulate others to perform more in-depth studies on these interesting animals in the neglected areas of ecology and behavior. APPENDIX A

MODIFIED METHYL-RED INDICATOR USED FOR KJELDAHL TITRATIONS

0.125 g methyl-red

0.0825 g methylene blue

100 ml ethanol

167 APPENDIX B

FATTY ACIDS OF SAGUARO WOOD

168 169

Table B.l. Fatty acids in total chloroform:methanol (2tl, v/v) extracts from saguaro skeletons of three age categories — Values are percentages of total peak area.

Sample No. Group 1 C16: 0 C16:1 Group 2 C18: 0 C18:1 C18:2

Caitegory A 1 4, 24 33. 85 4.49 3.49 27.80 22.05 4. 09 4 5. 70 34.44 2. 01 3.36 29.32 24.70 1.40 6 3. 97 33,18 1.78 1. 62 26.15 29. 07 4.26 7 5.16 32.31 2, 31 2. 00 30,22 22. 31 5. 91 8 4,33 25,32 3. 05 1.76 21, 83 33,93 9.71 9 6,66 30,49 1,72 2. 92 29.50 22, 25 6,48 Mean s;oi 31.58 2.56 2. 54 27, 47 25. 72 5.31 SE 0.42 1,37 0.43 0.35 1,27 1.97 1.14

Category B 1 5. 57 30,56 2.50 2.70 30. 04 22.25 6.39 2 2. 81 22. 96 2.73 1.61 17.58 43.55 8.76 3 3. 05 32.59 2. 68 4.73 21. 88 30. 93 3.51 5 4. 07 28,92 1. 07 2.79 18.29 36.18 8.71 6 6. 43 36. 02 1.19 3,61 26. 86 20.79 5,12 7 6. 75 33.27 1.12 4.21 22.36 25.54 6. 81 8 4. 97 32.13 0. 94 2. 70 24.90 26.59 7.78 9 5,59 29,85 1.41 .2.72 24,40 28.61 7.42 X 4. 91 30,79 1.71 3.13 23.41 29,31 6.81 SE 0, 52 1,36 0.28 0. 35 1,56 2.66 0.64

Category C 2 4, 70 25.71 6.56 4,46 14,41 30.49 12,75 3 3.17 23. 01 2.46 2.60 19. 81 36.67 13.52 4 3, 87 24. 19 2. 44 1. 80 16.97 36. 27 9. 97 5 5, 13 25,11 1. 56 2.18 17. 95 38. 00 10.09 6 3. 93 20.69 2.97 2, 18 16.41 39.37 14.37 7 3. 93 26. 35 4. 37 3.19 26.26 27. 62 8.32 9 6.09 29. 97 0. 90 2. 86 24.55 28. 89 8.10 X 4.40 25.00 3 .04 2.75 19.48 33.90 11. 02 SE 0.37 1. 09 0. 72 0. 33 1. 66 1,80 0.95 APPENDIX C

SUMMARY OF RESULTS FROM THE INTERMEDIATE SIZED FEEDING TRIAL

Values for response variables represent sums and means for 3 Marginitermes hubbardi (Banks) colonies which provided termites for the experiment.

170 171

Table C.l. Number of termites which survived to the end of the experiment, by temperature.

Temperature Group Repli­ °C Size cates Sum Mean Std. Dev. Variance

24 5 3 7 2.33 1.53 2.33 10 3 16 5. 33 4. 73 22. 33 20 3 43 14. 33 1.53 2. 33 50 3 119 39. 67 5.51 30.33 Mean 12 185 15. 42 15.67 ' 245.54

26 5 3 7 2. 33 2.08 4.33 10 3 23 7. 67 0.58 0.33 20 3 48 16. 00 1. 00 1.00 50 3 127 42. 33 3.51 12.33 Mean 12 205 17. 08 16.16 260.99

28 5 3 6 2. 00 1.73 3.00 10 3 23 7.67 4.04 16. 33 20 3 40 13.33 2.51 6.33 50 3 85 28.33 6.66 44.33 Mean 12 154 12, 83 10. 85 117.61

30 5 3 3 1. 00 1,73 3. 00 10 3 24 8. 00 1.00 1.00 20 3 43 14.33 1.15 1.33 50 3 107 35. 67 3.51 12.33 Mean 12 177 14.75 13.65 186.57

32 5 3 14 4. 67 0. 58 0.33 10 3 18 6. 00 1.00 1.00 20 3 40 13. 33 2,31 5.33 50 3 103 34. 33 6. 81 46.33 Mean 12 175 14.58 12.78 163.36

Total Population 60 896 14. 93 13.55 183.52 Table C.2. Number of termites which survived to the end of the experiment, by group size.

Group Temperature Repli- Size °C cates Sum Mean Std. Dev. Variance

24 3 7 2.33 1.53 2.33 26 3 7 2.33 2. 08 4.33 28 3 6 2. 00 1.73 3.00 30 3 3 1.00 1. 73 3.00 32 3 14 4.67 0.58 0.33 Group 15 37 2.47 1.85 3.41

24 3 16 5.33 4.73 22.33 26 3 23 7.67 0.58 0.33 28 3 23 ' 7.67 4.04 16. 33 30 3 24 8.00 1.00 1.00 32 3 18 6.00 1. 00 1.00 Group 15 104 6.93 2. 66 7.07

24 3 43 14,33 1.53 2. 33 26 3 48 16.00 1. 00 1.00 28 3 40 13.33 2. 52 6.33 30 3 43 14.33 1.16 1.33 32 3 40 13.33 2.31 5. 33 Group 15 214 14,27 1. 83 3.35

24 3 119 39. 67 5,51 30. 33 26 3 127 42.33 3.51 12.33 28 3 85 28.33 6,66 44.33 30 3 107 35. 67 3.51 12.33 32 3 103 34.33 6.81 46.33 Group 15 541 36.07 6.75 45.50

Total Population 60 896 14.93 13,55 183.52 173

Table C.3. Percentage of original group which survived to the end of the experiment, by temperature.

Temperature Group Repli­ °C Size cates Sum Mean Std. Dev. Variance

24 5 3 140 46.67 30.55 933.33 10 3 160 53. 33 47.26 2233.33 20 3 215 71.67 7. 64 58. 33 50 3 238 79.33 11. 02 121.33 Mean 12 753 62. 75 28,28 ' 799.84

26 5 3 140 46.67 41.63 1733.33 10 3 230 76. 67 5.77 33.33 20 3 240 80. 00 5.00 25.00 50 3 254 "84. 67 7. 02 49.33 Mean 12 864 72.00 . 24.02 576.91

28 5 3 120 40. 00 34. 64 1200.00 10 3 230 76.67 40. 42 1633.33 20 3 200 66.67 12. 58 158.33 50 3 170 56. 67 13. 32 177.33 Mean 12 720 60, 00 27. 86 776.18

30 5 3 60 20. 00 34. 64 1200.00 10 3 240 80. 00 10.00 100.00 20 3 215 71. 67 5.77 33.33 50 3 214 71.33 7. 02 49.33 Mean 12 729 60.75 29. 47 868.39

32 5 3 280 93.33 11. 55 133.33 10 3 180 60. 00 10. 00 100.00 20 3 200 66. 67 11. 55 133.33 50 3 206 68. 67 13. 61 185.33 Mean 12 866 72,17 16,57 274.52

Total Population 60 3932 65. 53 25,39 644.39 174

Table C.4. Percentage of original group which survived to the end of the experiment, by group size.

Group Temperature Repli­ Size °C cates Sum Mean Std. Dev. Variance

5 24 3 140 46. 67 30.55 933.33 26 3 140 46. 67 41. 63 1733.33 28 3 120 40. 00 34.64 1200.00 30 3 60 20. 00 34.64 1200.00 32 3 280 93. 33 11.55 • 133.33 Group 15 740 49. 33 36. 93 1363.81

10 24 3 160 53. 33 47. 26 2233.33 26 3 230 76.67 5.77 33.33 28 3 230 ' 76. 67 40.42 1633.33 30 3 240 80. 00 10. 00 100.00 32 3 180 60. 00 10. 00 100.00 Group 15 1040 69.33 26.58 706.67

20 24 3 215 71. 67 7.64 58.33 26 3 240 80. 00 5. 00 25. 00 28 3 200 66. 67 12. 58 158.33 30 3 215 71. 67 5.77 33.33 32 3 200 66. 67 11.55 133.33 Group 15 1070 71.33 9.16 83.81

50 24 3 238 79.33 11. 02 121.33 26 3 254 84. 67 7. 02 49.33 2? 3 170 56. 67 13,32 177.33 30 3 214 71.33 7. 02 49.33 32 3 206 68. 67 13.61 185.33 Group 15 1082 72. 13 13,49 181.98

Total Population 60 3932 65.53 25.39 644.39 175

Table C.5, Mean weight (mg, fresh) of termite groups at the beginning of the experiment, by temperature.

Tempera- Group Repli- ture °C Size cates Sum Mean Std. Dev. Variance

24 5 3 166.80 55. 60 5, 86 34. 41 10 3 304.70 101.57 20. 16 406.32 20 .3«•» 591.10 197.03 10. 19 103.72 50 3 1450.60 483.53 44. 47 1977.58 Mean 12 2513,20 209.43 174.98 30617.96

26 5 3 160,40 53. 47 4, 97 24.66 10 3 299.10 99.70 1. 30 1.69 20 3 543.70 181.23 11. 51 132.57 50 3 1441,60 .480,53 38,43 1476.65 Mean 12 2444.80 203,73 174. 47 30441.24

28 5 3 157.40 52,47 7. 82 61.17 10 3 276,50 92.17 18. 74 351.30 20 3 563,50 187,83 16. 34 267.00 50 3 1393,40 464.47 47. 53 2258.65 Mean 12 2390.80 199.23 169. 58 28756.68

30 5 3 159.40 53.13 1. 60 2.57 10 3 323.60 107.87 3. 17 10. 02 20 3 569.80 189,93 13. 04 169.97 50 3 1428.80 476.27 35.17 1237.05 Mean 12 2481.60 206.80 171. 02 29248.48

32 5 3 157.50 52.50 9. 88 97.56 10 3 300.60 100.20 5,10 26. 04 20 3 552.40 184.13 13. 93 194.10 50 3 1390,60 463,53 44. 30 1962.70 Mean 12 2401.10 200.09 167. 55 28074.03

Total Population 60 12231.50 203,86 165.67 27447,90 176

Table C.6. Mean weight (rag, fresh) of termite groups at the beginning of the experiment, by group size.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std, Dev. Variance

5 24 3 166.80 55. 60 5. 87 34.41 26 3 160.40 53. 47 4. 97 24. 67 28 3 157.40 52. 47 7. 82 61.17 30 3 159.40 53.13 1. 60 2.57 32 3 157.50 52. 50 9. 88 97.56 Group 15 801.50 53.43 5. 74 32.90

10 24 3 304.70 101.57 20. 16 406.32 26 3 299.10 99. 70 1. 30 1.70 28 3 276,50 92.17 18. 74 351.30 30 3 323.60 107.87 3. 17 10. 02 32 3 300.60 100.20 5.10 26.04 Group 15 1504.50 100.30 11. 85 140.50

20 24 3 591.10 197.03 10. 19 103.72 26 3 543.70 181.23 11. 51 132.57 28 3 563.50 187.83 16. 34 267.00 30 3 569.80 189,93 13. 04 169.97 32 3 522.40 184.13 13. 93 194.10 Group 15 2820.50 188.03 12. 46 155.21

50 24 3 1450.00 483.53 44. 47 1977.58 26 3 1441.60 480.53 38. 43 1476.65 28 3 1393.40 464.47 47. 53 2258.65 30 3 1428.80 476.27 35. 17 1237.05 32 3 1390.60 463,53 44. 30 1962.70 Group 15 7105,00 473,67 36,69 1345,79

Total Population 60 12231.50 203.86 165. 67 27447,90 Table C,7. Mean weight (mg, fresh) of termite groups at ti..= end of the experiment, by temperature.

