Data from: Adaptation for crypsis versus conspicuous social signalling following transitions across an extreme ecotone
Data files
Mar 10, 2025 version files 59.34 MB
-
README.md
22.83 KB
-
YaoOrd_JEB-2024-00274R1_background_dat.csv
203.20 KB
-
YaoOrd_JEB-2024-00274R1_BackgroundImages.zip
21.29 MB
-
YaoOrd_JEB-2024-00274R1_blennyULTRAtree.txt
5.15 KB
-
YaoOrd_JEB-2024-00274R1_body_dat.csv
130.39 KB
-
YaoOrd_JEB-2024-00274R1_BodyImages.zip
27.98 MB
-
YaoOrd_JEB-2024-00274R1_dorsalfin_dat.csv
159.22 KB
-
YaoOrd_JEB-2024-00274R1_FinImages.zip
9.55 MB
Mar 10, 2025 version files 59.34 MB
-
README.md
23.92 KB
-
YaoOrd_JEB-2024-00274R1_background_dat.csv
203.20 KB
-
YaoOrd_JEB-2024-00274R1_BackgroundImages.zip
21.29 MB
-
YaoOrd_JEB-2024-00274R1_blennyULTRAtree.txt
5.15 KB
-
YaoOrd_JEB-2024-00274R1_body_dat.csv
130.39 KB
-
YaoOrd_JEB-2024-00274R1_BodyImages.zip
27.98 MB
-
YaoOrd_JEB-2024-00274R1_dorsalfin_dat.csv
159.22 KB
-
YaoOrd_JEB-2024-00274R1_FinImages.zip
9.55 MB
Abstract
A key selection pressure in most habitats is predation, and a common strategy adopted by prey is crypsis through background matching. Many marine blenny fishes are in the process of a dramatic transition across one of the world’s most extreme ecotones: the invasion of land across the intertidal zone. We investigated the impact of this transition on body crypsis versus the conspicuousness of visual signals across 56 blenny taxa relative to 59 biologically relevant backgrounds, and as viewed by conspecifics and four representative fish and avian predators. We computed 33 colour and 23 pattern indices from standardised digital photographs of six individuals for each taxa (median sample). Six of these indices were selected for detailed analysis following phylogenetic Principal Component Analysis. While phylogenetic regressions revealed some aspects of body crypsis appeared to have changed adaptively with the progressive transition to land (specifically a reduction in body colour saturation), colonisation was primarily facilitated by a generalist form of crypsis. That is, the colours and patterns of aquatic blennies were already well matched to the range of terrestrial backgrounds where amphibious and terrestrial species were observed out of water. Predation appears to have been an important selection pressure constraining the colour and pattern of the dorsal fins used in social communication, which also matched visual backgrounds. Our data implies anti-predator strategies that translate well across habitats and different predator regimes will facilitate colonisation by either reducing predation risk or allowing species to persist long enough to respond adaptively to environmental change.
https://doi.org/10.5061/dryad.c59zw3rkh
Description of the data and file structure
Data files are .csv and can be opened using any text, spreadsheet or statistical program.
Missing values are indicated by NA.
All data files have the following columns unless specified otherwise.
Data are averages computed across all available images for a taxon.
location
- Geographic location where taxa were sampled; abbreviations on sites are given in Dryad Repository readme files associated with Ord & Cooke (2016; https://doi.org/10.5061/dryad.67g0t) and Morgans & Ord (2014; https://doi.org/10.5061/dryad.kp815).
Background_ID
- only in ‘background_dat’, this column specifies the image used to extract six representative regions for colour analysis; the label is simply a unique identify name based on an abbreviation of the location name in the previous column.
Species_Location
- species name an geographic location where population was sampled (not relevant for ‘background_dat’ file)
visual_model
- the type of visual model that images were processed through, including a model for the vision of blue tits, peacocks, coral trout, triggerfish and gobies
chroma_colour_number_average
- index of colour; here, the mean total number of colours
chroma_RNLx_average
- index of colour; here, the mean RNLx, defined as Receptor Noise Limited (RNL) measure of the red:green component channel for individual clusters.
chroma_RGratio_average
- index of colour; here, the mean R/G ratio. This is the value of the Red colour channel (a unit of scale from 0 to 255) and Green colour channel (a unit of scale from 0 to 255)
chroma_lw_mean_average
- index of colour; here, the mean long wave (lw) reflectance on a scale of 0 to 1
chroma_mw_mean_average
- index of colour; here, the mean medium wave (mw) reflectance on a scale of 0 to 1
chroma_sw_mean_average
- index of colour; here, the mean short wave (sw) reflectance on a scale of 0 to 1
chroma_RNLY_average
- index of colour; here, the mean RNLy , defined as Receptor Noise Limited (RNL) measure of the yellow:blue component channel for individual clusters.
