Monday, August 15, 2005

Final Paper!

Defense mechanisms of Nyssodesmus python (Polydesmidae)

Sheiphali Gandhi

Pott College of Science and Engineering, University of Southern Indiana


The purpose of this study was to examine the effects different characteristics of Nyssodesmus python had on the defense mechanisms of the organism. The defense mechanisms of N. python include: curling into a spiral and chemical defense. It was predicted that sex would not affect defense mechanisms, whereas size and level of calcification would. Populations were found in two locations: La Estación Biológica Monteverde and San Gerardo on different sides of the continental divide. One hundred and twenty millipedes were collected, 42 in Monteverde and 78 in San Gerardo. They were artificially threatened, and the time spent in a protective position and the release of toxin were recorded. The size of the individuals and level of calcification were also recorded. T-tests and chi-squared tests showed that there were significant differences in the defense mechanisms of males and females. Regression analysis showed a significant trend between size of an individual and the time it remained in the protected position, but the relationship may not be due to causation. The millipedes of different populations had some distinctive morphological and behavioral characteristics; additionally, there was a significant difference in their length of time spent in a curled position.


El objetivo de este estudio fue examinar los efectos de las diferentes característcas de Nyssodesmus python sobre los mecanismos de defensa de estos organismos. Los mecanismos de defensa de N. python incluyen: doblarse sobre sí mismos y la defensa química. Se predijo que el sexo no afectaría a los mecanismos de defensa, mientras el tamaño y el nivel de calcificación si lo harían. Los individuos se encontraron en dos lugares: La Estación Biológica Monteverde y San Gerardo, en lados diferentes de la división continental. Ciento veinte milpiés se estudiaron en total, 42 en Monteverde y 78 en San Gerardo. Se amenazaron artificialmente, y se registraron el tiempo que pasaron en una posición protectiva y la presencia de la secreción de toxina. También se anotaron el tamaño de los individuos y el nivel de calcificación. Los experimentos indicaron una relación significativa entre el sexo de un milpiés y sus mecanismos de defensa. El análisis de regresión mostró una relación significativa entre el tamaño del individuo y el tiempo que pasó en una posición protectiva. Los milpiés de poblaciones diferentes mostraron diferencias morfológicas y etológicas características; hubo una diferencia significativa en el tiempo en que permanecieron en una posición protectiva.


Nyssodesmus python, commonly known as the large forest-floor millipede, is usually found on the Caribbean slope of Costa Rica in patches of uncut understory. They feed primarily on rotting wood, like most other Polydesmidae, and are important decomposers in the rainforest ecosystem (Heisler 1983).

Adult N. python have twenty body segments, nineteen of which display a pair of horizontal keels that are large and flat, resembling a large isopod. Individuals are dull light yellow with two dark brown longitudinal stripes running down their backs. Sex can be easily determined because on the ventral side of the seventh body segment of male individuals, the first pair of legs is modified into a set of small curved “gonopods.” These structures are used in the transfer of sperm to the female millipede. Additionally, female N. python at 70-100 mm in length are larger than males, who measure on average 65-90 mm (Heisler 1983).

Nyssodesmus python molts throughout its entire life cycle even after it has stopped growing; after each molt, an adult N. python is soft and unpigmented. Approximately a month passes until the process of calcification and pigmentation is complete (Heisler 1983). The calcified exoskeleton plays an essential role in resisting the large pressures developed while burrowing. It is thought that calcium salts increase resistance without making the keel unwieldy. In general, the exoskeleton of males is stronger and more resistant than females. Also, as the body mass of females increases, the strength of their exoskeleton increases as well; while this trend does exist in male exoskeletons, it is much less apparent (Borrell 2004).

N. python has two forms of protection: an extremely rigid exoskeleton and a toxic chemical excretion. When threatened, it curls into a spiral, protecting its vulnerable underside, and secretes a toxin from its hindgut (Heisler 1983). The predominant components of their cyanogenetic defense secretions are hydrogen cyanide and benzaldehyde, together with other compounds such as phenol, benzoic acid, benzoyl cyanide, and mandelonitrile (Kuwahara et. al. 2002). Their characteristic secretion glands are known as oxadenes. They open laterally on the individual diplosegments to exude the repugnant liquid (Wright 1999). The composition of each toxin secretion is species-specific (Kuwahara et. al. 2002), and the secretion of N. python is commonly known for its pleasing cherry-almond scent. This liquid can be expelled violently up to a distance of 30 cm (Heisler 1983).