Tempera­ Group Repli­ ture °C Size cates Sum Mean Std, Dev. Variance

24 5 3 68.70 22. 90 20. 54 421.83 10 3 179.70 59. 90 56. 41 3181.63 20 3 505.00 168.33 23 „ 47 550.92 50 3 1412.70 470.90 74. 76 5589.19 Mean 12 2166.10 ISO.51 188.55 35552.19

26 5 3 85.10 28.37 24. 91 620.72 10 3 271.90 90. 63 7. 30 5: .34 20 3 515.20 171.73 26. 43 69E.62 50 3 1864.50 621.50 33.60 1129.24 Mean 12 2736.70 228.06 244. 06 59563.42

28 5 3 '4.20 24. 73 25. 27 638.36 10 3 2.24. 00 74. 67 36.38 1323.25 20 3 5;!3.20 174.40 26. 13 6.82. 84 50 3 12 ,'i 5. 0 0 411.67 80, 84 6534.26 Mean 12 20!>6. 40 171.37 160. 73 25834.00

30 5 3 26. 00 8. 67 15. 01 225.33 10 3 295.90 98. 63 22. 53 507.66 20 3 536.40 178.80 27. 21 740.59 50 3 1692.00 464.00 90. 05 8109.43 Mean 12 2550.30 212.53 224.98 50615.53

32 5 3 131.60 43. 87 3. 15 9. 92 10 3 171.30 57. 10 8. 06 64. 89 20 3 418.60 139.53 33. 59 1128.20 50 3 1087.90 362.63 87. 01 7570.16 Mean 12 1809.40 150.78 139, 22 19380.96

Total Population 60 11318.90 188.65 150. 78 36397.80 178

Table C.8. Mean weight (mg, fresh) of termite groups at the end of the experiment, by group size.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance \

24 3 58. 70 22.90 20. 54 421. 83 26 3 85,10 28. 37 24. 91 620. 72 28 3 74, 20 24. 73 25. 27 638. 36 30 3 26. 00 8.67 15. 01 225. 33 32 3 131.60 43. 87 3. 15 9.92 Group lrj 385.60 25,71 20. 25 410. 04

24 3 179. 70 59. 90 56. 41 3181.63 26 3 271. 90 90.63 7.30 53. 34 28 3 224. 00 74. 67 36. 38 1323. 25 30 3 295.90 98.63 22. 53 507.66 32 3 171. 30 57.10 0. 06 64. 89 Group 15 1142. 80 76. 19 31. 95 1021. 06

24 3 505. 00 168. 33 23.47 550.92 26 3 515. 20 171. 73 26,43 298. 52 28 3 523. 20 174. 40 26,13 682.84 30 3 536. 40 178. 80 27. 21 740. 59 32 3 418.60 139. 53 33,59 1128. 20 Group 15 2498. 40 166, 56 27. 41 751. 23

24 3 1412. 70 470,90 74,76 5589.19 26 3 1864. 50 621. 50 33.60 1129. 24 28 3 1235. 00 411. 67 80. 84 6534. 26 30 3 1692. 00 564. 00 90. 05 8109,43 32 3 1087. 90 362.63 87. 01 7570. 16 Group 15 7292,10 486.14 117,75 13865,39

Total Population 60 11318.90 188,65 190.78 36397.80 179

Table C.9, Mean weight Gmg) of a single individual at the beginning of the experiment.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std, Dev. Variance

5 24 3 33.36 11.12 1.17 1.38 26 3 32. 08 10. 69 0. 99 0.99 28 3 31.48 10. 49 1. 56 2.45 30 3 31.88 10.63 0. 32 0.10 32 3 31.50 10. 50 1. 98 ' 0. 90 Group 15 160.30 10. 69 1.15 1.32

10 24 3 30.47 10.16 2. 02 4.06 26 3 29.91 9. 97 0.13 0.02 28 3 27. 65 9.22 1. 87 3. 51 30 3 32.36 10.79 0.32 0.10 32 3 30.06 10.02 0.51 0.26 Group 15 150.45 10,03 1.19 1.41

20 24 3 29.56 9. 85 0.51 0.26 26 3 27.19 9. 06 0. 58 0.33 28 3 28.18 9.39 0. 82 0.67 30 3 28.49 9.50 0.65 0.43 32 3 27. 62 9.21 0.70 0. 49 Group 15 141,03 9,40 0.62 0.39

50 24 3 29.01 9. 67 0.89 0.79 26 3 28, 83 9. 61 0.77 0.59 28 3 27. 87 9.29 0. 95 0. 90 30 3 28.58 9. 53 0,70 0. 50 32 3 27,81 9,27 0. 89 0.79 Group 15 142,10 9.47 0,73 0, 54 V 180

Table C.10. Mean weight (mg) of a single individual at the end of the experiment.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 13.74 4,58 4.11 16. 87 26 3 17. 02 5.67 4. 98 24. 83 28 3 14. 84 4. 95 5.05 25. 54 30 - 3 5.20 1. 73 3. 00 9.01 32 3 26.32 8.77 0.63 0.40 Group 15 77,12 5.14 4.05 16.40

10 24 3 17.97 5,99 5. 64 31. 82 26 3 27.19 9.06 0.73 0.53 28 3 22.40 7.47 3. 64 13.23 30 3 29.59 9. 86 2.25 5.07 32 3 17,13 5.71 0. 81 0.65 Group 15 114.28 7.62 3. 20 10.21

20 24 3 25.25 8. 42 1.17 1.38 26 3 25.76 8.59 1,32 1.75 28 3 26.16 8.72 1.31 1.71 30 3 26. 82 8,94 1.36 1.85 32 3 20. 93 6.98 1. 68 2. 82 Group 15 124.92 8,33 1.37 1.88

50 24 3 28.25 9.42 1.50 2.24 26 3 37.29 12.43 0. 67 0. 45 28 3 24.70 8.23 1. 62 2. 61 30 3 33. 84 11.28 1. 80 3. 24 32 3 21.76 7.25 1. 74 3.03 Group 15 145.84 9,72 2.36 5.55

Total Population 60 462.16 7.70 3. 30 10.88 181 Table C.11. Per cent weight change of termite group during . .beginning wt t- final wt. , experiment ( V beg^nning wt ), by temperature.

Tempera*- Group Repli- ture °C Size cates Sum Mean Std. Dev. Variance

24 5 3 1.74 0.58 0.37 0.14 10 3 1,42 0.47 0.48 0.23 20 3 0.44 0.15 0.12 0.01 50 3 0.09 0. 03 0. 07 0.01 Mean 12 3,68 0.31 0.35 0.13

26 5 3 1.43 0.47 0.48 0. 23 10 3 0.27 0.09 0. 08 0. 01 20 3 0.17 0. 06 0. 09 0. 01 50 3 -0. 91 -0.30 0.17 0. 03 Mean 12 0. 96 0. 08 0.36 0.13

28 5 3 1.71 0.57 0. 42 0.17 10 3 0. 63 0. 21 0.29 0.09 20 3 0.20 0. 07 0.16 0.02 50 3 0.34 0.11 0.15 0. 02 Mean 12 2.88 0.24 0.31 0.10

30 5 3 2. 50 0.83 0.30 0.09 10 3 0,25 0.08 0.23 0.05 20 3 0.16 0.05 0.17 0. 03 50 3 -0.55 -0.18 0.14 0.02 Mean 12 2.36 0.20 0. 44 0.19

32 5 3 0. 43 0.14 0,19 0.04 10 3 1. 30 0, 43 0. 06 0.00 20 3 0. 69 0.23 0. 23 0. 05 50 3 0. 67 0.22 0.12 0.02 Mean 12 3.08 0.26 0.18 0.03

Total Population 60 12.96 0. 22 0. 34 0.11 182

Table C.12. Per cent weight change of termite group during . -beginning wt - final wt, , experiment ( * beginning"^ »' ^ froup size.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 1.74 0.58 0.37 0-14 26 3 1.43 0.48 0.48 0.23 28 3 1. 71 0.57 0.42 ' 0.17 30 3 2.50 0. 83 0.29 0.09 32 3 0.43 0.14 0.19 0.04 Group 15 7.80 0.52 0.38 0.15

10 24 3 1.42 0. 47 0.48 0.23 26 3 0.27 0.09 0.08 0.01 28 3 0.63 0.21 0.29 0.09 30 3 0.25 0.08 0.23 0.05 32 3 1.30 0.43 0. 06 0.00 Group 15 3.86 0.26 0.29 0.08

20 24 3 0.44 0.15 0.12 0.01 26 3 0.17 0.06 0.09 0.01 28 3 0. 20 0. 07 0.16 0.02 30 3 0.16 0. 05 0.17 0.03 32 3 0. 69 0. 23 0.23 0.05 Group 15 1.66 0.11 0.15 0.02

50 24 3 0. 09 0. 03 0.07 0.01 26 3 -0,91 -0.30 0.17 0.03 28 3 0.34 0.11 0.15 0.02 30 3 -0. 55 -0.18 0.14 0.02 32 3 0.67 0.22 0.12 0.12 Group 15 -0.35 -0,02 0. 23 0.05

Total Population 60 12,96 0.22 0.34 0,11 Table C.13. Biomass days (BT), fay temperature.

Temperature °c Group Size Replicates Sum Mean Std. Dev. Variance

24 5 3 6976 2325.33 1567.70 2457692.33 10 3 16786 5595.33 4863.54 0. 23712E+08 20 3 41444 13814.67 1841.75 3392041.33 50 3 115838 38612.67 7158.95 0. 51250E+08 Mean 12 181044 15087.00 15331.81 0. 23506E+09 26 5 3 7580 2526.67 2007.23 4028950.33 10 3 22656 7552.00 526.52 277225.00 20 3 42878 14292.67 1801.96 3247052.33 50 3 136379 45459.67 1288.54 1660337.33 Mean 12 209493 17457.75 17487.62 0.30581E+09 28 5 3 6745 2248.33 1895.92 3594517.33 10 3 20211 6737.00 3587.29 0.12868E+08 20 3 40505 13501.67 2122.44 4504754.33 50 3 91951 30650.33 7049.45 0.49694E+08 Mean 12 159412 13284,33 11832.87 0.14001E+09 30 5 3 3166 1055.33 1407.88 1982114.33 10 3 25054 8351,33 1309.59 1715045.33 20 3 42551 14183.67 1425.40 2031777.33 50 3 120488 40162.67 5582.14 0.31160E+08 Mean 12 191259 15938.25 15610.80 0.24369E+09 32 5 3 12612 4204.00 528.64 279457.000 10 3 16433 5477.67 852,32 726462.333 20 3 35893 11964.33 1666.28 2776472.33 50 3 i08071 36023.67 10784.57 0.11630E+09 Mean 12 173009 14417.42 14178,80 0.20103E+09 Total :Population 60 914217 15236.95 14556.22 0.21188E+09 Table C.14. Biomass days (BT), by group size.

Temperature Group Size °C Replicates Sum Mean Std. Dev. Variance

5 24 3 6976 2325.33 1567.70 2457692.33 26 3 7580 2526.67 2007.23 4028950.33 28 3 6745 2248.33 1895.92 3594517.33 30 3 3166 1055.33 1407.88 1982114.33 32 3 12612 4204.00 528.64 270457.00 Group 15 37079 2471.93 1688.82 2852096.21 10 24 3 16786 5595.33 4869.54 0.23712E+08 26 3 22656 7552.00 526.52 277225,00 28 3 20211 6737,00 3587.29 0.12868E+08 30 3 25054 8351.33 1309.60 1715045.33 32 3 16433 5477.67 852.33 726462.33 Group 15 101140 6742.67 2633.27 6934149.10 20 24 3 41444 13814,67 1841.75 3392041.33 26 3 42878 14292,67 1801.96 3247052.33 28 3 40505 13501.67 2122.44 4504754.33 30 3 42551 14183.67 1425.40 2031777.33 32 3 35893 11964.33 1666.28 2776272.33 Group 15 203271 13551.40 1742.81 3037399.83 50 24 3 115838 38612.67 7158.95 0.51250E+08 26 3 136379 45459.67 1288.54 1660337.33 28 3 91951 30650.33 7049.45 0.49694E+08 30 3 120488 40162.67 5582.14 0.31160E+08 32 3 108071 36023.67 10784.57 0.11630E+09 Group 15 572727 38181.80 781-7.19 0.61108E+08 Total Population 60 914217 15236.95 14556.22 0.21188E+09 185

Table C.15. Number of supplementary reproductives present at the end of the experiment, by temperature.

Tempera- Group Repli­ ture °C Size cates Sum Mean Std. Dev. Variance

24 5 3 2.00 0.68 0.58 0.33 10 3 4. 00 1.33 1.16 1.33 20 3 8.00 2.67 0.58 0. 33 50 3 6.00 2.00 1.00 1.00 Mean 12 20.00 1. 67 1.07 • 1.15

26 5 3 3.00 1.00 1. 00 1. 00 10 3 5.00 1.68 0.58 0.33 20 3 7.00 2. 33 0.58 0.33 50 3 9.00 ' 3.00 1.00 1.00 Mean 12 24.00 2.00 1.05 1.10

28 5 3 3.00 1.00 1.00 1.00 10 3 3.00 1.00 1.00 1.00 20 3 6.00 2.00 0.00 0.00 50 3 9.00 3. 00 1.00 1.00 Mean 12 21. 00 1.75 1.14 1.30

30 5 3 2. 00 0.68 1.16 1.33 10 3 9. 00 3.0 0 1. 00 1. 00 20 3 6.00 2. 00 1.00 1.00 50 3 7.00 2.33 0.58 0. 33 Mean 12 24,00 2.00 1.21 1.50

32 5 3 5.00 1.67 0.58 0.33 10 3 5.00 1. 67 0. 58 0.33 20 3 7.00 2.33 0.58 0.33 50 3 6.00 2.00 0. 00 0. 00 Mean 12 23.00 1. 92 0.52 0.27 1 • o o O O H Total Population 60 112,00 1.87 i— 186

Table C.16. Number of supplementary reproductives present at the end of the experiment, by group size.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 2 0.67 0.58 0.33 26 3 3 1.00 1.00 1. 00 28 3 3 1.00 1.00 1.00 30 3 2 0.67 1.16 1.33 32 3 5 1. 67 0.58 ' 0.33 Group 15 15 1.00 0.85 0.71

10 24 3 4 1. 33 1.16 1.33 26 3 5 1.67 0.58 0. 33 28 3 3 1.00 1. 00 1. 00 30 3 9 3.00 1.00 1. 00 32 3 5 1.67 0.58 0.33 Group 15 26 1.73 1.03 1.07

20 24 3 8 2.67 0.58 0. 33 26 3 7 2.33 0.58 0.33 28 3 6 2. 00 0.00 0. 00 30 3 6 2. 00 1. 00 1.00 32 3 7 2.33 0.58 0.33 Group 15 34 2.27 0.59 0.35

50 24 3 6 2. 00 1.00 1.00 26 3 9 3.00 1.00 1.00 28 3 9 3.00 1.00 1. 00 30 3 7 2.33 0. 58 0. 33 32 3 6 2. 00 0. 00 0. 00 Group 15 37 2.47 0. 83 0.70

Total Population 60 112 1.87 1.00 1.00 187

Table C.17. Final number of supplementary reproductives v initial number of termites in group.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 0.40 0.13 0. 12 0.01 26 3 0. 60 0. 20 0. 20 0.04 28 3 0. 60 0. 20 0. 20 0.04 30 3 0.40 0.13 0.23 0.05 32 3 1.00 0.33 0.12 0.01 Group 15 3.00 0. 20 0.17 0.03

10 24 3 0.40 0.13 0.12 0.01 26 3 0.50 0.17 0.06 0.00 28 3 0.30 0.10 0.10 0.01 30 3 0. 90 0.30 0.10 0. 01 32 3 0. 50 0.17 0. 06 0. 00 Group 15 2.60 0.17 0.10 0. 01

20 24 3 0.40 0.13 0. 03 0.00 26 3 0.35 0.12 0. 03 0.00 28 3 0.30 0.10 0. 00 -0.00 30 3 0.30 0.10 0. 05 0.00 32 3 0.35 0.12 0. 03 0.00 Group 15 1,70 0.11 0. 03 0.00

50 24 3 0.12 0.04 0.02 0. 00 26 3 0.18 0.06 0.02 0.00 28 3 0.18 0. 06 0. 02 0. 00 30 3 0.14 0.05 0.01 0.00 32 3 0.12 0. 04 0. 00 -0.00 Group 15 0.74 0.05 0.02 0. 00

Total Population 60 8. 04 0,13 0.11 0. 01 188

Table C.18. Number of larvae (from original group) present at the end of the experiment, by temperature.