chroma_RNLSaturation_average
- index of colour; here, the mean RNL saturation, designed as the distance (in receptor noise limited space, Vorobyev & Osorio 1998) of each colour from grey (the achromatic point).
chroma_DmaxChromaticity_average
- index of colour; here, the mean Dmax chromaticity. Dmax is defined as the most stimulated pairwise opponent channel (a colour channel which describes chromaticity, i.e. both hue and saturation) of each colour relative to grey (the achromatic point).
chroma_Sc_average
- index of colour; here, the mean Sc defined as Simpson colour diversity. This describes how evenly different colours in an image are represented, in terms of area coverage. The higher the number of different colours (number of discrete colour clusters), the higher the potential maximum value of Sc (up to the number of different colours).
chroma_Jc_average
- index of colour; here, the mean Jc defined as relative Simpson colour diversity. As above, but controlling for the number of different colours, so describes how evenly the colours are represented (in terms of area) independent of the number of colours.
chroma_Hc_average
- index of colour; here, the mean Hc defined as Shannon colour diversity. Also referred to as “entropy”, this is similar to Sc above, but uses a different diversity index.
chroma_Qc_average
- index of colour; here, the mean Qc defined as relative Shannon colour diversity. As above, but controlling for the number of different colours, so describes how evenly the colours are represented (in terms of area) independent of the number of colours, with a maximum value of 1.
chroma_MSsat_average
- index of colour; here, the mean MSsat defined as weighted mean of pattern RNL saturation contrast. The mean RNL saturation of the image (weighted by the area of each colour). RNL saturation is the distance (in receptor noise limited space, Vorobyev & Osorio 1998) of each colour from grey (the achromatic point). E.g. a fairly uniformly bright “red” image will have a high overall saturation, and a grey image will have a low value. Images with extremely varied colours (e.g. red, green and blue equally weighted) will also have a low mean, but high variance (see below).
chroma_sSsat_average
- index of colour; here, the mean sSsat defined as weighted standard deviation of pattern RNL saturation contrast. Standard deviation of the RNL saturation of the image (weighted by the area of each colour). Higher values imply more variance in RNL saturation values, and a more “colourful” image relative to grey. See also the VCA:MS value, which measures the internal differences between colours (i.e. colourful relative to each-other, instead of relative to grey).
chroma_CVSsat_average
- index of colour; here, the mean CVSsat defined as weighted CoV of pattern RNL saturation. Coefficient of variation of RNL saturation in the image (based on the above two values). Higher values imply more variance, while controlling for the mean value.
chroma_MS_average
- index of colour; here, the mean MS defined as weighted mean of pattern RNL chromaticity contrast. This is the average colour difference between all of the colours in the image, weighted by the area covered by each colour. Contrasts use the Vorobyev & Osorio 1998 Receptor Noise Limited method. Higher values imply larger internal colour differences between colours in the image. Therefore an image with lots of different colours will have a high value, irrespective of how far those colours are from the grey point. E.g. if the entire image is fairly uniformly bright “red” this value will be low, because all of the internal colours are similar. However, the saturation of the image (e.g. MSsat above) will be high.
chroma_sS_average
- index of colour; here, the mean sS defined as weighted standard deviation of pattern RNL chromaticity contrast. The standard deviation of RNL colour differences between all of the colours in the image, weighted by the area covered by each colour. Higher values imply larger variance in internal colour differences, e.g. lots of similar and dissimilar colours.
chroma_CVS_average
- index of colour; here, the mean CVS defined as weighted CoV of pattern RNL chromaticity contrast. Coefficient of variation of internal RNL colour differences in the image (based on the above two values). Higher values imply more variance, while controlling for the mean value.
achroma_Luminance_average
- index of brightness; here, the mean luminance on a scale of 0 to 1.
achroma_dbl_mean_average
- index of brightness; here, the mean dbl defined as luminance component channel created by the average of LW and MW channels for individual clusters.
achroma_ML_average
- index of brightness; here, the mean ML defined as weighted mean of pattern luminance contrast. The mean luminance of the image (weighted by the area of each colour).
