GARIMA KAKKAR UNIVERSITY OF FLORIDA

Transcription

1 TEMPORAL AND SPATIAL ASSESSMENT OF MOLTING IN WORKERS OF COPTOTERMES FORMOSANUS SHIRAKI: AN APPROACH TO SPEED UP THE COLONY ELIMINATION WITH THE USE OF CHITIN SYNTHESIS INHIBITOR BAITS By GARIMA KAKKAR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2015

2 2015 Garima Kakkar

3 To termite biologists and her If seasons can change, so can I

4 ACKNOWLEDGMENTS The list is long, but I would like to begin by thanking my major advisor Dr. Nan Yao Su who gave me the opportunity to work on this project. His constant support, and encouragement boosted me and kept me focused towards my goal. Special thanks to graduate committee members Drs. Robin M Giblin-Davis, William Kern, and Henry Hochmair for their unlimited and timely support. My sincere thanks to subterranean termite lab members Aaron Mullins, Hou-Feng Li, Ron Pepin, and Thomas Chouvenc who played an important role in my project behind the scenes. I would like to thank Sarah Kern for her help and support in every possible way throughout this program. My sincere gratitude to Mun-Wye Chung, Tiago Carrijo, Angelica Moncada, Kelly Ugarelli, Veena Sivaramakrishnan, Gurpreet Kaur, Stephanie Osario, Sarah Bernard, Ruth Passernig, Levente Juhasz, Sreten, Majid Alivand, Meike Saskia Kruger, Chintan Shukla, and Du He for their friendship, without them these four years would not have been easy. My gratitude to my father Kuldeep Singh Kakkar, who once asked if I will ever do Ph.D. Thanks for sowing the seed of Ph.D. in my mind. Thanks to my mother who worked hard so that I could fulfil my dreams, thanks to my mother-in-law for covering my needs during the toughest few months of my life. My sincere gratitude to my brother for the motivation. Last but not the least, I thank my partner Dr. Vivek Jha for his unconditional love and support, his time for listening to my termite stuff even when I did not bother to listen to his whitefly and thrips projects. Thanks mate for putting up with me for 15 years now. I am grateful to the almighty for this day and that I am able to write these acknowledgments. Thanks to everyone. 4

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS... 4 LIST OF FIGURES... 7 LIST OF ABBREVIATIONS... 9 ABSTRACT CHAPTER 1 LITERATURE REVIEW DETERMINING MOLTING INCIDENCE IN FORMOSAN SUBTERRANEAN TERMITES (ISOPTERA: RHINOTERMITIDAE) BY POST-ECDYSIS SCLEROTIZATION Introduction Materials and Methods Results Overall Observations Sclerotization of the Primary Point of Articulation of the Mandible Sclerotization of the Secondary Point of Articulation of the Mandible Sclerotization of the Left Mandible, Covered by the Labrum Sclerotization of the Mandibles, Without Labrum Width of Sclerotization of the Apical Tooth Discussion TEMPORAL ASSESSMENT OF MOLTING IN WORKERS OF FORMOSAN SUBTERRANEAN TERMITES (ISOPTERA: RHINOTERMITIDAE) Introduction Materials and Methods Molting Frequency of Workers in a Juvenile Colony Molting Frequency of Workers from Foraging Populations Results Molting Frequency of Workers in a Juvenile Colony Molting Frequency of Workers from Foraging Populations Discussion FASTING PERIOD AND TIME FOR MORTALITY Introduction Materials and Methods Results

6 Discussion SPATIAL ASSESSMENT OF MOLTING IN COPTOTERMES FORMOSANUS WORKERS (ISOPTERA: RHINOTERMITIDAE) Introduction Material and Methods Colony Rearing Site of Molting Nest Fidelity Results Study-1 Site of Molting Study-2 Nest Fidelity Discussion CONCLUSIONS LIST OF REFERENCES BIOGRAPHICAL SKETCH

7 LIST OF FIGURES Figure page 2-1 Frontal view of the head capsule and left mandible of a worker of C. formosanus Lateral view of the teneral stage in a worker of C. formosanus ( Jackknife position) Frontal view of the head capsules of C. formosanus at 0 h, 4 h, 8 h, 16 h, 20 h, 24 h, 36 h post-ecdysis and at intermolt stage Progression of the index of sclerotization for three variables from 0 h postecdysis until 36 h post-ecdysis at 4 h intervals when labrum was present Progression of the index of sclerotization of mandible teeth and for the width of sclerotized region for the apical tooth of the mandible (µm) from 0 h to36 h post-ecdysis at 4 h intervals when labrum was removed Planar arena (24 x 24 x 0.6 cm in thickness) filled with moistened sand for molting frequency of workers in a juvenile colony Average percentage of molting per day in three, 4-year old juvenile colonies Days taken to complete molting cycle 1 and 2 using field collected foraging population of three colonies ( A, B, and C) at 27 C Mean cumulative percentage of workers molted in three colonies for two cycles at 27 C and 21 C The extended foraging arena- The foraging arena was composed of 6 small arenas connected to each other by a 6 m long coiled Tygon tubing to form a linear distance of 30 m (a) Picture of white worker that died in the jackknife posture, (b) white worker that died with exuviae-wrapped posture, (c) blue worker that died in jackknife posture, (d) blue worker that died in exuviae-wrapped posture Average number of molting workers (white= unfed on treatment and blue= fed on treatment) in extended foraging setup for control (a), and noviflumuron (b) treatments during a 9 wk study Average number of workers (white= unfed and blue= fed)) showing molt inhibitory effect of noviflumuron (jackknife or exuviae-wrapped) in extended foraging setup for (a) control, and (b) noviflumuron treatment during a 9 wk study

8 4-5 Percentage of worker mortality recorded in 0-30 m extended foraging setup for (a) control, and (b) noviflumuron treatment during a 9 wk study Central planar arena (60 x 60 x 0.9 cm in thickness) filled with moistened sand and extended in three directions (X, Y, and Z) a) Picture of worker in premolt stage with separated exuviae from the epidermis, b) worker in molting stage in a jackknife posture, c) worker in newmolt/postmolt stage Central planar arena (60 x 60 x 0.9 cm in thickness) filled with moistened sand and extended 15 m in one direction through Tygon tubes Percentage of workers in four chronological categories of molting in the extended foraging arena Box plot of various distances. The lower and upper boundary of the box indicate the 25 th and 75 th percentile respectively

9 LIST OF ABBREVIATIONS CSI JHA JHM MAC Chitin Synthesis Inhibitor Juvenile hormone analog Juvenile hormone mimic Molt Accelerating Compound 9

10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TEMPORAL AND SPATIAL ASSESSMENT OF MOLTING IN WORKERS OF COPTOTERMES FORMOSANUS SHIRAKI: AN APPROACH TO SPEED UP THE COLONY ELIMINATION WITH THE USE OF CHITIN SYNTHESIS INHIBITOR BAITS Chair: Nan Yao Su Major: Entomology and Nematology By Garima Kakkar December 2015 Reduction in the time for elimination of colonies of subterranean termites with baits may decrease the cost of their control and yield high economic benefits. Because the time to molt for workers affect the lethal time of a chitin synthesis inhibitor (CSI) bait, premature molt initiation in workers and disrupting ecdysis using CSI can be a potential method for speeding up the elimination process. Before testing compounds for premature molt initiation in workers of Coptotermes formosanus Shiraki, it is imperative to determine the frequency of molting amongst workers and ensure that speeding up the mortality (= molting) will not cause aversion to bait in response to corpses of termites that died close to the bait station due to acceleration of the molting process. A methodology to distinguish recently molted workers from non-molting workers was developed based on the changes of sclerotization of the mouthparts. Using this technique, studies assessing time, and site of molting in workers of C. formosanus were conducted. Molting frequency amongst lab-raised juvenile colonies and time lapse between two consecutive molts for second and third instar workers was determined. 10

11 The estimated time served as a standard for subsequent studies focused on testing efficacy of molt accelerating compounds (MAC) in workers. Second and third instar workers in the foraging populations of field colonies were found to molt at an interval of 43 and 45 d, respectively. On treating the field collected foraging population with noviflumuron (0.5%) in extended arenas in the laboratory, workers in their fasting period during the initial 10 d after baiting did not acquire the lethal dose and thus molted successfully at the end of the fasting period. This resulted in extension of the time to mortality by another molt cycle i.e., extra days. The molting site fidelity by workers in a colony ensured that speeding up the time for mortality will not result in an inhibitory cascade of dead termites around the bait stations. Thus, speeding up the activity of CSI baits with the addition of MAC will not lead to secondary repellency. 11

12 CHAPTER 1 LITERATURE REVIEW Termites are a group of eusocial insects, belonging to the order Isoptera. The common name, termite has been derived from the Latin word termes, which means woodworm (Potter 1997). There are over 3,106 described species in this order and ~ 10% of these are known as important pests (Edwards and Mill 1986). The list of termite pests has been further narrowed to 80 species that are known to cause severe economic damage to structures, agriculture and forests (Rust and Su 2012), with 38 species belonging to subterranean termites of family Rhinotermitidae. Worldwide, these subterranean termites account for ~ $32 billion spent annually on their control and damage repair. Some of the important subterranean termite pests in the United States are Reticulitermes flavipes (Kollar), R. virginicus (Banks), R. hageni Banks, R. hesperus (Banks), Heterotermes aureus (Snyder), Coptotermes formosanus Shiraki, and C. gestroi (Wasmann). C. formosanus is an adventive pest in the United States. Its initial introduction has been linked to the transportation of infested material from Asia after World War II (Su and Tamashiro 1987). In North America, its first established population was reported from Hawaii (Tamashiro et al. 1973), whereas in the Continental US, the first established population was confirmed from Charleston South Carolina and later in adjoining states including Alabama, Florida, Georgia, Louisiana, Mississippi, North Carolina, Tennessee, and Texas (Woodson et al. 2001, Su 2003, CABI 2014) Subterranean termites of the family Rhinotermitidae, including C. formosanus mainly lives in soil and nest underground where the colony is composed of a main nest and/or satellite nests interconnected by a gallery system (King and Spink 1969). A 12

13 single colony may contain more than a million termites with galleries extending up to 100 m (Su and Scheffrahn 1988, Grace et al. 1989). Inside these galleries, subterranean termites forage for food and enter structures from the surrounding soil. Because of the large population size and cryptic nature of subterranean termites, it is hard to detect the invasion until there are external signs of infestation above ground. To date, liquid termiticides and baiting systems are the two main strategies in use for control of subterranean termites and these can be used as both prophylactic and remedial measures. Conventionally, repellent termiticides are applied to the soil providing a chemical barrier and preventing termites from entering the structure. These are applied in the soil by drenching or injecting the soil surrounding newly-constructed structures, inside the foundations, chimney bases, pipes, under filled porches and terraces. Termiticides with a repellent action in use today are mainly pyrethroids that include, permethrin, cypermethrin, and bifenthrin. The non-repellent termiticides around a building foundation and spot treatment upon infestation kills termites that come into contact with the treated area (Gahlhoff and Koehler 2001, Thorne and Breisch 2001, Su and Scheffrahn 1990a). These insecticides serve the purpose of both prophylactic and curative control by excluding termites from the structures. Although largely used, the liquid termiticides have certain limitations which make the treated structures vulnerable to future infestations such as through the gaps between the termiticide- treated areas which can provide a pathway to access and infest the structures. These gaps may be formed due to improper application, disturbances in the soil that create regions benign for termites, or termiticide degradation in soil over time (Su and Scheffrahn 1990b, Koehler et al. 2000). Another 13

