The Vitruvian spider: Segmenting and integrating over different body parts to describe ecophenotypic variation

Abstract

Understanding what drives the existing phenotypic variability has been a major topic of interest for biologists for generations. However, the study of the phenotype may not be straightforward. Indeed, organisms may be interpreted as composite objects, comprising different ecophenotypic traits, which are neither necessarily independent from each other nor do they respond to the same evolutionary pressures. For this reason, a deep biological understanding of the focal organism is essential for any morphological analysis. The spider genus Dysdera provides a particularly well-suited system for setting up protocols for morphological analyses that encompass a suit of morphological structures in any nonmodel system. This genus has undergone a remarkable diversification in the Canary Islands, where different species perform different ecological roles, exhibiting different levels of trophic specialization or troglomorphic adaptations, which translate into a remarkable interspecific morphological variability. Here, we seek to develop a broad guide, of which morphological characters must be considered, to study the effect of different ecological pressures in spiders and propose a general workflow that will be useful whenever researchers set out to investigate variation in the body plans of different organisms, with data sets comprising a set of morphological traits. We use geometric morphometric methods to quantify variation in different body structures, all of them with diverse phenotypic modifications in their chelicera, prosoma, and legs. We explore the effect of analyzing different combined landmark (LM) configurations of these characters and the degree of morphological integration that they exhibit. Our results suggest that different LM configurations of each of these body parts exhibit a higher degree of integration compared to LM configurations from different structures and that the analysis of each of these body parts captures different aspects of morphological variation, potentially related to different ecological factors.

Keywords: Araneae; Canary Islands; Dysderidae; geometric morphometrics; integration.

Figure 1

Analyzed landmark configurations of the Dysdera spider species and digitized landmarks. Blue dots represent fixed landmarks and green dots represent semilandmarks. (a) Dysdera silvatica in a typical silk retreat under the rock; (b) different body parts analyzed in the Dysdera species of this study (highlighted in gray); (c) ventral view of the chelicera (Q1); (d) lateral view of the chelicera (Q2); (e) ventral view of the fang (Q3); (f) dorsal view of the prosoma (C1); (g) lateral view of the prosoma (C2); (h) lateral view of the leg 1 femur (L11); (i) lateral view of the leg 1 tibia (L14); (j) lateral view of the leg 4 femur (L41); (k) lateral view of the leg 4 tibia (L44).

Figure 2

Different ecological groups selected for the study. (a) Generalist species; (b) Onsicophagous species; (c) cave‐dwelling species with troglomorphic adaptations; (d) species adapted to intertidal environments.

Figure 3

Pairwise integration values (a) and overall levels of integration (b) of landmark (LM) configurations related to the same morphological structure (white) versus LM configurations of different morphological structures (gray).

Figure 4

Phenotypic space of species means for the different subset combinations. (a) Subset of all landmark (LM) configurations combined. (b) Subset of all chelicera LM configurations (Q1, Q2, and Q3). (c) Subset of all prosoma LM configurations (C1 and C2). (d) Subset of all leg LM configurations (L11, L14, L41, and L44).

Figure 5

Phenotypic space of all specimens analyzed for the different subset combinations. (a) Subset of all landmark (LM) configurations combined. (b) Subset of all chelicera LM configurations (Q1, Q2, and Q3). (c) Subset of all prosoma LM configurations (C1 and C2). (d) Subset of all leg LM configurations (L11, L14, L41, and L44).

Figure 6

Deformation grids depicting shape differences between the minimum and maximum extremes of principal component (PC) axes for all the different LM configurations combined (top), and the subsets of the chelicera, carapace and legs (bottom) in PC1 (a) and PC2 (b).