achroma_sL_average
- index of brightness; here, the mean sL defined as weighted standard deviation of pattern luminance contrast. Standard deviation of the luminance of the image (weighted by the area of each colour). Higher values imply more variance in luminance values.
achroma_CVL_average
- index of brightness; here, the mean CVL defined as weighted CoV of pattern luminance contrast. Coefficient of variation of luminance in the image (based on the above two values). Higher values imply more variance, while controlling for the mean value.
achroma_MSL_average
- index of brightness; here, the mean MSL defined as weighted mean of RNL luminance contrast. This is the average RNL luminance difference between all of the colours in the image, weighted by the area covered by each colour. Contrasts use the Siddiqi et al. 2004 method. Higher values imply larger luminance differences between colours in the image.
achroma_sSL_average
- index of brightness; here, the mean sSL defined as weighted standard deviation of RNL luminance pattern contrast. The standard deviation of RNL luminance differences between all of the colours in the image, weighted by the area covered by each colour. Higher values imply larger variance in internal luminance differences, e.g. lots of similar and dissimilar luminance differences.
achroma_CVSL_average
- index of brightness; here, the mean CVSL defined as weighted CoV of RNL luminance pattern contrast. Coefficient of variation of internal RNL luminance differences in the image (based on the above two values). Higher values imply more variance, while controlling for the mean value.
achroma_BML_average
- index of brightness; here, the mean BML defined as weighted mean of luminance boundary strength. The average luminance difference (based on Michelson contrasts) of boundaries between colours across the image, weighted by the relative frequency of transitions between each potential colour combination. Higher values imply higher luminance differences across colour boundaries.
achroma_BsL_average
- index of brightness; here, the mean BsL defined as weighted standard deviation of luminance boundary strength. The standard deviation of luminance differences (based on Michelson contrasts) of boundaries between colours across the image, weighted by the relative frequency of transitions between each potential colour combination. Higher values imply higher variance in luminance differences across colour boundaries.
achroma_BCVL_average
- index of brightness; here, the mean BCVL defined as weighted CoV of luminance boundary strength. . Coefficient of variation of luminance differences of boundaries in the image (based on the above two values). Higher values imply more variance, while controlling for the mean value.
achroma_BMSL_average
- index of brightness; here, the mean BMSL defined as weighted mean of RNL luminance boundary strength. The average RNL luminance difference (using the Siddiqi et al. 2004 method) of boundaries between colours across the image, weighted by the relative frequency of transitions between each potential colour combination. Higher values imply higher luminance differences across colour boundaries.
achroma_BsSL_average
- index of brightness; here, the mean BsSL defined as weighted standard deviation of RNL luminance boundary strength. The standard deviation of RNL luminance differences of boundaries between colours across the image, weighted by the relative frequency of transitions between each potential colour combination. Higher values imply higher variance in luminance difference across colour boundaries.
achroma_BCVSL_average
- index of brightness; here, the mean BCVSL defined as weighted CoV of RNL luminance boundary strength. Coefficient of variation of RNL luminance differences of boundaries in the image (based on the above two values). Higher values imply more variance, while controlling for the mean value.
pattern_St_average
- index of pattern; here, the mean St defined as Simpson transition diversity. This describes the regularity of transitions between different colours in an image. St is equal to the number of different colours when all transitions are equally frequent.
pattern_Jt_average
- index of pattern; here, the mean Jt defined as relative Simpson transition diversity. As above, but controlling for the number of different colours, so describes how evenly the colours transition independent of the number of colours.
pattern_Ht_average
- index of pattern; here, the mean Ht defined as Shannon transition diversity. This describes the regularity of transitions between different colours in an image (using the Shannon rather than Simpson index). Ht is equal to the number of different colour clusters when all transitions are equally frequent.
pattern_Qt_average
- index of pattern; here, the mean Qt defined as relative Shannon transition diversity. As above, but controlling for the number of different colours, so describes how evenly the colours transition independent of the number of colours, with a maximum value of 1.
pattern_Scpl_average
- index of pattern; here, the mean Scpl defined as Simpson colour pattern complexity. This combines Jt and Jc (colour diversity and transition diversity respectively), such that colour patterns with uniform spacing and coverage (such as a chessboard) would have a maximum value of 1.
pattern_Qcpl_average
- index of pattern; here, the mean Qcpl defined as Shannon colour pattern complexity. As above, using Shannon instead of Simpson indices.
pattern_C_average
- index of pattern; here, the mean C defined as pattern complexity. This describes the geometric complexity of colour patterns based on the number of samecolour and different-colour transitions. Patterns with more complex structures will have higher values.