Tempera­ Group Repli­ ture °C Size cates Sum Mean Std. Dev. Variance

24 5 3 0.00 0.00 0.00 0.00 10 3 0. 00 0.00 0.00 0.00 20 3 1. 00 0.33 0.58 0.33 50 3 4. 00 1.33 0. 58 0. 33 Mean 12 5.00 0.42 0.67 ' 0.45

26 5 3 0.00 0.00 0.00 0.00 10 3 1.00 0.33 0.58 0.33 20 3 5.00 1.68 2.08 4.33 50 3 3.00 1.00 1.00 1.00 Mean 12 9.00 0.75 1.22 1.48

28 5 3 0. 00 0.00 0.00 0.00 10 3 2. 00 0.68 1.16 1.33 20 3 3. 00 1. 00 1.00 1.00 50 3 4.00 1.33 1.16 1.33 Mean 12 9.00 0.75 0. 97 0.93

30 5 3 0.00 0.00 0.00 0.00 10 3 0.00 0.00 0. 00 0.00 20 3 1. 00 0. 33 0. 58 0.33 50 3 6.00 2. 00 1.00 1.00 Mean 12 7.00 0.58 1. 00 1.00

32 5 3 2.00 0.68 0.58 0.33 10 3 2.00 0. 68 0. 58 0.33 20 3 6. 00 2.00 1. 00 1.00 50 3 4.00 1.33 1. 53 2.33 Mean 12 14.00 1.68 1.03 1.06 o Total Population 60 44.00 0.73 0.99 CO 189

Table C.19. Number of larvae (from original group) present at the end of the experiment, by group size.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std. Dev. Variam

5 24 3 0 0. 00 0.00 0. 00 26 3 0 0. 00 0. 00 0.00 28 3 0 0.00 0. 00 0.00 30 3 0 0.00 0.00 0.00 32 3 2 0.67 0.58 • 0.33 Group 15 2 0.13 0.35 0.12

10 24 3 0 0. 00 0. 00 0.00 26 3 1 0.33 0.58 0.33 28 3 2 0.67 1.16 1.33 30 3 0 0.00 0.00 0. 00 32 3 2 0.67 0.58 0.33 Group 15 5 0.33 0.62 0.38

20 24 3 1 0.33 0.58 0.33 26 3 5 1. 67 2.08 4.33 28 3 3 1.00 1.00 1.00 30 3 1 0.33 0. 58 0.33 32 3 6 2.00 1.00 1.00 Group 15 16 1.07 1.22 1.50

50 24 3 4 1.33 0. 58 0.33 26 3 3 1. 00 1.00 1.00 28 3 4 1.33 1.16 1.33 30 3 6 2. 00 1.00 1.00 32 3 4 1,33 1.53 2.33 Group 15 21 1,40 0.99 0.97

Total Population 60 44 0.73 0.99 0.98 190

Table C.20, Final number of larvae (from original group) by initial number of termites in group.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 0.00 0.00 0.00 0.00 26 3 0.00 0. 00 0. 00 0.00 28 3 0.00 0. 00 0.00 0.00 30 3 0. 00 0.00 0.00 0.00 32 3 0.40 0.13 0.12 0.01 Group 15 0.40 0.03 0.07 0.01

10 24 3 0.00 0.00 0. 00 0.00 26 3 0.10 0. 03 0.06 0.00 28 3 0.20 0.07 0.12 0.01 30 3 0.00 0.00 0. 00 0.00 32 3 0.20 0.07 0. 06 0.00 Group 15 0.50 0.03 0.06 0.00

20 24 3 0.05 0.02 0. 03 0.00 26 3 0.25 0.08 0.10 0.01 28 3 0.15 0.05 0.05 0.00 30 3 0.05 0.02 0.03 0.00 32 3 0.30 0.10 0.05 0.00 Group 15 0. 80 0.05 0. 06 0.00

50 24 3 0.08 0.03 0.01 0.00 26 3 0. 06 0. 02 0.02 0.00 28 3 0. 08 0.03 0.02 0.00 30 3 0.12 0.04 0. 02 0.00 32 3 0,08 0.03 0. 03 0.00 Group 15 0.42 0,03 0. 02 0.00

Total Population 60 2.12 0.04 0.06 0.00 191

Table C.21, Number of young nymphs present at the end of the experiment, by temperature.

Tempera- Group Repli- ture JC Size cates Sum Mean Std. Dev. Variance

24 5 3 4 1.33 1.53 2.33 10 3 12 4.00 3. 61 13.00 20 3 27 9.00 1.73 3.00 50 3 95 31.67 6.81 46.33 Mean 12 138 11.50 12.96 • 167.91

26 5 3 3 1,00 1. 00 1.00 10 3 14 4.67 0.58 0. 33 20 3 31 10.33 2.08 4.33 50 3 52 17.33 3.79 14.33 Mean 12 100 8.33 6.72 45.15

28 5 3 2 0. 67 0.58 0.33 10 3 14 4. 67 3. 22 10.33 20 3 23 7.67 3.22 10.33 50 3 43 14.33 7.10 50. 33 Mean 12 82 6.83 6.34 40.15

30 5 3 1 0.33 0.58 0.33 10 3 10 3.33 0.58 0. 33 20 3 31 10.33 2.08 4.33 50 3 58 19.33 4. 04 16. 33 Mean 12 100 8.33 7.89 62.24

32 5 3 5 1.67 0.58 0.33 10 3 8 2, 67 1. 53 2. 33 20 3 24 8. 00 1,00 1. 00 50 3 58 19.33 4.16 17.33 Mean 12 95 7.92 7.59 57. 54

Total Population 60 515 8.58 8.49 72.01 192

Table C.22. Number of young nymphs present at the end of the experiment, by group size.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 4 1.33 1.53 2.33 26 3 3 1.00 1.00 1. 00 28 3 2 0. 67 0.58 0.33 30 3 1 0. 33 0.58 0.33 32 3 5 1.67 0. 58 0.33 Group 15 15 1.00 0. 93 0. 86

10 24 3 12 4.00 3. 61 13. 00 26 3 14 4. 67 0.58 0.33 28 3 14 4. 67 3. 22 10.33 30 3 10 3.33 0.58 0.33 32 3 8 2. 67 1.53 2.33 Group 15 58 3. 87 2.10 4.41

20 24 3 27 9. 00 1.73 3.00 26 3 31 1.33 2.08 4.33 28 3 23 7.67 3.22 10.33 30 3 31 10.33 2. 08 4.33 32 3 24 8.00 1.00 1. 00 Group 15 136 9.07 2.15 4.64

50 24 3 95 31. 67 6. 81 46. 33 26 3 52 17.33 3.79 14.33 28 3 43 14.33 7.10 50.33 30 3 58 19.33 4.04 16.33 32 3 58 19. 33 4.16 17.33 Group 15 306 20. 40 7. 63 58.26

Total Population 60 515 8.58 8.49 72.01 193 Table C.23. Final number of young nymphs * by initial number of termites in group.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 0.80 0.27 0.31 0.09 26 3 0.60 0.20 0. 20 0.04 28 3 0.40 0.13 0.12 0. 01 30 3 0. 20 0. 07 0.12 0. 01 32 3 1.00 0.33 0.12 0. 01 Group 15 3.00 0. 20 0.19 0.03

10 24 3 1. 20 0.40 0.36 0.13 26 3 1.40 0.47 0.06 0.00 28 3 1.40" 0.47 0.32 0.10 30 3 1.00 0.33 0.06 0.00 32 3 0. 80 0.27 0.15 0.02 Group 15 5.80 0.39 0.21 0. 04

20 24 3 1.35 0.45 0.09 0.01 26 3 1.55 0. 52 0.10 0.01 28 3 1.15 0.38 0.16 0.03 30 3 1. 55 0. 52 0.10 0.01 32 3 1.20 0.40 0.05 0.00 Group 15 6. 80 0.45 0.10 0.01

50 24 3 1.90 0.63 0.14 0.02 26 3 1.04 0. 35 0.08 0.01 28 3 0. 86 0.29 0.14 0.02 30 3 1.16 0.39 0.08 0.01 32 3 1.16 0. 39 0.08 0.01 Group 15 6.12 0. 40 0.15 0. 02

Total Population 60 21.72 0. 36 0.19 0.04 194

Table C.24. Number of old nymphs present at the end of the experiment, by temperature.

Tempera- Group Repli­ ture °C Size cates Sum Mean Std. Dev. Variance

24 5 3 1 0. 33 0.58 0.33 10 3 0 0.00 0.00 0.00 20 3 5 1.67 1.53 2.33 50 3 11 3. 67 2. 89 8.33 Mean 12 17 1.42 2. 07 4. 27

26 5 3 0 0.00 0. 00 0.00 10 3 1 0.33 0. 58 0.33 20 3 2 0.67 1.16 1.33 50 3 60 20.00 6. 56 43.00 Mean 12 63 5. 25 9.34 87.30

28 5 3 0 0.00 0.00 0. 00 10 3 1 0. 33 0.58 0.33 20 3 5 1.67 2.08 4.33 50 3 25 8.33 5.51 30.33 Mean 12 31 2.58 4.34 18.81

30 5 3 0 0.00 0.00 0.00 10 3 2 0.68 1.16 1.33 20 3 2 0.68 1.16 1.33 50 3 33 11.00 8.89 79.00 Mean 12 37 3.08 6.14 37.72

32 5 3 0 0. 00 0. 00 0.00 10 3 0 0. 00 0.00 0.00 20 3 0 0.00 0. 00 0.00 50 3 0 0.00 0.00 0.00 Mean 12 30 10. 00 9. 85 - 97.00

Total Population 60 178 2. 97 6.03 36.34 195

Table C.25. Number of old nymphs present at the end of the experiment, by group size.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 1.00 0.33 0.58 0.33 26 3 0.00 0.00 0. 00 0.00 28 3 0.00 0.00 0.00 0. 00 30 3 0.00 0.00 0.00 0. 00 32 3 0.00 0.00 0.00 0. 00 Group 15 1. 00 0.07 0. 26 0. 07

10 24 3 0.00 0.00 0.00 0. 00 26 3 1.00 0.33 0.58 0.33 28 3 1.00 0. 33 0. 58 0.33 30 3 2.00 0. 67 1.16 1. 33 32 3 0.00 0.00 0. 00 0. 00 Group 15 4.00 0.28 0. 59 0.35

20 24 3 5. 00 1. 67 1. 53 2.33 26 3 2. 00 0. 68 1.15 1.33 28 3 5. 00 1.68 2. 08 4.33 30 3 2.00 0. 67 1.16 1. 33 32 3 0.00 0.00 0. 00 0.00 Group 15 14. 00 0. 93 1.33 1.78

50 24 3 11. 00 3.67 2. 89 8. 33 26 3 60.00 20.00 6. 56 43.00 28 3 25. 00 8.33 5.51 30. 33 30 3 33. 00 11.00 8. 89 79.00 32 3 30. 00 10.00 9. 85 97. 00 Group 15 159.00 10.60 8.20 67.26

Total Population 60 178.00 2. 97 6. 03 36.34 196

Table C.26. Final number of old nymphs * by initial number of termites in group.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 0.20 0.07 0.12 0.01 26 3 0.00 0.00 0.00 0.00 28 3 0.00 0. 00 0.00 0.00 30 3 0.00 0. 00 0. 00 0.00 32 3 0.00 0.00 0.00 0. 00 Group 15 0.20 0.01 0.05 0.00

10 24 3 0.00 0. 00 0.00 0.00 26 3 0.10 0.03 0.06 0.00 28 3 0.10 0.03 0.06 0.00 30 3 0.20 0. 07 0.12 0.01 32 3 0.00 0.00 0.00 0. 00 Group 15 0.40 0.03 0.06 0.02

20 24 3 0. 25 0.08 0.08 0.01 26 1 0.10 0.03 0.06 0.00 28 3 0. 25 0.08 0.10 0.01 30 3 0.10 0. 03 0.06 0.00 32 3 0.00 0.00 0. 00 0.00 Group 15

50 24 3 0. 22 0.07 0. 06 0. 00 26 3 1.20 0. 40 0.13 0.02 28 3 0.50 0.17 0.11 0.01 30 3 0.66 0. 22 0.18 0.03 32 3 0.60 0. 20 0. 20 0.04 Group 15 3.18 0. 21 0.16 0.03

Total Population 60 4.48 0.07 0.12 0. 02 197

Table C.27. Number of soldiers present at the end of the experiment, by temperature.

