Perkins BP, LSU Biological Sciences Undergraduate
May 9, 2008

The nuclear envelope is the double membrane that surrounds the nucleus and separates the genetic material and nucleoplasm from the cytosol in eukaryotic cells. It is composed of the inner and outer nuclear membranes (each consisting of a lipid bilayer), nuclear pore complexes, and the nuclear lamina. The nuclear envelope is also known to have some associations with chromatid. The space between the inner and outer nuclear membranes is called the perinuclear space (Mounkes et al., 2003). The outer membrane is covered in ribosomes and is continuous with the rough endoplasmic reticulum. The inner membrane is associated with the nuclear lamina (Wehnert and Bonne, 2002).

The major components of the nuclear lamina, the nuclear lamins, are Type V intermediate filaments (Mounkes et al., 2003). Lamins are evolutionarily conserved nuclear-specific intermediate filaments that support many nuclear functions that include maintaining nuclear shape, DNA replication, regulation of gene expression, transcription elongation by RNA Polymerase II, nuclear positioning, segregation of chromosomes, meiosis, and apoptosis (Wiesel et al., 2008). In metazoan cells there are four major lamin proteins encoded by three genes. The A-type lamins (lamins A and C) are different splice variants of the same gene, LMNA. The B-type lamins (lamins B1 and B2) are the products of two separate genes, LMNB1 and LMNB2 respectively (Meshorer and Gruenbaum, 2008).

All B-type lamins and lamin A (not lamin C) are synthesized with a conserved CaaX (Cysteine – aliphatic amino acid – aliphatic amino acid ‑ any amino acid) motif in their carboxyl tails that is subject to post-translational modifications (Liu and Zhou, 2008). The maturation of the prelamin begins with the farnesylation of the cysteine in the CaaX motif by a farnesyl transferase. Next, the last three residues on the carboxyl tail (aaX) are cleaved off by either Ras-converting enzyme (Rce1) or Zinc metalloproteases related to Ste24p (Zmpste24). As a result, the cysteine undergoes methylation by the isoprenylcysteine carboxyl methyltransferase. Prelamin A then undergoes a final processing step in which the 15 amino acids of the carboxyl tail, including the farnesyl group, are cleaved off by Zmpste24 (Fig. 1; Meshorer and Gruenbaum, 2008). The second cleavage of 15 amino acids at the carboxyl tail of prelamin A results in the removal of the farnesyl residue and subsequent dissociation of mature lamin A from the nuclear membrane (Mattioli et al., 2008). B‑type lamin, however, does not go through a second cleavage by Zmpste24 leaving B-type lamin constitutively farnesylated and associated with the nuclear membrane.

A-type and B-type lamins also differ based on the types of cells tin which they are expressed. All cell types express at least one of the B-type lamins, whereas some cells simply do not express, or express at significantly lower levels, the A-type lamins (e.g. cells of the early embryo, embryonic stem cells, stem cells in the immune and haematopoietic systems, and cells in the neuroendocrine system). The birth of mice with null mutations in their LMNA genes which were phenotypically indistinguishable from their wild type siblings during their first two weeks of postnatal development supported the hypothesis that A-type lamins are not essential to the overall development process (Mounkes et al., 2003). However, mutations in the human LMNA gene are the cause of at least 11 different heritable diseases that are commonly known as laminopathies (Wiesel et al., 2008).

The Laminopathies

One of the laminopathies that has drawn a great deal of attention from researchers is Hutchison‑Gilford progeria syndrome (HGPS). Hutchison-Gilford progeria syndrome is a disease in which the physical aspects of aging are accelerated (Fig. 2; Scaffidi et al., 2005). Most of the patients who suffer from Hutchison-Gilford progeria syndrome have a point mutation in the LMNA gene. This mutation results in the translation of a lamin A lacking 50 amino acids, including the second cleavage site in prelamin A. (Figure 1; Meshorer and Gruenbaum, 2008). The mutant protein (Called LAD50) is therefore constitutively farnesylated and incorporates abnormally into the nuclear lamina. This grouping of LAD50 proteins leads to mechanical defects, thickening of the lamina, loss of peripheral heterochromatin, and increased DNA damage, among other things. If farnesyl transferase inhibitors are administered, the cellular phenotypes can be treated. Also, mutations in the protein that performs the second cleavage (Zmpste24) lead to similar cellular phenotypes. This gives evidence for the hypothesis that the permanent farnesylation of the LAD50 protein is what leads to the diseases associated with this lamin A mutant. (Meshorer and Gruenbaum, 2008).

