Fig. 4 in Age and body size of Salamandrella keyserlingii (Caudata: Hynobiidae): a difference in altitudes, latitudes, and temperatures

Fig. 4 Fittest equations (rational function approximation or second order polynomial regression) between latitude (x-axis: °N)/temperature in descending order (x-axis: °C) and maximum body size (y-axis: mm). No relationship was found between latitude and M-TOL (solid circle: n 018, R 2 0.043, P 0.4116, y intercept 144.876, slope 0.414) M-SVL 0000– or (open circle: n 024, R 2 00.020, P00.5097, y intercept 077.240, slope 0 –0.175). Rational function approximations indicated an intraspecific tendency to decrease M-TOL (Equation a: y ¼ 4: 98338 x 454: 274 þ 16386: 0= x R2 ¼ 0: 202> 0: 043 regression coefficient for a given linear ' regression) with increased latitude from 43 to 57°N and then increase the size from 57 to 69°N (U shaped curve), unlike M-SVL (Equation b: y ¼ 0: 351151 x þ 21: 1776 þ 1478: 83= x R2 ¼ 0: 022 0: 020). A ' significant relationship was not found between temperature and M-TOL (n 18, R 2 0.127, P 0.1464, y intercept 124.426, slope 0.857), and 00000 – neither relationship nor trend was found between temperature and M-SVL (n 22, R 2 0.001, P 0.8804, y intercept 68.330, slope 0.048). Poly00000 – nomial regressions also appeared to decrease M-TOL (Equation c: y ¼ 0: 142278 x2 þ 1: 91652 x þ 121: 877 R2 ¼ 0: 304> 0: 127 ' regression coefficient for a given linear regression) with decreased temperature from 8 to –7 °C and then increase the size from –7 to –15 °C (U shaped curve), but they did not show any trend for M-SVL (Equation d: y ¼ 0: 00965508 x2 þ 0: 0919399 x þ 68: 1711 R2 ¼ 0: 003 0: 001) '