We have discovered that the spectrum of the well-known dwarf nova EM Cyg is contaminated by light from a K2–5V star ( in addition to the K-type mass donor star ) .
The K2–5V star contributes approximately 16 per cent of the light from the system and if not taken into account has a considerable effect upon radial velocity measurements of the mass donor star .
We obtain a new radial velocity amplitude for the mass donor star of K _ { 2 } = 202 \pm 3 \thinspace \hbox { $ \hbox { km } \thinspace \hbox { s } ^ { -1 } $ } , which compares with the value of K _ { 2 } = 135 \pm 3 \thinspace \hbox { $ \hbox { km } \thinspace \hbox { s } ^ { -1 } $ } obtained in Stover , Robinson & Nather ’ s classic 1981 study of EM Cyg .
The revised value of the amplitude combined with a measurement of rotational broadening of the mass donor \thinspace \hbox { $v \thinspace \hbox { sin } i$ } = 140 \pm 6 \thinspace \hbox { $ \hbox { km } % \thinspace \hbox { s } ^ { -1 } $ } , leads to a new mass ratio of q = M _ { 2 } / M _ { 1 } = 0.88 \pm 0.05 .
This solves a long standing problem with EM Cyg because Stover et al. ’ s measurements indicated a mass ratio q > 1 , a value which should have led to dynamically unstable mass transfer for the secondary mass deduced by Stover et al .
The revised value of the mass ratio combined with the orbital inclination i = 67 \pm 2 \degr leads to masses of 0.99 \pm 0.12M _ { \odot } and 1.12 \pm 0.08 M _ { \odot } for the mass donor and white dwarf respectively .
The mass donor is evolved , since it has a later spectral type ( K3 ) than its mass would imply .

We discuss whether the K star could be physically associated with EM Cyg or not , and present the results of the spectroscopic study .