Unlike the pluripotent and rapidly growing embryonic stem cells, that do not express A-type lamins, somatic (adult) stem cells are tissue specific and highly dependent on their niche, from which they receive signals that influence the stem cells’ fates. The signaling mechanisms that are used by the different stem cell niches converge on four main pathways (Notch, Wnt, TGFβ, and Sonic hedgehog; Meshorer and Gruenbaum, 2008). The Notch signaling pathway is a major regulator of stem cell fate that has been implicated in premature aging. To study the molecular mechanisms that make the LAD50 mutant produce progeroid phenotypes, two separate stable cell lines, one with the LAD50 protein and the other with the wild type protein, were created and the gene expression of the two cell lines was compared. In the cells expressing LAD50, downstream signaling via the Notch pathway was induced by the Notch co‑activator Ski‑interacting protein which loses its grip due to the progerinated lamina produced by the LAD50 proteins. Because Notch is a major regulator of stem cell fate, and because they are the main tissue group that is affected by progeria, the mesenchymal stem cells were chosen to be studied in this experiment. In addition to activation of the Notch pathway, the mesenchymal cells responded to the mutant LAD50 protein by displaying sporadic, undirected differentiation along all three germ layers. This result shows the involvement of mesenchymal stem cell regulation in the pathology of Hutchison-Gilford progeria syndrome through the downstream activation of Notch signaling (Meshorer and Gruenbaum, 2008).

Emery-Dreifuss muscular dystrophy (EDMD) is another laminopathy which is usually inherited as either an X‑linked form of muscular dystrophy or an autosomal-dominant form. However, rare cases of autosomal recessive transmission have been reported (Wehnert and Bonne, 2002). Emery-Dreifuss muscular dystrophy is generally characterized by three symptoms which include early onset contractures (chronic loss of joint motion typically in elbows or achilles tendon), very slow progressive muscle weakness and degeneration involving the upper arms and lower legs, and cardiac disease in adult life (Wehnert and Bonne, 2002). It is caused by mutations in the STA gene which encodes emerin. Mutations in the STA gene almost always lead to the displacement of emerin from the nuclear envelope. (Mounkes et al., 2003). Emerin is anchored to the inner nuclear membrane by its carboxy-terminal tail, the remainder of the molecule resides in the nucleoplasm. It has several serine protein kinase sites and seems to have a role in the organization of the nuclear membrane during cell division (Fig. 3; Wehnert and Bonne, 2002). Emerin is known to interact with nuclear lamins and many of the lamin associated proteins including Barrier-to-autointegration factor (BAF) and Btf, a death promoting transcriptional repressor. The exact consequences of a loss of emerin from the nuclear envelope on the activities of these emerin-binding proteins remains to be established but it is clear that a lack of emerin due to a mutation in the STA gene is the cause of Emery‑Dreifuss muscular dystrophy (Mounkes et al., 2003).

Disease Causing Mechanisms

It has been sufficiently demonstrated in the above examples that mutations in nuclear lamin can lead to a diverse array of diseases. These laminopathies are a group of heritable diseases that are the product of mutations in genes that code for A‑type lamin and lamina-associated proteins. They include Emery-Dreifuss muscular dystrophy, Pelger-Huet anomaly, Limb girdle muscular dystrophy type 1B, dilated cardiomyopathy, Dunnigan’s familial partial lipodystrophy, mandibuloacral dysplaysia, Hutchison‑Gilford progeria syndrome, and more (Mounkes et al., 2003). The laminopathies show a wide variety of inheritance patterns and mechanisms of causing laminopathies. Some of the diseases resulting from mutations in laminal genes are tissue specific while others are harmful to multiple tissue types, all leading to the question: Why is it that mutations affecting the nuclear lamina give rise to such a diverse group of diseases? (Worman and Courvalin, 2004).

