We present Doppler images of the young K5V - K7V rapid rotator LO Peg from seven nights of continuous spectroscopy obtained in 1998 from July 04 to July 10. The images reveal the presence of a strong polar cap with appendages extending to mid-latitudes, but no starspots are seen below 15\degr. We briefly discuss the distribution of spots in light of recent flux transport simulations which are able to reproduce the observed latitude dependence. With the full timeseries of spectra, of which 314 are useful, many phases are observed three times over the seven nights of observations. Using starspots as tracers of a solar-like latitudinal differential rotation in our image reconstructions, we find that the equatorial regions complete one more rotation than the polar regions every {181 $\pm$ 35} d. LO Peg is the second coolest star for which such a measurement has been made using indirect imaging methods. The degree of latitudinal shear is less than that seen in G and early K dwarfs, suggesting a trend in which differential rotation decreases with stellar mass in (pre-)main sequence objects.

From Zeeman Doppler images of the young, rapidly-rotating K0 dwarf AB Doradus, we have created a potential approximation to the observed radial magnetic field and have evolved it over 30 days due to the observed surface differential rotation, meridional flow and various diffusion rates. Assuming that the dark polar cap seen in Doppler images of this star is caused by the presence of a unipolar field, we have shown that the observed differential rotation will shear this field to produce the observed high-latitude band of unidirectional azimuthal field. By cross-correlating the evolved fields each day with the initial field we have followed the decay with time of the cross-correlation function. Over 30 days it decays by only 10 percent. This contrasts with the results of Barnes et al (1998), who show that on this timescale the spot distribution of He699 is uncorrelated. We propose that this is due to the effects of flux emergence changing the spot distributions.