Tempera- Group Repli- ture °C Size cates Sum Mean Std. Dev. Variance

24 5 o 0 0.00 0.00 0.00 10 3 0 0.00 0.00 0. 00 20 3 2 0. 67 0. 58 0. 33 50 3 3 1.00 0.00 0.00 Mean 12 5 0. 42 0.52 " 0.27

26 5 3 1 0.33 0.58 0. 33 10 3 2 0. 67 0. 58 0.33 20 3 3 1. 00 0. 00 0.00 50 3 3 ' 1. 00 0.00 0.00 Mean 12 9 0.75 0.45 0.21

28 5 3 1 0.33 0.58 0. 33 10 3 1 0.33 0. 58 0.33 20 3 3 1.00 0.00 0.00 50 3 4 1.33 0.58 0.33 Mean 12 9 0.75 0.62 0.39

30 5 3 0 0. 00 0.00 0.00 10 3 3 1.00 0.00 0. 00 20 3 3 1.00 0.00 0.00 50 3 3 1. 00 0.00 0.00 Mean 12 9 0.75 0.45 0. 21

32 5 3 2 0. 67 0.58 0.33 10 3 3 1. 00 0.00 0. 00 20 3 3 1. 00 0.00 0.00 50 3 5 1. 67 0.58 0.33 Mean 12 13 1. 08 0. 52 0.27

Total Population 60 45 0.75 0.54 0.29 198

Table C.28. Number of soldiers present at the end of the experiment, by group size.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 0 0.00 0.00 0.00 26 3 1 0. 33 0.58 0.33 28 3 1 0.33 0.58 0.33 30 3 0 0.00 0.00 0. 00 32 3 2 0. 67 0.58 • 0.33 Group 15 4 0.27 0.46 0.21

10 24 3 0 0.00 0.00 0.00 26 3 2 0.67 0.58 0.33 28 3 1 ' 0.33 0.58 0. 33 30 3 3 1. 00 0.00 0.00 32 3 3 1.00 0.00 0.00 Group 15 9 0.60 0.51 0.26

20 24 3 2 0. 67 0.58 0.33 26 3 3 1.00 0. 00 0.00 28 3 3 1. 00 0.00 0.00 30 3 3 1.00 0.00 0.00 32 3 3 1.00 0.00 0. 00 Group 15 14 0.93 0.26 0.07

50 24 3 3 1.00 0.00 0. 00 26 3 3 1. 00 0. 00 0.00 28 3 4 1. 33 0.58 0.33 30 3 3 1.00 0. 00 0.00 32 3 5 1.67 0.58 0.33 Group 15 18 1.20 0. 41 0.17

Total Population 60 45 0.75 0.54 0.29 199

Table C.29. Final number of soldiers t by initial number of termites in group.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 0. 00 0. 00 0. 00 0.00 26 3 0. 20 0. 07 0.12 0.13 28 3 0. 20 0. 07 0.12 0.13 30 3 0.00 0.00 0. 00 0.00 32 3 0.40 0.13 0.12 0.01 Group 15 0. 80 0.05 0.09 0.01

10 24 3 0.00 0.00 0.00 0.00 26 3 0.20 0.07 0.06 0. 00 28 3 0.10 0. 03 0. 06 0.00 30 3 0.30 0.10 0.00 -0. 00 32 3 0.30 0.10 0. 00 -0. 00 Group 15 0. 90 0.06 0.05 0.00

20 24 3 0.10 0.03 0. 03 0.00 26 3 0.15 0.05 0.00 -0. 00 28 3 0.15 0.05 0.00 -0.00 30 3 0.15 0.05 0.00 -0.00 32 3 0.15 0.05 0.00 -0.00 Group 15 0.70 0.05 0.01 0.00

50 24 3 0. 06 0.02 0.00 -0.00 26 3 0. 06 0.02 0. 00. -0.00 28 3 0.08 0.03 0.01 0.00 30 3 0.06 0.02 0.00 -0.00 32 3 0.10 0. 03 0.01 0.00 Group 15 0.36 0.02 0.01 0.00

Total Population 60 2.76 0.05 0.05 0.00 200

Table C.30. Number of eggs present at the end of the experiment', by temperature.

Tempera­ Group Repli­ ture °C Size cates Sum Mean Std. Dev. Variance

24 5 3 0. 00 0. 00 0.00 0. 00 10 3 1. 00 0. 33 0.58 0.33 20 3 0.00 0.00 0.00 0.00 50 3 4.00 1. 33 1.53 2.33 Mean 12 5. 00 0. 42 0.90 0. 81

26 5 3 0. 00 0. 00 0.00 0.00 10 3 2. 00 0.67 1.16 1.33 20 3 4. 00 1.33 2.31 5.33 50 3 6.00 ' 2.00 1.73 3. 00 Mean 12 12. 00 1. 00 1.54 2.36

28 5 3 0. 00 0.00 0.00 0.00 10 3 0. 00 0. 00 0. 00 0.00 20 3 5.00 1.67 1.16 1.33 50 3 28. 00 9.33 10.21 104.33 Mean 12 33.00 2. 75 5.96 35.48

30 5 3 0. 00 0.00 0.00 0. 00 10 3 1.00 0. 33 0.58 0.33 20 3 3. 00 1.00 1.00 1. 00 50 3 3.00 1.00 1.00 1.00 Mean 12 7.00 0.58 0.79 0.63

32 5 3 0. 00 0.00 0.00 0. 00 10 3 0.00 0.00 0.00 0.00 20 3 9. 00 3.00 3.00 9. 00 50 3 24.00 8.00 5.29 28.00 Mean 12 33. 00 2.75 4.29 18.39

Total Population 60 90.00 1.50 3.44 11. 85 201

Table C.31. Number of eggs present at the end of the experiment, by group size.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 0 0.00 0. 00 0.00 26 3 0 0.00 0. 00 0.00 28 3 0 0.00 0.00 0.00 30 3 0 0. 00 0.00 0.00 32 3 0 0.00 0.00 0. 00 Group 15 0 0.00 0.00 0.00

10 24 3 1 0.33 0.58 0.33 26 3 2 0. 67 1.16 1.33 28 3 0 ' 0. 00 0. 00 0.00 30 3 1 0.33 0.58 0.33 32 3 0 0.00 0.00 0.00 Group 15 4 0.27 0.59 0.35

20 24 3 0 0.00 0.00 0.00 26 3 4 1. 33 2.31 5.33 28 3 5 1. 67 1.16 1. 33 30 3 3 1.00 1.00 1.00 32 3 9 3.00 3.00 9.00 Group 15 21 1.40 1. 85 3.40

50 24 3 4 1.33 1.53 2.33 26 ' 3 6 2. 00 1.73 3.00 28 3 28 9.33 10.21 104.33 30 3 3 1. 00 1. 00 1.00 32 3 24 8.00 5.29 28. 00 Group 15 65 4.33 5.79 33.52

Total Population 60 90 1.50 3.44 11. 85 202

Table C.32. Final number of eggs * by the initial number of termites in group.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 0.00 0. 00 0. 00 0.00 26 3 0. 00 0. 00 0. 00 0.00 28 3 0. 00 0. 00 0. 00 0. 00 30 3 0. 00 0.00 0.00 0.00 32 3 0. 00 0.00 0.00 0. 00 Group 15 0. 00 0.00 0.00 0.00

10 24 3 0.10 0. 03 0.06 0.00 26 3 0.20 . 0. 07 0.12 0.01 28 3 0. 00 0.00 0.00 0.00 30 3 0.10 0.03 0.06 0. 00 32 3 0. 00 0. 00 0.00 0. 00 Group 15 0.40 0.03 0.06 0.00

20 24 3 0. 00 0.00 0. 00 0.00 26 3 0.20 0. 07 0.12 0.01 28 3 0.25 0.08 0.06 0.00 30 3 0.15 0.05 0.05 0.00 32 3 0.45 0.15 0.15 0.02 Group 15 1. 05 0.07 0.09 0.01

50 24 3 0. 08 0. 03 0. 03 0.00 26 3 0.12 0.04 0.04 0. 00 28 3 0. 56 0.19 0.20 0.04 30 3 0. 06 0. 02 0.02 0.00 32 3 0.48 0.16 0.11 0. 01 Group 15 1.30 0.09 0.12 0.01

Total Population 60 2.75 0.05 0.09 0.01 203

Table C.33. Number of larvae produced which survived to the end of the experiment, by temperature.

Tempera­ Group Repli­ ture °C Size cates Sum Mean Std,. Dev. Variance

24 5 3 0., 00 0,,00 0..00 0..00 10 3 0.. 00 0,.00 0.. 00 0,.00 20 3 0..00 0., 00 0..00 0., 00 50 3 0., 00 0,.00 0.,00 0..00 Mean 12 0., 00 0.,00 0.,00 0.,00

26 5 3 0.,00 0., 00 0., 00 0., 00 10 3 0..00 0.,00 0,. 00 0..00 20 3 1..00 0..33 0..57 0..33 50 3 0.,00 ' 0., 00 0., 00 0.,00 Mean 12 1.,00 0..83 0..29 0.,08

28 5 3 0., 00 0., 00 0.,00 0., 00 10 3 0., 00 0..00 0., 00 0., 00 20 3 0., 00 0..00 0., 00 0.,00 50 3 0., 00 0.,00 0.,00 0.,00 Mean 12 0., 00 0., 00 0.,00 0.,00

30 5 3 0., 00 0., 00 0., 00 0.,00 10 3 0., 00 0., 00 0., 00 0.,00 20 3 7., 00 2.,33 4., 04 16.,33 50 3 21. 00 7., 00 8., 89 79.,00 Mean 12 28.,00 2.,33 5.,12 26.,24

32 J 3 0., 00 0., 00 0., 00 0., 00 10 3 0., 00 0., 00 0., 00 0., 00 20 3 3., 00 1..00 1..73 3.,00 50 3 23., 00 7.,67 6.,11 37.,33 Mean 12 26. 00 2.,17 4., 30 18.,52

Total Population 60 55.00 0.92 3.09 9.57 204

Table C.34. Number of larvae, produced which survived to the end of the experiment, by group size.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 0 0. 00 0. 00 0.00 26 3 0 0. 00 0.00 0. 00 28 3 0 0. 00 0. 00 0. 00 30 3 0 0. 00 0. 00 0. 00 32 3 0 0.00 0. 00 0.00 Group 15 0 0. 00 0. 00 0. 00

10 24 3 0 0. 00 0.00 0.00 26 3 0 0.00 0. 00 0.00 28 3 0 ' 0. 00 0.00 0.00 30 3 0 0. 00 0. 00 0. 00 32 3 0 0. 00 0. 00 0. 00 Group 15 0 0.00 0.00 0.00

20 24 3 0 0. 00 0. 00 0. 00 26 3 1 0. 33 0.58 0.33 28 3 0 0. 00 0.00 0.00 30 3 7 2. 33 4.04 16.33 32 3 3 1.00 1.73 3.00 Group 15 11 0.73 1.91 3.64

50 24 3 0 0.00 0.00 0.00 26 3 0 0. 00 0. 00 0.00 28 3 0 0. 00 0. 00 0. 00 30 3 21 7.00 8.89 79.00 32 3 23 7.67 6.11 37.33 Group 15 44 2. 93 5. 52 30.50

Total Population 60 55 0. 92 3.10 9.57 205

Table C.35. Final number of new larvae v by initial number of termites in group.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 0.00 0.00 0.00 0. 00 26 3 0.00 0.00 0. 00 0. 00 28 3 0. 00 0.00 0.00 0.00 30 3 0.00 0.00 0. 00 0.00 32 3 0.00 0.00 0.00 • 0.00 Group 15 0.00 0.00 0.00 C.00

10 24 3 0.00 0. 00 0. 00 0.00 26 3 0.00 0. 00 0. 00 0.00 28 3 0.00 0.00 0.00 0. 00 30 3 0. 00 0.00 0.00 0.00 32 3 0.00 0. 00 0. 00 0.00 Group 15 0.00 0. 00 0.00 0.00

20 24 3 0.00 0.00 0.00 0. 00 26 3 0.05 0.02 0.03 0.00 28 3 0. 00 0.00 0. 00 0. 00 30 3 0.35 0.12 0. 20 0.04 32 3 0.15 0.05 0. 09 0.01 Group 15 0.55 0.04 0.10 0.01

50 24 3 0. 00 0.00 0.00 0.00 26 3 0.00 0. 00 0.00 0.00 28 3 0.00 0.00 0.00 0.00 30 3 0.42 0.14 0.18 0.03 32 3 0.46 0.15 0.12 0. 02 Group 15 0.88 0.06 0.11 0.01

Total Population 60 1. 43 0. 02 0.08 0.01 206 Table C.36. Biomass of dead termites (mg dry wt) present at the end of the experiment, by temperature.