One hypothesis for the disease causing mechanism of these mutations involves defects in lamin synthesis caused by mutations in genes encoding the lamins. Although in metazoan cells B-type lamins are the products of two separate genes, in Caenorhabditis elegans B-type lamins are the product of a single gene. Most biological roles of mammalian lamins are evolutionarily conserved in C. elegans. In addition, the single C. elegans lamin probably functions both as A-type and B-type lamin. Recently, it has been shown that the B-type C. elegans lamin (Ce-lamin) is able to form stable intermediate filaments in vitro. Therefore, by inserting mutations into conserved residues in Ce-lamin that cause diseases when mutated in human lamin A, it can be investigated how mutations in specific residues of Ce-lamin corresponding to laminopathic disease causing mutations in the human LMNA gene affect filament and paracrystal assembly in vitro and lamin organization and dynamics in vivo. The assembly of the filaments and paracrystals can be seen using negative staining electron microscopy techniques. Fourteen laminopathic disease causing mutations were studied in this way; of the fourteen studied, only six of them showed evidence of disruption to the assembly of Ce-lamin filaments or paracrystals. This finding suggests that although defects in filament assembly may be one of the mechanisms that leads to laminopathies, it cannot be the only mechanism involved in causing these diseases (Wiesel et al., 2008).

The “mechanical stress” hypothesis of laminopathic disease diversity states that abnormalities in nuclear structure that are caused by mutations in lamin encoding genes can make the cell more susceptible to cellular damage that is caused by physical stress (Worman and Courvalin, 2004). Mutated cells under such conditions of stress show a significant redistribution of the proteins that are associated with the nuclear envelope and an increased fragility of the nucleus. In effect, cells that undergo large amounts of mechanical stress begin to function improperly. It would follow from the same line of reasoning, however, that the mechanical stress model is less likely to account for the defects in white adipose tissues that are associated with the diseases mandibuloacral dysplasia and Dunnigan’s familial partial lipodystrophy because of the lack of mechanical stress that is placed on these tissues on a consistant basis (Mounkes et al., 2003). Therefore, there must be yet another mechanism for the causal agent of these laminopathies.

The “gene expression” hypothesis on the diversity of laminopathic diseases suggests that LMNA mutations may affect the structure and function of the nucleus which can in turn have an effect on gene expression (Worman and Courvalin, 2004). A class of proteins known as the Nespirins localize to the nuclear envelope and interact with emirin and A-type lamins. These Nespirins appear to be integral to anchoring the nucleus to the interphase cytoskeleton. Such an interaction may be important to the correct transmission of mechanically induced signaling from the cell surface to the nucleus that could be potentially disrupted at many levels, including at the level of a nuclear envelope that is functioning improperly because of mutations in genes encoding laminal proteins (Mounkes et al., 2003).

Although there is not a common molecular mechanism for disease causing that unites all of the laminopathies, the models that we have thus far discovered and suggested are nonexclusive to each other, giving us room for models that can be used together to help find treatments for laminopathic diseases. Even without a single common mechanism of action among the laminopathies, many scientists are coming up with ideas on how to treat cells with LMNA mutations and people who suffer from crippling laminopathies. As was illustrated above, prelamin A processing is altered in some LMNA‑linked diseases, including Hutchison‑Gilford progeria syndrome, which leads to permanent farnesylation of the mutated lamin A protein. Permanently farnesylated lamin A is toxic to the cell and leads to cell dysfunction and diseased phenotypes (Meshorer and Gruenbaum, 2008). Accumulation of prelamin A in laminopathic cells causes severe heterochromatin defects, but chromatin organization and function can be recovered by treatment with mevinolin, an inhibitor of the hydromethyl–glutaryl-synthase which is indirectly impairing prelamin A farnesylation, in combination with the inhibitor of histone deacethylases trichostatin A (Mattioli et al., 2008). The cellular phenotypes of LMNA mutations can also be reversed through the use of farnesyl transferase inhibitors preventing the permanent farnesylation of the mutated lamin A. This again showcases the toxicity of the constitutively farnesylated LAD50 mutant (Meshorer and Gruenbaum, 2008).

In Conclusion

Whith all of the papers written about the nuclear envelope in general and the nuclear lamina and laminopathies it seems as though we have a grasp of much of the information dealing with mutations in laminal genes and laminopathies caused by them. But, the nuclear lamina is a complicated network that likely plays a role in DNA replication, chromatin organization, spatial arrangement of nuclear pore complexes, nuclear growth, mechanical stabilization of the nucleus, anchorage of the nuclear envelope proteins (Wehnert and Bonne, 2002), and many other things. There is such a plethora of protein-protein interactions, protein-nucleic acid interactions, protein‑enzyme interactions, activator and supressor relationships to genes that can possibly get mutated and produce proteins that interact inderictly with almost every other organelle in the cell that finding the exact purpose and reason for every protein interaction or the molecular mechanism of disease causing for all the laminopathies will take a long time to come. But, as long as people are suffering from degenerative laminopathic diseases, and as long as scientist continue to search for the answers to life’s questions and people’s problems, we will be making new discoveries about the cell, the nucleus, the nuclear envelope, the lamina, laminopathies for years to come.

