Biology of Blaptica dubia (Blattodea: Blaberidae) 
 
by 
 
Hao Wu 
 
 
 
 
A thesis submitted to the Graduate Faculty of 
Auburn University 
in partial fulfillment of the 
requirements for the Degree of 
Master of Science 
 
Auburn, Alabama 
August 3, 2013 
 
 
 
 
Keywords: Blaptica dubia, life history, instar determination,  
temperature-dependent development  
 
 
Copyright 2013 by Hao Wu 
 
 
Approved by 
 
Arthur G. Appel, Chair, Professor of Entomology 
Xing Ping Hu, Co-chair, Professor of Entomology 
Peng Zeng, Associate Professor of Statistics 
 
  
ii 
 
Abstract 
 
 
The Dubia Cockroach, Blaptica dubia Serville, has become a popular food choice among 
amphibian and reptile enthusiasts. However, there has been little scientific research conducted on 
this species. My research consists of three topics: instar determination, temperature-dependent 
development, and parental care. 
Instar determination is fundamental to both basic entomological research and its 
application. A new method using Gaussian mixture models to determine the number of instars in 
this species was developed. Application of the method is illustrated by analysis of data collected 
on B. dubia. The analysis indicates that there are seven instars in B. dubia and that the growth 
ratio follows the Brooks-Dyar rule. The growth ratio of pronotal length, pronotal width, and head 
width are 1.26, 1.24, and 1.19, respectively. Since B. dubia shares a similar growth pattern with 
other paurometabolous insects, this method may be applicable to other species as well. 
Temperature-dependent development of the nymphs of B. dubia was described using data 
collected from constant-temperature laboratory experiments. Simple linear regression was used 
for the data from each instar. Degree-days required to complete each instar were estimated as 
457.5, 668, 1031 .1317, 1515, 2071 for instars 1-7, respectively. The results could be used to 
control the development rate and find the optimal rearing conditions for B. dubia. 
The social behavior of cockroaches is less studied compared to other insect groups. Most 
information about cockroaches is from brief notes made during field studies. For parental care, 
most of these studies have been focused on the family Cryptocercidae and no studies have been 
iii 
 
conducted on the parental care behavior of the Blaberidae. The objective of this study was to 
measure the possible effects of parental care on the offspring of B. dubia. In this experiment, no 
parental care effects were detected although parental care behavior was observed in B. dubia. 
iv 
 
Acknowledgments 
 
 
I thank my major professor, Dr. Arthur Appel for supporting my research and guidance in 
the past two years, especially the great patience and encouragement. I am grateful to my co-
advisor, Dr. Xing Ping Hu, who offered me a great opportunity to complete my study, and 
provided me kind suggestions and endless support. I also would like to thank Dr. Peng Zeng, my 
committee member, who was always available when I had problem in statistics. 
I also want to thank our entire lab group including Dr. Zhonglin Yuan, Dr. Xiangli Dong, 
Charles Stephen, Yao Xu, Yuan Zeng, Zachary Devries, and Znar Barwary, who provided a lot 
of wonderful ideas and helped me get through the research problems by sharing and discussing. 
Ms. Marla Eva also helped me a lot in preparing the experiments. 
  
v 
 
Table of Contents 
Abstract ........................................................................................................................................... ii 
Acknowledgments.......................................................................................................................... iv 
List of Tables ................................................................................................................................ vii 
List of Figures .............................................................................................................................. viii 
Chapter 1 Introduction .................................................................................................................... 1 
Instar determination .................................................................................................................... 3 
Parental care ................................................................................................................................ 4 
Temperature on insect development ........................................................................................... 5 
Reference cited............................................................................................................................ 7 
Chapter 2 Instar Determination of Blaptica dubia (Blattodea: Blaberidae) using Gaussian 
Mixture Models ............................................................................................................................. 10 
ABSTRACT .............................................................................................................................. 10 
Introduction ............................................................................................................................... 10 
Material and Methods ............................................................................................................... 13 
Insect Rearing ....................................................................................................................... 13 
Measurements ....................................................................................................................... 13 
Statistical Model ................................................................................................................... 13 
Results and Discussion ............................................................................................................. 14 
Model Selection .................................................................................................................... 14 
Cluster Analysis .................................................................................................................... 15 
Result Assessment .................................................................................................................... 15 
vi 
 
References Cited ....................................................................................................................... 18 
Chapter 3 Temperature-dependent development of Dubia Cockroaches (Blattodea: Blaberidae) 27 
ABSTACT ................................................................................................................................ 27 
Introduction ............................................................................................................................... 27 
Material and Methods ............................................................................................................... 28 
Insect Rearing ....................................................................................................................... 28 
Experiment Design................................................................................................................ 29 
Data Analysis ........................................................................................................................ 29 
Results ....................................................................................................................................... 29 
Discussions ............................................................................................................................... 30 
References Cited ....................................................................................................................... 32 
Chapter 4 The parental care of Dubia cockroaches (Blattodea: Blaberidae) ................................ 41 
ABSTACT ................................................................................................................................ 41 
Introduction ............................................................................................................................... 41 
Material and Methods ............................................................................................................... 43 
Insect Rearing ....................................................................................................................... 43 
Experiment Design................................................................................................................ 43 
Measurements ....................................................................................................................... 43 
Data Analysis ........................................................................................................................ 44 
Results ....................................................................................................................................... 44 
Discussion ................................................................................................................................. 45 
References Cited ....................................................................................................................... 47 
vii 
 
List of Tables 
Table 2-1. The means and 95% simultaneous T2-intervals of three characters in each instar (mm) 
of B. dubia. .................................................................................................................................... 21 
Table 2-2. Linear regression models for three characters in the seven instars of B. dubia. ......... 22 
Table 3-1. Development time of each instar in different temperature. (mean (SD) n ) ................ 34 
Table 3-2. Linear regression models of temperature-dependent development in the seven instars 
of B. dubia. .................................................................................................................................... 40 
Table 4-1. Effect of parental care on head width (mm) of 7th instar. (mean ? SD) ...................... 49 
Table 4-2. Effect of parental care on pronotal length (mm) of 7th instar. (mean ? SD) ............... 50 
Table 4-3. Effect of parental care on pronotal width (mm) of 7th instar. (mean ? SD) ................ 51 
Table 4-4.  Effect of parental care on body weight (g) of 7th instar. (mean ? SD) ....................... 52 
Table 4-5. Effect of parental care on development time (day) from birth to adult. (mean ?  SD) 53 
viii 
 
List of Figures 
Figure 2-1. The seven instars of pronotal length (mm) and head width (mm) of B.dubia. .......... 23 
Figure 2-2. The seven instars of pronotal width (mm) and head width (mm) of B.dubia. ........... 24 
Figure 2-3.  The seven instars of pronotal width (mm) and pronotal length (mm) of B.dubia. ... 25 
Figure 2-4. The linear regression of three characters with 95% CI region in each instar. Pronotal 
length (Length), Pronotal width (Width), Head width (Head). ..................................................... 26 
Figure 3-1. The linear regression of development rate and temperature (20?C, 25?C, 30?C, 35?C) 
of 1st instar..................................................................................................................................... 35 
Figure 3-2. The linear regression of development rate and temperature (20?C, 25?C, 30?C) of 2nd  
instar. ............................................................................................................................................. 36 
Figure 3-3. The linear regression of development rate and temperature (20?C, 25?C, 30?C) of 3rd  
instar. ............................................................................................................................................. 37 
Figure 3-4. The linear regression of development rate and temperature (20?C, 25?C, 30?C) of 4th 
instar. ............................................................................................................................................. 38 
Figure 3-5. The linear regression of development rate and temperature (20?C, 25?C, 30?C) of 5th 
instar. ............................................................................................................................................. 39 
 