pattern_PT_average
- index of pattern; here, the mean PT defined as average patch size. The average patch size of each colour based on horizontal and vertical “run lengths” of each colour.
pattern_PTHrz_average
- index of pattern; here, the mean PTHrz defined as average horizontal patch size. The average “run length” (in pixels) of horizontal colour patches.
pattern_PTVrt_average
- index of pattern; here, the mean PTVrt defined as average vertical patch size. The average “run length” (in pixels) of vertical colour patches.
pattern_Asp_average
- index of pattern; here, the mean Asp defined as aspect ratio. 0 implies horizontally elongated patterns, 0.5 indicates equal circular or random patterns, 1 indicates vertically elongated patterns.
pattern_MDmax_average
- index of pattern; here, the mean MDmax defined as weighted mean of pattern Dmax contrast. The mean Dmax contrast of the image (weighted by the area of each colour). Dmax is the most stimulated pairwise opponent channel (a colour channel which describes chromaticity, i.e. both hue and saturation) of each colour relative to grey (the achromatic point).
pattern_sDmax_average
- index of pattern; here, the mean sDmax defined as weighted standard deviation of pattern Dmax contrast. Standard deviation of the Dmax contrast of the image (weighted by the area of each colour). Higher values imply more variance in Dmax chromaticity values, and a more “colourful” image.
pattern_CVDmax_average
- index of pattern; here, the mean CVDmax defined as weighted CoV of pattern Dmax contrast. Coefficient of variation of Dmax contrast in the image (based on the above two values). Higher values imply more variance, while controlling for the mean value.
pattern_BMDmax_average
- index of pattern; here, the mean BMDmax defined as weighted mean of Dmax boundary strength. The average Dmax chromaticity difference (based on Michelson contrast) of boundaries between colours across the image, weighted by the relative frequency of transitions between each potential colour combination. Higher values imply higher Dmax chromaticity differences across colour boundaries. Dmax is the most stimulated pairwise opponent channel (a colour channel which describes chromaticity, i.e. both hue and saturation) of each colour relative to grey (the achromatic point).
pattern_BsDmax_average
- index of pattern; here, the mean BsDmax defined as weighted standard deviation of Dmax boundary strength. The standard deviation of Dmax chromaticity differences of boundaries between colours across the image, weighted by the relative frequency of transitions between each potential colour combination. Higher values imply higher variance in Dmax differences across colour boundaries.
pattern_BCVDmax_average
- index of pattern; here, the mean BCVDmax defined as weighted CoV of Dmax boundary strength. Coefficient of variation of Dmax chromaticity differences of boundaries in the image (based on the above two values). Higher values imply more variance, while controlling for the mean value.
pattern_BMSSat_average
- index of pattern; here, the mean BMSSat defined as weighted mean of RNL saturation boundary strength. The average RNL saturation difference (i.e. the distance of each colour from the grey point, compared using Michelson contrast) of boundaries between colours across the image, weighted by the relative frequency of transitions between each potential colour combination. Higher values imply higher RNL saturation differences across colour boundaries.
pattern_BsSSat_average
- index of pattern; here, the mean BsSSat defined as weighted standard deviation of RNL saturation boundary strength. The standard deviation of RNL saturation differences of boundaries between colours across the image, weighted by the relative frequency of transitions between each potential colour combination. Higher values imply higher variance in RNL saturation across colour boundaries.
pattern_BCVSSat_average
- index of pattern; here, the mean BCVSSat defined as weighted CoV of RNL saturation boundary strength. Coefficient of variation of RNL saturation differences of boundaries in the image (based on the above two values). Higher values imply more variance, while controlling for the mean value.
pattern_BMS_average
- index of pattern; here, the mean BMS defined as weighted mean of RNL chromaticity boundary strength. The average RNL colour difference (based on Vorobyev & Osorio 1998) of boundaries between colours across the image, weighted by the relative frequency of transitions between each potential colour combination. Higher values imply higher colour differences across colour boundaries.
pattern_BsS_average
- index of pattern; here, the mean BsS defined as weighted standard deviation of RNL chromaticity boundary strength. The standard deviation of RNL colour differences of boundaries between colours across the image, weighted by the relative frequency of transitions between each potential colour combination. Higher values imply higher variance in colour difference across colour boundaries.
pattern_BCVS_average
- index of pattern; here, the mean BCVS defined as weighted CoV of RNL chromaticity boundary strength. Coefficient of variation of RNL colour differences of boundaries in the image (based on the above two values). Higher values imply more variance, while controlling for the mean value.
behavioral_rating_mean
- index of amphibiousness, ranging from exclusively aquatic (0) through to exclusively terrestrial (4) (not relevant for ‘background_dat’ file)
behav_cat
- a categorisation of ‘behavioral_rating’ into three groups of aquatic, amphibious and terrestrial (not relevant for ‘background_dat’ file)
Files and variables
File: YaoOrd_JEB-2024-00274R1_blennyULTRAtree.txt
Description:
The phylogeny used in ancestor state reconstructions, phylogenetic PCA and phylogenetic regression analyses.