14 limitation of liquid termiticides is that their lethal time depends on the lethal dose received by termites, which kills them early before the toxicant could be passed to the nestmates (Su and Scheffrahn 1988). Nevertheless, liquid termiticides are the predominant means used by the industry for control of subterranean termites and takes up to 77% of the market share (Anonymous 2002). To overcome the problems related to fast-acting or repellent termiticides, the slow-acting and non-repellent termiticides like chlorfenapyr, chlorantraniliprole, and imidacloprid, were introduced and became popular for termite control in the past few years (Neoh et al. 2014). These termiticides are moderate-to-less toxic and do not cause lethal effect to termites with immediate contact (Parman and Vargo 2010). However, in laboratory experiments (Potter and Hillery 2002, Su 2005, Saran and Rust 2007) on C. formosanus or Reticulitermes spp., individuals exposed to the termiticides (fipronil or thiamethoxam) either directly or through social contact were found to move ~5 m away from the treatment zone and died before passing the toxicant to the other individuals. The presence of dead termites around the treated area induced secondary repellency and lead to the division of the colony population into two sub-groups. The second control strategy available for subterranean termites is the baiting system incorporated with a non-repellent and slow-acting toxicant. Randall and Doody (1934) first reported the use of arsenic dust as a slow-acting toxicant. Later, Esenther and Gray (1968) proposed the use of baits impregnated with dechlorane (another slowacting toxicant) to eliminate a R. flavipes colony. With no subsequent studies thereafter on baits, interest in the use of such baits was renewed in 1980 s (Su et al. 1982, Jones 1984), and in 1995, chitin synthesis inhibitor (hexaflumuron) incorporated baits were 14

15 developed and commercialized for the control of subterranean termites. Baits incorporated with hexaflumuron (a chitin synthesis inhibitor (CSI) that is slow-acting with dose-independent lethal time) were tested in the laboratory and field and were found to be effective in eliminating colonies of R. flavipes and C. formosanus (Su and Scheffrahn 1993, Su 1994). Later, numerous other studies demonstrated the potency of CSI baits for elimination of subterranean termite colonies (Grace and Su 2001). The two chemical groupings of insecticides used as active ingredients in baits for subterranean termite control are CSI and metabolic inhibitors (MI). At a proper concentrations, both groups are slow-acting and non-repellent toxicants, which satisfies the basic requirement for the success of a control program for subterranean termites. However, baits incorporated with slow-acting MI (sulfuramid and hydramethylnon) did not eliminate field colonies of subterranean termites (Su and Scheffrahn 1991, Pawson and Gold 1996, Ripa et al. 2007). Similar to the non-repellent termiticides, the lethal time of MI baits is dependent on the dose ingested by the termites, which results in faster death of workers with higher doses before the toxicant is spread to the healthy workers in a colony (Su and Scheffrahn 1998). Commercially, these baits are used in combination with liquid applications. The CSIs used in baits for subterranean termite control include, diflubenzuron, chlorfluazuron, bistrifluron, triflumuron, hexaflumuron. These are insect growth regulators (IGR) that interfere with the formation of cuticle and affect the molting process of workers. Worker is an important caste of this eusocial group and it comprises the major population in a colony. A worker in a colony are involved in foraging, repair and maintenance of the nest, feeding and grooming brood and soldiers in the nest 15

16 (Krishna 1969). They grow by developing a new exoskeleton under their old skeleton, and upon reaching maturity, the old cuticle is shed and the individual (pharate) develops a new and flexible cuticle that allows expansion. Because workers are involved in maintenance of the nest and feeding their nest mates, their presence is critical for the survival of a colony (Kofoid 1946). By killing workers in a colony during the molting process, CSIs disrupt the social balance of termite colonies which eventually leads to the colony collapse. Unlike non-repellent termiticides and MIs, CSIs do not kill the insect until it molts and thus the procedure of eliminating a C. formosanus colony of > 1 million individuals may take two to nine months after baiting (Eger et al. 2012). Besides CSIs, the other class of IGRs tested on termites is juvenile hormone analogs (JHAs) or juvenile hormone mimics (JHMs). These are also known to disrupt the colony homeostasis by inducing excessive soldier production in a colony (Hrdy and Krecek 1972). High soldier: worker proportions lead to starved individuals in a colony and eventually the colony collapses. Because different termite species maintain different proportions of workers and soldiers in a colony, impact of juvenoids (JHA and JHM) on a colony can be highly variable. Studies suggest that juvenoids are effective in inducing high soldier proportion amongst species with naturally low soldier proportions, such as in Reticulitermes spp., (Su and Scheffrahn 1990c), and failed to cause similar effects on Coptotermes spp., which have high soldier proportions (Su 2003). Another class of molting hormones tested on termites for their control is ecdysone agonists. Ecdysone when applied to termites is known to induce molting amongst workers of lower termites (Luscher and Karson 1958). Halofenozide, one of the several commercially available ecdysone agonists is known as a molt accelerating compound 16

17 because of its premature-molt inducing properties in insects. When tested as a potential bait-active ingredient for termite control (Monteagudo 2004), it was found to have nonrepellent properties. Further experimentation indicated that workers of R. flavipes and C. formosanus when exposed to halofenozide were induced to molt and died without molting successfully (Su et al. 2011). Although these results suggest a potential use of ecdysteroids as bait toxicants, it is important to ensure that lethal time of ecdysone agonists is not dose dependent to avoid issues of secondary repellency, commonly seen in other available control measures. Currently, CSI bait is considered the most successful method for elimination of subterranean termite colonies. But, to maintain the suppression and continuous availability of the active ingredient to termite colonies in an area with high pest pressure, baits have to be quarterly (3-mo) replenished and monitored for a long time for control, which increases the cost of implementation. Thus, any reduction in the time for elimination can be economically beneficial (Su and Scheffrahn 1998). However, considering that the slow process of colony elimination, avoids secondary repellency and supports the spread of toxicant to a large population over a distance of several meters, expediting the colony elimination can reduce the performance of baiting system. Total time for colony elimination using CSI has been divided into three segments: 1) bait interception time, 2) lethal-dose acquisition time, and 3) lethal time. The bait interception time is the time taken by a foraging population of a colony to locate the bait station and recruit others to the bait. The time taken to discover the bait station is highly variable among colonies and it largely depends on the foraging pattern of a colony and population pressure in the area. 17

18 Lethal-dose acquisition is the time spent for the majority of termites to acquire lethal dose. It is highly variable in the natural habitat of termites depending upon the consumption rate or behavior of a colony (Su and La Fage 1984). Size of the colony also affects the acquisition time, as for a large population size, more time will be taken to administer lethal dose by workers in a colony. Tunnel length between the bait station and the main nest can also affect time to administer the lethal dose. The ingestion of lethal dose can be either by trophallaxis or direct feeding on the bait (Sheets et al. 2000), where distance will affect the time travelled by workers to reach the toxicant. Furthermore, the number of bait stations intercepted by a colony can be another factor affecting the time for lethal dose acquisition. The number of intercepted baits along with the competing food resources will affect the speed at which a toxicant will be spread in a colony. Workers of C. formosanus are known to feed randomly from the available food resources intercepted by a colony, ensuring that each worker will feed at least once from the bait station when given enough time (Su et al. 1984). But depending on the competing food resources, it may take considerable time which can affect the total time for colony elimination. The third segment of the total time for elimination is contributed by lethal time, which is the time taken for workers with an acquired lethal dose of CSI to die. Because, time of mortality is a function of time of molting, mortality will be observed when a worker undergoes the molting process. Thus, unlike other segments, time taken for completion of this segment remains constant. In the past, the slow process of colony elimination with the use of CSI baits has raised concerns (Evans and Iqbal 2014, Raina et al. 2008, Su et al. 2011), and several 18

19 attempts have been made to improvise the bait efficacy. However, the focus has been mainly on the selection of better active ingredient for baits (Osbrink et al. 2011), or on increasing bait acquisition with the use of arrestants or feeding stimulants (Chen and Henderson 1996, Reinhard et al. 2002a, b). Some of the studies have focused on increasing bait palatability and improving bait matrices to make baits durable (Rojas and Morales-Ramos 2001, Su 2007, Thoms et al. 2009, Eger et al. 2012, Eger et al. 2014). Another attempt at enhancing the effect of CSI in baits was the combination of ecdysteroids with CSI in baits (Su et al. 2011). The two compounds when tested on subterranean termites in the laboratory were found to have an enhanced effect compared with the CSI or MAC alone. These resulted in accelerated/premature molting in the workers under the effect of MAC, which at the time of molting had poorly formed new cuticle under the effect of CSI, resulting in failed molting leading to death. However, before testing the potential of the combination of these IGRs, it is important to understand the biology of molting in workers and determine the contribution of molting time towards the total time of colony elimination. This information will help to determine if accelerating time to molt before impacting workers with CSI will have any significant effect on the total time of colony elimination. Study on the interplay between the lethal dose of CSI and its effect on termite molting biology is also important because the component analysis will help determine the weak points in the process of molting, which could be exploited for control purposes. Secondary repellency is a major limitation of all of the control methods (except CSI baits) for termites in use. Thus, before speeding up the molting process, it is important to ensure if workers leave the foraging site or bait stations (in case of CSI bait treated colonies) before molting and 19

20 determine the point in their cycle at which they may leave the foraging site. Answers will ensure that the efficacy of CSI baits that have been accelerated with the use of MAC will not be compromised. With the overall goal of speeding up the activity of a CSI-based baiting program, a series of experiments were conducted for spatial and temporal assessment of molting amongst workers of C. formosanus, where specific objectives of the project were: 1) development of methodology for determination of molting incidence in workers of C. formosanus by post ecdysis sclerotization, 2) temporal assessment of molting in C. formosanus workers in laboratory conditions, 3) determination of the impact of acquisition of lethal dose of CSI and fasting period on time for mortality in workers, and 4) Spatial assessment of molting in C. formosanus workers in laboratory conditions. 20

21 CHAPTER 2 DETERMINING MOLTING INCIDENCE IN FORMOSAN SUBTERRANEAN TERMITES (ISOPTERA: RHINOTERMITIDAE) BY POST-ECDYSIS SCLEROTIZATION Introduction Insect cuticle made up of chitin, forms the semi-rigid exoskeleton of insects, which provides shape, protection from environment and a substrate for muscle attachment to the insects (Andersen 1979, Chapman 2009). However, these benefits come at the expense of energy spent on producing new cuticle for growth and metamorphosis. During molting, the hard and rigid cuticle is shed and is replaced with an underlying soft and flexible cuticle that allows expansion for growth (Laufer 1983). The new cuticle after molting (post-ecdysis) becomes tough and dark to be functional again and this process is known as tanning or sclerotization. Because the chitin is the building block of the cuticle that makes up the exoskeleton system, any disruption in its formation during molting can lead to the death of individuals (Xing et al. 2013, 2014). Chitin synthesis inhibitors (CSI) classified under insect growth regulators (Branes 1997) are a widely used group of insecticides (Candy and Kilby 1962, Nishioka et al. 1979, Branes 1997, Mommaerts et al. 2006). They interfere with chitin biosynthesis, which is essential for reproduction, growth, and development of insects (Cohen 1987, Reynolds 1987, Muthukrishnan et al. 2012). In the United States, CSI incorporated baits are used for elimination of subterranean termite colonies including two important pests, C. formosanus and R. flavipes (Rust and Su 2012). The process of elimination of termite colonies using these baits can take 2-9 months depending upon the termite species, age and size of the colony (Eger et al. 2012). Although the slow acting CSIs are important for the spread of toxicant in a colony of millions of individuals, reduction in time taken for colony elimination with CSI baits may decrease the cost of subterranean 21