Tempera- Group Repli­ ture °C Size cates Sum Mean Std. Dev. Variance

24 5 3 17. 90 5.97 5.40 29.10 10 3 21. 00 7.00 8.27 68. 32 20 3 3. 60 1.20 1.20 1.44 50 3 0.00 0.00 0. 00 0.00 Mean 12 42.50 3.54 5. 27 • 27.76

26 5 3 13. 80 4. 60 6.47 41.88 10 3 3. 20 1.07 1.16 1.34 20 3 3.90 1.30 2.25 5.07 50 3 0. 00 0.00 0.00 0.00 Mean 12 20.90 1.74 3.47 12. 01

28 5 3 12. 00 4.00 3.86 14.89 10 3 6.70 2.23 3.53 12.44 20 3 0. 00 0.00 0. 00 0. 00 50 3 27.60 9.20 15.50 240.37 Mean 12 46.30 3. 86 7. 83 61.24

30 5 3 9.50 3.17 1. 02 1.04 10 3 1. 20 0.40 0.35 0.12 20 3 1. 90 0. 63 0.93 0. 87 50 3 0. 50 0.17 0.15 0. 02 Mean 12 13.10 1. 09 1.40 1.97

32 5 3 3. 00 1.00 1.73 3.00 10 3 12. 60 4.20 4.10 16.84 20 3 3. 60 1.20 2.08 4.32 50 3 15.10 5. 03 8. 63 74.50 Mean 12 34.30 2. 86 4.63 21.41

Total Population 60 157.10 2.62 4.93 24.32 207

Table C.37. Biomass of dead termites (mg dry wt) present at the end of the experiment, by group size.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 17.90 5.97 5.39 29.10 26 3 13. 80 4. 60 6.47 41. 88 28 3 12. 00 4.00 3. 86 14.89 30 3 9.50 3.17 1. 02 1.04 32 3 3. 00 1.00 1.73 3.00 Group 15 56.20 3. 75 3.97 15.76

10 24 3 21.00 7.00 8. 27 68. 32 26 3 3.20 1.07 1.16 1.34 28 3 6.70 ' 2.23 3.53 12.44 30 3 1.20 0.40 0.34 0.12 32 3 12.60 4.20 4.10 16.84 Group 15 44.70 2.98 4.50 20.26

20 24 3 3.60 1.20 1. 20 1. 44 26 3 3. 90 1.30 2.25 5. 07 28 3 0.00 0. 00 0. 00 0. 00 30 3 1.90 0. 63 0. 93 0.86 32 3 3. 60 1.20 2.08 4. 32 Group 15 13. 00 0. 87 1.39 1.93

50 24 3 0. 00 0.00 0. 00 0.00 26 3 0. 00 0.00 0. 00 0.00 28 3 27.60 9. 20 15. 50 240.37 30 3 0. 50 0.17 0. 15 0.02 32 3 15.10 5.03 8. 63 74.50 Group 15 43. 20 2. 88 7. 73 59.67

Total Population 60 157.10 2.62 4. 93 24.32 208

Table C.3 8. Carton (mg dry wt) produced during the experiment.

Tempera­ Group Repli­ ture °C Size cates Sum Mean Std. Dev. Variance

24 5 3 3. 90 1.30 2.25 5. 07 10 3 2.50 0. 83 1.44 2.08 20 3 0. 80 0.27 0.38 0.14 50 3 18. 80 6.27 7. 56 57.21 Mean 12 26.00 2.17 4.24 " 17.99

26 5 3 0.00 0.00 0.00 0. 00 10 3 10.60 3.53 0. 70 0.49 20 3 34.90 11.63 7.79 60.70 50 3 33. 90 '11.30 7.32 53.56 Mean 12 79.40 6.62 6.95 48.24

28 5 3 5.70 1. 90 1.67 2.78 10 3 48. 60 16. 20 12.45 155.07 20 3 33.10 11.03 6. 03 36.40 50 3 106.20 35. 40 22.87 523.03 Mean 12 193.60 16.13 17.15 294.00

30 5 3 2.30 0.77 1.33 1.76 10 3 46. 00 15.33 14. 04 197.00 20 3 72. 20 24.07 3.59 12.90 50 3 55.70 18.57 19.56 382.57 Mean 12 176.20 14.68 13.75 189.11

32 5 3 0.00 0. 00 0. 00 0. 00 10 3 10. 00 3,33 4. 54 20. 58 20 3 40. 20 13.40 6.40 40. 99 50 3 96.10 32. 03 18.64 347.37 Mean 12 146.30 12.19 15.62 244.06

Total Population 60 621.50 10.36 13. 26 175.69 209

Table C.39. Carton (mg dry wt) present at the end of the experiment.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance # 24 3 3.90 1.30 2.25 5.07 26 3 0.00 0.00 0.00 0.00 28 3 5.70 1. 90 1.66 2.77 30 3 2.30 0. 77 1.33 1.76 32 3 0. 00 0. 00 0. 00 0.00 Group 15 11.90 0. 79 1. 40 1.96

10 24 3 2.50 0. 83 1.44 2. 08 26 3 10.60 3. 53 0.70 0.49 28 3 48.60 '16.20 12.45 155.07 30 3 46.00 15.33 14.04 197.00 32 3 10. 00 3.33 4. 54 20.58 Group 15 117.70 7. 85 9. 97 99.46

20 24 3 0. 80 0.27 0.38 0.14 26 3 34.90 11.63 7.79 60.70 28 3 33.10 11. 03 6. 03 36.40 30 3 72.20 24.07 3. 59 12.90 32 3 40. 20 13.40 6. 40 40.99 Group 15 181.20 12. 08 9.11 82.94

50 24 3 18. 80 6. 27 7. 57 57.21 26 3 33.90 11. 30 7.32 53.56 28 3 106.20 35. 40 22. 87 523.03 30 3 55.70 18.57 19.56 382.57 32 3 96.10 32. 03 18. 64 347.37 Group 15 310.70 20. 71 18. 25 333.20

Total Population 60 621.50 10.36 13. 26 175.69 210

Table C.40. Material not utilized (NU) (mg dry wt) present at the end of the experiment, by temperature.

Tempera- Group Repli- ture °C Size cates Sum Mean Std. Dev. Variance

24 5 3 4.70 1. 57 2.71 7.36 10 3 0. 00 0. 00 0.00 0.00 20 3 0.00 0. 00 0. 00 0. 00 50 3 0.00 0.00 0.00 0.00 Mean 12 4.70 0.39 1.36 1.84

26 5 3 0.00 0. 00 0.00 0.00 10 3 0.90 0.30 0.27 0.07 20 3 3. 90 1.30 1. 54 2.37 50 3 0. 00 0.00 0. 00 0.00 Mean 12 4.80 0. 04 0.87 0.76

28 5 3 9.40 3.13 5. 43 29. 45 10 3 34.00 11.33 19. 63 385.33 20 3 17. 30 5.77 9.99 99.76 50 3 1.60 0. 53 0. 50 0. 25 Mean 12 62.30 5.19 10.54 111.05

30 5 3 7. 00 2. 33 4.04 16.33 10 3 16. 30 5.43 4. 69 22.01 20 3 6.90 2.30 3.98 15.87 50 3 0. 90 0. 30 0.52 0.27 Mean 12 31.10 2.59 3.69 13.58

32 5 3 5.10 1.70 2. 61 6. 79 10 3 15.10 5. 03 8.29 68.70 20 3 0.00 0. 00 0. 00 0. 00 50 3 0.40 0.13 0. 23 0.05 Mean 12 20. 60 1. 72 4. 27 18.22

Total Population 60 123.50 2.06 5.51 30.32 211 Table C.41. Material not utilized (NU) (mg dry wt) present at the end of the experiment, by group size.

Group Tempera­ Repli­ Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 4. 70 1.57 2.71 7.36 25 3 0. 00 0. 00 0.00 0.00 28 3 9.40 3.13 5.43 29.45 30 3 7.00 2.33 4. 04 16.33 32 3 5.10 1.70 2.61 6.79 Group 15 26.20 1.75 3.12 9.71

10 24 3 0.00 0.00 0.00 0.00 26 3 0. 90 0.30 0.27 0.70 28 3 34. 00 11.33 19. 63 385.33 30 3 16.30 5. 43 4.69 22. 01 32 3 15.10 5. 03 8.29 68.70 Group 15 66.30 4. 42 9.29 86.38

20 24 3 0.00 0. 00 0. 00 0.00 26 3 3.90 1.30 1. 54 2.37 28 3 17. 30 5.77 9.99 99. 76 30 3 6.90 2.30 3. 98 15. 87 32 3 0. 00 0. 00 0.00 0. 00 Group 15 28.10 1. 87 4.66 21.72

50 24 3 0.00 0.00 0.00 0. 00 26 3 0.00 0.00 0.00 0. 00 28 3 1. 60 0.53 0.50 0. 25 30 3 0. 90 0.30 0.52 0. 27 32 3 0.40 0.13 0.23 0. 05 Group 15 2. 90 0.19 0.36 0.13

Total Population 60 123.50 2. 06 5.51 30.32 212

Table C.42. Saguaro wood (mg dry wt) consumed by termites, by temperature.

Tempera­ Group Repli­ ture °C Size cates Sum Mean Std. Dev. Variance

24 5 3 441 147.00 79. 37 6300.00 10 3 656 218.67 60. 07 3608.33 20 3 1106 368.67 84. 01 7058.33 50 3 3071 1023.67 181. 48 32933.33 Mean 12 5274 438.50 374. 36 1'40143.18

26 5 3 531 177.00 34. 64 1200.00 10 3 1021 340.33 55. 08 3033.33 20 3 1836 612.00 35. 00 1225.00 50 3 5071 1690.33 352. 33 124133.33 Mean 12 8459 704.92 634. 83 403006.63

28 5 3 761 253.67 25. 17 633.33 10 3 1291 430.33 177. 86 31633.33 20 3 2191 730.33 49. 33 2433.33 50 3 4221 1407.00 284.78 81100.00 Mean 12 8464 705.33 481. 42 231760.61

30 5 3 746 258.67 94.12 8858.33 10 3 1501 500.33 85.05 7233.33 20 3 2738 912.67 120. 40 14496.33 50 3 6546 2182.00 224. 89 50575.00 Mean 12 11531 960.92 786. 29 618250.45

32 5 3 1066 355.33 27. 54 758.33 10 3 1356 452.00 44. 44 1975.00 20 3 2998 999.33 35. 30 1246.33 50 3 6301 2100.33 648. 56 420633.33 Mean 12 11721 976.75 775. 92 602058.02

Total Population 60 45449 757.48 641.77 411866.42 213

Table C.43. Saguaro wood (rag dry wt) consumed by termites, by group size.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

24 3 411 147. 00 79. 37 6300.00 26 3 531 177. 00 34.54 1200. 00 28 3 761 253.67 25..17 633.33 30 3 746 248. 67 94.12 8858. 33 32 3 1066 355. 33 27. 54 " 758. 33 Group 15 3545 236.33 90. 18 8131.67

10 24 3 656 218.67 60.07 3608. 33 26 3 1021 340. 33 55. 08 3033. 33 28 3 1291 430. 33 177. 86 31633. 33 30 3 1501 500.33 85. 05 7233. 33 32 3 1356 452. 00 44. 44 1975.00 Group 15 5825 388. 33 131. 83 17380.24

20 24 3 1106 368. 67 84. 01 7058.33 26 3 1836 612. 00 35. 00 1225.00 28 3 2191 730. 33 49. 33 2433. 33 30 3 2738 912. 67 120. 40 14496.33 32 3 2998 999. 33 35.30 1246. 33 Group 15 10869 724.60 239. 59 57404. 40

50 24 3 3071 1023.67 181. 48 32933. 33 26 3 5071 1690. 33 352. 33 124133. 33 28 3 4221 1407. 00 284. 78 81100. 00 30 3 6546 2182. 00 224. 89 50575.00 32 3 6301 2100. 33 648. 56 420633. 33 Group 15 25210 1680.67 549. 09 301501.67

lulation 60 45449 757. 48 641. 77 411866. 42 Table C.44. Fecal pellets (g) produced by termites, by temperature.

Temperature °C Group Size Replicates Sum Mean Std. Dev. Variance

24 5 3 72.,30 24.10 24.,47 598.68 10 3 145..50 48. 50 32.,39 1049. 23 20 3 364..70 121. 57 36.,63 1341.64 50 3 958.. 80 319.60 152.. 13 23143.71 Mean 12 1541..30 128 . 44 139.,44 19442.40 26 5 3 73..70 24. 57 17.. 94 321.96 10 3 297,,10 99. 03 11.. 85 140.42 20 3 523..90 174.63 20..90 436. 58 50 3 1951..40 650. 47 129.,54 16779. 84 Mean 12 2846., 10 237.18 261..53 68398.03 28 5 3 87.. 00 2 .00 18..38 337. 63 10 3 261.. 90 87. 30 47..77 2281. 77 20 3 690..30 230. 10 23..05 531. 16 50 3 1535..80 511.93 122..16 14921. 84 Mean 12 2575..00 214. 58 203..17 41276.80 30 5 3 42..10 14.03 16,.47 271. 37 10 3 372.. 80 124. 27 33.. 78 1140.90 20 3 809..00 269. 67 40..48 1638. 72 50 3 2472.. 60 824. 20 82,.53 6810. 67 Mean 12 3695..50 308. 04 328,.09 107640.18 32 5 3 199..10 66. 37 6.. 77 45. 80 10 3 148..20 74.10 14..71 216. 32 20 3 755..10 251. 70 73..59 5415.91 50 3 2251..70 750.57 279,.26 77984.08 Mean 12 3354..10 304.92 323,.83 104868.67 Total Population 60 14013..00 237. 51 259,.70 67444.09 Table C.45. Fecal pellets (g) produced by termites, by group size.