Figures (Click Thumbnails for a Clearer Image):

Figure 1. Processing of lamin A in normal and HGPS cells. (left) The process of maturation of prelamin A. The first three steps are common to all CAAX proteins, including all B-type lamins. Inhibition of the second or third steps results in toxic lamin A accumulation, causing HGPS, restricted dermopathy (RD), or mandibuloacral dysplasia (MAD). The fourth step involves cleavage of 15 amino acids away from the terminal cysteine by Zmpste24. (right) The processing of prelamin A in the most common HGPS mutation, which deletes amino acids 607 – 656 (progerin / LAD50), including the second cleavage site of lamin A by Zmpste24 (Meshorer and Gruenbaum, 2008).

Figure 2. Hutchinson-Gilford Progeria Syndrome. HGPS is a childhood disorder caused by mutations in one of the major architectural proteins of the cell nucleus. (bottom, right) In HGPS patients the cell nucleus has dramatically aberrant morphology. (top, right) The uniform cell nucleus shape typically found in healthy individuals (Scaffidi et al., 2005).

Figure 3. Schematic view of the nuclear envelope. The nuclear envelope is composed of two lipid bilayer membranes, the nuclear pore complexes and the nuclear lamina. The outer nuclear membrane is continuous with the endoplasmic reticulum. The inner nuclear membrane is separated from the outer nuclear membrane by the perinuclear space, except at the nuclear pore complexes, where the outer and the inner nuclear membrane are connected. Underlying the inner membrane is the fibrous nuclear lamina. It is composed of two types of lamin proteins: A-type lamins (lamins A/C) and B-type lamins (B1 and B2). They interact with chromatin, BAF, and HP1 as well as with other proteins of the inner nuclear membrane as lamin B receptor LBR), lamina-associated proteins (LAPs), emerin, MAN1, and nurim. (Wehnert and Bonne, 2002).


Liu B, & Zhou Z (2008). Lamin A/C, laminopathies and premature ageing. Histology and Histopathology, 23, 747-763

Mattioli E, Columbaro M, Capanni C, Santi S, Maraldi NM, D’Apice MR, Novelli G, Riccio M, Squarzoni S, Foisner R, & Lattanzi G. (2008). Drugs affecting prelamin A processing: effects on heterochromatin organization. Experimental. Cell Research., 314, 453-462

Meshorer E, & Gruenbaum Y. (2008). Gone with the Wnt/Notch: stem cells in laminopathies, progeria, and aging. Journal of Cell Biology, 181, 9-13

Mounkes L, Kozlov S, Burke B, & Stewart CL. (2003). The laminopathies: nuclear structure meets disease. Current Opinion in Genetics & Development, 13, 223-230

Scaffidi P, Gordon L, & Misteli T. PLoS Biol 3(11):e395. (2005). PLoS BiolWehnert MS, & Bonne G. (2002). The nuclear muscular dystrophies. Seminars in Pediatric Neurology., 9 (2), 100-107

Wiesel N, Mattout A, Melcer S, Melamed-Book N, Herrmann H, Medalia O, Aebi U, & Gruenbaum Y. (2008). Laminopathic mutations interfere with the assembly, localization, and dynamics of nuclear lamins. Proceedings of the National Academy of Sciences U.S.A., 105, 180-185

Worman HJ, & Courvalin JC. (2004). How do mutations in lamins A and C cause disease? The Journal of Clinical Investigation, 113, 349-351

B. P. Perkins, LSU Biology Undergrad


The competitive exclusion principle can be paraphrased in four words: Complete competitors cannot coexist (Hardin, 1960). Although some find this to be an over-simplified maxim that may cause some ecologists to overlook more important underlying evidence, it does raise some interesting questions in the curious mind (Cole, 1960). The principle states that if two distinct populations use the same resources, live sympatrically, and if one population is even slightly better at translating energy to reproductive success than the other; then the fitter population will eventually drive the less fit population to extinction (Hardin, 1960). However, there is another option for the less fit population. Character displacement is an evolutionary process that involves a directional selection toward niche divergence. Sympatric species whose niches originally overlap will have selective pressures on them that cause reduced niche overlap and allow the species to coexist (Molles, 2007).