22 termite control and promote the acceptance of baits to homeowners. One of the potential methods to reduce the time taken for colony elimination using CSI baits is to reduce the intermolt period in the worker caste with the use of molt accelerating compounds like ecdysone agonists (Monteagudo 2004). However, one limitation of some pesticides for termite control is the cause for primary or secondary repellency that may reduce the traffic of termites to a treated area and ultimately prevent colony control (Su 2005). Investigating where, when, and how the molting process occurs in a subterranean termite colony is therefore key to determine the potential impact of molt accelerating compounds for control purposes. In order to conduct temporal and spatial assessments of molting in workers and evaluate effectiveness of molt accelerating compounds, it is essential to be able to identify recently molted individuals. However, Dyar (1890) pointed out that, it is no difficult thing to overlook a molt or even to think one has occurred when it has not and this is particularly true for termite workers that have a soft body and relatively light sclerotization. There is therefore a high chance for under or overestimation of molting incidence in subterranean termites, especially when ecdysis is difficult to observe. The limitation of the study of molting in subterranean termites is mainly due to their cryptic behavior and asynchronous molting amongst overlapping generations in a colony (Haverty and Howard 1979, Xing et al. 2013). Molting incidence amongst many other insects can be determined by the presence of exuviae of the molted individuals under observation (Swezey 1905, Singh and Mabbett 1976) but in C. formosanus and other termite species, protein rich exuviae are eaten upon molting by other nestmates (Grassé 1949, Xing et al. 2013) in an attempt to sanitize the colony or recycle nitrogen 22

23 in the colony (La Fage and Nutting 1978), thereby dispelling the evidence of molting. Adding to this problem is the occurrence of stationary molts in C. formosanus, where workers molt into another worker instar. A classic methodology used for determination of molting incidence amongst immature stages of insects is the measurement of head capsule (Dyar 1890). Dyar law (1890) suggests that head capsule and other body parts of larvae grow in geometric progression with each molting and this growth is constant for the species. However, in case of C. formosanus, the head capsule distribution of C. formosanus worker instars is not discrete (Higa 1981) and such method cannot be used to determine if an individual has recently molted. Increment in the antennal segments with each molt amongst termites has been observed in the past (Higa 1981, Raina et al. 2004, Chouvenc and Su 2014). Counting antennal segments pre and post-ecdysis can be a possible method to determine the occurrence of molting, but the process of counting antennal segments of soft bodied termites under the microscope can cost the vigor of termite workers which in result may never molt or die early during the experiment. Thus, the lack of reliable visual cue to determine molting occurrence in subterranean termites remains problematic. In the past, insect cuticle sclerotization based on biochemical changes was reported to be highly variable, depending on the insect species and their developmental stages (pre-ecdysis/post-ecdysis) (Andersen 1981, Andersen et al. 1996). However, limited information is available on the changes in the cuticle sclerotization upon ecdysis in termites. Raina et al. (2008) observed that C. formosanus worker mandibles are lightpink in color soon after molting and becomes fully sclerotized within 2 d of molting. Mandibles in workers have an important role in chewing wood and excavating tunnels 23

24 (Indrayani et al. 2007, Li and Su 2009). Termite mandibles are dicondylic with two points of articulation (primary and secondary), which forms the plane of attachment for the mandibles. To hold heavy sclerotized mandibles and allow their movement, the two articulation points are also sclerotized. Considering that these sclerotized regions on the body undergo melanization post-ecdysis, studying the progression of sclerotization can be a useful method for determination of molt frequency in laboratory experiments. Comparisons of sclerotization levels between recently molted individuals and intermolt workers over time will be useful in determining the stage at which molted workers are indistinguishable from the intermolt stage. Based on the preliminary description of the sclerotization process in termites by Raina et al. (2008), we investigated the morphological changes post-ecdysis in C. formosanus workers in order to establish a method to easily distinguish newly-molted workers, even when ecdysis is not observed. The specific objective of the study was to observe and describe the changes of sclerotization of the mouthparts and other regions at four hour- interval post-ecdysis. Materials and Methods Termites were collected from three field colonies of C. formosanus in Broward County, FL, by using the method of Su and Scheffrahn (1986). Collected termites were processed (Tamashiro et al. 1973) and kept at 27 o C in1-liter plastic containers with pieces of moist wood (Picea sp.) in the laboratory. Raina et al. (2008) suggested that foraging termites do not molt for the first d post-collection from the field, thus, termites were kept in the containers for at least 10 d before the experiment. Ten days after collection, groups of ~500 workers from each colony were placed in a Petri dish (diameter: 9.2 cm, height: 2.1 cm) containing moist Nile Blue A filter paper (0.05% wt: 24

25 wt) on the bottom for 24 h. Because termites preparing to molt stop to feed 6 d before ecdysis (Raina et al. 2008), workers about to undergo molting can be identified by the lack of Nile Blue A in their body (Xing et al. 2014). Workers without blue color were transferred to a Petri dish (diameter: 9.2 cm) containing a moist filter paper and observed for molting every 2 h. Upon ecdysis, workers were separated and later transferred to vials containing 80% alcohol at 0, 4, 8, 12, 16, 20, 24, 28, 32, 36 h postecdysis. For each time interval, three workers were arbitrarily collected from each of the three colonies, making a total of nine workers per hour. The index of sclerotization of the cuticle of workers post-ecdysis was estimated over time by measuring the intensity of the darkness of the mouthparts at specific areas: 1) the primary articulation points of the mandibles (left and right), 2) the secondary articulation points of the mandibles (left and right), and 3) the teeth of the mandibles from the molar plate to the apical tooth (left and right) (Figure 2-1). However, the partially translucent labrum settles on top of the mandibles which can obstruct the visibility of their teeth. Thus, two series of measurements were taken. First, specimens were placed under the microscope, one at a time, to take a picture from a frontal point of view of the mandibles (with labrum). In this case, intensity of sclerotization of the primary articulation point, the secondary articulation point and the teeth on the left mandible was recorded, as the left mandible overlaps the right mandible and can limit the view of apical and marginal teeth of the right mandible. Second, the labrum of each individual was removed using dissecting scissors and fine tipped forceps for a direct view of the mandibles, where measurements were taken for the width of the sclerotized region (µm) for the apical tooth of mandibles and intensity of sclerotization of both left 25

26 and right mandibles. Thus, the measure of sclerotization amongst workers at different time intervals was determined per individual for a total of five variables (three with labrum and two without labrum) (Figure 2-1). Images of all the samples were taken from the frontal point of view using a Leica DFC425 digital camera mounted on a Leica M205 C stereo microscope (Leica Microsystems GmbH, Wetzlar, Germany) at a focal length of 35 mm (at 50x magnification). To ensure consistency, all the pictures were taken under identical illumination, using flat incident light and six bottom LED lights of the three illumination arches of Leica LED 5000 MCI illumination system. The movable illuminator arcs were set at 45 o on each side and 70% brightness for all the images to make sure that all pictures were comparable and standardized. The measure of sclerotization of the selected regions was determined using GIMP software (2.8.14). The GIMP software provided a darkness value of the selected areas and the index of sclerotization was calculated by converting the hexadecimal/ html value into a decimal system, where 0=white (html color ffffff) and 100=black (html color ). A paired t-test was performed to check for color symmetry of the major location (left and right) of each variable and no significant difference was found. As a result, both sides of measurements of each variable were combined for each individual and used as a single variable for further analysis. Measurements were also taken for non-molting workers (intermolt stage, n=9) from the three colonies and used as a standard for comparisons with termites at different hours post-ecdysis. These values were then compared amongst different time intervals post-ecdysis using a t- test (HolmBonferroni method, controlling for family-wise 26

27 error, α=0.05) for each variable to determine the time post-ecdysis at which the variable was no longer significantly different from intermolt individuals. In addition, as none of the variables had a linear distribution, a Spearman correlation test (ρ, R project 2015) was used to describe the gradual increase of sclerotization over time and to determine at when a given variable reached a plateau (α=0.05). The Spearman test was first applied to the whole variable (0 h to 36 h) to confirm a positive correlation between time and sclerotization. The test was then restricted to the 4 h to 36 h time interval, then 8 h to 36 h time interval, etc. until the correlation was no longer significant, indicating an absence of increasing sclerotization. The effect of colonies on intensity of sclerotization was tested using ANOVA with Tukey s HSD test (SAS Institute 2009). All tests were run at α = Results Overall Observations Immediately after ecdysis, workers had soft and wrinkled white cuticle. In termites, the jackknife posture is formed mainly to help shed the exuviae during the molting process (Su and Scheffrahn 1993) when the termite bends the head ventrally. As a result, the distance between the pronotum and the head extends dorsally to allow the individual to initiate the exuviae shedding (Figure 2-2). This stage of workers with soft and wrinkled cuticle, and extended region between the head and the pronotum is generally known in insects as the teneral stage (Neville 1983). We observed that the head of newly-molted termite remained protracted until ~ 2 h post-ecdysis as workers <2 h post-ecdysis were observed walking with an extended head-pronotum distance. For all variables, the index of sclerotization was the weakest just after ecdysis. It gradually became more and more sclerotized within the 36 h observation period post- 27

28 ecdysis as seen for primary point of articulation and secondary point of articulation in Figure 2-3 (a-d, i-l) and for mandibular teeth and width of sclerotization of the apical tooth in Figure 2-3 (e-h, m-p). The colony of origin was not a significant factor for the four indices of sclerotization variables and nor for the width of the sclerotized area of the mandibles post-ecdysis (α = 0.05). Sclerotization of the Primary Point of Articulation of the Mandible Immediately after ecdysis (0 h) the sclerotization of the primary point of articulation of the mandible was visible for workers (Figure 2-3a) but the index of sclerotization was significantly lower than workers in intermolt stage (t-test, p<0.001) (Figure 2-4). There was a significant increase of sclerotization over time (0 h to 36 h) (ρ=0.62, P<0.001) as seen in Figure 2-3 (a-d, i-l); however, the index of sclerotization of the primary point of articulation reached a plateau at 24 h, with no significant increase after 24 h (ρ=0.22, P<0.06) (Figure 2-4). While the increase in the index of sclerotization was significant for the first 24 h post-ecdysis, pairwise comparisons of each time interval showed that after 8 h the index of sclerotization was not different from intermolt workers (t-test, α=0.05). This was due to the high variability of the index of sclerotization for some individuals between 8 to 24 h. Sclerotization of the Secondary Point of Articulation of the Mandible Similar to the primary point of articulation, the secondary point of articulation was lightly sclerotized at 0 h post-ecdysis. (Figure 2-3a) and the index of sclerotization was also significantly lower than workers in intermolt stage (t-test, p<0.001) (Figure 2-4). Between 0 h to 36 h post-ecdysis, there was an increase in sclerotization (ρ=0.68, P<0.001) (Figure 2-3a-d, i-l) but unlike the previous observation with the primary point of articulation, the index of sclerotization of the secondary point of articulation did not 28

29 reach a plateau within the first 36 h (Figure 2-4), with still a significant increase between 32 h to 36 h (ρ=0.58, P<0.001). In addition, pairwise comparisons of each time interval were all significantly different to intermolt individuals (t-test, α=0.05), except at 36 h where the sclerotization of the secondary point of articulation was not different from intermolt workers (P=0.96). Sclerotization of the Left Mandible, Covered by the Labrum The cuticles of the labrum covers the mandibles, buffering the visible sclerotization of the left mandible underneath. The apical tooth and marginal teeth were lightly sclerotized just after ecdysis (0 h) with a light orange coloration and no sclerotization was observed on the molar plate or on the marginal teeth (Figure 2-3a). Between 0 h to 36 h post-ecdysis, there was an increase of sclerotization of the mandible (ρ=0.97, P<0.001) (Figure 2-3a-d, i=l). The gradual sclerotization process of the apical tooth did not reach a plateau until 28 h (28 h to 36 h, ρ=0.62, P=0.07) (Figure 2-4). This result was confirmed by pairwise comparisons of each time interval and only values before 28 h were significantly different from intermolt individuals (t-test, α=0.05). Sclerotization of the Mandibles, Without Labrum With the removal of the labrum, both mandibles were exposed and the sclerotization was directly observed (Figure 2-3e). As previously observed in individuals with a labrum, the mandibles gradually sclerotized from 0 h to 36 h (ρ=0.91, P <0.001) (Figure 2-3e-h, m-p), and the index of sclerotization reached a plateau at 28 h postecdysis (28 h to 36 h, ρ=0.04, P=0.75) (Figure 2-5). However, unlike the observation with the labrum in place, the index of sclerotization of the mandible was different from intermolt individuals at all-time intervals, even at 36 h post-ecdysis (t-test, α=0.05).this 29