Research has been done on the reproductive behavior of this species by Hyatt (1993) and Arnold (1998), but little else is known about the natural history of N. python (Heisler 1983). The present study explored significant trends in defense mechanisms dependent on sex, size, and population. It was predicted that sex would not affect defense mechanisms, whereas size and level of calcification would. The hypothesis was based on the fact that little distinction exists between the morphology of males and females, whereas a clear disparity can be seen in size and calcification.


Study site

Nyssodesmus python were collected in two locations: the forest close to La Estación Biológica Monteverde and the forest in San Gerardo. The forest close to La Estación Biológica is on the Pacific slope in the lower montane wet forest life zone with an elevation of 1400-1800 m (Haber et al. 2000). San Gerardo sits at approximately 1300 meters in elevation on the Atlantic slope in the premontane rain forest life zone. Collection occurred close to trails due to accessibility (Fig. 1).

Data collection

The millipedes from the forest above La Estación Biológica Monteverde were collected in plastic bags and brought to the station by CIEE students. The millipedes collected in San Gerardo by Alan Masters were placed in a plastic container containing leaf litter and brought back to La Estación Biológica Monteverde to be tested. All subjects were collected between July 21, 2005 and August 1, 2005. They were kept at the biological station in terraria containing leaf litter collected on station property and covered with plastic wrap to maintain a moist environment. The millipedes from the two life zones were kept in separate terraria. The millipedes were left alone for one day to allow for acclimation. Multiple terraria were maintained so that simultaneous acclimation periods could be used for millipedes collected on different days. No more than ten millipedes were kept in a terrarium at the same time.

After the acclimation period, the millipedes were taken out of the terraria individually. The subject was then placed on a table and was artificially threatened by roughly turning it over onto its dorsal side. The amount of time the millipede remained curled in a spiral was recorded. The presence of toxin was determined by the presence of a sweet/cyanide smell and recorded. The length and width of the millipede were measured in millimeters by holding the millipede on it ventral side against a ruler on the table. These measures were multiplied to produce a size index for the comparison of different individuals. The new layer of exoskeleton, after molting, is unpigmented; therefore, the amount of calcification of the exoskeleton can be determined by its pigmentation. This was rated on a scale from one to ten with ten being the darkest and one the lightest. Calcification was recorded for only the Monteverde population because the pigmentation trend is only known for this population of millipedes (Heisler 1983).

Statistical analysis

Parametric tests were used to compare values. The log of the time of protection was taken to generate a normal data set. Unpaired t-tests were used to compare the protection time and size index of millipedes found in different locations. A chi-squared test was used to determine if the secretion of toxin differed between populations at the two sites. Unpaired t-tests were used to compare the curling time of males and females. Because there is a significant size difference between the two sexes (Heisler 1983) (Fig. 2), separate tests compared the size index and the protection time of each sex using simple regression analysis. Similarly, separate unpaired t-tests were run comparing the size index and presence of toxin of each sex. A chi-squared test was run to compare the presence of toxin for both sexes. Lastly, simple regression analysis was used to see if the protection time was a function of calcification.


A total of 120 Nyssodesmus python were collected. Close to La Estación Biológica Monteverde, 42 millipedes were collected, whereas 78 were collected in the San Gerardo forest. There were 46 females collected (16 in Monteverde and 30 in San Gerardo), while 74 males were collected (26 in Monteverde and 48 in San Gerardo). The ratio of females to males for each population was approximately 5 to 8.

Effect of sex

The protection time for the different sexes was found to be significantly different (p=0.0034). The mean protection time for females was 266.322±48.530 seconds, the mean protection time for males was 141.597±22.641 seconds. There was also a significant relationship between the sex of a millipede and the secretion of toxin (X2=10.781, DF=1, p=0.0010). Females were more likely to secrete toxin than males.