The file is a .txt file that can be opened using any text, spreadsheet or statistical program.
The tree is given in the phylip format.
Taxon names correspond to those in the ’Species_Location’ columns in data files.
File: YaoOrd_JEB-2024-00274R1_background_dat.csv
Description:
Data used to quantify the colouration and pattern of visual backgrounds in sites frequented by amphibious and terrestrial blennies. These data were compared with data found in ‘body_dat’ and ‘dorsalfin_dat’.
Data files are .csv and can be opened using any text, spreadsheet or statistical program.
Missing values are indicated by NA.
All data files have the following columns unless specified otherwise.
Data are averages computed across all available images for a taxon.
File: YaoOrd_JEB-2024-00274R1_body_dat.csv
Description:
Data used to test the extent to which body colouration and pattern have shifted with the transition to land in the context of visually-oriented predators (birds and fish).
Data files are .csv and can be opened using any text, spreadsheet or statistical program.
Missing values are indicated by NA.
All data files have the following columns unless specified otherwise.
Data are averages computed across all available images for a taxon.
File: YaoOrd_JEB-2024-00274R1_dorsalfin_dat.csv
Description:
Data used to test the extent to which the colouration and pattern of the first dorsal fin have shifted with the transition to land in the context of visually-oriented predators (birds and fish) and a representative conspecific.
Data files are .csv and can be opened using any text, spreadsheet or statistical program.
Missing values are indicated by NA.
All data files have the following columns unless specified otherwise.
Data are averages computed across all available images for a taxon.
File: YaoOrd_JEB-2024-00274R1_BackgroundImages.zip
Description:
The cropped regions of photographs from which indices of colour, luminance and pattern for visual backgrounds were directly computed. We consider these images the raw source of all the work presented in this paper. PLEASE NOTE: full images are available upon request. Just email Terry Ord (t.ord@unsw.edu.au) outlining how you would like to use the image library and we'd be happy to provide access and support your project.
File is a compressed folder of images .zip .
Images are labelled based on location, site within location, and region cropped from background within that site. These correspond to the 'Location' and 'Background_ID' in 'YaoOrd_JEB-2024-00274R1_background_dat' data file.
File: YaoOrd_JEB-2024-00274R1_BodyImages.zip
Description:
The cropped regions of photographs from which indices of colour, luminance and pattern for the body regions of individuals for each blenny species were directly computed. We consider these images the raw source of all the work presented in this paper. PLEASE NOTE: full images are available upon request. Just email Terry Ord (t.ord@unsw.edu.au) outlining how you would like to use the image library and we'd be happy to provide access and support your project.
File is a compressed folder of images .zip .
Images are labelled based on species, collection location, sex (F, female; M, male), and site. These correspond to the 'Species_Location' column in 'YaoOrd_JEB-2024-00274R1_body_dat' data file.
File: YaoOrd_JEB-2024-00274R1_FinImages.zip
Description:
The cropped regions of photographs from which indices of colour, luminance and pattern for the first dorsal fin regions of individuals for each blenny species were directly computed. We consider these images the raw source of all the work presented in this paper. PLEASE NOTE: full images are available upon request. Just email Terry Ord (t.ord@unsw.edu.au) outlining how you would like to use the image library and we'd be happy to provide access and support your project.
File is a compressed folder of images .zip .
Images are labelled based on species, collection location, sex (F, female; M, male), and site. These correspond to the 'Species_Location' column in 'YaoOrd_JEB-2024-00274R1_dorsalfin_dat' data file.
Access information
Other publicly accessible locations of the data:
- The phylogeny was originally developed by Ord & Cooke (2016) Evolution 70:1747-1759
Data was derived from the following sources:
- All photographs used were collected or compiled by Terry Ord (t.ord@unsw.edu.au), with images processed by Shizhi (Estella) Yao (estella.yao1@unsw.edu.au)
See paper for full details on data collection and methodogy.