1 
 
Chapter 1 Introduction 
Cockroaches are insects of the order of Blattodea. The order name for cockroaches 
sometimes is given as Blattaria. They are closely related to the praying mantids (Mantodea), and 
often are grouped with them (as suborders) to form the order Dictyoptera. Termites (Isoptera) 
can also be placed in the order Dictyoptera, and are considered by some to be social cockroaches.  
Cockroaches are mostly secretive, primarily nocturnal, typically ground-dwelling insects. 
Cockroaches originated in the Carboniferous, approximately 359.2 million years ago. Modern 
cockroaches are believed to appear in Cretaceous approximately about 100 million years ago 
(Grimaldi and Engel 2005). Most cockroach species live in tropical and subtropical areas and it 
is estimated there are about 4000 species worldwide (Roth 2003). Few of them, only about 30 
species, successfully enter into urban environment and become very import urban pests. 
In the 10th edition of Systema Natur? (Linnaeus 1758), 15 species of Blattodae  are 
named.  McKittrick (1964) examined the external genitalia, oviposition behavior, and crop 
structure in a wide variety of species. She suggested that cockroach evolution proceeded in two 
lines, one leading to the superfamily Blattoidea (families Cryptocercidae and Blattidae), and the 
other to the superfamily Blaberoidea (families Polyphagidae, Blattellidae, and Blaberidae). 
Based on the study of McKittrick (1964), Roth (2003) recognized six families of cockroaches: 
Polyphagidae, Cryptocercidae, Nocticolidae, Blattidae, Blattellidae, and Blaberidae; the majority 
of cockroaches fall into the latter three families. 
The family Blaberidae (1020 species) is the most recently evolved cockroach family and 
the one that has undergone the most extensive adaptive radiation. The group is primarily tropical 
and contains the largest cockroach species. Its members are generally found under logs, in 
2 
 
humus, etc., though some species are arboreal. A few species occasionally become associated 
with humans, such as, Pycnoscelis surinamensis Linnaeus (the Surinam cockroach), Rhyparobia 
maderae Fabricius (the Madeira cockroach), and Nauphoeta cinerea Olivier (the lobster 
cockroach). P. surinamensis may be found in greenhouses, in warmer climates, or outdoors 
where it can badly damage the roots of crops; it is also found in chicken houses and is known to 
be an intermediate host for the chicken eyeworm nematode (Gillott 2005). There are also some 
cockroaches species in the Blaberidae that are very popular pets. The Madagascar hissing 
cockroach (Gromphadorhina portentosa Schaum ), also known as the hissing cockroach, is one 
of the largest species of cockroach, reaching 5.1?7.6 cm as adult.  They are from the island of 
Madagascar off the African mainland. 
Blaptica dubia Serville (commonly known as the Dubia cockroach, Guyana spotted 
cockroach, or Orange-spotted cockroach) is a large (up to ca. 4.5 cm in length), sexually 
dimorphic blaberid cockroach. Males have wings, but they rarely fly. Because they lack 
developed arolium between their claws, neither adults nor nymphs can climb smooth surfaces. 
B.dubia is an ovoviviparous species and gives birth to live young. This reproduction pattern is 
typical of the family Blaberidae of Blattodea (Mckittrick 1964, Roth 2003). More importantly, 
this species is used by pet lovers as a food source for many reptiles and amphibians. Compared 
to other foods such as crickets, B. dubia contains a higher percentage of protein, is easier to 
maintain, and unlike other cockroaches, produces little odor. Consequently, they have become an 
increasingly popular food among amphibian and reptile enthusiasts. 
My study is focused on the instar determination, parental care, and temperature 
development of the B. dubia. 
3 
 
Instar determination  
Growth of insect larva is continuous, but the growth of the exoskeleton is not (Gullan and 
Cranston 2005).  Heavily sclerotized structures, such as headcapsules, remain approximately the 
same size during a stadium. The most measurable changes in size occur following molts. These 
special characters can be used to differentiate instars. There is a long history of entomologists 
trying to quantify changes in the size of body parts for a wide range of insects. One of the earliest 
attempts in this area was that of Dyar, who developed a rule from observations on the caterpillars 
of 28 species of Lepidoptera in 1890. Dyar's measurements showed that the width of the head 
capsule increased by a ratio (range 1.3-1.7) that was constant for a given species. (Dyar 1890) 
A related empirical "law" of growth is Przibram's rule, which states that an insect's 
weight is doubled (2.09) during each instar, and at each molt all linear dimensions are increased 
by a ratio of 1.26 (Wigglesworth 1984, Solow and Faber 1995). However, the growth of most 
insects shows no general agreement with this rule, which assumes that the dimensions of a part 
of the insect body should increase at each molt by the same ratio as the body as a whole, which is 
not the case in most situations (Gullan and Cranston 2005). Besides, the food consumption of 
different individuals and the differences between their development stages also makes their 
weight vary greatly. These factors may be the primary reason that Przibram?s law is not as 
widely used as Dyar?s law. There is an area of research called allometry that addresses these 
problems. 
Morphological measurements can be viewed as normal distributions in each instar. Thus, 
samples from a given population of insects are a mixture of several normal distributions, one for 
each instar. This assumption is reasonable because measurements from individuals from the 
same ontogenetic stage are often distributed normally (Sokal and Rohlf 1994). In recent years, 
4 
 
the development of technology makes many computer intensive analytical methods possible. 
Parental care 
Parental care involves any sort of parental behavior that increases the survival of the 
offspring. Parental care has been described in most animal phyla, but is especially well 
developed in numerous species of insects, crustaceans, and vertebrates. The amount and type of 
parental care provided to offspring is quite diverse. The most common parental care is to assist 
the offspring in living through harsh physical conditions, making resources available, which are 
otherwise difficulty or impossible to obtain, and decreasing risk of predation on eggs or young 
(Clutton-Brock and Scott 1991, Glazier 2002). Although levels of complexity are various, 
parental care can be divided into three major categories: (a) those that physically protect the 
young from danger, (b) those that protect resources vital to offspring, and (c) those that facilitate 
offspring feeding (Glazier 2002).  
Parental care in insects ranges from covering eggs with a protective coating to remaining 
to feed and protect young, to forming eusocial societies. A variety of forms are widespread in 
various insect orders; parental care is most developed in Hemiptera, Coleoptera, Thysanoptera, 
Embiidina, Hymenoptera and Isoptera . 
Most cockroaches show some form of parental care in the broad sense, because even the 
most primitive cockroaches have ootheca (egg cases) to protect the young before they hatch. If 
only post-hatch parental care is considered, most of the information on cockroaches is from brief 
notes made during field work (Bell et al. 2007) and more of these are focused on species of 
Cryptocercidae (Seelinger and Sellinger 1983, Nalepa 1987, 1988, Park and Choe 2003). There 
is currently no information concerning B.dubia. 
5 
 
Temperature on insect development 
Most insects are ectothermic, i.e., the body temperature directly varying with the 
environmental temperature. Heat is the main factor impacting the growth rate and development 
when food is unlimited. The effect of temperature on insect development is rather a basic topic in 
insect biology. In the last century, numerous studies developed appropriate developmental rates 
functions for phonological or population models that can be used under a variety of conditions to 
predict important events in insect life cycles, or to estimate insect abundance for specific control 
strategies. The development of specific functions for each developmental stage is determined by 
rearing insects at range of constant temperatures (Damos and Savopoulou-Soultani 2012). 
The earliest and still widely used model was the linear degree-days model (Arnold 1959). 
The calculations of lower developmental thermal thresholds and degree-day have long been used 
for the estimation of developmental times under different constant temperatures. This method 
has the advantage of simplicity and allows estimation of the developmental threshold and the 
degree-day requirement of insect species. The principle and the detailed methods are well 
described in general entomology textbooks (e.g., Gullan and Cranston 2005). 
While the degree-days method, which assumes a linear relationship between 
developmental rate and temperatures, is often suitable for intermediate temperatures, this model 
does not consider nonlinearity at high and low temperatures. Therefore, it will produce 
considerable error when temperature conditions are toward the extremes under variable 
conditions.  
Various nonlinear models also have been developed to improve these results. Many 
nonlinear models have been developed by entomologists, such as, the Belehradek model, and the 
logistic equation of Davidson (Briere et al. 1999). More models are provided by Logan (Logan et 
6 
 
al. 1976), Schoolfield (Schoolfield et al. 1981), and Lactin (Lactin et al. 1995). 
  
