The slit normally defines the location of a given wavelength in the spectrum but if the slit is wider than the star image then I guess there could be some wavelength shift between exposures due to scintillation. To avoid scintillation problems (and uneven coverage of the slit due to guiding errors) I normally accumulate at least 1 min total exposure time for a spectrum (which for bright targets at low resolution can be in shorter sub exposures), though this is not based on any scientific analysis.(I see I only used 30 sec for eps Tau in this particular exercise)
https://britastro.org/sites/default/files/attachments/ALPY600_MILES_tests_20161105.pdf