The large majority of extragalactic star cluster studies done to date have essentially used two or three-passband aperture photometry , combined with theoretical stellar population synthesis models , to obtain age , mass and extinction estimates , and sometimes also metallicities .
The accuracy to which this can be done depends on the choice of ( broad-band ) passband combination and , crucially , also on the actual wavelengths and the wavelength range covered by the observations .
Understanding the inherent systematic uncertainties ( the main aim of this paper ) is of the utmost importance for a well-balanced interpretation of the properties of extragalactic star cluster systems . We simultaneously obtain ages , metallicities and extinction values for \sim 300 clusters in the nearby starburst galaxy NGC 3310 , based on archival Hubble Space Telescope observations from the ultraviolet ( UV ) to the near-infrared ( NIR ) .
We show that , for ages 6 \lesssim \log ( { age / yr } ) \lesssim 9 , and if one can only obtain partial coverage of the spectral energy distribution ( SED ) , an optical passband combination of at least four filters including both blue and red passbands results in the most representative age distribution , as compared to the better constrained ages obtained from the full UV–NIR SED coverage .
We find that while blue-selected passband combinations lead to age distributions that are slightly biased towards younger ages due to the well-known age–metallicity degeneracy , red-dominated passband combinations should be avoided . NGC 3310 underwent a ( possibly extended ) global burst of cluster formation \sim 3 \times 10 ^ { 7 } yr ago .
This coincides closely with the last tidal interaction or merger with a low-metallicity galaxy that likely induced the formation of the large fraction of clusters with ( significantly ) subsolar metallicities .
The logarithmic slope of the V -band cluster luminosity function , for clusters in the range 17.7 \lesssim { F 606 W } \lesssim 20.2 mag , is \alpha _ { F 606 W } \simeq - 1.8 \pm 0.4 .
The observed cluster system has a median mass of \langle \log ( m / M _ { \odot } ) \rangle \simeq 5.25 \pm 0.1 , obtained from scaling the appropriate model SEDs for known masses to the observed cluster SEDs .