30 indicates that additional sclerotization may take placeup to 36 h, but can only be observed if the labrum is removed. Width of Sclerotization of the Apical Tooth Just after ecydsis, the apical tooth showed signs of sclerotization (Figure 3e). However, the sclerotization was limited to the marginal area of the tooth. As time passed, not only did the index of sclerotization of this area increase (Figure 2-5), but the area of sclerotization expanded (Figure 2-3e-h, m-p). The width of sclerotization of the mandible gradually increased over time (0 h to 36 h, ρ=0.98, P <0.001), and it reached a plateau at 32 h post-ecdysis (32 h to 36 h, ρ=0.18, P =0.27) (Figure 2-5). This result was again confirmed by the pairwise comparisons of each time interval and only values before 32 h were significantly different from the intermolt individuals (t-test, α=0.05). Discussion Detecting the incidence of molting in termites is a challenging task due to their cryptic behavior, their asynchronous molting, and the low level of sclerotization in their soft body. In this study, we identified characters on the head capsule of C. formosanus workers that can be used to differentiate individuals that molted up to 36 h post-ecdysis, in comparison with intermolt workers. The sclerotization of the primary articulation point was fully sclerotized near 24 h post-ecdysis whereas the secondary point of articulation and the mandibles (with or without labrum) were fully sclerotized at h postmolting. These results indicate that progression of sclerotization of the cuticle varies for different regions of the mouthparts, which can be used for distinguishing recently-molted workers from intermolt workers. The primary point of articulation was the only area that displayed high variability in its index of sclerotization between 8 h to 24 h among all individuals, whereas for all 30

31 other areas, the variables were consistent among individuals. This suggests that if we rely only on the primary point of articulation, it is only possible to differentiate individuals that molted within 4h with certainty. Using the observation of the index of sclerotization of all other variables from the mouthparts, especially for the secondary point of sclerotization and the width of sclerotization of the mandible, it is possible to estimate the time of molting of a given termite worker within the first 36 h post-ecdysis, with a margin of error of ±4 h. While our approach allowed us to describe the progression of the post-molting sclerotization in termite workers using individuals in alcohol under the microscope, we were also able to visually recognize the same traits on live termites kept in planar arenas used for termite laboratory bioassays (Chouvenc et al. 2011) over a 36 h period. Thus, a trained eye can differentiate individuals that molted within 4 h, 12 h, 24 h and 36 h, on live termites with the use of a simple hand magnifier. Of course, the determination of the incidence of molting was more precise using dead individuals under the microscope, especially after the removal of the labrum to reveal the mandibles, but in an experiment that requires observation without destructive sampling, visual recognition is an acceptable compromise to obtain reasonable reliable molting data from a live group of termites. We found that unlike many other insects (Neville 1983), newly-molted C. formosanus workers in their teneral stage exhibited pre-ecdysis sclerotization of some mouthparts, at the primary point of articulation of the mandible in particular. Because termite workers go through successive stationary molts, individuals have to maintain the overall mandibular structure from one instar to the next. We suggest that in order to 31

32 extirpate the exuvia that forms a socket around the mandible during the jackknife position, the molting termite has to open the mandibles so that the exuvia around it can slide out. The light sclerotization of the primary point of articulation therefore allows the mandible movement and the shedding of the exuvia without imposing physical constraints on rest of the mouthparts that are soft and unsclerotized. Within 36 h post-ecdysis, the whole mandibular structure had regained its rigidity, motility and near-full sclerotization. This suggests that the newly-molted worker can rapidly regain its functionality for wood consumption, excavation and grooming, although Raina et al. (2008) showed that workers only regain symbiotic protozoans after 4 d post-ecdysis, implying that it can only participate in digestive activities much later. As workers that initiate the molting process enter a phase of fasting 10 d before ecdysis (Raina et al. 2008, Xing et al. 2013), it implies that in a termite colony, an individual has reduced activity for a period of 14 d, each time it molts from one instar to the next. The rapid sclerotization of the mouthparts post-ecdysis could be a way to regain some level of activity as early as possible, in order to reduce the burden of individual molting on the overall colony function. In a mature termite colony, such a burden may be negligible, as the large number of individuals can negate the cost of molting. However, in a young colony with a small number of workers, disabling a worker for 14 d because of the molting process may result in substantial cost on the growth of the colony (Chouvenc and Su 2014, Chouvenc et al. 2015). Therefore, the rapid sclerotization post-molting may be a way to compensate and regain some limited activity to participate in the colony activity. 32

33 In conclusion, this study elucidates the process of cuticular sclerotization over time in C. formosanus post-ecdysis. We confirmed that the process of sclerotization was complete within the first two days post-ecdysis, which corresponds to the observations made by Raina et al. (2008). The qualitative morphological changes of sclerotization we provide can save the tedious work involved in timing the ecdysis in insects and especially in a termite colony with asynchronous molting amongst overlapping generations. These results will serve as a basis for conducting future studies on assessments of the time lapse between two consecutive molts for workers of lower termites at different instar and in the evaluation of any untimely/early molting occurrences amongst workers while testing molt accelerating compounds. Furthermore, the detailed description of individuals at various hours post-ecdysis will help in determination of the site of molting and track the movement of workers post-molting within a colony for behavioral studies. 33

34 Figure 2-1. Frontal view of the head capsule and left mandible of a worker of C. formosanus showing a = primary point of articulation, b= secondary point of articulation, the different teeth of the mandible and the width of sclerotization of the apical tooth. 34

35 Figure 2-2. Lateral view of the teneral stage in a worker of C. formosanus ( Jackknife position). Note the distance of the head capsule from the pronotom, visible amongst workers at <2 h post-ecdysis. 35

36 Figure 2-3. Frontal view of the head capsules of C. formosanus at 0 h, 4 h, 8 h, 16 h, 20 h, 24 h, 36 h post-ecdysis and at intermolt stage, with and without the labrum. 36

37 Figure 2-4. Progression of the index of sclerotization for three variables from 0 h postecdysis until 36 h post-ecdysis at 4 h intervals when labrum was present. Mean±SE of the index of sclerotization over time (0=white, 100=black). The dotted line represents the average index of sclerotization of workers in intermolt stage (control base line). A) primary point of articulation, B) secondary point of articulation, C) mandible tooth (with labrum). 37

38 Figure 2-5. Progression of the index of sclerotization of mandible teeth and for the width of sclerotized region for the apical tooth of the mandible (µm) from 0 h to36 h post-ecdysis at 4 h intervals when labrum was removed. Mean±SE of the index of sclerotization over time (0=white, 100=black). A) mandible teeth (no labrum), where the dotted line represents the average index of sclerotization of mandible teeth of workers in intermolt stage (control base line). B) apical tooth sclerotized width (no labrum), where the dotted line represents the average width of sclerotized region (µm) for the apical tooth of mandibles of workers in intermolt stage. 38

39 CHAPTER 3 TEMPORAL ASSESSMENT OF MOLTING IN WORKERS OF FORMOSAN SUBTERRANEAN TERMITES (ISOPTERA: RHINOTERMITIDAE) Introduction Control of subterranean termites is challenging because of their cryptic nature that makes it hard to determine the spread of a field colony for treatment. Traditionally, liquid termiticides are used for controlling subterranean termites (Gold et al. 1996, Su 2005) and it holds a major share of the termite control market (Rust and Su 2012). Another control measure available for subterranean termites is baiting. The baits used for termite control are incorporated with non-repellent and slow-acting active ingredient like CSI. Su (1994) evaluated CSI, hexaflumuron (Dow AgroSciences, Indianapolis, IN) incorporated baiting system in the field and found it to be effective in eliminating colonies of C. formosanus and R. flavipes. Later, multiple other studies demonstrated the success of CSIs baits in eliminating colonies of many subterranean termite species (Su et al. 1995, Getty et al. 2000, Sajap et al. 2000, Grace and Su 2001, Husseneder et al. 2007, Osbrink and Cornelius 2013). CSIs are insect growth regulators that kill workers by disrupting the formation of a new cuticle at the time of molting (Su and Scheffrahn 1993). Molting is important for growth and colony health and each worker in a colony molts multiple times in its lifespan. As a result, in a baited colony workers that have acquired lethal dose of CSI are fated to die upon molting. CSIs are slow acting and lethal time to affect termites is independent of dose (Su 2005). Lethal time depends on timing of workers to molt and the process of elimination of a colony can take several months after baiting (Eger et al. 2012). Although the slow acting CSIs allow the horizontal transfer of the toxicant in a 39

40 colony, there is an incentive to reduce the duration of elimination time for economic purposes. Since the development of CSI baiting technology, there have been many attempts to improve CSI baits efficacy; and focus has been mainly on the selection of a better active ingredient for baits (Osbrink et al. 2011) and improving bait matrices to make baits more durable and palatable to termites (Su 2007, Thoms et al. 2009, Eger et al. 2012, Eger et al. 2014). Some of the work has focused on reducing time required for subterranean termites to discover commercial bait stations to make colony elimination faster (Swaboda 2004). However, not much research has been done on elucidating the time taken by workers for molting on which the lethal time of CSI for colony elimination largely depends. Any information on the timing of molting in termites may give insights into potential methods of reducing the overall time taken for colony elimination. In their laboratory study on C. formosanus, Raina et al. (2008) reported that an average of 1.01% workers in a foraging population molt each day, which implies that it may take ~100 days for all workers in a colony to complete one molt cycle. Another study suggests a worker may take as long as 7 months to molt again (Nakajima et al. 1963) and molting frequency may depend on the age of individuals (Raina et al. 2008, Chouvenc and Su 2014) but definitive information on termite molting period is still lacking. Considering the implications of worker molting frequency on subterranean termite control using CSI baits and lack of available information, there is a need to conduct a temporal assessment of molting events in the worker caste. In this study, we evaluated (1) the frequency of molting in workers from laboratory juvenile colonies of C. formosanus, and (2) the frequency of molting in workers collected from foraging sites of 40

41 mature field colonies. We also evaluated how temperature may affect molting frequency of workers and discussed its possible role in affecting time for elimination of a baited colony. Materials and Methods Molting Frequency of Workers in a Juvenile Colony Three, 4-yr old juvenile colonies of C. formosanus were used in this study. Colonies were initiated by collecting alates from swarming events of C. formosanus in New Orleans, LA. Paired dealates were placed in plastic cylindrical vials (8 x 9 x 2.5 cm) containing moistened soil and pieces of wood and kept at 27 o C. As the dealates reproduced and each colony grew in size, the vial containing termites was moved into a plastic box (17 x 12 x 7 cm) that provided larger space for colony expansion. In the box, each colony was provisioned with sufficient food and moistened soil for survival and growth. For ease of observation, one planar arena was used for each colony, totaling three planar arenas for three colonies. The planar arena as described by Chouvenc et al. (2011) was made of two clear sheets of Plexiglas (24 x 24 x 0.6 cm in thickness) and a spacer (Plexiglas laminate of 0.2 cm thickness) for maintaining the inner space of 0.2 cm between sheets. The planar arena was filled with 50 g of oven dried sand moistened with 15 ml of sterile deionized water. The upper sheet had an access hole (0.5 cm in diameter) on top of which a Plexiglas cup (4.5 cm in diameter and 3 cm in height) was fitted to form a termite introduction chamber (Su 2005). For introduction of termites from each box containing a colony of C. formosanus to the planar arena, a 10 mm hole was 41

