QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

First published online 24 November 2004
doi: 10.1242/dev.01547

This Article Summary Full Text Full Text (PDF) Supplementary Material Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Quigley, I. K. Articles by Parichy, D. M. Search for Related Content PubMed PubMed Citation Articles by Quigley, I. K. Articles by Parichy, D. M.

Evolutionary diversification of pigment pattern in Danio fishes: differential fms dependence and stripe loss in D. albolineatus

Ian K. Quigley^*, Joan L. Manuel^*, Reid A. Roberts^*, Richard J. Nuckels, Emily R. Herrington, Erin L. MacDonald and David M. Parichy

Section of Integrative Biology, Section of Molecular, Cell and Developmental Biology, Institute for Cellular and Molecular Biology, University of Texas at Austin, 1 University Station C0930, Austin, TX 78712, USA

View larger version (71K): [in a new window]   Fig. 1. Adult stripe development and loss in danios. (A) Adult D. rerio exhibit dark stripes of melanophores and iridophores (with few xanthophores), alternating with light stripes of xanthophores and iridophores (with few melanophores). (B) Adult D. albolineatus lack distinctive melanophore stripes and exhibit only a weak interstripe region posteriorly. (C) fms mutant D. rerio exhibit disrupted stripes in which metamorphic melanophores are reduced and xanthophores are absent. (D) Detail of stripes and interstripes in D. rerio. Dark cells are melanophores and yellow-orange cells are xanthophores (arrow). Reflective iridophores are found throughout, but are organized differently in interstripe regions (arrowheads); under this illumination, iridophores outside of the interstripe region are evident only by their blue iridescence over some melanophores. (E) Detail of pigment pattern in D. albolineatus adult. Melanophores are widely distributed over the flank. Adults exhibit reddish erythrophores (arrow) distributed widely over the flank and interstripe iridophores (arrowheads) form only a narrow, irregular band over part of the flank. Erythrophores are present in several other danios as well, although not in D. rerio. Adult fish are 25-30 mm in length. Scale bars in D: 200 µm for D,E.

View larger version (47K): [in a new window]   Fig. 2. Comparative analyses reveal fms-dependence of hybrid pigment pattern formation across Danio species. (A) Phylogenetic relationships. The traditional grouping of danios comprises at least two major clades, `Danio' and `Devario' (Kullander, 2001; Fang, 2003). Heterospecific danios with names in color were used for interspecific complementation tests with fmsj4e1 mutant D. rerio. Colors of names indicate fms-dependence of hybrid (or heterozygous) pigment pattern: red, strong fms-dependence; green, mild fms-dependence; blue, no apparent fms-dependence (see text). (B,E,H,K,N,Q) Danio wild-type pigment patterns. (C,F,I,L,O,R) Control hybrids between wild-type D. rerio and heterospecific danios. (D,G,J,M,P,S) Hybrids between fmsj4e1 mutant D. rerio and heterospecific danios. (B-D) Danio albolineatus lack distinctive stripes whereas control hybrids develop stripes with irregular borders; fms mutant hybrids lack stripes. (E-G) Danio aff. albolineatus resembles D. albolineatus but has a reduced interstripe region posteriorly and the anal fin lacks a melanophore stripe; these fish may represent a divergent population of D. albolineatus or a closely related species (Fang and Kottelat, 2000). Control hybrids develop stripes with irregular borders like D. albolineatus control hybrids; fms mutant hybrids lack distinctive stripes. (H-J) Danio `hikari' resembles D. kerri (Parichy and Johnson, 2001): melanophore stripes are broad and diffuse and include very few xanthophores; interstripe regions are narrow and irregular by comparison with D. rerio (Fig. 1A). Control hybrids develop distinctive melanophore stripes and interstripes with regular borders. Tester fms mutants have fewer xanthophores than controls. (K-M) Danio aff. kyathit resembles D. rerio initially but develops fissures in melanophore stripes as the fish grows; D. aff. kyathit also exhibit red erythrophores, particularly in the fins. Control hybrids lack stripe fissures, and fms mutant hybrids are not discernibly different (minor differences between control and tester hybrids are within the range of variation exhibited by different families within genotypes). (N-P) Danio choprae transiently develop horizontal stripes, then lose these stripes as a uniform pattern of melanophores emerges; vertical bars of melanophores arise in adults. Control hybrids develop and maintain horizontal stripes resembling those of D. rerio whereas fms mutant hybrids have fewer melanophores and xanthophores and less organized patterns. (Q-S) Danio aff. dangila adults have melanophore stripes interrupted by lighter spots and interstripes, and are morphologically indistinguishable from D. dangila (Parichy and Johnson, 2001). Control hybrids resemble D. rerio, and fms mutant hybrids exhibit no clear difference from controls. Data not shown: Danio nigrofasciatus exhibit well-defined stripes and control hybrids are intermediate between D. rerio and D. nigrofasciatus parental species (Parichy and Johnson, 2001; Quigley et al., 2004); tester hybrids do not differ from controls for either fmsj4blue (Parichy and Johnson, 2001) or fmsj4e1. Tester fmsj4e1 hybrids could not be obtained for D. kerri; tester fmsj4blue hybrids resemble control hybrids, similar to tester fmsj4blue hybrids with D. albolineatus, likely reflecting modifier loci in the fmsj4blue background (see text). Hybrids with species outside of the Danio clade (A) were not viable, consistent with previous observations (Parichy and Johnson, 2001). All fish are between 25-35 mm standard length, except D. dangila and its hybrids, which are 50-80 mm. GenBank accession numbers for 12S and 16S sequences used in phylogeny reconstruction were: AY707450, AY707456; U21372, U21381; AF322658, AF322663; AY707446, AY707452; AF322663; AY707449, AY707455; AY707447, AY707453; AF322656, AF322661; U21376, U21384; AF322659, AF322664; U21377, U21377; U21375, U21370; AY707448, AY707454; U21553, U21554; AF322660, AF322665; U21378, U21386; AY707445, AY707451; AY37481, AY37482; AY37483, AY37484.