7 
 
Reference cited 
Arnold, C.Y. 1959. The determination and significance of the base temperature in a linear heat  
 unit system. Proc. Amer. Soc. Hort. Sci. 74:430-445. 
Bell, W. J., L. M. Roth, and C. A. Nalepa. 2007. Cockroaches: ecology, behavior, and natural 
history. JHU Press. Baltimore, Maryland. 
Briere, J.-F., P. Pracros, A.-Y. Le Roux, and J.-S. Pierre. 1999. A novel rate model of  
 temperature-dependent development for arthropods. Environ. Entomol. 28:22-29. 
Clutton-Brock, T. H., and D. Scott. 1991. The evolution of parental care. Monographs in  
 behavior and ecology. Princeton University Press, Princeton, New Jersey. 
Damos, P., and M. Savopoulou-Soultani. 2012. Temperature-driven models for insect  
 Development and Vital Thermal Requirements. Psyche: A Journal of Entomology. 2012: 1?
13. 
Dyar, H. G., and R. N.Y. 1890. The number of molts of lepidopterous larvae. Psyche. 5: 420? 
 422. 
Gillott, C. (2005). Entomology. Springer. New York. 
Glazier. 2002. Parental Care. Encyclopedia of Evolution. Oxford University Press. 
Grimaldi, D., and M. S. Engel. 2005. Evolution of the insects. Cambridge University Press. 
Gullan, P. J., and P. Cranston. 2005. The Insects: An Outline of Entomology, 3rd ed. Wiley- 
 Blackwell, Malden, Massachusetts. 
Lactin, D. J., N. J. Holliday, D. L. Johnson, and R. Craigen. 1995. Improved rate model of 
 temperature-dependent development by arthropods. Environ. Entomol. 24: 68-75. 
Logan, J. A., D. J. Wollkind, S. C. Hoyt, and L. K. Tanigoshi. 1976. An analytic model for  
 description of temperature dependent rate phenomena in arthropods. Environ.  
8 
 
 Entomol. 5: 1133-1140. 
Mckittrick, F. A. 1964. Evolutionary studies of cockroaches. Cornell University Agricultural  
 Experiment Station Memoir. Ithica, New York. 
Nalepa, C. A. 1987. Life history studies of the woodroach Cryptocercus punctulatus scudder  
 and their implications for the evolution of termite eusociality. PhD thesis. North Carolina 
State University . 
Nalepa, C. A. 1988. Cost of parental care in the woodroach Cryptocercus punctulatus Scudder  
 (Dictyoptera: Cryptocercidae). Behav. Ecol. and Sociobiol. 23: 135?140. 
Park, Y., and J. Choe. 2003. Effects of parental care on offspring growth in the Korean wood- 
 feeding cockroach, Cryptocercus kyebangensis. J. Ethology. 21: 71?77. 
Roth, L. M. 2003. Systematics and phylogeny of cockroaches (Dictyoptera: Blattaria). Oriental 
 Insects. 37: 1?186. 
Schoolfield, R. M., P. J. H. Sharpe, and C. E. Magnuson. 1981. Non-linear regression of  
 biological temperature-dependent rate models based on absolute reaction-rate theory.  
 J. Theor Biol. 88: 719?731. 
Seelinger, G., and U. Sellinger. 1983. On the social organisation, alarm and fighting in the  
 primitive cockroach Cryptocercus punctulatus Scudder. Zeitschrift f?r Tierpsychologie. 61: 
 315?333. 
Sokal, R. R., and F. J. Rohlf. 1994. Biometry: The principles and practices of statistics in  
 Biological Research.. W. H. Freeman, New York, New York. 
Solow, A. R., and D. B. Faber. 1995. A test of Przibram?s Rule. J.  Zool. 237: 159? 
 161. 
Wigglesworth, V. B. 1984. Insect Physiology. Wiley, New York, New York.. 
9 
 
  
10 
 
Chapter 2 Instar Determination of Blaptica dubia (Blattodea: Blaberidae) using Gaussian 
Mixture Models 
ABSTRACT Instar determination is fundamental to both basic entomological research and its 
application. The cockroach, Blaptica dubia Serville, is a popular pet and an excellent feeder 
insect for many reptiles and amphibians. A new method using Gaussian mixture models to 
determine the number of instars in this species is developed. Application of the method is 
illustrated by analysis of data collected on B. dubia. The analysis indicates that there are seven 
instars in B. dubia and that the growth ratio follows the Brooks-Dyar rule. The growth ratio of 
pronotal length, pronotal width, and head width are 1.26, 1.24, and 1.19, respectively. Since B. 
dubia shares a similar growth pattern with other paurometabolous insects, this method may be 
applicable to other species as well. 
KEYWORDS Instar, Brooks-Dyar Rule, Model based clustering 
Introduction  
Instar determination is fundamental to both basic entomological research and application (Logan 
et al. 1998). Determining instar distributions from a given population is required for life table 
analysis, key factor analysis, and other important ecological investigations. The number of 
instars is useful in the development of phenology models, or in the refinement of existing models 
(Godin et al. 2002). In paleobiology, understanding the growth, evolution, and ecology of extinct 
species often depends on calibrating ontogeny with respect to either absolute time or relative 
developmental age (Hunt and Chapman 2001). In forensic studies, instar determination helps to 
estimate the minimum postmortem interval (Vel?squez and Viloria 2010).  
The number of instars varies widely across insect species. Instar number is frequently 
considered to be invariable within species, although intraspecific variability in the number of 
11 
 
instars is not an exceptional phenomenon (Esperk et al. 2007). In many species of Arthropoda, 
heavily sclerotized structures, such as the head capsule, remain approximately the same size 
during a stadium (Daly 1985). Measurable changes in size occur following molts. These 
sclerotized characters, particularly head capsule, have been used to differentiate instars. 
For instar determination, direct observation may be most accurate, but in many cases, 
observations are not possible or it may take too much time to obtain the results (in our 
observation, it takes about 6 months at 30 ? 2?C). There is a long history of entomologists 
quantifying changes in size of particular structures for a wide range of insects. It has long been 
supposed that the ratio of measurements of well-sclerotized structures in any instar to the 
equivalent measurement in the preceding instar tends to be constant throughout the life history 
(Hutchinson and Tongring 1984). Many different techniques have been proposed for estimating 
either the growth stage (instar) or the age of insects (Daly 1985). The original observations on 
arthropod sizes and developmental stages were made by W. K. Brooks on stomatopod larvae 
collected by the H.M.S. Challenger (Brooks 1886). In Brooks?s observation, the length of an 
instar is five-fourths of its predecessor. Dyar later independently came to the same conclusion of 
equivalent size ratios from observations on the caterpillars of 26 species of Lepidoptera in 1890. 
Dyar?s measurements showed that the width of the head capsule increased by a ratio, ranging 
from 1.3-1.7, that was constant for a given species (Dyar 1890): 
                                                       