42 drilled in the bottom of colony box. The box was then placed on top of the introduction chamber on the Plexiglas planar arena such that the opening drilled in the box lies on top of access hole of the planar arena (Figure 3-1A). To allow movement of termites, one end of a moist wood piece (Spruce, Picea spp.,) was inserted into the opening drilled in the box and the other end of the wood into the access hole of the Plexiglas planar arena. The wood piece served as a connection for termites from the box to expand in to the planar arena. Meanwhile, the lid of the colony box was removed for the soil in the box to dry, so that termites were forced to move into the planar arena. The box remained on top of the planar arena until all the termites moved into the planar arena. Migration of termites (including the primary reproductive pair) from the box into the planar arena took 4 to 6 d depending on the size of the colony. Once the colony moved into the planar arena, the box was removed and moist wood pieces (Spruce, Picea spp.,) were added to the introduction chamber. One week after all termites of a colony moved into the planar arena, termites in the planar arena were video recorded using a mounted camera (Model DP70, Olympus Optical Co., Ltd., Tokyo, Japan) for 24 h x 10 d. Videos were later viewed and daily rate of molting [(number of workers molting/total workers in the colony)*100] was determined. Molting Frequency of Workers from Foraging Populations Termites were collected from three field colonies of C. formosanus in Broward County, FL by using the method described by Su and Scheffrahn (1986). Collected termites were processed (Tamashiro et al. 1973) and kept at 27 o C in plastic containers with pieces of moist wood (Picea sp.) in the laboratory for a week before setting up the experiment. 42

43 Study was conducted in planar arenas made of two clear sheets of Plexiglas (12 x 12 cm) and a laminate spacer to maintain an inner space of 0.2 cm between the two sheets for movement of termites (Figure 3-1B). A 0.4 mm wide hole was made in the center of the top Plexiglas sheet for air flow and addition of water to maintain moisture during the experiment. Another hole (5 mm wide) was made on one top corner of the Plexiglas for introduction of termites into the planar arena. The planar arena was filled with 18 g of oven dried moistened sand (15 g sand + 3ml of sterile deionized water) along with a moistened cellulose absorbent pad (45 mm dia, 2 mm thick) which served as a food source. An area of 5 by 2 cm was left empty near the introduction hole for the movement of termites. The sheets and the spacer were held by a central screw and binder clips and planar arenas were sealed using hot glue to prevent moisture loss. Fifty workers (undifferentiated larvae of at least the third instar) plus seven soldiers of C. formosanus were added to each planar arena. Termites were introduced into the planar arena using a funnel and after introduction the hole was covered with a plastic cover slip and secured with a binder clip. The experiment was conducted at 21±0.5 o C and 27±0.5 o C. Three units were prepared for each colony for a total of 18 experimental units (nine for each temperature). The number of molting termites was visually monitored every 12 h and a picture of each planar arena was taken once a day to count the number of surviving workers over time. Because termites undergo asynchronous molting and workers are at different stages of their life cycle, it can be hard to determine molted workers unless ecdysis was observed. Based on the degree of sclerotization on workers mouth parts described in Chapter 2, molted workers were distinguished from intermolt stage workers to record molting incidence. Observations 43

44 were made for two molt cycles and up to the beginning of the third molt cycle (up to 88 d). Cycle day one was counted with the initiation of the first observed molting in each planar arena independently. The cycle was considered complete on the day that a cumulative 100 percent of molting amongst surviving workers in a planar arena was recorded. Similarly, the second molting cycle was completed with 200% cumulative molting percentage in an arena. Upon completion of two molt cycles planar arenas were opened and antennal segments of each worker were counted under the stereo microscope for determination of workers growth stages. Due to the slow rate of molting amongst termites in planar arenas placed at 21 o C, observations were stopped and planar arenas were opened at the time of opening planar arenas at 27 o C. The effect of colonies on molting frequency of workers in all the studies was tested using ANOVA with Tukey s HSD test. The square-root of the percentage of molting was subjected to an arcsine transformation before performing ANOVA. The time required for completing the first molting cycle and the second molting cycle for groups of workers at 27 o C was compared using a t-test. The overall molting rates at 21 o C and 27 o C were compared using a Cox-proportional hazard regression model, with temperature as a factor and day of molting as the variable. Also, comparisons between percentage of cumulative molting at two temperatures were made upon completion of cycle one (44 d) and cycle two (45d) in arenas at 27 o C using t-test. All tests were run at α = 0.05 (SAS Institute 2009). The data presented are the untransformed means. Results Molting Frequency of Workers in a Juvenile Colony The colony of origin was not a significant factor for the average daily rate of molting (F (2, 27) = 0.82, P =0.064). The average daily molting percentage in the three 44

45 colonies was recorded to be 1.7± 0.19 (Mean ± SE) when pooled across colonies (Figure 3-2). However, the daily percentage of workers molting in a colony was variable and it ranged between %, %, and % per day for colony 1, 2, and 3, respectively. Of all the molting events observed in the three colonies, only in three events (less than 1%) were workers found to molt into soldiers while the rest molted into workers. Molting Frequency of Workers from Foraging Populations All groups of 50 workers originating from field foraging populations had >80% survival after 100 d in the planar arenas. The proportion of molting termites was adjusted to surviving termites on a daily basis, and we observed that at 27ºC, it took 43.9±3.1 d (mean±sd) for 100% workers to molt at least once. The colony of origin was not a significant factor of molting activity (F (2, 8) =3.57, P=0.095) during this first molting cycle. It took an additional 45.6±3.5 d for the groups of workers to complete a second cycle of molting and again the colony of origin was not a significant factor of molting activity (F (2, 8) =0.98, P=0.42). There was no significant difference for the time required for 100% of workers to complete molting between the first molting cycle and the second molting cycle (t-test, P=0.43, Figure 3-3). In all of the planar arenas, workers molted into another worker instar for both the cycles (i.e., no newly-produced presoldiers or soldiers). Examination of workers in planar arenas at 27 o C upon completion of two molting cycles revealed that on average 93% of the workers had 14 antennal segments which indicates the presence of mostly 4th worker instars (W4) after two molting cycles (W2 W3 and W3 W4). 45

46 At 21 o C, termites from all three colonies took longer to molt than termites at 27 o C, as none of the arenas at 21ºC completed 100% molting workers within the duration of the experiment (Figure 3-4). While at 27ºC all workers completed their first molting cycle at 43.9 d in average, only 23% of workers achieved their molting at 21ºC. In addition, while all workers completed their second molting cycle at 27ºC within 90 d, 43% of workers still had not molted within 90 d at 21ºC. The molting rates of workers were significantly different between 27ºC and 21ºC, with a molting rate of 2.2% per d at 27ºC and a molting rate of 0.6% per d at 21ºC (Cox regression analysis, P<0.01). Discussion Molting is a physiological process by which insects grow and differentiate. Usually it is accomplished in the early stages of an individual, as it allows for growth. However, in Isoptera, especially the lower termites, workers have the ability to undergo periodic molting. The workers molt into higher worker instars with W7 as the highest known instar in C. formosanus (Shimizu 1962, Chouvenc and Su 2014) or until they differentiate into soldiers or reproductives. In our study with foraging populations from field colonies, workers underwent only stationary molts and there was no differentiation into soldiers. The likelihood of a worker molting into a soldier was inhibited by the presence of 15% soldier of the total population in each planar arenas, which is above the observed percentage of workers found in foraging population of C. formosanus (Haverty 1979). This ensured that the presented result on time lapse between two consecutive molts were based only on worker to worker molts. The average daily percentage of molting for the laboratory juvenile colonies in our study (1.7± 0.19 (± SE)) was higher than that reported by Raina et al. (2008). In their study, observations were limited to the foraging group of workers which comprised 46

47 old worker instars that did not molt as frequently as young instars (Raina et al. 2008) resulting in low molting frequency. Furthermore, their counts of recently-molted individuals based on sclerotization were made on alternate days. However, in our previous study on the degree of sclerotization in newly-molted workers of C. formosanus, it showed that workers at hr post ecydsis are indistinguishable from non-molting workers in a colony (chapter 2), suggesting that there are chances that not all the molting events were recorded in their study, which led to different molting frequencies between Raina et al. (2008) and our results. Nevertheless, Raina et al. (2008) suggested that molting incidence could be different if observations were made using the entire colony and not just the foraging population, which corroborates with the current study. Based on the molting rate (2.1%, 1.3%, and 1.7%) in the three laboratory juvenile colonies under study, the projected time period for workers in the colony to complete a cycle of molting is 47, 77, and 58 days, respectively. Because younger worker instars molt more frequently than the older worker instars, it is possible that depending on the age structure of a colony, young worker may molt multiple times in a certain time period, while old worker instars none at all. Thus for the colonies in our study with 2.1%, 1.3%, and 1.7% average molting rate per day, there is a possibility that not every individual would have molted in the projected time periods of 47, 77, and 58 days. Consequently, the actual time taken for all individuals in a colony to molt at least once will be longer than the projected time based on observed molting percentage. This variability based on the age structure of a colony can lead to the survival of old workers 47

48 for a longer duration than the young workers and thereby increasing the total time of colony elimination than the projected time, when treated with CSI baits. Use of CSI such as noviflumuron in baits is one of the methods available for elimination of subterranean termite colonies (Su and Scheffrahn 1993). CSI in baits, target the molting process of the worker caste and upon acquisition of the lethal dose of CSI, mortality is observed amongst workers that attempt to molt (Rust and Su 2012). Thus, the time of elimination of a colony upon administering the lethal dose from the baits largely depends on time taken by workers to molt. In the past, a study showed that the use of molting accelerating compounds (MACs) like ecdysone agonists along with CSI can shorten the lethal time when compared with CSI alone on a group of termites in laboratory (Su et al. 2011). The ecdysone agonist acts similar to the molting hormone, 20-hydroxyecdysone and initiates early molting by stimulating the apolysis and the formation of cuticle (Riddiford and Truman 1978, Wing et al. 1988), the process on which CSI acts. However, before testing the potential of these compounds for their commercial use in baits to reduce the time taken to molt, it is important to determine the time lapse between the two molts for a worker for comparisons. Nakajima et al. (1963) speculated that second worker instar (W2) may take months to molt into the sixth worker instar (W6), suggesting that workers W2 W3 W4 W5 W6 molt at an average time interval of 105 to 120 days. In a recent study, Chouvenc and Su (2014) suggested that workers of C. formosanus may take at least 210 days to molt into another instar (Chouvenc and Su 2014). However, these two studies were conducted under different conditions from the current study that may have led to variation in results. 48

49 In the current study we found that at 27 o C, both W2 and W3 took ~ 47 days to molt into next instar. This suggests that an incipient colony (0-7 months old, Chouvenc and Su 2014) comprising of W3 as the oldest worker instar and immature colony (8-26 months old, Chouvenc and Su 2014) comprising of W4 as the oldest worker instar would take at least 47 days to molt. Because mortality is a function of molting in CSI baited colonies, we assume that elimination of any similar aged colony will take at least 47 days following the ingestion of lethal dose of CSI. Nevertheless, these conclusions are based on molting frequency observed under controlled conditions at 27 o C. Nakajima et al. (1963) discussed the effect of season and temperature on molting frequency of workers of C. formosanus. They reported that depending on the cold or warm months of the year, W2 may stay in this stage for 3-5 months, W3 for 2-4 months, W4 for 2-6 months, and W5 for 1-7 months. Similarly, temperature was found to affect molting frequency in the current study as well. All the individuals (100%) in planar arenas placed at 27 o C molted in 47 days, whereas only 23% of the total individuals in planar arenas at 21 o C managed to molt during this period. Based on these results it can be concluded that temperature will affect the lethal time for colony elimination when treated with CSI baits, which is in agreement with the findings of Van den Meiracker et al. (2002) where C. formosanus and R. flavipes treated with hexaflumuron were found to have higher survivorship at temperature 20 o C than at 25 and 30 o C. To conclude, the average daily molting rate in a juvenile colony gave an estimate of the time in which the majority of the workers in a colony can molt. However, the ultimate time for all the workers in a colony to molt at least once will depend on the 49