Group Size Temperature °C Replicates Sum Mean Std. Dev. Variance

5 24 3 72.30 24.10 24. 47 598.68 26 3 73.70 24.57 17.94 321.96 28 3 87.00 29.00 18 .38 337.63 30 3 42.10 14.03 16. 47 271.37 32 3 199.10 66.37 6 .77 45.80 Group 15 474.20 31.61 23. 97 574.30 10 24 3 145.50 .4 8.50 32.39 1049.23 26 3 297.10 99.03 11.85 140.42 28 3 261.90 87.30 47.77 2281.77 30 3 372.80 124.27 33.78 1140.90 32 UO 148.20 74.10 14.71 216.32 Group 15 1225.50 87.54 38.05 1447.51 20 24 3 364.70 121.57 36.63 1341.64 26 3 523.90 174.63 20.90 436.58 28 3 690.30 230.10 23.05 531.16 30 3 809.00 269.67 40.48 1638.72 32 3 755.10 251.70 73.59 5415.91 Group 15 3143.00 209.53 67.11 4503.40 50 24 3 958.80 319.60 152.13 23143.71 26 3 1951.40 650.47 129.54 16779.84 28 3 1535.80 511.93 122.16 14921.84 30 3 2472.60 824.20 82.53 6810.67 32 3 2251.70 750.57 279.26 77984.08 Group 15 9170.30 611.35 233.44 54495.36 Total Population 60 14013.00 237.51 259.70 67444.09 216 C— FU Table C.46. Assimilation efficiency (-^—) of termite groups, by temperature.

Tempera­ Group Repli­ ture °C Size cates Sum Mean Std.. Dev. Variance

24 5 3 2..58 0., 86 0.. 07 0..01 10 3 2,.39 0., 80 0,.10 0..01 20 3 2,. 02 0..67 0.. 05 0.. 00 50 3 2.. 06 0., 69 0.,13 0..18 Mean 12 9.. 05 0., 75 0.,12 0.,01

26 5 3 2., 60 0., 87 0.,09 0.,01 10 3 2..12 0.,71 0.. 05 0., 00 20 3 2..14 0.,71 0.. 04 0., 00 50 3 1.. 85 ' 0., 62 0., 00 0..00 Mean 12 8.,70 0., 73 0.,11 0.,01

28 5 3 2.,67 0.,89 0..07 0., 00 10 3 2.,42 0., 81 0.,34 0., 00 20 3 2.,10 0.,69 0.. 01 0., 00 50 3 1., 91 0.,64 0..01 0., 00 Mean 12 9.,10 0.,75 0.,11 0.,01

30 5 3 2., 86 0.. 95 0., 04 0.. 00 10 3 2., 26 0.,75 0., 06 0., 00 20 3 2.,12 0.,71 0., 01 . 0., 00 50 3 1.,87 0., 62 0.. 00 0..00 Mean 12 9..10 0. 76 0.,13 0.,02

32 5 3 2., 44 0., 81 0., 02 0., 00 10 3 1., 66 0., 83 0., 02 0.,00 20 3 2., 24 0.,75 0.. 08 0..01 50 3 1. 94 0. 65 0.. 02 0.. 00 Mean 12 8., 28 0.,75 0., 08 0.,01

Total Population 60 44.18 0.75 0.11 0.01 217 C-FU Table C.47. Assimilation efficiency (——) of termite groups, by group size.

Group Tempera- Repli- Size ture °C cates Sum Mean Std. Dev. Variance

5 24 3 2.58 0.86 0. 07 0.01 26 3 2.60 0. 87 0.09 0.01 28 3 2.66 0. 89 0. 07 0.00 30 3 2. 86 0.95 0.04 0.00 32 3 2.44 0. 81 0. 02 0.00 Group 15 13.15 0. 88 0.07 0.01

10 24 3 2.39 0.79 0.10 0. 01 26 3 2.12 0.71 0. 05 0.00 28 3 2.42 0. 81 0.04 0. 00 30 3 2. 26 0.75 0. 06 0.00 32 3 1.66 0. 83 0.02 0. 00 Group 15 10. 85 0. 78 0. 07 0.01

20 24 3 2.02 0.67 0. 05 0.00 26 3 2.14 0. 71 0. 04 0. 00 28 3 2.06 0.69 0.01 0.00 30 3 2.12 0. 71 0. 01 0.00 32 3 2. 24 0.75 0. 08 0.01 Group 15 10.57 0. 71 0.05 0.00

50 24 3 2.06 0.69 0.13 0.02 26 3 1. 84 0. 62 0.00 0.00 28 3 1. 91 0.64 0.01 0.00 30 3 1. 87 0. 62 0.00 0. 00 32 3 1.94 0.65 0.02 0. 00 Group 15 9. 62 0. 64 0. 06 0.00

Total Population 60 44,18 0. 75 0.11 0. 01 REFERENCES

Alcorn, S. M. (1976). Professor of Plant Pathology, The University of Arizona, personal communica-cion.

Alcorn, S. M., and May, C. (1962). Attrition of a saguaro forest. Plant Disease Reporter, 46, 156-158.

Allison, L. E., Bollen, W. B., and Moodie, C. D. (1965). Total Carbon. In Methods of Soil Analysis, ed. C. A. Black, part 2, pp. 1346-1366. Madison-, Wisconsin: Monograph Number 9. The American Society of Agronomy.

Atwater, W. 0., and Benedict, F. G. (1905). A respiration calorimeter with appliances for the direct determina­ tion of oxygen. Washington, D. C.: Carnegie Institute of Washington, Publication 42. 193 pp.

Bartlet, J. C., and Iverson, J. L. (19 66). Estimation of fatty acid composition by gas chromatography using peak heights and retention time. Journal of Association of Official Analytical Chemists, 49, 21-27.

Basalingappa, S. (1970). The differential occurrence of fat in the various castes and undifferentiated instars of the termite Odontotermes assmuthi: The result of discriminate feeding. Journal of Animal Morphology and Physiology, 17, 106-110.

Beard, R. L. (1972). Ants as predators of Reticulitermes flavipes. Environmental Entomology, 2, 397-399.

Becker, G. (1964). Termiten-anlockende Wirkung einiger bei Basidiomyceten-Angriff in Holzentstehender Verbindungen. Holzforschung, 18, 168-172.

Becker, G. (1969). Rearing of termites and testing methods used in the laboratory. In Biology of Termites, ed. K. Krishna and F. M. Weesner, Vol. 1, pp. 351-385. New York: Academic Press.

218 219

Becker, G. (1971). Physiological influences on wood- destroying insects of wood compounds and substances produced by microorganisms. Wood Science and Technology, 5, 236-246.

Beckwith, T. D., and Rose, E. J. (1929). Cellulose digestion by organisms from the termite gut. Proceedings of the Society for Experimental Biology and Medicine, 27, 4-6.

Bender, A. E. , and Doell, B. Ii. (1957). Biological evalua­ tion of proteins: A new aspect. British Journal of Nutrition, 11, 140-148.

Benson, L. (19 69). The Cacti of Arizona. 3rd ed. Tucson: The University of Arizona Press. 218 pp.

Berry, J. W., Anmin, H., and Steelink, C. (1960). Constituents of the Saguaro (Carnegiea gigantea). I. Proximate analysis of the woody tissues. Journal of Organic Chemistry, 25, 1267-1268.

Bletchly, J. D. (1969). Seasonal differences in nitrogen content of Scots Pine (Pinus sylvestris) sapwood and their effects on the development of the larvae of the common furniture beetle (Anobium punctatum DeG. ). Journal, Institute of Wood Science, 4~7 43-47.

Bodenheimer, A. (1951). Insects as Human Food. The Hague: W. Junk. 352 pp.

Bodine, M. C., and Ueckert, D. N. (1975). Effect of desert termites on herbage and litter in a shortgrass ecosystem in West Texas. Journal of Range Manage­ ment, 28, 353-358.

Breznak, J. A., Brill, W. J., Mertins, J. W., and Coppel, H. C. (1973). Nitrogen fixation in termites. Nature, London, 244, 577-580.

Brian, M. V. (1971). Ants and Termites. In Methods of Study in Quantitative Soil Ecology: Population, Produc­ tion and Energy Flow, ed. J. Phillipson, pp. 247- 2 61. Oxford and Edinburgh: Blackwell Scientific Publications.

Brody, S. (1945). Bioenergetics and Growth. New York: Reinhold Publishing Corporation. 1023 pp. 220

Browning, B. L. (1952). Introduction and miscellaneous determinations. In Kocd Chemistry, L. E. Wise and E. C. Jahn, Vol. 2, pp. 1119-1137. New York: Reinhold Publishing Corporation.

Browning, B. L. (1969). Wood Chemistry. In Pulp and Paper Technology, ed. K. W. Britt, 2nd ed., pp. 3-12. New York: Van Nostrcind Reinhold Company.

Buchner, P. (1965). Endosymbiosis of Animals with Plant Microorganisms^ Revised English Version. New York, London, Sidney: Interscience Publishers. 909 pp.

Caldwell, R. L. (1966). A biochemical investigation of callus tissue in the saguaro cactus [Carnegiea gigantea (Engelm.) Britt. & Rose]. Doctoral Dissertation. Tucson: The University of Arizona. 121 pp.

Carlson, S. D. (1966). Respiration measurement of Tribolium confusum by ga.s chromatography. Journal of Economic Entomology, 59, 335-338.

Carter, F. L., Dinus, L. A., and Smythe, R. V. (1972a). Effect of wood decayed by Lenzites trabea on the fatty acid composition of the eastern subterranean termite, Reticulitermes flavipes. Journal of Insect Physiology, 18, 1387-1393.

Carter, F. L., Dinus, L. A., and Smythe, R. V. (1972b). Fatty acids of the eastern subterranean termite, Reticulitermes flavipes (Isoptera: Rhinotermitidae). Annals of the Entomological Society of America, 65, 655-658.

Christie, W. M. (1973). Lipid Analysis. Oxford: Pergamon Press. 327 pp.

Cleveland, L. R. (1923). Symbiosis between termites and their intestinal protozoa. Proceedings of the National Academy of Sciences, U.S.A., 9, 424-428.

Cleveland, L. R. (1924). The physiological and symbiotic relationships between the intestinal protozoa of termites and their host, with special reference to Reticulitermes flavipes Kollar. Biological Bulletin, Marine Biological Laboratory, Woods Hole, 46, 178- T2T. 221

Cleveland, L. R. (1925a). The effects of oxygenation and starvation on the symbiosis between the termite, Termopsis, and its intestinal flagellates. Biological Bulletin, Marine Biological Laboratory, Woods Hole, 48, 309-326.

Cleveland, L. R. (1925b). The ability of termites to live perhaps indefinitely on a diet of pure cellulose. Biological Bulletin, Marine Biological Laboratory, Woods Hole, 48, 289-293.

Cleveland, L. R. (1926). Symb.fosis among animals with special reference to termites and their intestinal flagellates. The Quarterly Review of Biology, 1, 51-60.

Cleveland, L. R. (1928). Further, observations and experi­ ments on the symbiosis between termites and their intestinal protozoa. Biological Bulletin, Marine Biological Laboratory, Woods Hole, 54, 231-237.

Cmelik, S. H. W. (1969a). The neutral lipids from various organs of the termite Macrotermes goliath. Journal of Insect Physiology, 15, 839-849.

Cmelik, S. H. W. (1969b). Composition of the neutral lipids from termite queens. Journal of Insect Physiology, 15, 1481-1487.

Cmelik, S. H. W. (1972). Some properties of the neutral lipids from the workers of the termite Macrotermes falciger. Insect Biochemistry, 2, 361-366.

Coaton, W. G. H. (194 8). Cryptotermes brevis, a new wood- borer problem in South Africa. Bulletin of the Department of Agriculture and Forestry of South Africa, 290, 1-18. k ' Coaton, W. G- II. (1951). The snouted harvester termite; natural mortality as an aid to chemical control. Farming in South Africa, 26, 263-367.

Collins, M. S. (1951). Variations in the fat body of Reticulitermes flavipes (Kollar). Anatomical Record, 3, 477.

Cook, S. F. (1932). The respiratory gas exchange in Termopsis nevadensis. Biological Bulletin, Marine Biological Laboratory, Woods Hole, 63, 246-257. 222

Cook, S. F. (1943). Nonsymbiotic utilization of carbo­ hydrates by the termite, Zootermopsis angusticollis. Physiological Zoology, 16, 123-128.

Cook, S. F., and Smith, R. E. (1942). Metabolic relations in the termite-protozoa symbiosis: Temperature effects. Journal of Cellular and Comparative Physiology, 4, 95-110.

C3te, W. A., Jr. (1968). Chemical composition of wood. In Principles of Wood Science and Technology, I Solid Wood, ed. F. F. P. Kollmann and W. A. Cote, Jr., pp. 55-78. New York: Springer-Verlag.

Cowling, E. B. (1961). Comparative biochemistry of the decay of sweetgum sapwood by white-rot and brown-rot fungi. United States Department of Agriculture Technical Bulletin No. 1258. 7 9 pp.

Cowling, E. B., and Merrill, W. (1966). Nitrogen in wood and its role in wood deterioration. Canadian Journal of Botany, 44, 1539-1554.

Dadd, R. H. (1973). Insect nutrition: Current developments and metabolic implications. Annual Review of Entomology, 18, 381-420.

Damaschke, K.. , and Becker, G. (1964). Kor^elation der Atmungsintensitat von Termiten zu Anderrungen der Impulsfolgefrequenz der Atmospherics. Zeitschrift flir Naturforschung, 19b, 157-160.