This experiment looks at the apparent coexistence of three orb-weaver spider species in Louisiana: Gasteracantha cancriformis, Nephila clavipes, and Leucauge sp. Based on the exclusion principle explained above, if the three species inhabit the same ecological niche, the most successful species of spider should drive the others to extinction. Because there is more than one species of spider living sympatrically, it must be assumed that they partition the resources in some way that reduces niche overlap and allows for coexistence. The only question that remains is that of how the spiders actually divide the resources amongst themselves.

Enders (1974) showed that coexisting spider species can partition their niches by building webs at different heights in the forest to catch different types of prey. To test if the orb-weaver spiders that are being studied use this method of niche partitioning, the average web height of each species will be compared. If the spiders being tested use this method of vertical stratification for reducing niche overlap we will expect to see a significant difference in the web heights measured. Craig (1989) lumps understory orb-weaver spiders in Panama into two alternative foraging modes. Large spiders build one large web per feeding period whereas small spiders build a few smaller webs. To test if the orb-weavers in Louisiana use these alternative methods of foraging we will compare the average web area of each species. A significant difference in web area among species would be evidence supporting this method of niche partitioning. Barghusen et al. (1997) showed that the web of the common house spider is more efficient at capturing flies when the strand density is increased. Although orb weaver spiders are being studied, it can still be determined if they use alternative strand densities as a method of reducing niche overlap. To test this we will compare the strand and radii density of the spider webs measured to look for a significant difference among the species. A significant difference supports the hypothesis that they use this method of niche partitioning.


Field work.—Random samples of orb-weaver spiders were taken at two locations in Baton Rouge, Louisiana (Ben Hur Experimental Forest and Bluebonnet Swamp) over a two year time. One sample was taken during the fall semester of 2006 and the other was taken during the fall semester of 2008. Measurements were taken of the spider size (as measured by length), the height of the center of the web off of the ground, the longest length across a diameter of the web, the length of the diameter perpendicular to the longest diameter, the number of strands per five centimeters of web (strand density), and the number of radii around the web. Data was collected on three species of orb-weaver spiders (Gasteracantha cancriformis, Nephila clavipes, and Leucauge sp.).

Data analyses.—First the data was compiled from the measurements taken in the fall of 2006 and the fall of 2008. The web area was then calculated by multiplying together the two diameter lengths. The radii density was found by dividing the number of radii per web by the area of the web. Bar graphs were made showing the average spider length (to be used as a reference point for comparing spiders), average web height, average web area, strand density, and radii density with 95% confidence intervals. Finally, ANOVA and T-tests were performed on each group of data (spider length, web height, web area, strand density, and radii density) to determine if there are in fact any significant differences among the three species of spider.

Results (Boring and Technical, feel free to skip to the Discussion)

Figure 1 is a representation of the average sizes of the spiders measured. There is a significant difference in size among the three orb-weaver species (F2,318 = 145.11, P = 1.66*10 45). Nephila clavipes is significantly larger than both Gasteracantha cancriformis (t293 = 15.51, tcrit = 1.97, P = 4.86*10-40) and Leucauge sp. (t151 = 7.14, tcrit = 1.98, P = 3.66*10-11). Leucauge sp. is significantly smaller than Gasteracantha cancriformis (t192 = -6.14, tcrit = 1.97, P = 4.73*10-9).

The average web heights for each species of spider measured are presented in Figure 2. There is a significant difference in the web height among the three species of orb-weaver spiders (F2,318 = 40.17, P = 2.79*10-16). Nephila clavipes have significantly higher webs off of the ground than both Gasteracantha cancriformis (t293 = 5.06, tcrit = 1.97, P = 7.33*10-7) and Leucauge sp. (t151 = 7.83, tcrit = 1.98, P = 7.81*10 13). Leucauge sp. have significantly lower webs to the ground than Gasteracantha cancriformis (t192 = -6.40, tcrit = 1.97, P = 1.16*10-9).

Figure 3 is a representation of the average web areas of the spiders. There is a significant difference in web area among the three species of orb-weaver spiders (F2,318 = 19.10, P = 1.47*10-8). Nephila clavipes have significantly larger web area than both Gasteracantha cancriformis (t293 = 5.38, tcrit = 1.97, P = 1.54*10-7) and Leucauge sp. (t151 = 3.44, tcrit = 1.98, P = 0.0007). There is not a significant difference between the web areas of Gasteracantha cancriformis and Leucauge sp. (t192 = 1.74, tcrit = 1.97, P = 0.08).