50 proportion of the highest worker instars present in a colony. Time lapse between two consecutive molts determined using foraging populations helps elucidate part of the story, but not all. Future studies are needed to determine the time lapse between molts for higher worker instars (W4 and higher). Because the time taken for workers to molt is an important segment of the total time taken for elimination of C. formosanus colonies (Su et al. 2011), information on time for molting will give insights into potential methods of reducing time for colony elimination using CSI baits. 50

51 Figure 3-1 Planar arena (24 x 24 x 0.6 cm in thickness) filled with moistened sand for molting frequency of workers in a juvenile colony. Box on top of the arena contains 4-year old juvenile colony of C. formosanus. Four wood blocks under the planar arena were for the support. The scale represents length of the planar arena. B) Planar arena (12 x 12 x 0.6 cm in thickness) filled with moistened sand and contains a cellulose absorbent pad for daily molting frequency of foraging population of a field collected colony. The scale represents length of the planar arena. 51

52 Average daily percentage of molting colony 1 colony 2 colony 3 Figure 3-2. Average percentage of molting per day in three, 4-year old juvenile colonies. No significant difference in the percentage of molting amongst the colonies. 52

53 Figure 3-3. Days taken to complete molting cycle 1 and 2 using field collected foraging population of three colonies ( A, B, and C) at 27 C. 53

54 Figure 3-4. Mean cumulative percentage of workers molted in three colonies for two cycles at 27 C and 21 C, with upper (mean+sd) and lower (mean-sd) limit line. The 100% molted workers represent completion of cycle 1 and 200% cumulative molting represent completion of cycle 2. 54

55 CHAPTER 4 FASTING PERIOD AND TIME FOR MORTALITY Introduction The order Isoptera has over 3,100 described species, of which 363 species (11.7%) are considered important structural pests (Krishna et al. 2013). Although few in number, these are responsible for causing economic loss of $40 billion/annum worldwide (Rust and Su 2012). The subterranean termites of family Rhinotermitidae which represent only 1.2% of the total termite species are responsible for nearly 80% ($32 billion) spent annually on control and damage repairs worldwide. Amongst several subterranean termite pests in the US, C formosanus has been an economically important species since its discovery in the continental US in 1957 (Edwards and Mill 1986, Chambers et al. 1988, Rust and Su 2012). They have a subterranean habitat with an extensive interconnected gallery system, long foraging distances (King and Spink 1969), which makes it challenging to apply treatment for their control. In the last two decades, remedial and preventative control of subterranean termites has included baits incorporated with slow acting CSIs. These products have been shown to successfully eliminate subterranean termite colonies (Su 1994, Sajap et al. 2000, Grace and Su 2001). The slow-acting CSI are effective because of their doseindependent lethal time, which ensure that termites do not die upon ingesting the lethal dose (Rust and Su 2012). The mortality among workers with acquired lethal dose is observed at the time of molting (Su and Scheffrahn 1993), where time taken to molt for C. formosanus can be as long as days for second and third worker instars (Chapter 2) and even longer for an older worker instar (Raina et al. 2008), suggesting that elimination can be a lengthy process. The slow action of a CSI is critical, in that it 55

56 ensures the toxicant is spread in a colony of millions of individuals, yet, reduced time for colony elimination may decrease the cost of subterranean termite control using CSI baits. Increased use of CSI-based baiting technology could yield economic and psychological benefits to the homeowners. Thus, studies to understand events contributing to the total time taken by CSI-based baiting systems for colony elimination in pursuit of acclerating the colony elimination is justified. Su et al. (2011) divided the total time taken for colony elimination using CSI baits into three sections. These are: 1) Bait interception time - the time taken to discover CSI baits stations in the field. It is highly variable among colonies and depends on the foraging pattern of a colony that will affect the probability of discovering baits. 2) Lethaldose acquisition time- the time spent in the spread of the toxicant in a colony, which depends on the size of the colony, distance between nests and the discovered bait stations, and the number of discovered bait stations. 3) Lethal time- the time taken for workers with an acquired lethal dose of CSI to molt. Because a CSI kills workers by disrupting the formation of cuticle at the time of molting (Su and Scheffrahn 1993, Xing et al. 2014), time of mortality is as a function of the time of molting. As a result, the lethal time for a CSI is considered to be the time taken by workers to undergo a molt cycle, but it is not always clear if the lethal effect of a CSI will be observed in the current molt cycle or in a following cycle. In a laboratory study on the molting process in workers of C. formosanus, Raina et al. (2008) reported that workers undergo a fasting period of around 10 d before ecdysis and expel their gut contents 6 d before ecdysis. This implies that for a CSI to express its potency in the coming molt cycle, the lethal dose must be acquired before 56

57 the initiation of fasting period. The objective of the current study was to determine how the fasting period and lethal dose acquisition affects the time for mortality in workers treated with noviflumuron. We also observed and described mortality and molt-inhibitory symptoms in foraging population under the effect of noviflumuron over time. Materials and Methods Foraging populations of C. formosanus were collected from three field colonies in Broward County, FL, with the method described by Su and Scheffrahn (1986). Collected termites were processed (Tamashiro et al. 1973) and kept at 27 ± 0.5 o C in1-l plastic containers with pieces of moist wood (Picea sp.) in the laboratory for one week before testing. The 2-D foraging setup was composed of six planar arenas (24 x 24 cm and 1.4 cm in thickness). The planar arena was made of two clear sheets of Plexiglas (24 x 24) cm and laminate spacer on four sides to maintain an inner space of 0.2 cm between the two sheets for movement of termites, as described by Su (2005). The upper sheet of the planar arena had an access hole (0.5 cm in diameter) on top, upon which a Plexiglas cup (4.5 cm in diameter and 3 cm in height) was fitted to form an introduction chamber for termites (Su 2005).The two sheets of Plexiglas and the laminate spacer were bolted together and the planar arena was filled with ~ 80 g of oven dried moistened sand (65 g sand + 15 ml of sterile deionized water). The six planar arenas were connected to each other by 6 m long (12 pieces of 0.5 m long tubes) coiled Tygon tubing (0.6 cm in diameter) to form a linear foraging distance of 30 m (Figure 4-1). Five thousand workers soldiers from a colony were divided into ~six groups of 958 termites and each group was introduced to one of the six planar arenas as described by Su (2005). Upon release, termites in each planar arena were 57

58 provisioned with pieces of moist wood (Picea sp.) and left undisturbed for a week at 27± 0.5 o C to allow them to disperse and connect through a 30 m long setup. For each colony, two extended foraging setups were prepared where one setup was treated with cellulose pellets containing 0.5% noviflumuron with Nile blue A (0.05% wt: wt) and the other with the control (cellulose pads with 0.05% w:w Nile blue A). Treatments were added to the treatment chamber (6 cm in diameter and 8 cm in height) that was inserted one week after allowing termites to connect through the foraging arenas. The treatment chamber was added between the Tygon tubes connecting the corner most planar arena (on one side of the extended foraging arena) to the adjacent planar arena (Figure 4-1). The experiment was replicated using 3 different colonies of origin totaling 6 extended arena units (2 (trt + control) X 3).The corner most planar arena which was closest to the treatment chamber was referred to as the 0 m arena, and the others as 6 m, 12 m, 18 m, 24 m and 30 m in reference to the treatment chamber. Once every day, visual counts of workers in the process of molting, and those recently molted based on the degree of sclerotization of mouthparts (described in chapter 2) were made in both control and treatment units. Number of dead workers with any signs of molt inhibitory expression were also counted. These workers exhibit the typical jackknife and exuvia wrapped postures (Figure 4-2a, b), which are also formed during the normal molting process but when affected by noviflumuron workers that attempt to molt die jammed in the exuviae as they could not finish molting successfully. While these workers remain in the posture, there is excessive grooming by the nestmates and as a result, their appendages (antennae and legs) are consumed by nestmates, distinguishing them from normal molting workers (Su 2005). The data 58

59 presented is for the weekly counts of molting workers and the inhibitory effect on molting. A week after adding each treatment, images of planar arenas in each setup was taken once a week for both treatment and control units for nine weeks. Images were later used to count the number of dead and living workers (blue and white) observed in each planar arena each week as described by Su (2005). Because termites feed on corpses of nestmates, counts included only freshly dead workers. Comparisons of percent observed mortality was made between control and treatment setups. At the end of 9 wks, the extended foraging setups were disassembled and the number of living termites was counted. The effect of colonies on worker mortality, molting and molt inhibition was tested using ANOVA with Tukey s HSD test. Because no colony effects were seen, all data were pooled together for analysis. The square-root of percentage of mortality was subjected to arcsine transformation before performing ANOVA. Comparisons for the molting and mortality count between control and noviflumuron treated groups were made using t-test at α = 0.05 (SAS Institute 2009). The data presented are the untransformed means. Results Termites were found in the treatment chambers within 4 h of attaching them to the foraging arenas. Fed termites were recognized by the presence of blue color from Nile Blue A absorbed in their fat bodies and these were initially concentrated in the planar arenas closer to the treatment chamber. It took around 4 wk for ~ 90% of the 59

60 workers to feed from the treatment chamber in both control and treatment groups when dyed termites were found throughout the extended foraging setup. The colony of origin was not a significant factor affecting the molting and mortality of workers in both control and treatment groups (α = 0.05). The number of workers molting/wk in control groups (12.14±0.78, Mean ± SE) was significantly higher (t-test = 7.6, P < ) than the workers molting/wk in noviflumuron-treated groups (1.42±1.15, Mean ±SE) (Figure 4-3), with the counts of molted individuals including both white and blue-stained workers. During the first week, only unfed white workers molted successfully in the control group. The first molting incidence amongst blue workers in control group was on the 11 th day after placing the treatment chamber and over time the count of blue workers undergoing molting increased with the increase in total number of blue workers in the arenas (Figure 4-3a). Unlike control extended foraging setups, only unfed white workers molted successfully in the noviflumuron-treated groups. White workers were found to molt between wk 1 and wk 5 of the study, and the majority of them molted within the first week in the treatment group (Figure 4-3b). During the second week (11 d), the inhibitory effect on molting from noviflumuron was predominant resulting in mortality amongst workers (Figure 4-4b), while no such effect was observed in the control group (Figure 4-4a). The molt inhibitory effect of noviflumuron was first observed amongst white (unfed) workers and later in blue workers during the second week (Figure 4-4b). These workers began to induce peristaltic movements to partially displace exuviae from the distal end of the abdomen and died in this state (exuviae-wrapped posture), while few managed to reach an advanced stage of molting before death, called as jackknife posture (Figure 4-60

61 2a, c). Towards the end of the study (wk 8 and 9), some of the blue workers were still found in jackknife and exuviae-wrapped postures, while white workers with molt inhibitory effects of noviflumuron were absent (Figure 4-4b). From the final counts of workers from extended foraging setups, the mean worker mortality in the treated groups (100%) was significantly higher than the control group (15.6 %) (t-test = 97.8, P < ). During wk 2, mortality was observed in all the planar arenas of the noviflumuron-treated group, indicating that the toxicant had spread to the farthest arena by that time (Figure 4-5). Termites died throughout the extended foraging setup without concentrating in any particular arena. However at 7 wk, the planar arena closest (0 m) to the treatment chamber was the first to attain 100% mortality. Over the 9 wk period, there was an increase in the percentage of workers dead at various distances, while no mortality was observed in the control groups at any time of the study (Figure 4-5). This could be due to the low mortality in the control group and the propensity for necrophagy or burials by nestmates which could have reduced their observability in weekly images. Discussion Results suggest that within 9 wk, noviflumuron eliminated 5000 workers and 750 soldiers in the 30 m long foraging arena. Because all the blue-stained workers in the noviflumuron-treated groups died eventually, the 4 wk period for the workers to acquire color is indicative of the time at which the foraging population acquired a lethal dose. The absence of mortality in the first week amongst blue workers that fed on noviflumuron bait was due to the low number of blue workers attempting to molt as C. formosanus workers are not expected to molt after feeding for at least 10 days in preparation of the molt (Raina et al. 2008). This information concurred with the results of 61