View larger version (93K): [in a new window]   Fig. 3. Mild and strong fms-dependence of hybrid pigment patterns for D. `hikari' and D. choprae. (A,B) Danio `hikari' hybrids. (A) Wild-type control hybrids develop distinctive melanophore stripes and interstripe regions (arrow, A), whereas fms mutant hybrids have about half as many xanthophores as controls (B). (C-F) Danio choprae hybrids at larval stages. (C,E) Wild-type control hybrids have numerous xanthophores both on the body (arrow, C) and in the fins (arrowheads, E). (D,F) fms mutant hybrids have fewer xanthophores than controls. (G,H) Adult D. choprae control hybrids (G) have more xanthophores (arrows) than tester fms mutant hybrids (H). Scale bars: in A, 500 µm for A,B,G,H); in C, 150 µm for C-F.

View larger version (72K): [in a new window]   Fig. 4. Segregation analysis supports an essential role for fms in D. albolineatus xD. rerio hybrid stripe development. (A,B,D,E) Tester fmsj4e1/+ hybrids exhibit either of two easily distinguishable phenotypes, one in which stripes are well organized (A,B), and another in which stripes are weakly organized with irregular borders and fewer xanthophores (D,E). Differences are evident during pigment pattern metamorphosis (A,D) but are most apparent in adults (B,E). Details showing well-organized (C) and weakly organized (F) stripes in hybrid adults. Pigment cell complements in hybrids with weak stripes were significantly reduced compared with hybrids with strong stripes, with 84% of the number of melanophores (F1,16=8.76, P<0.01) and only 14% of the number of xanthophores (F1,16=20.85, P<0.0005). The reduction in melanophore numbers is comparable to that observed in homozygous fmsj4e1 mutant D. rerio and fms174 mutant D. rerio at restrictive temperature; however, neither D. rerio mutant allele retains xanthophores (Parichy et al., 2000b; Parichy and Turner, 2003a). (G,H). Primer extension genotyping for fms alleles. (G) In D. rerio, an extension primer adjacent to the fmsj4e1 lesion yields additions of two nucleotides in homozygous wild-type individuals, additions of four nucleotides in homozygous fmsj4e1 individuals, and additions of both two and four nucleotides in heterozygous fms+/fmsj4e1 individuals. (H) Primer extension genotyping for tester fmsj4e1/+ hybrids that have been classified as having `strong' or `weak' stripes. All hybrids with strong stripes carry the D. rerio fms+(wik) wild-type allele and the D. albolineatus fms allele, resulting in two nucleotide additions to the extension primers. All hybrids with weak stripes carry the D. rerio fmsj4e1 mutant allele and a D. albolineatus fms allele, resulting in the addition of four and two nucleotides to the extension primer, respectively. Scale bars: in A, 200 µm for A,D; in B, 500 µm for B,E; in C, 200 µm for C,F.