Darnell, R, M. (1968). Animal nutrition in relation to secondary production. American Zoologist, 8, 83-93.

DeFoliart, G. R. (1975). Insects as a source of protein. Bulletin of the Entomological Society of America, 21, 161-163.

Dickman, A. (1931). Studies on the intestinal flora of termites with reference to their ability to digest cellulose. Biological Bulletin, Marine Biological Laboratory, Woods Hole, 61, 85-92.

Dimmitt, M. A. (1974). Research Associate, Cabrillo College, Aptos, California, personal communication. 223

D.uBois, E. F. (1936). Basal Metabolism in Health and Disease. Philadelphia: Lee and Febiger. 494 pp.

Eggum, B. 0. (.1970). Nutritional evaluation of proteins by laboratory animals. In Proceedings of the Inter­ national Biological Programme (IBP) and Wenner-Gren Center Symposium held in Stockholm, September, 1968, Evaluation of Novel Protein Products, ed. A. E. Bender et al., pp. 117-128. Oxford: Pergamon Press.

Emerson, A. E. (1955). Geographical origins and dispersions of termite genera. Fieldiana: Zoology, 37, 465- 521.

Engelmann, M. D. (1961). The role of soil arthropods in the energetics of an old-field community. Ecological Monographs, 31, 221-238.

Engelmann, M. D. (1966). Energetics, terrestrial field studies, and animal productivity. In Advances in Ecological Research, ed. J. B. Cragg, Vol. 3, pp. 73-115. London and New York: Academic Press.

Evans, F. C. (1956). Ecosystem as the basic unit in ecology. Science, 123, 1127-1128.

Folch, J. M., Lees, M., and Stanley, G. H. S. (1957). A simple method for the isolation and purification of total lipides from animal tissues. Journal of Biological Chemistry, 226, 497-509.

Frost, D. V. (1959). Methods of measuring the nutritive value of proteins, protein hydrolyzates, and amino acid mixtures. The repletion method. In Protein and Amino Acid Nutrition, ed. A. A. Albanese, pp. 225-279. New York and London: Academic Press.

Gay, F. J., Greaves, T., Holdaway, F. G., and Wetherly, A. H. (1955). Standard laboratory colonies of termites for evaluating the resistance of timber, timber preservatives, and other materials to termite -attack. Commonwealth Scientific and Industrial Research Organization Bulletin No. 277. Melbourne: Commonwealth Scientific and Industrial Research Organization. 224

Ghidini, G. M. (1939). Studi sulle termiti. 6. Richerche sul quoziente respiratorio nella diverse caste di Reticulitermes lucifugus. Revista Di Bio.logia Coloniale Rome, 2, 385-399.

Gilbert, L. I. (1967). Lipid metabolism and function in insects. In Advances in Insect Physiology, ed. J. W. L. Beament et al., Vol. 4, pp. 69-211. London and New York: Academic Press.

Gilmour, D. (1940). The anaerobic gaseous metabolism of the termite Zootermopsis nevadensis Hagen. Journal of Cellular and Comparative Physiology, 15, 331-342.

Golley, F. B. (1S61). Energy values of ecological materials Ecology, 42, 581-584.

Gordon, H. T. (1968). Quantitative aspects of insect nutrition. American Zoologist, 8, 131-138.

Grasse, P. P. (1949). Ordr.x des Isopceres on Termites. In Traite de Zoologie, ed. P. P. Grasse, Vol. IX, pp. 408-544. Paris: Masson et Cie.

Green, J. W. (1963). Wood cellulose. In Methods in Carbohydrate chemistry, ed. R. L. Whistler, Vol. 3, pp. 9-21. New York and London: Academic Press.

Gregory, K. F., and Winter, H. C. (1965). Data reduction with constant-pressure respirometers. Analytical Biochemistry, 11, 519-531.

Hammel, H. T., and Hardy, J. D. (1963). A gradient type of calorimeter for measurement of thermo-regulatory responses in the dog. In Temperature: its Measure­ ment and Control in Science and Industry, ed. J. D. Hardy, Vol. 3, part 3, pp. 31-42. New York: Reinhold Publishing Corporation.

Harris, G. C. (1952). Wood resins. In Wood Chemistry, ed. L. E. Wise and E. C. Jahn, Vol. 1, pp. 590-617. New York: Reinhold Publishing Corporation.

Harris, L. E. (1966). Biological energy interrelationships and glossary of energy terms. National Academy of Sciences, National Research Council Publication 1411 Washington: Printing and Publishing Office, National Academy of Sciences. 35 pp. 225

Harvey, P. A. C1934). II. Life history of Kalotermes minor. In Termites and Termite Control, ed. C. A. Kofoid et al., 2nd ed., pp. 217-233. Berkeley: University of California Press.

Haverty, M. I. (1974). The significance of the subterranean termite, Heterotermes aureus (Snyder), as a aetritivore in a desert grassland ecosystem. Doctoral Dissertation. Tucson: The University of Arizona. 89 pp.

Haverty, M. I., La Fage, J. P., and Nutting, W. L. ,(1974). Seasonal activity and environmental control of foraging of the subterranean termite, Heterotermes aureus (Snyder) , in a desert grassland"! Life Sciences, 15, 1091-1101.

Haverty, M. I., and Nutting, W. L. (1974). Department of Entomology, The University of Arizona, unpublished data.

Haverty, M, I., and Nutting, W. L. (1975a). Density, dispersion, and composition of desert termite foraging populations and their relationship to superficial dead wood. Environmental Entomology, 4, 480-486.

Haverty, M. I., and Nutting, W. L. (1975b). Natural wood preferences of desert termites. Annals of the Entomological Society of America, 68, 533-536.

Haverty, M. I., and Nutting, W. L. (1975c). A simulation of wood consumption by the subterranean termite, Heterotermes aureus (Snyder), in an Arizona desert grassland. Insectes Sociaux, 22, 93-102.

Haverty, M. I., Nutting, W. L., and La Fage, J. P. (1974). Department of Entomology, The University of Arizona, unpublished data.

Haverty, M. I., Nutting, W. L., and La Fage, J. P. (1975). Density of colonies and spatial distribution of foraging territories of the desert subterranean termite, Heterotermes aureus (Snyder). Environ­ mental Entomology, 4, 105-109. 226

Hebrant, F. (1970). Etude du flux energetique chez deux especes du genre Cubitermes Wasmann (Isoptera, Termitinae), termites humivores des savanes tropicales de la region ethiopienne. Doctoral Dissertation. Louvain: Universite Catholique de Louvain. 227 pp.

Hendee, E. C. (1933). The association of the termites, Kalotermes minor, Reticulitermes hesperus, and Zootermopsis angusticollis with fungi. University of California Publications in Zoology, 39, 111-134.

Hendee, E. C. (1935). The role of fungi in the diet of the common damp-wood termite, Zootermopsis angusticollis. Hilgardia, 9, 499-525.

Hewitt, P. H., and Nel, J. J. C. (1969a). Toxicity and repellency of Chrysocoma tenuifolia (Berg) (Compositae) to the Harvester termite Hodotermes mossambicus (Hagen) (Hodotermitidae). Journal of the Entomological Society of Southern Africa, 32, 133-136.

Hewitt, P. H., and Nel, J. J. C. (1969b). The influence of group size on the sarcoscmal activity and the behavior of Hodotermes mossambicus alate termites. Insect Physiology, 15, 2169-2177.

Hewitt, P. H., Nel, J. J. C., and Schoeman, I. (1972). Some aspects of the biochemistry and behavior of Trinervitermes trinervoides (Sjostedt) reproductives. Journal of the Entomological Society of Southern Africa, 35, 111-117.

Honigberg, B. M. (1970). Protozoa associated with termites and their role in digestion. In Biology of Termites, ed. K. Krishna and F. M. Weesner, Vol. ^t pp. 1-36. New York and London: Academic Press.

Hungate, R. E. (1936). Studies on the nutrition of Zootermopsis. I. The role of bacteria and molds in cellulose decomposition. Zentralblatt fur Bakteriologie, Parasltenkunde und Infektionskrank- heiten. 2. Abt., 94, 240-249.

Hungate, R. E. (1938). Studies on the nutrition of Zootermopsis. II. The relative importance of the termite and the protozoa in wood digestion. Ecology, 19, 1-25. 227

Hungate, R. E. (1939). Experiments on the nutrition of Zootermopsis. III. The anaerobic carbohydrate dissimilation by the intestinal protozoa. Ecology, 20, 230-245.

Hungate, R. E. (1941). Experiments on the nitrogen economy of termites. Annals of the Entomological Society of America, 34, 467-489.

Hungate, R. E. (1943). Quantitative analyses on the cellulose fermentation by termite protozoa. Annals of the Entomological Society of America, 36, 730- 739.

Hungate, R. E. (1955). Mutualistic intestinal protozoa. In Biochemistry and Physiology of Protozoa, ed. S. H. Hunter and A. Lwoff, Vol. 2, pp. 149-199. New York: Academic Press.

Jermyn, M. A. (1955). Cellulose and hemicelluloses. In Modern Methods of Plant Analyses, ed, K. Paech and M. V. Tracey, pp. 197-225. Berlin, Gottingen, Heidelberg: Springer-Verlag.

Johnson, D. B., Moore, W. E., and Zank, L. C. (1961). The Spectrophotometry determination of lignin in small wood samples. Tappi, 11, 793-798.

Johnson, K. A., and Whitford, W. G. (1975). Foraging ecology and relative importance of subterranean termites in Chihuahuan desert ecosystems. Environmental Entomology, 4, 66-70.

Katzin, L. K., and Kirby, H., Jr. (1939). The relative weights of termites and other protozoa. The Journal of Parasitology, 25, 444-445.

Keister, M., and Buck, J. (1974). Respiration: Some exogenous and endogenous effects on rate of respiration. In The Physiology of Insecta, ed. M. Rockstein, 2nd ed., Vol. 6, pp. 469-509. New York and London: Academic Press.

Kim, J—0, and Kohout, F. J. (1975). Multiple regression analysis: Subprogram regression. In Statistical Package for the Social Sciences, ed. N^ HT Nie et al., 2nd ed., pp. 320-367. New York: McGraw-Hill Book Company. 228

Kirk, T. K. (1971). Effects of microorganisms on lignin. Annual Review of Phytopathology, 9, 185-210.

Kirk, T. K., and Harkin, J. M. (1972). Lignin biodegrada- tion and the bioconversion of wood. American Institute of Chemical Engineers, Annual Meeting, Paper No. 30a, 11 pp.

Konicek, J., Suurkuusk, J., and Wadso, I. (1971). A precise drop heat capacity calorimeter for small samples. Chemica Scripta, 1, 217-220.

Krishna, K. (1969). Introduction. In Biology of Termites, ed. K. Krishna and F. M. Weesner, Vol. 1, pp. 1-17. New York: Academic Press.

La Fage, J. P. (1975). Department of Entomology, The University of Arizona, unpublished data.

La Fage, J. P.f Crowder, L. A., and Watson, T. F. (1974). Amino acids and total nitrogen of diapause and non-diapause larvae of the pink bollworm, Pectinophora gossypiella. Annals of the Entomological Society of America, 67, 472-474.

La Fage, J. P., and Nutting, W. L. (In Press). Nutrient dynamics in termites. In Production Ecology of Ants and Termites, ed. M. V. Brian. London: Cambridge University Press.

La Fage, J. P., Nutting, W. L. , and I-Iaverty, M. I. (1973). Desert subterranean termites: A method for studying foraging behavior. Environmental Entomology, 2, 954-956.

Lasker, R. (1959). Cellulose digestion in insects. In Marine Boring and Fouling Organisms, ed. D. L. Ray, pp. 348-355. Seattle: University of Washington Press.

Lasker, R. , and Giese, A. C. (1956). Cellulose digestion by the silverfish Ctenolepisma lineata. Journal of Experimental Biology, 33, 542-553.

Lee, K. E., and Wood, T. G. (1971). Termites and Soils. London and New York: Academic Press. 251 pp.

Lehninger, A. L. (1970). Biochemistry. New York: Worth Publishers, Inc. 833 pp. 229

Leung, W.-T., W (1968). Food Composition Table for Use in Africa. United States Department of Health, Education and Welfare, Maryland, and Food and Agriculture Organization United Nations, Rome, ix + 3 06 pp.

Levi, M. P., Merrill, W., and Cowling, E. B. (1968). Hole of nitrogen in wood deterioration. VI. Mycelial fractions and model nitrogen compounds as substrates for growth of Polyporus versicolor and other wood- destroying and wood-inhabiting fungi. Phyto­ pathology, 58, 626-634.

Light, S. F. (1934). The southern and mountain dry-wood termites, Kalotermes hubbardi and Kalotermes marginipennis. In Termites and Termite Control, ed. C. A. Kofoid et al., 2nd ed., pp. 266-268. Berkeley: University of California Press.

Lincoln, F. B., and Mulay, A. S. (1929). The extraction of nitrogenous materials from pear tissues. Plant Physiology, 4, 233-250.

Lindeman, R. L. (1942). The trophic-dynamic aspect of ecology. Ecology, 23, 399-418.

Little, T. M., and Hills, F. J. (1972). Statistical Methods in Agricultural Research. Berkeley: University of California Agricultural Extension. 242 pp.

Lund, E. E. (1930). The effect of diet upon the intestinal fauna of Termopsis. University of California Publications in Zoology, 36, 81-96.

Ltischer, M. (1956). Die Luftneuerung im Nest der Termite Macrotermes natalensis (Haviland). Insectes Sociaux, 3, 273-276.