Effect of size

The millipede’s size index was found to have a significant effect on the protection time between sexes (males:p=0.0410, r2=0.057; females:p=0.0109, r2=0.138), but due to a low r2 value denoting a weak fit of the regression relationship, this may not be caused by millipede size. The trend showed that smaller millipedes tended to remain in the protected position for a longer span of time than a larger millipede of the same sex (Fig. 3 and Fig. 4). It was found that the size of a millipede did not have a significant relation to the secretion of toxin within each sex.

Results of different locations

The difference between the secretion of toxin in the two populations of millipedes was not significant (X2=0.001, DF=1, p=0.9693), as well as their differences in size (p=0.7893). There was a significant difference in the mean time of each group in the protected position (p=0.0018). The millipedes from Monteverde stayed in a spiral for 137.618±26.921 seconds, whereas the millipedes from San Gerardo averaged 217.295±33.311 seconds.

The two populations of millipedes showed divergent morphologies. The coloration of the group found in Monteverde consisted of yellow- and brown-striped keels (Fig. 5), whereas the keels of millipedes found in San Gerardo were black with yellow coloration on the outer edges (Fig. 6). A smell difference was also noted between the two groups. The millipedes on the Pacific slope secreted the characteristic cherry-almond smell, whereas the millipedes from the Atlantic slope secreted a substance that had a sweet smell that was neither as pleasing nor as strong as the secretion of the other group.

Other results

Calcification ratings were recorded for the millipedes found on the Pacific slope. Thirty-five of the 42 millipedes had complete calcification based on their pigmentation. It was observed that three millipedes had a rating of 1, one millipede had a rating of 2, one millipede had a rating of 4, and two had a rating of 8. There was no apparent relationship between the calcification of the exoskeleton of the millipedes found in Monteverde and its protection time.



Females remained, on average, in the protected position for a longer period of time and were more apt to secrete toxin, in comparison to males. The exoskeletons of male Nyssodesmus python are stronger than that of females (Borrell 2004). This may have been a contributing factor to this significant difference between the two sexes. Less protective measures would be necessary for the survival of the males since their exoskeletons provide them with better protection; therefore, the females with weaker calcification must compensate by using a longer time period in the spiral position or by relying on cyanogenetic secretions to deter predators.


This slight negative relationship may be explained by the results of a study that found that as the size of female millipedes increased, the strength of the cuticle increased as well (Borrell 2004). It would not be necessary for large female millipedes to take as many defensive precautions as smaller female millipedes. Exoskeleton strength also increases with size in males, but the relationship was not nearly as defined as it was with females (Borrell 2004). This was mirrored in the relationship between protection time and the size of male millipedes. Therefore, although there is a significant relationship between the protection time of males, the relationship may not be strong due to the fact that the strength of their exoskeleton does not significantly increase with size.


The two populations are comparable because they have very similar ratio between the numbers of males to females. In terms of patterns of toxin secretion and size, the two populations of millipedes differed little, but the distinctions between the protective times of the two groups were significantly different. It may have been due to the fact that the San Gerardo millipedes were moved to a much greater distance and to a new environment that caused them to remain in a protective position for a longer span of time. No prior research has been done on the San Gerardo population in particular, so no inferences can be made of this difference in times spent in the protective position.

The two populations have characteristics that may grant them to be considered as different species or subspecies. First of all, morphologically, the coloration of the species was very different. According to some definitions, this would be enough justification to separate the populations taxonomically. Additionally, the secretions of the two populations smelled very different. Within the family Polydesmidae, the compositions of the secretions are species-specific (Kuwahara et. al. 2002), further supporting that the two populations should be distinguished taxonomically.


Molting causes fluctuations in the exoskeleton strength of N. python (Heisler 1983), but these fluctuations may not impact the defense mechanisms of the organisms as much as the gradual increases of calcification due to size (Borrell 2004). So, although calcification due to molting cycles did not show a significant relation, it says little of the overall calcification of an individual. Additionally, only seven millipedes were found to have exoskeleton strength differences due to a recent molt. This small sample size may have hindered the soundness of the data found.