View larger version (117K): [in a new window]   Fig. 5. A temperature-sensitive allele uniquely implicates fms in melanophore stripe disorganization and xanthophore reduction in hybrids. Tester fms174 mutant D. rerio xD. albolineatus hybrids were reared at 28.5°C through middle metamorphosis, then were transferred either to 24°C or 33°C until adult pigment patterns had formed. (A,B) fms174 hybrids reared at 24°C develop well-organized stripes and numerous xanthophores on the body (A) and fins (B, arrows). (C,D) fms174 hybrids reared at 33°C develop poorly organized melanophore stripes and have fewer xanthophores on the body (C) and fins (D, arrows). Pigment patterns of control wild-type hybrids do not differ between 24°C and 33°C (data not shown). Scale bars: in A, 1 mm for A,C; in B, 150 µm for B,D.

View larger version (117K): [in a new window]   Fig. 6. Development of adult pigment patterns in D. rerio (A-F) and D. albolineatus (G-L). Panels shown are of selected days from a complete image series for individual, representative larvae. (A) In D. rerio, pigment pattern metamorphosis is marked by the differentiation of metamorphic melanophores (arrow, showing one of many) over the myotomes. The white box in each image delineates the horizontal myoseptum; the single arrowhead in A indicates one of several early larval melanophores that have persisted along the horizontal myoseptum through the beginning of metamorphosis. Double arrowhead indicates deep melanophores along the dorsal aspect of the neural tube; triple arrowhead, deep melanophores lining the dorsal surface of the peritoneum. sb, swim bladder. (B) As metamorphosis proceeds, additional metamorphic melanophores (small arrows) develop over the myotomes, and iridophores and xanthophores differentiate ventral to the horizontal myoseptum in the prospective primary interstripe region (large arrow). (C) By middle stages of pigment pattern metamorphosis, melanophores have started to organize into stripes in the region of the prospective dorsal primary melanophore stripe (p1D) and prospective ventral primary melanophore stripe (p1V). (D) As the dorsal and ventral melanophore stripes become increasingly distinctive, additional late metamorphic melanophores differentiate already within these stripes (e.g. arrow). (E) Near the completion of pigment pattern metamorphosis, distinctive dorsal and ventral primary melanophore stripes (1D, 1V) border a well-defined interstripe region. (F) Pigment pattern metamorphosis is completed with the development of scales and scale-associated melanophores (s). In D. albolineatus, melanophores typically do not persist along the horizontal myoseptum from earlier stages, and instead melanophores, initially deeper between the myotomes, migrate to the surface (arrowhead, data not shown). (H) Metamorphic melanophores (e.g. small arrow) differentiate widely scattered over the myotomes, as in D. rerio, but fewer iridophores (large arrow) develop ventral to the horizontal myoseptum. (I,J) As metamorphosis proceeds, additional metamorphic melanophores appear over the myotomes, yet these cells typically do not migrate far from their site of differentiation. Relatively few late metamorphic melanophores appear within the regions where stripes form in D. rerio (compare with D). (K) Near the completion of metamorphosis, melanophores remain relatively dispersed compared with D. rerio, and only a weak pattern of melanophore stripes (arrowheads) borders an irregular and narrow interstripe region on the posterior trunk. (L) At the end of metamorphosis, D. albolineatus have far fewer sub-dermal melanophores than D. rerio, and these cells are widely distributed where the interstripe region develops in D. rerio. The interstripe region extends only to the middle of the flank (arrow) and reddish erythrophores have started to differentiate within this region. Images shown have been rescaled across stages to maintain the same approximate field of view. Larval standard lengths in mm, (A-F) 7.1, 7.7, 8.3, 9.5, 12.0, 13.3, (G-L) 8.1, 9.3, 10.1, 11.0, 12.8, 13.4. Scale bars: in A, 500 µm for A,G; in F, 1 mm for F,L.

View larger version (29K): [in a new window]   Fig. 7. Quantitative analyses of melanophore morphogenesis in D. rerio and D. albolineatus. (A) In regions where melanophore stripes develop in D. rerio, melanophore numbers are dramatically reduced in D. albolineatus and this deficit becomes more pronounced through pigment pattern metamorphosis. (B) In the middle of the flank, where the first interstripe region forms in D. rerio, melanophore numbers are increased in D. albolineatus, reflecting the absence of a distinctive and persistent melanophore-free region in the anterior of the flank. Despite the increased melanophore numbers, the overall complement of melanophores is dramatically reduced overall compared with D. rerio. (C) Analyses of twice daily image series reveal differential appearance ('births') and disappearance ('deaths') of melanophores between species. Shown are cumulative mean births and deaths recorded in each of three developing larvae through middle stages of pigment pattern metamorphosis, with `births' defined as the appearance of new melanophores (either by differentiation or proliferation) and `deaths' defined by the unambiguous loss of melanophores (see Materials and methods). Melanophore births were not significantly different between species (F1,107=1.87, P=0.2) after controlling for variation among individuals (F4,107=2.72, P<0.05) and across days (F21,107=5.43, P<0.001). By contrast, melanophore deaths were significantly greater in D. albolineatus than D. rerio (F1,81=3.04, P<0.005) after controlling for variation among individuals (F4,81=2.10, P=0.09) and across days (F29,81=2.47, P<0.001). Error bars are omitted for clarity. (D) Total melanophore movements are reduced in D. albolineatus compared with D. rerio, both during early and middle metamorphosis (left), and through later metamorphosis (right). Shown are mean (±s.e.m.) distances moved by individual melanophores, with distances expressed as percentages of the flank height. Left, species differences were significant (F1,696=14.23, P<0.0005; n=368, 344 melanophores in D. rerio and D. albolineatus) after controlling for variation associated with individuals (nested within species: F4,696=2.60, P<0.05), and anteroposterior region nested within individuals (F10,696=5.19, P<0.0001). Right, species differences were significant (F1,708=8.45, P<0.005; n=291, 419 melanophores in D. rerio and D. albolineatus) whereas inter-individual differences were not significant (P=0.2). (E,F) Directional movements of melanophores were significantly reduced in D. albolineatus compared with D. rerio. Each point represents a single melanophore followed from its first appearance to its final position at the end of the series or until it was lost (n=368, 344 melanophores in D. rerio and D. albolineatus, respectively). Plots show the initial dorsoventral positions at which melanophores first appeared, and the subsequent changes in dorsoventral positions by the end of the images series. The dorsal-most position on the flank is assigned a relative value of 0, and the ventral-most position on the flank is assigned a value of 1. Regression slopes are estimated separately for dorsal and ventral regions of the flank because of differences in shape and growth. (E) In D. rerio, melanophores that initially appear in more dorsal regions of the flank tend to move ventrally whereas melanophores that initially appear in more ventral regions of the flank tend to move dorsally (partial regression coefficients±s.e.m. for relationship between starting dorsoventral position and arcsine-transformed movements for dorsal and ventral regions of the flank, respectively: -0.17±0.02, -0.14±0.04). (F) In D. albolineatus, directional movements were significantly reduced compared with D. rerio in dorsal regions of the flank (F1,399=6.43, P<0.05), although species differences were not detectable in ventral regions (F1,399=1.40, P=0.2), as assessed by the magnitude of starting position xspecies interactions [partial regression coefficients for dorsal and ventral, respectively: -0.10±0.02, -0.10±0.02; after controlling in both dorsal and ventral analyses for variation among individuals (nested within species, both P<0.005), variation among the three examined anteroposterior regions of the flank nested within individuals (both P<0.0005), main effects of species (P=0.7, P<0.005, respectively), and starting position independent of species (both P<0.0001)].

View larger version (118K): [in a new window]   Fig. 8. Altered melanophore lineage development in D. albolineatus. (A-H) Melanophore precursor abundance does not differ dramatically between species as revealed by distributions of cells expressing tyrosinase, dopachrome tautomerase (dct) and mitfa (Lister et al., 1999; Kelsh et al., 2000; Camp and Lardelli, 2001). tyrosinase and dct encode enzymes required for melanin synthesis, whereas mitfa encodes a transcription factor essential for melanophore specification. (A-D) In situ hybridization for tyrosinase mRNA in D. rerio (A,B) and D. albolineatus (C,D) during middle stages of pigment pattern metamorphosis (equivalent to larvae shown in Fig. 6C,I). (A) In D. rerio, numerous tyrosinase+ melanophores and unmelanized melanophore precursors (arrowhead) are observed in the region of the prospective dorsal primary melanophore stripe. (B) Higher magnification image of a different larvae showing tyrosinase+ melanophores (arrow) and melanophore precursors (arrowhead). (C) In D. albolineatus, fewer melanophores are present but tyrosinase+ melanophore precursors (arrowhead) are not obviously reduced in number. (D) Higher magnification of a different larva showing tyrosinase+ melanophores (arrowhead) and unmelanized melanophore precursor (arrowhead). (E-H) Expression of other melanophore lineage markers also is similar between D. rerio and D. albolineatus. (E,F) Danio rerio exhibit unmelanized cells (arrowheads), and some melanized cells (arrow) expressing dct (E) and mitfa (F). (G,H) In D. albolineatus, the numbers of unmelanized cells expressing dct (G) and mitfa are similar to D. rerio. Results for kit and sox10-expressing cells were similar (data not shown) (Parichy et al., 1999; Dutton et al., 2001). (I-L) Tyrosinase-expressing melanophore precursors revealed by treating fixed larvae with the essential precursor for melanin synthesis, L-dopa (Camp and Lardelli, 2001; McCauley et al., 2004). Images in I and K are prior to L-dopa incubation, and images in J and L are the same fields of view after treatment with L-dopa for 5 hours. (I,J) In D. rerio, only a few tyrosinase+ melanophore precursors are revealed by L-dopa incubation (arrowheads). (K,L) In D. albolineatus, however, numerous tyrosinase+ melanophore precursors are revealed by L-dopa treatment (arrowheads show only a few of these cells). These cells exhibit morphologies typical of melanoblasts and recently differentiated melanophores (inset). The increased number of L-dopa stained, tyrosinase+ cells as compared with molecular markers in D. albolineatus may reflect perduring protein in the absence of transcriptional activity. (M-T). Melanophores and melanophore precursors frequently are lost in D. albolineatus. (M-O) The same region of a D. albolineatus larvae imaged at 12-hour intervals reveals the transient appearance of several melanophores (arrowheads). Yellow-orange cells are xanthophores; these and other melanophores do not change positions between images. (P) High magnification image of D. albolineatus reveals melanin-containing debris (arrow) typical of melanophore death. Arrowhead, melanophore precursor that acquired melanin following L-dopa incubation of this larva. (Q-S) Cross-sections reveal the locations of melanophores and tyrosinase+ melanophore precursors in D. rerio (Q) and D. albolineatus (R,S). (Q) In D. rerio, few melanophores or melanophore precursors are located within the plane of the epidermis; one such melanophore is indicated by the arrowhead. e, epidermis; m, myotome. Arrow indicates iridophores of the developing interstripe region. (R) In D. albolineatus, numerous melanophores and tyrosinase+ melanophore precursors occur within the plane of the epidermis (arrowheads), although some melanophores are found subdermally, as in D. rerio (arrow). (S) Higher magnification of bracketed region in R, showing a melanin-containing extrusion body typical of teleost melanophore death. Arrowhead indicates bounding membrane. n, neuromast. Melanin-containing debris in P also is superficially located, as revealed by hexagonal outlines of adjacent epidermal cells. (T) High magnification image of whole-mount larva, showing extrusion body containing melanin granule (arrow) and staining for dct mRNA (arrowhead). Scale bars: in A, 80 µm for A,C; in B, 40 µm for B,D; in E, 60 µm for E-H; in I, 80 µm for I-L; in P, 20 µm for P; in Q, 60 µm for Q,R; in S, 20 µm for S,T.

View larger version (150K): [in a new window]   Fig. 9. Excess xanthophores in D. albolineatus compared with D. rerio. (A-H) Xanthophores in D. rerio (A,B,F), D. albolineatus (C,D,G), and wild-type D. rerio xD. albolineatus hybrids (E,H). (A) In D. rerio, few xanthophores (arrowhead) are visible during early stages of metamorphosis (stage equivalent to Fig. 6A) and most initially are found ventral to the horizontal myoseptum (box). Arrow indicates iridophores in the prospective primary interstripe region. (B) During middle stages of pigment pattern metamorphosis in D. rerio (e.g. Fig. 6C), xanthophores occur in the developing interstripe region (arrow) and a few faint xanthophores can be seen along the dorsal myotomes. (C) In D. albolineatus, xanthophores (arrowheads) are widely dispersed over the flank during early pigment pattern metamorphosis (e.g. Fig. 6G). (D) During later pigment pattern metamorphosis in D. albolineatus (e.g. Fig. 6I), xanthophores (arrowheads) persist widely scattered over the flank. Reddish erythrophores have started to develop in the interstripe region (arrow) and persist into the adult. Lineage relationships of erythrophores to xanthophores are unclear. (E) In hybrids between wild-type D. rerio and D. albolineatus, excess xanthophores (arrowheads) develop over the flank compared with D. rerio, and reddish erythrophores develop in the interstripe region (arrow). (F-H) Higher magnification images of larval D. rerio, D. albolineatus, and D. rerio xD. albolineatus hybrids shown in B-E. (I-L) In situ hybridization for early markers of the xanthophore lineage (Parichy et al., 2000b; Ziegler et al., 2000). Shown is GTP cyclohydrolase I (gch), which encodes an enzyme required for synthesizing pteridine pigments of xanthophores. Results for xanthine dehydrogenase (xdh), encoding a second pteridine synthesis enzyme were similar (data not shown). (I,J) D. rerio larvae; (K,L) D. albolineatus larvae. (I,K) In both species, gch+ cells occur over the myotomes during early metamorphosis. (J,L) At an optical plane medially within the same larvae shown in I and K, relatively few gch+ cells are observed in D. rerio (J), whereas many gch+ cells are present in D. albolineatus (L). (M,N) In situ hybridization for fms mRNA does not reveal clear differences in the numbers or distributions of fms+ cells between D. rerio (M) and D. albolineatus (N). Scale bars: in A, 100 µm for A,C; in B, 160 µm in B,D,E; in F, 60 µm for F-H; in I, 60 µm for I-L; in M, 80 µm for M,N.

View larger version (125K): [in a new window]   Fig. 10. Differential sensitivities across danios to melanophore reduction during hybrid pigment pattern development. An abbreviated phylogeny is shown on the left and wild-type danio pigment patterns are shown in A,C,E,G,I,K,M. Representative duchamp mutant hybrids are shown in B,D,F,H,J,L,N; schematics illustrate melanophore distributions (dorsal scale-associated melanophores are omitted for clarity). (A,B) Heterozygous duchamp mutant D. rerio (B) exhibit spots and fewer melanophores than wild type (A). Xanthophore and iridophore deficits are not apparent. (C,D) duchamp hybrids for D. kyathit develop spots of melanophores, although these are somewhat less regular than in D. rerio, as is the wild-type D. kyathit stripe pattern. (E,F) duchamp hybrids for D. nigrofasciatus develop well-organized spots or even complete stripes. (G,H) duchamp hybrids for D. albolineatus exhibit a more severe melanophore reduction than observed in heterozygous duchamp mutant D. rerio or other hybrids, and these melanophores remain widely dispersed over the caudal flank. As in duchamp mutant D. rerio, gross deficits in xanthophore numbers were not apparent during pigment pattern metamorphosis (data not shown). (I,J) duchamp hybrids for D. `hikari' develop intermediate patterns, in which more melanophores are present than in D. albolineatus hybrids, but melanophore patterns range from weak clustering, to reticulation, to more uniform dispersion. (K-M) duchamp hybrids for the more distantly related D. choprae and D. dangila develop spots similar to the D. rerio species group. Scale bars: in A, 1 mm for A-L; in M, 0.5 mm for M,N.

View larger version (124K): [in a new window]   Fig. 11. Melanophore and iridophore patterns in tester duchamp mutant hybrids. (A,B) Different illumination of the same fields of view reveals melanophore organization (A,C) and iridophore organization (B,D). (A,B) Tester duchamp hybrids for D. dangila exhibit clusters of melanophores (arrows) with iridophores (arrowheads) organized around these clusters, as for hybrids of the D. rerio species group. (C,D) Tester duchamp hybrids for D. albolineatus typically do not form clusters of melanophores (arrows), and iridophores (arrowheads) are poorly organized. Scale bars: in A, 200 µm for A,B; in C, 200 µm for C,D.