Lusk, G. (1915). A respiration calorimeter for the study of disease. Archives of Internal Medicine, 15, 793-

Mannesmann, R. (1969). Vergleichende Untersuchungen liber den Einfluss der temperatur auf die Darm-Symbionten von Termiten und iiber die regulatorischen Mechanismen bei der Symbiose, Part 1. Zeitschrift filr Angewandte Zoologie, 56, 385-440. 230

Mannesmann, R. (1972a). Temperature effects on hindgut- flagellates of Coptotermes formosanus Shiraki and Reticulitermes virginicus (Banks). Material und Organismen, 7, 205-214.

Mannesmann, R. (1972b). A comparison between cellulolytic bacteria of the termites Coptotermes formosanus Shiraki and Reticulitermes virginicus (Banks). International Biodeterioration Bulletin, 8, 104-111.

Marton, J. (Symposium Chairman). (1966). Lignin Structure and Reactions, Advances in Chemistry Series .59, ed. R. F. Gould, Washington: American Chemical Society Publications. 267 pp.

Matsumoto, T. (1976). The role of termites in an equatorial rain forest ecosystem of West Malaysia. Oecologia (Berlin), 22, 153-178.

Mauldin, T. K., Smythe, R. V. , and Baxter, C. E. (1972). Cellulose catabolism and lipid synthesis by the subterranean termite, Coptotermes formosanus. Insect Biochemistry, 2, 209-217.

Maynard, L. A. (1937). Animal Nutrition. New York: McGraw-Hill. 4 83 pp.

McBee, R. H. (1959). Termite cellulase. In Marine Boring and Fouling Organisms, ed. D. L. Ray, pp. 342-347. Seattle: University of Washington Press.

McBrayer, J. F. (1973). Exploitation of deciduous leaf litter by Apheloria montana (Diplopoda: Eurydesmidae). Pedobiologia, 13, 90-98.

McBrayer, J. F. , Reichle, D. E., and Witkamp, M. (1974). Energy flow and nutrient cycling in a cryptozoan food-web. Environmental Sciences Division, Oak Ridge National Laboratory Publication No. 575. Oak Ridge, Tennessee: National Technical Information Service. 78 pp.

McMahan, E. A. (1962). Laboratory studies of colony establishment and development in Cryptotermes brevis (Walker) (Isoptera: Kalotermitidae). Proceedings of the Hawaiian Entomological Society, 18, 145-153.

McMahan, E. A. (1966). Studies of termite wood-feeding preferences. Proceedings, Hawaiian Entomological Society, 19, 239-250. 231

McNeill, S., and Lawton, J. Ii. (1970). Annual production and respiration in animal populations. Nature, London, 225, 472-474.

Merrill, A. L. , and Watt, B. K. (1955). Energy Value of Foods. Agricultural Research Service, United States Department of Agriculture, Agriculture Handbook No. 74. Washington, D. C.: United States Government Printing Office. 105 pp.

Montalenti, G. (1927). Sull' allevamento di termiti senza i protozoi del 1'ampolla cecale. Accademia Nazionale dei lincei, Rome, Classe di Scienze Fische, Matematiche e Naturali, Rendiconti, Series 6, 6, 529-532.

Moore, B. P. (1969). Biochemical, studies in termites. In Biology of Termites, ed. K. Krishna and F. M. Weesner, Vol. 1, pp. 407-432. New York: Academic Press.

Munck, L. (1970). Mice as test animals for screening protein quality. In Proceedings of the Inter­ national Biological Programme (IBP) and Wenner-Gren Center Symposium held in Stockholm, September, 1968, Evaluation of Novel Protein Products, ed. A. E. Bender et al., pp. 143-147.

Nagin, R. (1972). Caste determination in Neotermes jouteli (Banks). Insectes Sociaux, 19, 39-61.

Nel, J. J. C-, Hewitt, P. H., and Joubert, L. (1971). Growth in relation to colony size in the harvester termite Hodotermes mossambicus (Hagen). Phyto- phylactica, 3, 1-6.

Niering, W. A., and Whittaker, R. H. (1963). Vegetation of the Santa Catalina Mountains. Progressive Agricul­ ture in Arizona, 15, 4-6.

Noirot, C. (1969). Formation of castes in the higher termites. In Biology of Termites, ed. K. Krishna and F. M. Weesner, Vol. 1, pp. 311-350. New York: Academic Press.

Noirot, C., and Noirot-Timothee, C. (1969). The digestive system. In Biology of Termites, ed. K. Krishna and F. M. Weesner, Vol. T~, pp. 49-88. New York: Academic Press. 232

Nutting, W. L. (1969). Flight and colony foundation. In Biology of Termites, ed. K. Krishna and F. M. Weesner, Vol. 1, pp. 233-282. New York: Academic Press.

Nutting, W. L. (1970). Composition and size of some termite colonies in Arizona and Mexico. Annals of the Entomological Society of America, 63, 1105-1110.

Odum, E. P., Marshall, S. G., and Marples, T. G. (1965). The caloric content of migrating birds. Ecology, 46, 901-904.

Odum, H. T. (1957). Trophic structure and productivity of Silver Springs. Ecological Monographs, 27, 55-112.

Odum, H. T., and Odum, E. P. (1955). Trophic structure and productivity of a windward coral reef community on Eniwetok Atoll. Ecological Monographs, 25, 291-320.

Oser, B. L. (1959). An integrated essential amino acid index for predicting the biological value of proteins. In Protein and Amino Acid Nutrition, ed. A. A. Albanese, pp. 281-295. New York and London: Academic Press.

Oser, B. L., ed. (1965). Hawk's Physiological Chemistry. 14th ed. New York: McGraw-Hill Book Company, The Blakiston Division. 1472 pp.

Paris, 0. H., and Sikora, A. (1967). Radiotracer analysis of the trophic dynamics of natural isopod popula­ tions, In Secondary Productivity of Terrestrial Ecosystems, ed. I<. Petrusewicz, Vol. 2, pp. 741-771. Warszawa and Krakow: Panstwowe Wydawnictwo Naukowe.

Peakin, G. J. (1973). The measurement of the costs of maintenance in terrestrial poikilotherms: A compari­ son between respirometry and calorimetry. Experientia, 29, 801-802.

Pence, R. J. (1956). The tolerance of the drywood termite, Kalotermes minor Hagen to desiccation. Journal of Economic Entomology, 49, 553-554.

Petrusewicz, K. (1967). Concepts in studies on the secondary productivity of terrestrial ecosystems. In Secondary Productivity of Terrestrial Ecosystems, ed. K. Petrusewicz, Vol. 1, pp. 17-4 9. Warszawa and Krakow: Panstwowe Wydawnictwo Naukowe. 233

Petrusewicz, K., and Macfadyen, A. (1970). Productivity of Terrestrial Animals. Philadelphia: F. A. Davis Company. 190 pp.

Phillipson, J. (1966). Ecological Energetics. New York: St. Martin's Press. 57 pp.

Prat, H. (1954). Analyse micro-calorimetrique des variations de la thermogenese chez divers insectes. Canadian Journal of Zoology, 32, 172-197.

Prat, H. (.1975) . Faculte des Sciences de Marseile Laboratorie de Botanique, personal communication.

Prouty, L. R. , Barrett, M. J., and Hardy, J. D. (1949). A simple calorimeter for the simultaneous determination of heat loss and heat production in animals. The Review of Scientific Instruments, 20, 357-363.

Rennie, P. J. (1965). The determination of nitrogen in woody tissue. Journal of the Science of Food and Agriculture, 16, 629-638.

Ritter, G. J. , and Barbour, J. II. (1935). Effect of pre- treatments of wood on the lignin determination. Industrial and Engineering Chemistry, Analytical Edition, 7, 238-240.

Ritter, H. T. M. (1955). A study of the biology of certain intestinal associates of Retlculitermes flavipes. Doctoral Dissertation. Bethlehem, Pennsylvania: Lehigh University. 166 pp.

Robinson, T. (1963). The Organic Constituents of Higher Plants, Their Chemistry and Interrelationships. Minneapolis: Burgess Publishing Company. 306 pp.

Rogozinski, M. (1964). A rapid quantitative esterification technique for carboxylic acids. Journal of Gas Chromatography, 2, 136-137.

Rohrmann, B. (1975). Entomologist, U. S. Department of Agriculture, Forest Service, Forestry Sciences Laboratory, Covalis, Oregon, personal communication.

Ruddle, K. (1973). The human use of insects: examples from the Yukpa. Biotropica, 5, 94-101. 234

Ruelle, J. (1962). Etude de guelgues variables du micro- climat du nid de Macrotermes natalensis (Hav.) (Isoptera, Termitidae) en rapport avec le declenchement de 11essaimage. Doctoral Disserta­ tion, Kinshasa. 120 pp.

Saeki, I., Sumimoto, M. , and Kondo, T. (1973). The termiticidal substances from the wood of Chaxnaecyparis pisifera D. Don. Holzforschung, 27, 93-96.

Schubert, W. J. (1965). Lignin Biochemistry. New York and London: Academic Press. 131 pp.

Schiirch, A. F. , Lloyd, L. E. , and Crampton, E. W. (1950). The use of chromic oxide as an index for determining the digestibility of a diet. Journal of Nutrition, 629-636.

Scurfield, G., and Nicholls, P. W. (1970). Amino-acid composition of wood proteins. Journal of Experi­ mental Botany, 21, 857-868.

Sebrell, W. H., ed. (1963). Evaluation of protein quality. National Academy of Sciences, National Research Council Publication 1100. Washington: Printing and Publishing Office, National Academy of Sciences. 74 pp.

Sellers, W. D., ed. (1960). Arizona Climate. Tucson: The Institute of Atmospheric Physics, University of Arizona.

Snyder, T. E. (1954). Order Isoptera, the Termites of the United States and Canada. Nev York: Technical Bulletin of the National Pest Control Association. 64 pp.

Stecher, P. G., cd. (196 8). The Merck Index, an Encyclopedia of Chemicals and Drugs. 8th ed. Rahway, N. J.: Merck & Co., Inc. 1713 pp.

Steel, R. G. D., and Torrie, J. H. (1960). Principles and Procedures of Statistics. New York: McGraw-Hill Book Company. 481 pp.

Strickland, J. D. H., and Parsons, T. R. (1968). A Practical Handbook of Seawater Analysis. Fisheries Research Board of Canada Bulletin No. 167. Ottawa: Fisheries Research Board of Canada. 311 pp. 235

Tassoni, J. P. (1951). The nitrogen to protein conversion factors for the pupae of the moth, Telea polyphemus Cramer. Anatomical Record, 3, 446.

Teal, J. M. (1957). Community metabolism in a temperate cold spring. Ecological Monographs, 27, 283-302.

Teotia, J. S., and Miller, B. F. (1973). Fly pupae as a dietary ingredient for starting chicks. Poultry Science, 52, 1830-1835.

Tetrault, P. A., and Weis, W. L. (1937). Cellulose decomposition by a bacterial culture from the intestinal tract of termites. Journal of Bacteriology, 33, 95-96.

Tihon, L. (1946). A propos des termites au point de vue alimentaire. Bulletin Agricole du Congo Beige, 37(4), 865-868.

Trager, W. (1932). A cellulase from the symbiotic intestinal flagellates of termites and of the roach, Cryptocercus punctulatus. Biochemical Journal, 26, 1762-1771.

Tsoumis, G. (1968). Wood as Raw Material. Oxford: Pergamon Press. 27 6 pp.

Umbreit, W. W., Burris, R. H., and Stauffer, J. F., ed. (1972). Manometric & Biochemical Techniques. 5th ed. Minneapolis: Burgess Publishing Company. 375 pp.

Usher, M. B. (1974). Runway building and food exploration strategy of Pseudacanthotermes militaris (Insecta: Isoptera). Biotropica, 6~, 154-157.

Watson, J. A. L., and Gay, F. J. (1970). The role of grass- eating termites in the degradation of a mulga ecosystem. Search, 1, 43. n Weber, C. W. (1976). Professor of Poultry Science, The / University of Arizona, personal communication.

Whitford, W. G. (1976). Professor of Biology in the Department of Biology, New Mexico State University, Las Cruces, N. M., personal communication. 236

Whitney, W. K., and Ortman, E. E. (1962). Gas chromatography for measuring insext respiration. Journal of the Kansas Entomological Society, 35, 323-332.

Wiegert, R. G. (1968). Thermodynamic considerations in animal nutrition. American Zoologist, 8, 71-81.

Wiegert, R, G., and Coleman, D. C. (1970). Ecological significance of low oxygen consumption and high fat accumulation by Nasutitermes costalis (Isoptera: Termitidae). Bioscience, 20, 663-665.

Wilkinson, W. (1962). Dispersal of alates and establish­ ment of new colonies in Cryptotermes havilandi (SjSstedt) (Isoptera, Kalotermitidae). Bulletin of Entomological Research, 53, 265-286.

Wilson, E. 0. (1971). The Insect Societies. Cambridge, Massachusetts: The Belknap Press of Harvard University Press. 548 pp.

Wise, L. E. (1952). Miscellaneous extraneous components of wood. In Wood Chemistry, ed. L. E. Wise and E. C. Jahn, Vol. 1, pp. 63 8-6 88. New York: Reinhold Publishing Corporation.

Wolcott, G. N. (1946). Factors in the natural resistance of woods to termite attack. Caribbean Forester, 7, 121-134.

Wood, T. G. (In Press). The food and feeding habits of termites. In Production Ecology of Ants and Termites, ed. M. V. Brian. London: Cambridge University Press.