Future research

In order to improve the validity of this study for future replications, some adjustments must be made. Differences between adult and juvenile millipedes were not considered. Juveniles have fewer segments because segments are added with age (Heisler 1983). The exoskeleton of juveniles is generally weaker than that of the adults (Borrell 2004); therefore, this may have adversely affected the data. Results derived from this study indicated that smaller millipedes took more protective precautions; when in fact, this could be attributed to the age of the millipede. A more intensive study should be performed to obtain more robust results to test the difference in cuticle strength between sexes and its effect on defense mechanisms. The strength of the exoskeleton of each individual could be tested to determine its relationship to defense mechanism. This can be done by following methods used by B.J. Borrell, in which fracture force and density of keels were determined (2004). It is necessary to clarify the relationship between the populations of the opposite slopes. Morphologically, the population from Monteverde matches known natural history of Nyssodesmus python (Heisler 1983). Little is known of the San Gerardo population because there is no known existing literature describing its characteristics, or even stating if it is a subspecies of N. python. No tests were run on the possible reproductive interactions between the two groups to confirm relationship. Another population of Polydesmidae millipedes was found within the Monteverde collection site. In terms of coloration, they resemble those found in San Gerardo, but in general, they are smaller than those found in San Gerardo (Appendix 1).


I would like to acknowledge all the people who helped me with my data collection of the Monteverde population: Gina Rozinka, Mary Oppold, Stephanie Place, Kent Melchiors, Nathaniel Talbot, and Carlos Guindon. An appreciative hug goes to Alan Masters who provided the entire San Gerardo population. I am still astounded by his thoughtfulness and generosity. I would not have smelled like hydrogen cyanide for a week without all of you. Thank you to Javier Méndez and Carlos Guindon for all of your professorly advice and explanations of statistics. For helping me with my spur of the moment questions throughout the process, I appreciate all the help of my teaching assistants, Maria Jost and Nathaniel Talbot. Plus, they danced around the station, made me smile, and made sure that I didn’t measure my millipedes in millipedes. Additionally, the teaching staff was amazing in their efforts of trying to figure out the taxonomic information of the San Gerardo population of millipedes. The support of all my fellow CIEE students helped me get through the exhilarating process of the scientific method, and the group wallowing alleviated my problems more than I could have ever imagined. Kathy Rockwell was a huge support and help for taking me to the doctor twice when both of my feet decided to stop being functional. Thanks to telepathy, I was able to understand all my good times with my friend Meg Smith, without whom, I would be lost in the Costa Rica. All the people back home were a great support during hard times, especially my sister Sonali and my friends at USI. I will never be able to forget the concern of la familia Cruz Solís of my plight of finding a large sample size. It was wonderful to come home to such kind people. And, of course, thank you to the staff at La Estación Biológica Monteverde for the Costa Rican brewed coffee that always made my day and night.

Arnold, N. 1998. Mating behavior of millipedes, Nyssodesmus python. Educational Abroad Program, spring 1998. Monteverde, Costa Rica.

Borrell, B. J. 2004. Mechanical properties of calcified exoskeleton from the neotropical millipede, Nyssodesmus python. Journal of Insect Physiology. 50:1121-1126.

Haber, W. A., Zuchowski W. and Bello, E. 2000. An Introduction to Cloud Forest Trees Monteverde, Costa Rica, p. 11. Mountain Gem Publications, Monteverde de Puntarenas, Costa Rica.

Heisler, I. L. 1983. Nyssodesmus python (Milpies, Large Forest-floor Millipede). In D. H. Janzen (Ed). Costa Rican Natural History, pp. 747-749. University of Chicago Press, Chicago, Illinois.

Hyatt, P. 1993. The mating behavior of the tropical millipede Nyssodesmus python. CIEE, summer 1993. Monteverde, Costa Rica.

Kuwahara, Y., Omura, H., and Tanabe, T. 2002. 2-Nitroethenylbenzenes as natural products in millipede defense secretions. Naturwissenschaften. 89:308-310.

Wright, Johnathon C. 1999. Myriapoda (Including Centipedes and Millipedes). Encyclopedia of Life Sciences. John Wiley & Sons, Ltd: Chichester. (

No comments: