Fig. 5. Temperature shift experiments reveal temporal requirements for Fms activity in promoting xanthophore development and melanophore stripe formation in homozygous fms^174A mutants. (A,B) Mean densities (±95% confidence intervals) of xanthophores present in the adult pigment patterns of individuals following temperature shift at the sizes indicated. Orange line in A and B indicates the mean density of xanthophores in control individuals reared at 24°C throughout development; xanthophores were absent in control individuals reared at 33°C throughout development. n, sample sizes for each size class. Note that only midpoints of size classes are indicated, and that ranges of sizes per class vary (1 mm per class at sizes <8 mm, when rapid changes occur during pigment pattern metamorphosis; 2 mm per class at sizes 8 mm, reflecting slower changes at late metamorphic and juvenile stages; DMP and JMT, manuscript in preparation). (A) Temperature up-shift ablates xanthophores through middle stages of pigment pattern metamorphosis though residual xanthophores persist in individuals shifted during late pigment pattern metamorphosis or beyond. Mean densities of xanthophores were significantly reduced in upshifted individuals as compared to sibling controls left at 24°C (F_1,148=500.0, P<0.0001). (B) Temperature down-shift allows substantial xanthophore recovery through middle stages of pigment pattern metamorphosis, but less marked recovery during later metamorphic and juvenile stages (see text for details). Mean densities of xanthophores for downshifted individuals were significantly greater overall than sibling controls left at 33°C (F_1,74=9.27, P<0.005). (C) Melanophore organization is correlated with xanthophore density. Reduced variation in nearest neighbor distances between melanophores is associated with increased xanthophore densities in both temperature upshift (red points) and downshift (green points) experiments. Red diamond indicates the mean for individuals completely lacking xanthophores at 33°C (upshift and control, pooled); green diamond, the mean for control individuals reared exclusively at 24°C. Note that variability in melanophore nearest neighbor distances is increased among individuals with partially disrupted stripes, whereas the most severe phenotypes at 33°C have somewhat lower coefficients of variation, reflecting a more uniform dispersion of melanophores once xanthophores and stripes have been lost. Regression shown includes only individuals with partial xanthophore deficits compared to controls. Individual and pooled values shown are based on 98,643 melanophores, 47,067 xanthophores. (D) Complete melanophore stripes are more common when Fms activity is provided by temperature downshift prior to late metamorphic stages. Shown are percentages of individuals downshifted at different sizes that exhibited complete melanophore stripes (as defined by 1 200 µm gap per side). Thus, individuals shifted at sizes 8 mm SL typically exhibited more broken stripe patterns (²=65.2, d.f.=13, P<0.0001). The orange line indicates the percentage of individuals with complete stripes among controls reared at 24°C.