However, the sizes of different instars may overlap, making instar determination more difficult. 
Morphological measurements can be viewed as normal distributions in each instar. Thus, 
samples from a given population of insects are a mixture of several normal distributions, one for 
each instar. The probability density function of a mixture of G normal distributions can be 
12 
 
written as: 
                                                    
 
where   is the mixing proportion,  , ,  is the  p-dimensional normal density 
function , with mean vector , and covariance matrix   . This assumption is 
reasonable because measurements of individuals from the same ontogenetic stage are often 
distributed normally (Hunt and Chapman 2001). 
Distribution based methods have been adopted by many entomologists for a variety of 
species (Logan et al. 1998, Hunt and Chapman 2001, Hammack et al. 2003, Gullan and Cranston 
2005, Delbac et al. 2010). Almost all of these methods are based on the widths of head capsules. 
There are some efforts to use different characters to determine instars (McClellan and Logan 
1994, Vel?squez and Viloria 2010), but the characters are used independently. Since different 
characters of insects are highly correlated, it is more statistically appropriate to adopt a 
multivariate approach.  
Blaptica dubia Serville (commonly known as the Dubia cockroach, Guyana spotted 
cockroach, or Orange-spotted cockroach) is a large (up to ca. 4.5 cm in length), sexually 
dimorphic blaberid cockroach that is often kept as a pet. More importantly, this species is used as 
a food source for many reptiles and amphibians. Compared to other food sources such as 
crickets, B. dubia contains a higher percentage of protein, is easier to maintain, and produces 
little odor. Consequently, they have become an increasingly popular food among amphibian and 
reptile enthusiasts. In B. dubia, the head capsule width, pronotal width, and body length are the 
most convenient characters to measure in all nymphal instars. These three characters are used 
13 
 
together in the analysis to determine instars. Since the three characters can be seen as a 
multivariate normal distribution in each instar, they form a Gaussian mixture when combined 
(Hastie et al. 2003). As a result, a new method using Gaussian mixture models to determine the 
number of instars in this species is developed. The results are assessed by the Brooks-Dyar rule 
and verified by direct observations on the development of B. dubia in the laboratory and 
corroborated with results of a previous study (Hintze-Podufal and Nierling 1986).  
Material and Methods 
 Insect Rearing. Blaptica dubia were reared in clear gallon glass jars (3.8 L, 15.24 cm 
width by 25.4 cm height) with cardboard harborage at 30 ? 2?C. They were exposed to a 
photoperiod of 12:12 (L:D) h and supplied with water and dry dog Chow (Purina? Dog Chow?, 
Ralston Purina, St. Louis, MO) ad libitum. 
Measurements. A total of 1925 nymphs of B. dubia were measured from Jan 2012 to 
Apr 2012. Head capsule width (the widest distance between the two compound eyes) was 
measured to 0.01 mm using a microscope (Leica MZ 6, Solms, Germany) with an ocular 
micrometer. The length and width of the pronotum were measured to 0.01mm using an electronic 
digital caliper (Model: 62379-531, Control Company, Friendswood, TX). 
Statistical Model. In model-based clustering, it is assumed that data are generated by a 
mixture of underlying probability distributions , in which each component k represents a 
different group or cluster  (Johnson and Wichern 2007). If there are G clusters, the observation of 
variables is modeled as arising from the mixture distribution:  
 
where   is the mixing proportion,  , ,  is the  p-dimensional normal density 
14 
 
function , with mean vector , and covariance matrix   . The Gaussian mixture 
model for one observation with G components is: 
 
Inference is based on the likelihood, which for N objects and a fixed number of clusters 
G is (Johnson and Wichern 2007): 
                    
The Bayesian Information Criterion (BIC), 
 
Where Lmax is the maximum of  likelihood function, N is the number of observations, G is 
number of component, and p is number of variables. It has been used to determine the number of 
clusters (Fraley and Raftery 1998). In instar determination, the BIC gives a quantitative method 
to determine the number of instars for a given insect. 
Fraley and Raftery (2012) developed the R (R Development Core Team 2012) package 
?mclust? for model-based clustering based on a parameterized Gaussian mixture model (Fraley 
and Adrian 1998). The software package ?mclust? was used in the data analysis. 
Results and Discussion 
Model Selection. After comparing the BIC of every cluster (sub-model) and all the 
parameterized models, the unconstrained model with seven clusters was selected because it has 
the largest BIC value, -6232.165 (range -25282.88 ~  -6232.165). Therefore, there are seven 
instars in B. dubia when reared as described above.  
15 
 
Cluster Analysis. Gaussian mixture model based clustering was conducted based on the 
model selection using the BIC. Figs. 1-3 show the cluster results of three characters. From the 
scatter plots, there are seven instars in B. dubia and the characters of each instar cluster together. 
The means and 95% simultaneous -intervals (similar to confidence intervals, but for 
multivariate data) of the three characters in each instar are shown in Table 1. The pronotal length 
is in the range of 2.41 to 9.37 mm; the pronotal width is between 4.18 mm and 15.17 mm and the 
range of head width is 1.79 to 5.04 mm, for first and seventh instars, respectively (Table 1). 
Based on the result in the cluster analysis, predictions can be made on new observations, using 
the function provided in the mclust package.  
Result Assessment. According to the Brooks-Dyar rule (Dyar  1890):   
  where   and  . This equation can also be written 
as: 
 
Which is equivalent to:  
                                                            , 
a linear function. We can, therefore, assess the cluster results using linear regression models. The 
constant growth ratio is . From Fig. 4 and Table 2 the linear relationships of the three 
characters agree with the Brooks-Dyar rule. In our analysis, the growth ratios of pronotal length, 
pronotal width, and head width are 1.26, 1.24, and 1.19, respectively (Table 1). The growth rates 
of pronotal length and width are different and the shape of the pronotum changes slightly during 
development. According to Cole (1980), the median growth ratio for hemimetabolous insects is 
1.27 , and the mean is 1.27 ? 0.011 (n = 50, ? = 0.05). For holometabolous insects the median is 
16 
 
1.52, and the mean is 1.55 ? 0.033 (n = 55, ? = 0.05). This indicates that all instars are 
represented in the samples and that the model-based approach gives reasonable results.  
In another study, the development of B. dubia was monitored at the same conditions as 
stated above [30 ? 2?C,12:12 (L:D) h]. We directly observed seven instars in B. dubia over a 
period of 6 months.  In addition, Hintze-Podufal and Nierling (1986) also indicated that there are 
seven instars in B. dubia reared at 28 ? 2?C. 
Although the results of our analyses support the Brooks-Dyar rule, they do not rely on it. 
Model-based clustering can also be used for other species even though their growth may not 
follow Brooks-Dyar rule (Bliss and Beard 1954, Klingenberg and Zimmermann 1992), as long as 
there are size differences between different instars. 
All of the studies about instar determination that use statistical techniques are based on 
the normal distribution of the given characters in each instar (Gullan and Cranston 2005, Hunt 
and Chapman 2001, Godin et al. 2002, Logan et al. 1998, Delbac et al. 2010). Our approach uses 
the multinormal distribution of characters in each instar and this is the only assumption in our 
analysis. Distributions were examined visually using the bivariate plot of the characters. 
Compared to the widely adopted method of McClellan and Logan (1994), model-based 
clustering provides a better approach to solve the instar determination problem. Instar 
determination using Gaussian mixture models successfully solved the problem of determining 
the number of instars using the BIC. In the McClellan and Logan (1994) method, the number of 
instars is based on direct observations and the initial value of the mean and variance of each 
instar are based on this observation. This approach may produce some bias because different 
observers may have various initial values. 
The clustering method based on Gaussian mixture models described above illustrates that 
17 
 
it can be successfully used to determine instars in B. dubia. Since no characters in our analysis 
are specific to B. dubia, the same cluster methods should be applicable to other species as well. 
18 
 
References Cited 
 
Bliss, C. I., and R. L. Beard. 1954. The growth of the head capsule in individual milkweed  
 bugs. Ann. Entomol. Soc. Am. 47: 5. 
Brooks, W. K. 1886. Report on the Stomatopoda collected by H. M. S. Challenger during the  
 years 1873-76. In The Voyage of H.M.S. Challenger. Zool. 45:5. 
Cole, B. J. 1980. Growth ratios in holometabolous and hemimetabolous Insects. Ann. Entomol.  
 Soc. Am. 73: 489-491. 
Daly, H. V. 1985. Insect morphometrics. Ann. Rev. Entomol. 30: 415-438. 
Delbac, L., P. Lecharpentier, and D. Thiery. 2010. Larval instars determination for the  
 European Grapevine Moth (Lepidoptera: Tortricidae) based on the frequency distribution of  
 head-capsule widths. Crop Protect. 29: 623-630. 
Dyar, H. G. 1890. The number of molts of lepidopterous larvae. Psyche. 5: 420-422. 
Esperk, T., T. Tammaru, and S. Nylin. 2007. Intraspecific variability in number of larval  
 instars in insects. J. Econ. Entomol. 100: 627-645. 
Fraley, C., and A. E. Raftery. 1998. How many clusters? Which clustering method? Answers  
 via model-based cluster analysis. The Computer J. 41: 578 - 588. 
Fraley, C., and A. E. Raftery. 2012. MCLUST Version 4 for R: Normal mixture modeling for  
 model-based clustering, classification, and density estimation. Technical Report no.597,  
 Department of Statistics, University of Washington, USA. 
Godin, J., P. Maltais, and S. Gaudet. 2002. Head capsule width as an instar indicator for larvae  
 of the cranberry fruitworm (Lepidoptera: Pyralidae) in Southeastern New Brunswick. J.  
 Econ. Entomol. 95: 1308-1313. 
Gullan, P. J., and P. Cranston. 2005. The Insects: An Outline of Entomology, 3rd ed. Wiley- 
19 
 
