![The UV spectral slope ($\beta$, where $f_\lambda \propto \lambda ^{\beta }$) plotted against the [O iii]$\lambda 5007$ EW. Left panel: The observed $\beta$ slope. Centre panel: $\beta$ slope corrected for dust attenuation. Right panel: $\beta$ slope plotted against [O iii]$\lambda 5007$ EW, both corrected for dust attenuation (taking into account the differential extinction of the nebular line relative to the continuum as in Calzetti et al. 2000). In each panel, we mark galaxies as diamonds where the $\beta$ slope was measured directly from the prism spectroscopy with diamonds, and plot as squares those where we rely on the spectrum at a slight longer rest-frame wavelength and photometry for our $\beta$ measurement (typically for those only at $z\lt 4$). High- and low-redshift subsamples are identified as blue and orange data points, respectively. In all panels, we include a shaded region to mark the EELG EW threshold of [O iii]$\lambda 5007 \gt 750$Å, and the measured quantities for our four stacks are shown in large circles (coloured to match Fig. 3). In the right panel, we show in bold the individual galaxies where their classification changes from non-EELG to EELG due to the dust correction.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/mnras/535/2/10.1093_mnras_stae2430/1/stae2430fig9.jpeg?Expires=1749770935&Signature=bkX70ZaW5KnSzcYvri6RuzVMPsjZ83aUL0Q52QsqDULlRzs-QLGwAAWDtQTijUqJQWfe0pHvYsTfQRVFekQwvgrr~NWNQohN1YNoQjYZdNj7RwfNeV~gqMbCEhT4SDnkHwYpyy~C2mnLeGLOsX9pK7ApX0y9XHkIgsxOWcyGw-~QCuTfzOUXN4U2EHY3qpLqJtsDf8Xie195RhI9T3SO5GhtRVV8drwT4TyrXyywOoi8xrGzjQ4qA199nGIKn2ld5TO9WiSB2saTv-ni6gqHzoBUmT12-n5X0KdCaNzGePuQBxepPErQnGrGunRP5CLQjdDiUQcinqEM5dCSGCUB7A__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
The UV spectral slope (|$\beta$|, where |$f_\lambda \propto \lambda ^{\beta }$|) plotted against the [O iii]|$\lambda 5007$| EW. Left panel: The observed |$\beta$| slope. Centre panel: |$\beta$| slope corrected for dust attenuation. Right panel: |$\beta$| slope plotted against [O iii]|$\lambda 5007$| EW, both corrected for dust attenuation (taking into account the differential extinction of the nebular line relative to the continuum as in Calzetti et al. 2000). In each panel, we mark galaxies as diamonds where the |$\beta$| slope was measured directly from the prism spectroscopy with diamonds, and plot as squares those where we rely on the spectrum at a slight longer rest-frame wavelength and photometry for our |$\beta$| measurement (typically for those only at |$z\lt 4$|). High- and low-redshift subsamples are identified as blue and orange data points, respectively. In all panels, we include a shaded region to mark the EELG EW threshold of [O iii]|$\lambda 5007 \gt 750$|Å, and the measured quantities for our four stacks are shown in large circles (coloured to match Fig. 3). In the right panel, we show in bold the individual galaxies where their classification changes from non-EELG to EELG due to the dust correction.
