29 glitches detected at Urumqi Observatory

Glitches detected in pulsar timing observations at the Nanshan radio telescope of Urumqi Observatory between 2002 July and 2008 December are presented. In total, 29 glitches were detected in 19 young pulsars, with this being the first detection of a glitch in 12 of these pulsars. Fractional glitch amplitudes range from a few parts in 10(-11) to 3.9 x 10(-6). Three 'slow' glitches are identified in PSRs J0631+1036, B1822-09 and B1907+10. Post-glitch recoveries differ greatly from pulsar to pulsar and for different glitches in the same pulsar. Three small glitches in PSRs B0402+61, B0525+21 and J1853+0545 show evidence for normal post-glitch recovery, but for PSRs B0144+59 and B2224+65 the spin frequency nu continually increases relative to the pre-glitch solution for hundreds of days after their small glitches. Most large glitches show some evidence for exponential post-glitch recovery on time-cales of 100-1000 d, but in some cases, e. g. PSR B1758-23, there is little or no recovery, i.e. no detectable increase in |(nu)over dot| at the time of the glitch. Beside exponential recoveries, permanent increases in slowdown rate are seen for the two large glitches in PSRs B1800-21 and B1823-13. These and several other pulsars also show a linear increase in.. following the partial exponential recovery, which is similar to the Vela pulsar post-glitch behaviour. However, the rate of increase, (nu) over dot is an order of magnitude less than for the Vela pulsar. We present improved positions for 14 of the glitching pulsars. Analysis of the whole sample of known glitches shows that fractional glitch amplitudes are correlated with characteristic age with a peak at about 10(5) yr, but there is a spread of two or three orders of magnitude at all ages. Glitch activity is positively correlated with spin-down rate, again with a wide spread of values. For some (but not all) pulsars, there is a correlation between glitch amplitude and the duration of the following interglitch interval. In no case is there a correlation of glitch amplitude with the duration of the preceding interglitch interval.