62 the control group where the blue-stained workers were not found to molt until 11 d after treatment indicating workers were in their fasting period of 10 d that lapsed between feeding and molting. CSIs are slow acting toxicants that are known to cause mortality amongst worker of C. formosanus by disrupting the formation of a new cuticle during the molting process (Su and Scheffrahn 1993, Xing et al. 2014). As a result mortality amongst C. formosanus can be assumed to be a function of molting when treated with CSI. Correspondingly, the first incidence of mortality in this study was on the 11 d when counted from the first day of adding the treatment chamber containing noviflumuron incorporated cellulose pellets. These results indicate that mortality in the field will not begin until at least 10 days following the interception of a bait station. However, only one day of feeding on a lethal dose of CSI before the beginning of 10 d pre-molt fasting period can be effective in causing mortality amongst the workers. Because the workers that died at 11 d were in their first molt cycle after acquiring noviflumuron, it can be concluded that the lethal time is dependent on the time taken for the next molting cycle to occur as workers could not overcome the lethal effects of noviflumuron acquired even a day before undergoing molt preparation. In treated extended foraging setup, mortality was first spotted amongst white (unfed) and later in blue (fed) workers that died in postures (exuviae-wrapped or jackknife posture) characteristic of the molt inhibitory effect of noviflumuron. The appearance of white workers in these postures before the blue workers suggests that these workers were probably close enough to a molt that they were affected by a small dose of noviflumuron (not enough to dye workers) acquired either by feeding a minute 62

63 amount or via trophallaxis, or via contact. As a result, these individuals managed to undergo apolysis and initiate ecydsis but died during the molting process. Xing et al. (2013, 2014) reported that during molting, termites form a dorsal breach on the thorax due to the peristaltic movement. Following which the old cuticle slips over the new cuticle and extends out (Figure 4-2 b, d). Under the effect of noviflumuron, many workers in the current study died at this stage with the extended exuviae referred to as the exuviae-wrapped posture. Upon extending the exuviae from the distal end of the abdomen, workers form the jackknife posture and pull legs out of the exuviae for which they need firm muscle attachment (Xing et al. 2014). However, under the effect of noviflumuron in this study, the weak muscle reattachment to the poorly formed new cuticle before ecdysis led to the failed attempts of workers to complete the breach and pull the legs out in order to shed the old cuticle, which eventually led to the death of workers in the jackknife posture. Successful molting was restricted only to unfed workers during wk 1 because by wk 2 there was a radical decrease in the number of workers molting successfully and an increase workers mortality. This indicates that the lethal dose had spread earlier than 4 wk as hypothesized before on the basis of the presence of blue workers. Some of the white (unfed) termites in the group treated with noviflumuron were found to molt successfully during the first 10 d. These workers escaped the effect of noviflumuron because they did not acquire the lethal dose of noviflumuron due to their fasting period that coincided with the initial few days after baiting. Thus, these workers could be responsible for extending the time for elimination of a colony (Section 3 as explained by Su et al. 2011). Given that on average 1-2% workers can molt each day in 63

64 a colony (Raina et al. 2008, chapter 3) which can have as many as a million individuals (Su and Scheffrahn 1988, Rust and Su 2012), the first 10 d escapees (~ 100, ,000) will have the chance to survive at least until the next molt cycle which can be 43 d or longer depending on the worker instar (Chapter 3). In addition to the white workers molting in the noviflumuron-treated arenas, two blue workers (Fig. 4) at 7 wk were found molting successfully with no potent effect of noviflumuron. This could be due to the acquisition of low dose of noviflumuron containing Nile Blue A, which was enough to give blue color but not sufficient to disrupt the cuticle formation or due to the metabolism of the noviflumuron by the time workers reached the molting stage in the 7 wk after adding treatment, as the half-life of noviflumuron is 4 wk (Karr et al. 2009). Based on these results, it can be concluded that the threshold of dose may differ among workers and it may increase for workers that are not ready to molt immediately upon acquiring the lethal dose. In other words, there is interplay between lethal dose and time for acquisition of lethal dose for the mortality to occur, where critical time of acquisition of lethal dose impacts workers ability to accomplish various stages (peristaltic movements, jackknife, exuviae-wrapped) of the molting process. Nonetheless, all these workers eventually died except in a few cases where workers managed to metabolize noviflumuron before the beginning of the molt. In conclusion, this study shows that with the exception of successful molting events during the first 10 d after baiting, mortality amongst workers with lethal doses of noviflumuron in the next molt is inevitable. The majority of workers that attempted to molt after 10 d died suggesting that the toxicant in a group of 5000 workers takes only 10 d to spread amongst the foraging population. Because lethal time for colony 64

65 elimination using CSI is dependent on the time taken by the workers to molt, the premature initiation of molting amongst workers using molt accelerating compounds (ecdysone or juvenile hormone agonists) could prove useful in reducing the lethal time for elimination, provided the lethal dose is acquired before the onset of the fasting period. 65

66 Wood in introduction chamber Treatment chamber Figure 4-1. The extended foraging arena- The foraging arena was composed of 6 small arenas connected to each other by a 6 m long coiled Tygon tubing to form a linear distance of 30 m. Each small planar arena was made of two sheets of Plexiglas filled with moistened sand. From the left, between planar arena 1 and planar arena 2 is the treatment chamber. The scale shows the length of a single planar arena. 66

67 Figure 4-2. (a) Picture of white worker that died in the jackknife posture, (b) white worker that died with exuviae-wrapped posture, (c) blue worker that died in jackknife posture, (d) blue worker that died in exuviae-wrapped posture under the effect of noviflumuron 67

68 Figure 4-3. Average number of molting workers (white= unfed on treatment and blue= fed on treatment) in extended foraging setup for control (a), and noviflumuron (b) treatments during a 9 wk study. 68

69 Figure 4-4. Average number of workers (white= unfed and blue= fed)) showing molt inhibitory effect of noviflumuron (jackknife or exuviae-wrapped) in extended foraging setup for (a) control, and (b) noviflumuron treatment during a 9 wk study. 69

70 Figure 4-5. Percentage of worker mortality recorded in 0-30 m extended foraging setup for (a) control, and (b) noviflumuron treatment during a 9 wk study. 70

71 CHAPTER 5 SPATIAL ASSESSMENT OF MOLTING IN COPTOTERMES FORMOSANUS WORKERS (ISOPTERA: RHINOTERMITIDAE) Introduction In the last two decades, molting disruption in C. formosanus workers with CSI incorporated baits has been important for their control (Rust and Su 2012). These baits are successful in elimination of subterranean termite colonies because the active ingredients are non-repellent, slow acting and their lethal time is dose independent (Su et al. 1982, Su and Scheffrahn 1988a). The lethal time of CSIs depends on the time to molt by workers, therefore these workers upon ingestion of lethal dose may move away from bait station before the onset of death. Hence, there is no aversion to the treatment site amongst other workers and the active ingredient is transferred to the entire colony leading to the complete colony collapse. Although a large number of studies have demonstrated the success of CSI baits in elimination of termite colonies in field (Grace and Su 2001, Su 2003), none have reported on the site of mortality of workers in a colony. In an attempt to speed up the CSI bait activity by premature molt initiation, the biggest challenge is to ensure that termites move away from the bait station before the onset of death of the majority of the workers in a colony. Mortality of a worker in a CSIbaited colony is a function of molting, and information on site of molting will help determine if speeding up the bait activity with the use of molt accelerating compounds (MACs) such as juvenile hormone agonists or ecdysone agonists is feasible. Molting in insects is an extensively studied area of research, but studies in the past were largely concentrated on solitary insects and limited information is available on the social groups of insects. The cryptic habitat of subterranean social insects impede observation of individuals resulting in limited research on this group. Subterranean 71

72 termites have large colony size and asynchronous molting events owing to periodically laid eggs by queen that further diminishes the possibility of conducting studies on molting behavior. Raina et al. (2008) studied the physiological process of molting in C. formosanus and reported that termites undergo a fasting period of 10 d before ecdysis in preparation of the molt. They also found that workers in field-collected foraging populations did not molt until the 11 th d post-collection. No molting for the first 10 d suggests the absence of workers undergoing a fasting period from the collected foraging populations. Based on this information it was hypothesized that termites might leave the foraging site at least 10 d before ecydsis and molt away from the foraging site such as in location near the nest. The objective of this study was to conduct spatial assessment of molting in C. formosanus colonies in extended foraging setup and to determine if reproductives in the nest influenced the molting site fidelity in a colony. Material and Methods For determination of the site of molting in a colony, data was obtained from preserved specimens in alcohol from a study of N.Y. Su (unpublished) conducted for determination of relationship between colony size and foraging distance of a colony. From this experiment, the additional data was obtained pertaining to the objective (site of molting) of the current study. Following elaborates the setup and the termite colonies used for their study. Colony Rearing Three 6-yr old laboratory-raised colonies of C. formosanus were used. Colonies were established by collecting alates from swarms of C. formosanus in New Orleans, LA. Paired male and female dealates were introduced into a cylindrical vial (8x 9 x

73 cm) containing moistened soil and pieces of wood. These vials remained covered with a perforated cap to allow aeration and were placed at 27 ± 0.5 o C. As the dealates reproduced and colony grew in size, each vial containing termites was moved into a bigger box (30 x 22 x 17 cm) that provided larger space for colony expansion. The colonies were provisioned with sufficient wood (Picea spp.,) and moistened soil for survival and growth in these units. These rearing units were stored at 27 ± 0.5 o C in the laboratory. Site of Molting The extended foraging setup of Su et al. (unpublished data) was made of a nest box containing a 6-yr old lab raised colony connected to a Plexiglas (60 x 60 x 0.9 cm in thickness) planar arena, which was connected to a linear series of small planar arenas in three directions (Figure 5-1) to allow termites to forage to the maximum distance for up to 6 months. After 6 months, the setup was disassembled to count the numbers of each caste at various distance from the initial nest box. Termites from the nest, and each planar arena were preserved in separate glass vials containing 95% ethanol and labeled with their respective distance from the nest. Preserved samples were used in this study to determine the site of molting based on the degree of sclerotization of mouthparts. For spatial assessment of molting, workers were categorized based on chronological events between two molt cycles, described by Xing et al The categories were 1) premolting workers- included workers with a separated old cuticle from the epidermis (Figure 5-2a), 2) molting workers- workers in the process of molting recognized by the jackknife and exuviae-wrapped postures formed during the process (Chapter 4) (Figure 5-2b), 3) postmolting- workers including those that had recently molted and were between 0-36 h post-ecdysis (Figure 5-2c), 73

74 and 4) intermolt workers- including those between postmolting and premolting stage (Figure 5-2c). The postmolting and intermolt workers were distinguished by degree of the sclerotization of mouthparts post-ecdysis. Workers post-molting were divided into a 0-20 h post-ecdysis and a h post-ecdysis group using methodology described in Chapter 2. Similarly, workers in the intermolt stage were also recognized by the intensity of sclerotization of their mouthparts. Each worker from planar arenas at various distances, central panel and the box was carefully observed under the stereo microscope (Leica M 205 C) and categorized into the four groups. The count of jackknife and exuviae-wrapped posture was combined under a molting workers group, and the count of workers in 0-20 h and h post-ecdysis was combined under the postmolting workers group before analysis. The data was pooled for small planar arenas at every 5 m interval across three directions. Pooled data was used for statistical analysis and presented in graphs. The effect of colonies on distribution of workers was tested using ANOVA with Tukey s HSD test before combining the data. Distribution of the intermolt workers was compared with the other stages between the two molt cycles over the foraging distance from the nest box to the last planar arena at 45 m using Pearson's Chi-squared test with simulated p- value. Because many planar arenas had fewer than 5 individuals, p-values were computed by Monte-Carlo simulation. All statistical analyses were performed using R 3.0. The graph represents the percentage of workers from each group over the distance. Nest Fidelity For determination of the site of molting within the nest, the study was conducted using a new extended foraging setup constructed solely for this study. The setup was 74