There is no significant difference among the strand densities (Figure 4; F2,318 = 2.49, Fcrit = 3.02, P = 0.08), but there is a significant difference in the radii densities among the three species of orb weaver spiders (Figure 5; F2,318 = 45.04, Fcrit = 3.02, P = 6.00*10-18). Gasteracantha cancriformis and Nephila clavipes do not have a significant difference in radii density (t293 = 0.27, tcrit = 1.97, P = 0.79). Leucauge sp. has a significantly larger radii density than both Gasteracantha cancriformis (t192 = 7.94, tcrit = 1.97, P = 1.66*10-13) and Nephila clavipes (t151 = 6.94, tcrit = 1.98, P = 1.11*10-10).


The analysis of web heights shows that each species does build webs at different heights. The largest species (Nephila clavipes) had the highest webs and as spider size decreases (Gasteracantha cancriformis then Leucauge sp.) so does web height (Figure 1 & 2). This evidence supports the hypothesis that orb-weaver spiders use the vertical stratification method of niche partitioning. The spiders likely build their webs at different heights to catch different types of bugs, affectively reducing niche overlap and promoting coexistence. However, the web height could also be a function of web area. Larger webs will naturally need to be higher off of the ground than smaller webs; therefore, we must also take into consideration the analysis of web area to determine the method of niche partitioning for these spiders.

The web area analysis shows that Nephila clavipes have significantly larger webs than both Gasteracantha cancriformis and Leucauge sp., but there is not a difference between the web areas of the latter two species (Figure 3). This supports Craig’s (1989) observation of two alternative foraging modes in orb weaver spiders. The larger Nephila clavipes builds a larger web whereas the other two species build smaller webs. This allows the spiders to divide the resources and not drive one another to extinction. Because we did not count the number of webs each spider built there is more research that must be done to determine if the spiders use these exact two methods, but this is a step in that direction.

Finally, there was no difference between strand densities of the spiders (Figure 4), and only Leucauge sp. had a different radii density (Figure 5). This does not support the hypothesis that orb-weaver spiders use strand density as a method of niche partitioning. If web thread density was a factor used to decrease niche overlap then we would expect to see each species with a different strand or radii density, but that is not the case. The spiders must use some other method of promoting niche divergence in order to coexist in the same environment.

These data suggest a combination of methods that spiders use to minimize niche overlap. While, to the untrained observer, it seems that the spiders all use the same resources and thus must drive one another to extinction, upon closer observation they are not actually complete competitors . A combination of web height and size is seen to be used by the orb-weavers to divide resources, but the strand and radii densities did not seem to play a role. This all suggests a complicated relationship between many different factors that allow multiple species to live in the same environment. Further research into alternative foraging modes in spiders from different places could shed light on the orb weavers. It would also be interesting to compare these data with data from spiders of the same species from different locations. The methods found to be used by the spiders studied in this experiment could have come about solely as a function of the local fauna. The same species compared from a different location could give a different order in the stratification of the webs, or a whole new method of niche partitioning altogether. Whether, like Cole (1960) said, the competitive exclusion principle is a trite maxim or not; it opens the door for some interesting research in the field of Ecology.


BARGHUSEN, L., CLAUSSEN, D., ANDERSON, M., & BAILER, A. (1997). The effects of temperature on the web-building behaviour of the common house spider, Achaearanea tepidariorum Functional Ecology, 11 (1), 4-10 DOI: 10.1046/j.1365-2435.1997.00040.x

Cole, L. (1960). Competitive Exclusion Science, 132 (3423), 348-349 DOI: 10.1126/science.132.3423.348

Craig CL. (1989). Alternative foraging modes of orb weaving spiders. Biotropica, 21 (3), 257-264

Enders, F. (1974). Vertical Stratification in Orb-Web Spiders (Araneidae, Araneae) and a Consideration of Other Methods of Coexistence Ecology, 55 (2) DOI: 10.2307/1935219

Hardin, G. (1960). The Competitive Exclusion Principle Science, 131 (3409), 1292-1297 DOI: 10.1126/science.131.3409.1292

Molles M. 2007. Ecology: Concepts and Applications. 4th ed. McGraw Hill. NY. 309-317.

(Pictures thanks to Wikipedia)