        Blackwell, Malden, MA. 
Hammack, L., M. M. Ellsbury, R. L. Roehrdanz, and J. L. Pikul. 2003. Larval sampling and  
 instar determination in field populations of northern and western corn rootworm ( 
 Coleoptera: Chrysomelidae). J. Econ. Entomol. 96: 1153-9. 
Hastie, T., R. Tibshirani, and J. H. Friedman. 2003. The Elements of Statistical Learning.  
 Springer, New York. 
Hintze-Podufal, C.  and Nierling, U. 1986 Der Einfluss der Nahrung auf Entwicklung,  
Wachstum und Prareproduktionsphase von Blaptica dubia Stal (Blaberoidea, Blaberidae).  
Mitteilungen der Schweizerischen Entomologischen Gesellschaft. 59: 177-186. 
Hunt, G., and R. E. Chapman. 2001. Evaluating hypotheses of instar-grouping in arthropods: a  
 maximum likelihood approach. Paleobiol. 27: 466-484. 
Hutchinson, G. E., and N. Tongring. 1984. The possible adaptive significance of the Brooks- 
 Dyar rule. J. Theor. Biol. 106: 437-439. 
Johnson, R. A., and D. W. Wichern. 2007. Applied Multivariate Statistical Analysis, 6 ed. ed.  
 Prentice Hall, Upper Siddle River, New Jersey. 
Klingenberg, C. P., and M. Zimmermann. 1992. Dyar?s rule and multivariate allometric  
 growth in nine species of waterstriders (Heteroptera: Gerridae). J. Zool. 227: 453-464. 
Logan, J. A., B. J. Bentz, J. C. Vandygriff, and D. L. Turner. 1998. General program for  
 determining instar distributions from head capsule widths: Example analysis of mountain  
 pine beetle (Coleoptera: Scolytidae) data. Environ. Entomol. 27: 555-563. 
McClellan, Q. C., and J. Logan. 1994. Instar determination for the gypsy moth (Lepidoptera:  
 Lymantriidae) based on the frequency distribution of head capsule widths. Environ.  
 Entomol. 23: 6. 
R Development Core Team. 2012. R: A Language and Environment for Statistical Computing.  
20 
 
 Vienna, Austria. 
Vel?squez, Y., and ?. L. Viloria. 2010. Instar determination of the neotropical beetle 
Oxelytrum discicolle (Coleoptera: Silphidae). J. Med. Entomol. 47: 723-726. 
 
21 
 
 
Table 2-1. The means and 95% simultaneous T2-intervals of three characters in each instar (mm) of B. dubia. 
 Instar 1 Instar 2 Instar 3 Instar 4 Instar 5 Instar 6 Instar 7 Ratio 
N 127 178 146 280 394 408 285  
Pronotal length 2.41(2.37-2.45) 2.97(2.91-3.03) 3.82(3.77-3.87) 4.91(4.84-4.98) 6.17(6.11-6.23) 7.54(7.46-7.63) 9.39(9.30-9.47) 1.26 
Pronotal width 4.18(4.14-4.22) 5.21(5.13-5.30) 6.47(6.40-6.54) 8.01(7.89-8.13)  9.95(9.86-10.04) 12.28(12.15-12.41) 15.17(15.04-15.30) 1.24 
Head width 1.79(1.78-1.80) 2.08(2.06-2.11) 2.49(2.47-2.51) 2.97(2.94-3.00) 3.55(3.53-3.58) 4.24(4.21-4.28) 5.04(5.00-5.07) 1.19 
22 
 
 
Table 2-2. Linear regression models for three characters in the seven instars of B. dubia. 
Character Slope ? SE Ratio* Intercept ? SE R2 F (df = 
1,5) 
P ( > |t|) 
Pronotal Length 0.2278 ? 0.0039 1.26 0.6573 ? 0.0176 0.9988 4082 < 0.0001 
Pronotal Width 0.2136 ? 0.0011 1.24 1.2235 ? 0.0047 0.9999 108300 < 0.0001 
Head Width 0.1734 ? 0.0014 1.19 0.3979 ? 0.0064 0.9996 12990 < 0.0001 
*Ratio is eslope 
  
23 
 
 
Figure 2-1. The seven instars of pronotal length (mm) and head width (mm) of B.dubia. 
  
24 
 
 
Figure 2-2. The seven instars of pronotal width (mm) and head width (mm) of B.dubia. 
  
25 
 
 
Figure 2-3.  The seven instars of pronotal width (mm) and pronotal length (mm) of B.dubia. 
  
26 
 
 
Figure 2-4. The linear regression of three characters with 95% CI region in each instar. 
Pronotal length (Length), Pronotal width (Width), Head width (Head). 
  
27 
 
Chapter 3 Temperature-dependent development of Dubia Cockroaches (Blattodea: 
Blaberidae) 
ABSTACT Temperature-dependent development of the nymphs of the Dubia Cockroach, 
Blaptica dubia Serville, was described using data collected from constant-temperature laboratory 
experiments. Simple linear regression models were developed based the data from each instar. 
Degree-days required to complete a particular life stage were estimated as 457.5, 668, 1031, 
1317, 1515, 2071 for instars 1-7, respectively. The results could be used to control the 
developmentl rate of B. dubia and find the optimal temperature to rear B.dubia in a given 
situation. 
KEYWORDS Temperature-dependent development, Blaptica dubia 
Introduction 
Blaptica dubia Serville (commonly known as the Dubia cockroach, Guyana spotted 
cockroach, or Orange-spotted cockroach) is a large (up to ca. 4.5 cm in length), sexually 
dimorphic blaberid cockroach that is often kept as a pet. More importantly, this species is used as 
a food source for many reptiles and amphibians. Compared to other food sources such as 
crickets, B. dubia contains a higher percentage of protein, is easier to maintain, and produces 
little odor. Consequently, they have become an increasingly popular food among amphibian and 
reptile enthusiasts (Wu et al. 2013). Development of the temperature-dependent development 
model for B. dubia is essential to control the growth rate of B. dubia and find the optimal 
solution to rear this species under a given situation. 
The effect of temperature on development and growth of ectotherms has been well 
studied (Dixon et al. 2009). In the last century, numerous studies developed appropriate 
developmental rates functions for phonological or population models that can be used under a 
28 
 
variety of conditions to predict important events in insect life cycles, or insect abundance for 
specific control strategies. The development function for each stage is determined by rearing 
insects at range of constant temperatures (Briere et al. 1999, Trugill et al. 2005). The earliest and 
still most widely used model is the linear degree-days model. The calculations of lower 
developmental thermal thresholds and degree-day have long been used for the estimation of 
developmental times under different constant temperatures. This method has the advantage of 
simplicity and allows estimation of the developmental threshold and the degree-day requirement 
of insect species. The principle and the detailed methods are well described in Gullan and 
Cranston (2005). 
Various nonlinear models also have been developed to improve degree-day models. 
Many nonlinear models have been developed by entomologists, such as the Belehradek model 
and the logistic equation of Davidson (Briere et al. 1999). More models are provided by Logan et 
al. (1976), Schoolfield et al. (1981), and Lactin et al. (1995). 
While the degree-days method, which assumes a linear relationship between 
developmental rate and temperature, is often suitable for intermediate temperature, this model 
does not consider nonlinearity at high and low temperatures. Therefore, it will produce 
considerable error when temperature conditions are toward the extremes, especially under 
variable conditions (Gullan and Cranston 2005). However, in our analysis, degree days are still 
used due to the limited available data and in an effort to better compare our results with others. 
Material and Methods  
Insect Rearing:  Blaptica dubia were reared in clear gallon glass jars (3.8 L, 15.24 cm 
width by 25.4 cm height) with cardboard harborage at 30 ? 2?C. They were exposed to a 
photoperiod of 12:12 (L:D) h and supplied with water and dry dog Chow (Purina? Dog Chow?, 
29 
 
Ralston Purina, St. Louis, MO) ad libitum. 
Experiment Design:  Couples of adult females and males were kept separately in clear 
half gallon glass jar (1.9 L, Dimensions: 5.5" width x 6.25" height) in conditions stated above. 
After neonates were born, they were separated from their parents and divided into groups of 10 
nymphs. Six groups (60 nymphs) were set up in every temperature. In this study, temperature 
was set at 20?C, 25?C, 30?C, 35?C, and 40?C and exposed to a photoperiod of 12:12 (L:D). The 
period of each instar was recorded and the development time (days) was calculated. Simple 
linear regression models of development were developed. The relationship between development 
time and temperature can be expressed as: 
 
Where D is the development time and T is the termperature. Degree days are calculated based on 
the linear regression of development rate and time. 
Data Analysis:  Development rates for each instar were calculated using reciprocal of the 
average number of days (i.e., 1/days) required to complete a particular life stage. The 
relationship between development rate and temperature was described by a linear model (Gullan  
and Cranston 2005). Linear regression (Faraway 2004, Ramsey and Schafer 2012) was 
conducted on data collected for each instar using the R software package (R Development Core 
Team 2011). 
Results 
At 40?C, 100% of the nymphs survived the first day, but by day 4, there was 100% 
mortality. At 35?C, the nymphs began to die in 3rd instar and 100% died in 4th instar. At 15?C, 
after 277 days, one nymph molt into 2nd instar. After 467 days, some nymphs were still in the 1st 
instar and no nymphs had entered into 3rd instar. 
30 
 
 At 20?C, 25?C, 30 ?C, the nymph grew normally from 1st instar into adults. Table1 
shows the description statistics on the development time of each instar in 20?C, 25?C, 30 ?C, and 
35?C. Simple linear model are developed using the data available as shown in Table 4 and Figs. 1 
- 5. The degree days required to complete instars 1-7 were: 457.5, 668, 1031 1317, 1515, and 
2071, respectively. The lower development temperature threshold was 7.02?C, 11.56?C, 11.83?C, 
12.66?C, 13.65?C, 16.48?C, and 16.12?C for instars 1-7, respectively. 
Discussions 
In our analysis, only simple linear models were used due to the limited date available. 
There are many more temperature ?development models available (Damos and Savopoulou-
Soultani 2012) and probably new models can be developed if more data are available (Bentz et 
al. 1991). In our situation, the simple linear model is the only choice. As stated in above, the 
linear model works well between 20?C ~30?C. 
The same Degree-days methods was adopted in a temperature-dependent development 
study on Lacanobia subjuncta Grote and Robinson (Lepidoptera: Noctuidae ) by Doerr et al. 
(2002). In Doerr?s study, larvae successfully completed development of all instars in constant-
temperature regimes of 10.0 to 30.0?C. No larval development occurred at 37.5?C. Lower 
developmental threshold for larvae is 9.9?C, 7.8?C, 6.6?C, 5.1?C, 6.9?C, and 7.3?C for instars 1-
6, respectively. The degree days required to complete the six instars were 52.1, 53.2, 62.9, 72.5, 
50.0, and 142.9 for instars 1-6, respectively. 
Compared to Doerr et al.?s results, the lower development threshold for B. dubia is high 
and the degree-days for B.dubia are larger.  Both species develop fastest at 30?C. In both studies, 
the insects can grow normally only in a narrow temperature rage. The lower limit for B. dubia is 
higher than L. subjuncta, probably because it originates from Central and South America, 
31 
 
beginning in Costa Rica. It is common from French Guyana and Brazil to Argentina. 
One of the short coming of this study is that we did not have enough different 
temperatures to model. Although six temperatures were selected, no validate data were recorded 
at 15?C, 35?C, and 40?C. The only valid temperatures were 20?C, 25?C, and 30?C. If more data 
at other temperature were available, more accurate linear or none linear models could be 
developed. 
In order to fully understand the temperature-development pattern in B. dubia, it might be 
useful to compare constant temperature development data with fluctuating temperature 
development data (Doerr et al. 2002). However, in most situations, B. dubia are reared by pet 
lovers and are not present in the wild. As a result, there is probably not enough incentive to 
observe in fluctuating temperature experiment. 
The information presented here forms the basis for developing a more complete 
understanding of B.dubia. These data will be useful in control the development of B. dubia 
through temperature. It might also be useful to compare life history of different groups of 
cockroaches. 
  
32 
 
References Cited 
Bentz, B., J. Logan, and G. Amman. 1991. Temperature-dependent development of the  
 mountain pine beetle (Coleoptera?: Scolytidae) and simulation of its phenology, Can. 
 Entomol. 123: 1083-1094. 
Briere, J.-F., P. Pracros, A.-Y. Le Roux, and J.-S. Pierre. 1999. A novel rate model of  
 temperature-dependent development for arthropods. Environ. Entomol. 28: 22-29. 
Damos, P., and M. Savopoulou-Soultani. 2012. Temperature-driven models for insect  
 development and vital thermal requirements. Psyche: A J. Entomol. 2012: 1?13. 
Dixon, A. F. G., A. Hon?k, P. Keil, M. A. A. Kotela, A. L. ?izling, and V. Jaro??k. 2009.  
 Relationship between the minimum and maximum temperature thresholds for development  
 in insects. Funct. Ecol. 23: 257?264. 
Doerr, M. D., J. F. Brunner, and V. P. Jones. 2002. Temperature-dependent development of  
 Lacanobia subjuncta (Lepidoptera: Noctuidae). Environ. Entomol. 31: 995?999. 
Faraway, J. J. 2004. Linear Models with R, 1st ed. Chapman and Hall/CRC, Boca Raton, 
Florida 
Gullan, P. J., and P. Cranston. 2005. The insects: An outline of entomology, 3 ed. Wiley- 
 Blackwell, Malden, Massachusetts. 
Lactin, D. J., N. J. Holliday, D. L. Johnson, and R. Craigen. 1995. Improved rate model of  
 temperature-dependent development by arthropods. Environ. Entomol. 24: 68-75. 
Logan, J. A., D. J. Wollkind, S. C. Hoyt, and L. K. Tanigoshi. 1976. An analytic model for  
 description of temperature dependent rate phenomena in arthropods. Environ.  
 Entomol. 5: 1133-1140. 
33 
 
R Development Core Team. 2011. R: A language and environment for statistical computing. 
Ramsey, F., and D. Schafer. 2012. The statistical sleuth: a course in methods of data  
 analysis. Cengage Learning, Independence, Kentucky. 
Schoolfield, R. M., P. J. H. Sharpe, and C. E. Magnuson. 1981. Non-linear regression of  
 biological temperature-dependent rate models based on absolute reaction-rate theory 
J. Theor. Biol. 88: 719?731. 
Trugill, D. L., A. Honek, D. Li, and N. M. Straalen. 2005. Thermal time - concepts and utility 
Ann. Appl. Biol. 146: 1?14. 
Wu, H., A. G. Appel, and X. P. Hu. 2013. Instar determination of Blaptica dubia (Blattodea:  
 Blaberidae) using gaussian mixture models. Ann. Entomol. Soc. Am. 106: 323?328. 
34 
 
Table 3-1. Development time of each instar in different temperature. (mean (SD) n ) 
Instar 20?C 25?C 30?C 35?C 
1 38.8 (3.96) 28 25.4 (2.84) 36 20.08 (3.23) 38 17.46 (5.19) 59 
2 79.9 (5.931) 28 51.18 (3.79) 34 36.06 (4.63) 36 NA 
3 116.15 (8.60) 26 78.32 (4.83) 31 53.26 (7.28) 34 NA 
4 161.74 (13.91) 27 106.76 (7.10) 33 70.28 (8.41) 36 NA 
5 220.48 (24.54) 27 133.45 (7.71) 33 87.56 (8.91) 36 NA 
6 NA* 167.80 (10.48) 32 105.70 (17.00) 36 NA 
7 NA 233.20 (17.55) 25 149.20(15.69) 35  NA 
NA* = Not Applicable 
 
35 
 
 
 
Figure 3-1. The linear regression of development rate and temperature (20?C, 25?C, 30?C, 
35?C) of 1st instar.  
  
36 
 
 
Figure 3-2. The linear regression of development rate and temperature (20?C, 25?C, 30?C) 
of 2nd  instar.  
 
  
37 
 
 
Figure 3-3. The linear regression of development rate and temperature (20?C, 25?C, 30?C) 
of 3rd  instar.  
  
38 
 
 
Figure 3-4. The linear regression of development rate and temperature (20?C, 25?C, 30?C) 
of 4th instar.  
  
39 
 
 
Figure 3-5. The linear regression of development rate and temperature (20?C, 25?C, 30?C) 
of 5th instar.  
 
40 
 
Table 3-2. Linear regression models of temperature-dependent development in the seven instars of B. dubia.  
NA* = Not Applicable 
 
Instar Model R2  F P ( > |t|) Lower Development 
Threshold (?C) 
Degree days 
1st  instar Y = 0.0021X ? 0.0148 0.984 120 < 0.01 7.02 457.5 
2nd  instar Y = 0.0149X ? 0.0172 0.997 301 0.0366 11.56 668 
3rd  instar Y = 0.001023X ? 0.01203 0.989 90.8 0.0666 11.83 1031 
4th  instar Y = 0.000805X ? 0.010190 0.986 69 0.0763 12.66 1317 
5th  instar Y = 0.000689X ? 0.009398 0.993 151 0.0157 13.65 1515 
6th instar Y = 0.0007X ? 0.0117 NA* NA NA 16.48 1429 
7th instar Y = 0.000483X ? 0.007783 NA NA NA 16.12 2071 
41 
 
Chapter 4 The parental care of Dubia cockroaches (Blattodea: Blaberidae) 
ABSTACT:  The social behavior of cockroaches is less studied compared to that of other insect 
groups. Most of the information on cockroaches is from brief notes made during field 
observation. For parental care, most of observations are focused on species of the Cryptocercidae 
and no study has been conducted on the parental behavior of the Blaberidae. The objective of 
this study was to measure the possible effects of parental care on offspring of the Dubia 
cockroach, Blaptica dubia Serville. No parental care effects on growth or body mass were 
detected although parental care behavior was observed. 
KEYWORDS: Blaptica dubia, Parental care 
Introduction 
Parental care involves any sort of behavior that increases the survival of the offspring. 
Parental care has been described in most animal phyla, but is especially well developed in 
numerous species of insects, crustaceans, and vertebrates. There is great diversity in the amount 
of parental care provided to offspring by different species. The most common parental care is to 
assist the offspring in living through harsh physical conditions, making resources available to the 
offspring which are otherwise difficulty or impossible to obtain, and decreasing the risk of 
predation on eggs or young (Clutton-Brock and Scott 1991). Although there are various levels of 
complexity, parental care can be divided into three major behavior categories: (a) those that 
physically protect the young from danger, (b) those that protect resources vital to offspring, and 
(c) those that facilitate offspring feeding (Glazier 2002, Roldan and Soler 2011). 
Parental care in insects ranges from covering eggs with a protective coating to remaining 
42 
 
to feed and protect young, to forming eusocial societies. A variety of forms are widespread in 
various insect order; parental care is most developed in Hemiptera, Coleoptera, Thysanoptera, 
Embiidina, Hymenoptera and Isoptera . 
However, there is great diversity in the range of parental care in cockroaches. It is 
difficult to find other groups of animals that are as diversely social as cockroaches. Most 
cockroaches show some form of parental care in the broad sense, because even the most 
primitive cockroaches have oothecae or egg cases that protect the young before they hatch 
(Mckittrick 1964). If only the post-hatch parental care is considered, most of the information 
about cockroaches is from brief field notes (Bell et al. 2007). Most of these are focused on 
species of Cryptocercidae (Seelinger and Sellinger 1983, Nalepa 1987, 1988, 2010, Park and 
Choe 2003). 
Few laboratory observations of parental care behaviors have been focused on species of 
the Blaberidae. Until recently, a new form of maternal provisioning of newly hatched nymphs 
was described in the ovoviviparous hissing cockroach Gromphadorhina portentosa Schaum. 
Shortly after expelling the hatching egg case, the female exudes from her abdominal tip a 
whitish, translucent material on which neonates actively feed (Perry and Nalepa 2003). 
This study is focused on the parental behavior in Blaptica dubia Serville (commonly 
known as the Dubia cockroach, Guyana spotted cockroach, or Orange-spotted cockroach). B. 
dubia is a large (up to ca. 4.5 cm in length), sexually dimorphic blaberid cockroach that is often 
kept as a pet. It is ovoviviparous and exhibits post-hatch parental care behavior.  More 
importantly, this species is used as a food source for many reptiles and amphibians. Compared to 
other food sources such as crickets, B. dubia contain a higher percentage of protein, are easier to 
maintain, and unlike other cockroaches, produce little odor. As a result, they have become an 
43 
 
increasingly popular food among amphibian and reptile enthusiasts. The objective of the present 
study is to investigate the effects of parental care on nymphs of B.dubia. If nymphs show 
differences in development time, body weight, head length, or head capsule width between 
groups with or without parents, it is an indication that parental care has an effect on the 
development of B.dubia.  
Material and Methods  
Insect Rearing:  Blaptica dubia were reared in clear gallon glass jars (3.8 L, 15.24 cm 
width by 25.4 cm height) with cardboard harborage at 30 ? 2?C. They were exposed to a 
photoperiod of 12:12 (L:D) h and supplied with water and dry dog Chow (Purina? Dog Chow?, 
Ralston Purina, St. Louis, MO) ad libitum. 
Experiment Design: The development of nymphs with the presence or absence of 
parents was compared. In the control, there were 10 neonates, and in one test group, there is one 
extra adult female and in the other test group, there is one extra adult male. All the 30 neonates 
were from the same mother and the experiment was set up on the second day after the neonates 
were born. There were 6 replications in a randomized complete block design.  We monitored the 
development of the cockroaches daily. When the nymphs reached the 7th instar, the following 
characters were measured: the width of the head capsule, the pronotal length and width, the body 
weight, and development time. The development time from birth to adult were recorded. The 
survival number of cockroaches in each jar were counted. 
Measurements:  Head capsule width (the widest distance between the two compound 
eyes) was measured to 0.01 mm using a microscope (Leica MZ 6, Solms, Germany) with an 
ocular micrometer. The length and width of the pronotum were measured to 0.01mm using an 
electronic digital caliper (Model: 62379-531, Control Company, Friendswood, TX). Body weight 
44 
 
was measured to the nearest 0.1 mg using an electronic scale (Serial No 0074082, Denver 
Instrument Company). 
Data Analysis. The Shapiro-Wilk nomality test (Shapiro and Wilk 1965) was used to test 
the normality of the data collected. The mean and standard deviation of each variable was 
calculated. For each group, ANOVA were conduct after the normality tests (R Development 
Core Team 2011, Ramsey and Schafer 2012).  
 Results 
Effect of parental care on measurable characters 
In Tables 1~4, the mean of head capsule width, pronotal length, pronotal width, and body 
weight of 7th instars reared with an without adults are compared. The head width shows no 
difference between neonate only group, neonates with adult female and neonates, with adult 
male ( P = 0.89, 0.49, 0.87; n = 3 ). The pronotal length shows no difference between neonate 
only group, neonates with adult female, and neonates with adult male (P =0.89, 0.40, 0.62; n = 
3). The pronotal width shows no difference between neonate only group, neonates with adult 
female, and neonates with adult male (P = 0.77, 0.33, 0.66; n = 3). Body weight shows no 
difference between neonate only group, neonates with adult female, and neonates with adult 
male (P =  0.86, 0.076, 0.57; n = 3). In summary, the results of ANOVA indicate that there is no 
difference between the three groups. 
Effect of parental care on development time 
In Table 5, the mean development times of the three treatments are compared. The 
development time shows no differences between neonate only group, neonates with adult female, 
and neonates with adult male ( P =  0.810, 0.083, 0.330, 0.590, 0.830, 0.280; n = 5). The results 
of ANOVA indicate that there is no difference in development time between the three groups. 
45 
 
Effect of parental care on survival rate 
Survival rate is not affected by the presence of adult females or adult males in the 
nymphal stadium. The mean number of nymphs that survived from 1st to 7th instar in neonate 
only group, neonates with adult female and neonates with adult male group are 8.667( n = 6, SD 
=1.0328 ), 8.333( n = 6, SD =1.2111 ), 8.167 ( n = 6, SD = 0.9832), out of 10 neonates, 
respectively. The result Kruskal-Wallis rank sum test for the three group indicate there is no 
difference in the three groups (P =0.067, df = 2, Kruskal-Wallis ? 2 = 0.8009). 
Discussion 
In Park and Choe?s (2003) manipulation experiments with Cryptocercus kyebangensis 
Grandcolas, offspring separated from their parents could survive independently, but grow more 
rapidly when they remained with their parents. In particular, the effects of parental care on 
offspring growth were found to be stronger in groups with both parents than those with single 
parents. 
However, our results indicate that there is no difference in the head size, pronotal length 
and width, body weight, and developmental time of groups of nymphs with and without an adult. 
That is rather counter-intuitive. Why was a ?parental effect? not observed in our study? One 
possible explanation may be that: the rearing environment was so good (30?C is an excellent 
rearing temperature for B. dubia see Chapter 3), that any effects of parental behavior were 
hidden. If the temperature was very low, the B. dubia nymphs huddle around the mother 
decreasing heat loss. Since we only measured parental effects on the last instar (7th instar) it is 
possible that there were differences in the early stages that were hidden by later development. 
To get more accurate results, there are some possible ways to improve our experimental 
design: (a) change the temperate to one that is less favorable, such as 20?C or 25?C. (b) Redesign 
46 
 
the experiment and put two groups. One is neonate only and another is neonates and its parents 
(i.e., two adults). (c) Further observation of parental behavior of B. dubia is also necessary.  
Our result indicates that there is no difference in the head size, pronotal length and width, 
body weight, and developmental time. According to these results, nymphs alone are sufficient to 
initiate a new colony without the presence of adults. 
  
47 
 
References Cited 
Bell, W. J., L. M. Roth, and C. A. Nalepa. 2007. Cockroaches: ecology, behavior, and natural  
 history. JHU Press. Baltimore, Maryland. 
Clutton-Brock, T. H., and D. Scott. 1991. The evolution of parental care. Monographs in  
 behavior and ecology. Princeton University Press, Princeton, New Jersey. 
Glazier. 2002. Parental Care. Encyclopedia of Evolution. Oxford University Press. 
Mckittrick, F. A. 1964. Evolutionary studies of cockroaches. Cornell University Agricultural  
 Experiment Station Memoir, Ithica, New York. 
Nalepa, C. A. 1987. Life history studies of the woodroach Cryptocercus punctulatus scudder  
 and their implications for the evolution of termite eusociality. PhD thesis. North Carolina 
State University  
Nalepa, C. A. 1988. Cost of parental care in the woodroach Cryptocercus punctulatus Scudder 
 (Dictyoptera: Cryptocercidae). Behav. Ecol. and Sociobiol. 23: 135?140. 
Nalepa, C. A. 2010. Altricial development in subsocial cockroach ancestors: foundation for the  
 evolution of phenotypic plasticity in termites. Evol. & Devel. 12: 95?105. 
Park, Y., and J. Choe. 2003. Effects of parental care on offspring growth in the Korean wood- 
 feeding cockroach, Cryptocercus kyebangensis. J. Ethol. 21: 71?77. 
Perry, J., and C. A. Nalepa. 2003. A new mode of parental care in cockroaches. Insectes  
 Sociaux. 50: 245?247. 
R Development Core Team. 2011. R: A language and environment for statistical computing. 
Ramsey, F., and D. Schafer. 2012. The statistical sleuth: a course in methods of data  
 analysis. Cengage Learning, Independence, Kentucky. 
Roldan, M., and M. Soler. 2011. Parental-care parasitism: how do unrelated offspring attain 
48 
 
 acceptance by foster parents? Behav. Ecol. 22: 679?691. 
Seelinger, G., and U. Sellinger. 1983. on the social organisation, alarm and fighting in the  
 primitive cockroach Cryptocercus punctulatus scudder. Zeitschrift f?r Tierpsychologie. 61:  
 315?333. 
Shapiro, S. S., and M. B. Wilk. 1965. An analysis of variance test for normality (Complete  
 Samples). Biometrika. 52: 591?611. 
  
49 
 
 
Table 4-1. Effect of parental care on head width (mm) of 7th instar. (mean ? SD)  
Replicate Neonate only With adult female With adult male P value 
1 5.079?0.1451 5.062?0.1737 5.033?0.2315 0.89 
2 5.409?0.1384 5.367?0.1370 5.444?0.0967 0.49 
3 5.233?0.1175 5.215?0.1428 5.244?0.0760 0.87 
 
  
50 
 
Table 4-2. Effect of parental care on pronotal length (mm) of 7th instar. (mean ? SD)  
Replicate 
number 
Neonate only With adult female With adult male P value 
1 9.693?0.2693 9.732?0.3798 9.631?0.4897 0.89 
2 10.10?0.3466 10.25?0.2518 10.07?0.2362 0.40 
3 9.933?0.3678 9.787?0.3298 9.876?0.2452 0.62 
 
  
51 
 
Table 4-3. Effect of parental care on pronotal width (mm) of 7th instar. (mean ? SD)  
Replicate Neonate only With adult female With adult male P value 
1 15.22?0.2531 15.17?0.6773 15.37?0.6217 0.77 
2 16.25?0.3955 16.05?0.3808 15.87?0.7114 0.33 
3 15.49?0.5815 15.70?0.3813 15.63?0.4670 0.66 
 
  
52 
 
 
Table 4-4.  Effect of parental care on body weight (g) of 7th instar. (mean ? SD) 
Replication Neonate only With adult female With adult male P value 
1 2.070?0.2062 2.122?0.2615 2.046?0.2952 0.860 
2 2.644?0.1862 2.743?0.1750 2.487?0.2739 0.076 
3 2.057?0.2271 2.165?0.2291 2.163?0.2734 0.570 
  
53 
 
 
Table 4-5. Effect of parental care on development time (day) from birth to adult. (mean ?  
SD) 
Replication Neonate only With adult 
female 
With adult male P value 
1 142.9 ? 9.662 139.7 ? 11.715 141.4 ? 9.985 0.810 
2 165.3 ? 6.873 147.0 ? 13.476 158.1 ? 17.201 0.083 
3 202.5 ? 8.536 210.4 ? 34.122 222.5 ? 28.224 0.330 
4 167.6 ? 12.293 161.5 ? 19.206 160.3 ? 11.056 0.590 
5 130.4 ? 9.812 130.8 ? 9.203 132.9 ? 7.120 0.830 
6 175.7 ? 25.050 171.8 ? 15.106 158.8 ? 25.246 0.280