**Table 2:** Swift XRT temporal decay fits. Quoted errors are 90% confidence on 1 interesting parameter.
	GRB050126	GRB050219a
Parameter	Model 1: Powerlaw ( $f(t)\propto t^{-\alpha}$ )
$\alpha$	$\rm 2.52^{+7.5}_{-0.3}$	$\rm 2.50^{+0.16}_{-0.16}$
$\chi^{2}/\nu$	-	225.1/38
Cash $statistic/ndp^{\dagger}$	62.0/20	-
	Model 2: Broken powerlaw
$\alpha_{1}$	$\rm 2.52^{+0.50}_{-0.22}$	$\rm 3.17^{+0.24}_{-0.16}$
$T_{\rm break}$	$\rm 424^{+561}_{-120}$	$\rm 332.1^{+25.8}_{-21.6}$
$\alpha_{2}$	$\rm 1.00^{+0.17}_{-0.26}$	$\rm0.75^{+0.09}_{-0.07}$
$\chi^{2}/\nu$	-	74.6/36
Cash statistic/ndp	26.1/20	-
	Model 3: Offset powerlaw $f(t)\propto (t-t_{\rm a})^{-\alpha}$
$\alpha$	$\rm 1.08^{+0.09}_{-0.09}$	$\rm 1.10^{+0.09}_{-0.08}$
$t_{\rm a}$	$\rm 105.1^{+9.1}_{-11.3}$	$\rm 100.7^{+2.8}_{-4.0}$
$\chi^{2}/\nu$	-	114.1/37
Cash statistic/ndp	31.7/20	-
	Model 4: Gaussian + powerlaw
$t_{\rm g}$ (fixed)	0.0	0.0
$\sigma$	$\rm 89.1^{+17.8}_{-16.1}$	$\rm 77.9^{+6.2}_{-6.0}$
$\alpha$	$\rm 1.11^{+0.12}_{-0.11}$	$\rm0.81^{+0.09}_{-0.07}$
$\chi^{2}/\nu$	-	89.4/36
Cash statistic/ndp	26.4/20	-

$\textstyle \parbox{9cm}{ \noindent $\dagger$\space Ndp number of data points.}$