75 built using planar arenas made of two clear sheets of Plexiglas (60 x 60 x 0.9 cm in thickness) as described for the central planar arena above (Figure 5-3). The two clear sheets of Plexiglas (60 x 60 x 0.9 cm in thickness) separated by a Plexiglas laminates on 4 sides (5 cm in width and 0.2 cm in thickness) placed along the outer margins for maintaining an inner gap of 0.2 cm between sheets. The planar arena was partially filled with moistened sand and contained two blocks of moist piece of wood Picea sp. (5 x 0.5 x 0.2 cm), leaving the rest of the space for the termite colony. In the center of the upper sheet of the arena was an access hole, on top of which was a Plexiglas cup (4.5 cm in diameter and 7 cm in height) which was fitted to form a termite introduction chamber. Termites were introduced into the arena by placing the box (rearing unit) upon the introduction chamber (Figure 5-3). Unlike the setup used for Su et al. (unpublished study), in the current study termites including reproductives and brood were moved into the planar arena, which allowed the direct observation of molting incidence in workers in the foraging setup. On the bottom of the box a moist piece of wood (Picea spp.) piece was positioned so that one end reached the access hole of the Plexiglas planar arena while the other end connected with the box. The wood piece served as a connection for termites to the box to expand in to the central planar arena. The lid of the box was removed for the soil in the box to dry so that termites in the box were forced to move in to the central planar arena. Migration of termites (including the primary reproductive pair) from the box into the planar arena took ~ 2 wk depending on the size of the colony. Meanwhile, termites continued moving into the planar arena, the setup was extended in one direction and connected to three small planar arenas (24 x 24 cm and 1.4 cm in thickness), where each planar arena was connected through 5 m long coiled Tygon 75

76 tubes, making a total linear distance of 15 m from the original planar arena (Figure 5-3). Reproductives and brood (eggs/larvae or both) always stayed in the large central planar arena referred as nest. The box was removed upon introduction of the entire colony in to the central planar arena and moist wood pieces (Picea spp.,) were added to the introduction chamber. The arena was stored undisturbed in the dark at 27 ± 0.5 o C and termites were allowed to settle in the planar arena for a week. To determine if reproductives and eggs influence the nest fidelity of molting termites in a colony, location of thirty nonmolting workers and thirty molting workers in the central planar arena (nest) containing reproductives and eggs was determined (Figure 5-3). Molting and non-molting workers were only selected in the large central planar arena as they contained eggs and reproductive which together with other castes represent a nest, similar to the nest box in the first part of this study. The location of the non-molting workers was used to make comparisons between distance from molting individuals to eggs and reproductives vs. non-molting workers to eggs and reproductives to determine if the presence of eggs or reproductive in proximity affect the molting events. Direct observations were made to determine the location of molting individuals in the central planar arena and workers that were in jackknife posture were marked as molting workers. Comparisons were made for: 1) distance between egg mass and30 molting workers vs. 30 non-molting workers, 2) distance between reproductives and 30 molting workers vs. 30 non-molting workers, and 3) distance between 30 molting workers and egg mass vs. reproductives. The nonmolting workers location was determined to assess if the location of molting workers in relation to reproductives or eggs (dependent variables) is closer more than expected by 76

77 chance. Comparisons for the site of molting between above listed variables were made using t-test (using the program R-Project 3.0). Pairwise comparisons for each of the three combinations were adjusted by the Holm-Bonferroni method (α = 0.05). Results Study-1 Site of Molting There was a significant difference in the distribution of workers in the premolting stage compared with intermolt workers at various sites in colonies (χ 2 =998.09, P < 0.001), where intermolts were present in the nest box, central planar arena and in all the small planar arenas placed at 5 m interval (Figure 5-4). However, the distribution of workers at the premolting stage was limited to the nest box and central panel, while the majority of them (97%) were present in the nest box. Inside the nest box, intermolts were the most abundant (74%) followed by premolts that made up 22% of the population. There was a significant difference in the distribution of molting workers compared with intermolts (χ 2 =90.519, P < 0.001). The molting workers were limited to the nest because none of the workers in exuviae-wrapped and jackknife posture was found anywhere except the nest box. In one of the colonies, two exuviae from a molted worker were also found in the nest. The count of exuviae-wrapped and jackknife individuals was low and together represented 1.7% of the workers in the nest box. The percentage of molting individuals in the study represents individuals that were in the molting process at the time of placing termites in vials containing alcohol after disassembling the entire extended foraging arena setup. Corresponding to workers in the premolting and molting stages, the distribution of postmolts was also significantly different from intermolts (χ 2 =467.51, P < 0.001). Workers beyond 36 h post-ecdysis were indistinguishable from non-molted workers 77

78 (intermolts) and based on the degree of sclerotization of mouthparts, workers between 0-36 h post-ecdysis were limited to the nest box and the central planar arena. Seventy five percent of the total workers in post-molting stage were found in nest box and the remaining in the central planar arena. Their contribution of post-molting workers to the nest box population was only 1.3%. Study-2 Nest Fidelity The central planar arena contained workers, soldiers, eggs and the two reproductives (king and queen). Larvae were present only in one colony and they were close to the egg mass throughout the study period. Besides workers and soldiers, no other caste was present in any of the small planar arenas, while the molting workers were only observed in the central planar arena (the nest). The origin of the colony was not a significant factor for distance between molting workers and eggs (F 2, 87 = 0.65, P = 0.06) in the nest. When comparing the distance between the egg mass to non-molting workers with egg mass to molting workers, molting workers were significantly closer to the eggs than non-molting workers (t-test=15.3, P < 0.001), which indicates that molting event tend to occur in proximity of eggs (Figure 5-5a). The average distance between molting workers and eggs was 2.23 ± 1.2 cm (mean ± SD), where the farthest distance at which workers were found molting from the egg mass was 4.9 cm. However, no significant difference was found (t-test= 1.72, P = 0.08) on comparing the distance between molting individuals and reproductives (34.35±10.2) with the distance between nonmolting workers and reproductives (31.53 ±1.3 cm), indicating that workers molting close to the reproductives were only by chance (Figure 5-5b). Furthermore, the molting individuals were found to be significantly closer to the eggs than with reproductives (t- 78

79 test = 29.6, P < ), where the maximum distance between molting workers and eggs was ~5 cm and the distance between reproductives and molting workers varied widely depending on where egg mass and reproductives were in different colonies (Figure 5-5c). Discussion The presence of exuviae, and workers in jackknife and exuviae-wrapped postures in the nest box confirmed that workers of C. formosanus molt in the nest, where nest is referred to an area where reproductives rest with their brood. On examining the mandibles of individuals in the colony, some of the recently molted workers between h post-ecdysis were found to move out of the nest into the central planar arena (0-60 cm). Because workers post 36 h of ecydsis were indistinguishable from intermolts based on the degree of sclerotization of mouthparts, it was hard to track workers after 36 h of ecydsis in the colony. Workers of C. formosanus are known to acquire their gut fauna 4 d after ecdysis (Raina et al. 2008). Although by 36 h post-ecdysis their cuticle gets dark and mandibles look like typical intermolts (Chapter 2), the gut cuticle may take longer to develop for it to be able to hold the gut fauna and digest wood, which could explain the 4 d wait to reacquire the gut fauna. Another possibility is that it may take a few days for workers to stabilize the high volume of juvenile hormone in their body post-ecdysis, which has been associated with defaunation in termites (Haverty and Howard 1979). Considering that the first 4 d after ecydsis are crucial for reacquisition of gut fauna from nestmates, the workers stay close to the nest and may not resume foraging. Intermolts in the current study constituted 80% of the colony population where 72% of them were present in the nest box. Given that maintenance of the nest, feeding 79

80 immature and other castes, and grooming eggs and nestmates require a large work force, recovery of the majority of intermolts in the nest box in our study is reasonable. However, the presence of 19% of the worker population comprising premolts and postmolts found only in the nest box, suggests that molting may possibly affect the division of labor in workers of C. formosanus during their premolt and postmolt phase. Workers from foraging sites preparing to molt could be trading their roles with workers in the nest while the workers in the nest move to the foraging sites and help maintain labor homeostasis in the colony. Workers in C. formosanus colonies are known to forage around 100 m in the field (King and Spink 1969, Su and Scheffrahn 1988b). To feed a large colony with as many as millions of individuals (Rust and Su 2012), termites must employ an optimal foraging strategy in order to maximize the energy gain (Traniello and Leuthold 2000). The periodic return trips to the nest from foraging sites before molting can be a confounding factor to optimize foraging of C. formosanus. Thus, to overcome this and avoid loss of the labor force involved in procuring food at the foraging sites, the task switching strategy could be beneficial. Additional benefits in leaving the foraging sites before molting include; escape from predators at the time of shedding exuviae, efficient worker traffic in tunnels made exclusively for foraging, and help from nestmates in shedding exuviae during ecdysis. Termites are known to feed largely on cellulose, an abundant food source rich in carbon but deficient in nitrogen. To overcome the N deficiency in their diet, termites have adopted certain strategies (Higashi et al. 1990) and one of these is the post-molt recycling for recovery of N from the exuviae (Chapman 2013, Mira 2000). Nitrogen is 80

81 important for growth and reproduction in insects (McNeil and Southwood 1978, Mattson 1980), especially for the queen, which is the only reproducing body in a termite colony. Because workers are the sterile caste, it was assumed that workers molt in the nest to provision the queen with the N recycled from their exuviae. However, we could not confirm from our study if the exuviae was fed to the queen. On the contrary, workers were found to molt closer to the egg mass than reproductives in the nest box, where the average distance of molting workers from egg mass was 2.23 ± 0.12 cm (mean ± SE). Workers were observed to place eggs in a cluster as soon as they were laid by the queen in a nest. The eggs were often groomed by workers, which is an important behavior observed in other termites as well (Matsuura et al. 2000). In conclusion, we affirm workers in C. formosanus leave the foraging site before ecdysis and stay inside or close to the egg mass until at least 36 h post-molting in the nest. In light of the implications of this study, we suggest that if attempts are made to reduce the lethal time of colony elimination by initiating premature molting in workers with the use of MACs, regardless of how fast it is, there is a scope to reduce lethal time for colony elimination without any bait station aversion. The common problem with baits incorporated with toxicants (metabolic inhibitor) other than CSI is mortality of workers around the bait station (Su 2005). Because lethal time of these toxicants is dependent on the ingestion of dose, termites with high doses die around the bait station. The dead workers cause aversion to the traffic of workers towards the bait stations, resulting in blocked tunnels that connect the colony with the bait stations. However, the molting site fidelity confirmed from the current study, suggests that workers in CSI-baited colonies will die in the nest as they attempt to molt. Because termites in the premolt stage were 81

82 also found in the nest box, there should be enough time for workers to head back to the nest when stimulated to molt early under the effect of molting hormone agonists (MACs). 82

83 Figure 5-1. Central planar arena (60 x 60 x 0.9 cm in thickness) filled with moistened sand and extended in three directions (X, Y, and Z) through 5 m long Tygon tubes and attached to small planar arenas (24 x 24 x 0.6 cm in thickness) filled with moistened sand. Box on top of the arena contains 7-year old colony of C. formosanus. 83

84 Figure 5-2. a) Picture of worker in premolt stage with separated exuviae from the epidermis, b) worker in molting stage in a jackknife posture, c) worker in newmolt/postmolt stage recognized by its light color mandibles surrounded by workers in intermolt stage. 84

85 Figure 5-3. Central planar arena (60 x 60 x 0.9 cm in thickness) filled with moistened sand and extended 15 m in one direction through Tygon tubes. Three small planar arenas (24 x 24 x 0.6 cm in thickness) filled with moistened sand were attached at every 5 m along the linear foraging distance (= total of 15 m). 85