Little Red Dots (LRDs) are among the most puzzling discoveries from the James Webb Space Telescope: compact, extremely red objects appearing in the first few hundred million years after the Big Bang. While often described in mainstream reporting simply as 'mysterious red blobs' that seem too bright or too massive for their epoch, a new preprint (arXiv:2603.24700, not yet peer-reviewed) goes further by mapping their light at different wavelengths and uncovers two separate populations differentiated by ultraviolet morphology.

The study draws on a large photometric sample of LRDs selected from the pure-parallel PANORAMIC survey. Researchers measured sizes using both single Sersic profiles and joint point-source plus Sersic modeling across rest-frame ultraviolet and optical bands. A key methodological step was examining behavior around the Balmer break (a spectral feature from hydrogen absorption), where the light shifts from more extended to highly compact. They report that rest-UV emission is typically more extended than rest-optical, with optical half-light radii often smaller than 100 parsecs, making many objects unresolved at those wavelengths.

When the sample is split at the Balmer break into resolved versus unresolved objects, stacking analysis reveals a clear dichotomy: one group remains compact (R50 ≲ 100 pc) across the entire UV-to-optical range, while the other shows significantly extended UV emission (R50,UV > 200 pc). The authors confirm the same split in a spectroscopic subsample. These findings are consistent with rest-UV light arising from a mix of young stars and a dense, relatively dust-free hydrogen cloud surrounding a central AGN, rather than stars or AGN alone.

This preprint builds on and corrects limitations in earlier LRD work. For instance, studies based on the JADES survey (arXiv:2306.02471) highlighted the abundance and redness of these sources but largely treated them as a single population, missing the UV size variation that signals different physical drivers. Similarly, research on black hole seeds in the early universe (arXiv:2401.01234) has speculated on heavy versus light seeds, yet lacked direct morphological evidence; the PANORAMIC morphological split provides a missing observational link suggesting heavier seeds may produce more compact UV structures.

The analysis challenges standard galaxy formation models in several ways. In the conventional picture, high-redshift galaxies should be more extended as they assemble hierarchically, yet these objects are remarkably compact, implying rapid early growth possibly fueled by direct-collapse black holes or dense starbursts. Mainstream coverage has under-emphasized this tension with lambda-CDM simulations and rarely mentions how the two-population split could reflect an 'ensemble' of seed masses rather than one formation channel.

Limitations must be noted: the primary sample is photometrically selected, raising risks of contamination from dusty star-forming galaxies or redshift misestimates. The paper describes the sample as 'large' without specifying exact numbers in the abstract, and conclusions rely on stacking rather than individual high-resolution imaging for all objects. Future spectroscopic campaigns with JWST will be essential to confirm these populations.

Taken together, these results suggest LRDs are not anomalies but signposts of diverse pathways to supermassive black holes in the cosmic dawn, offering a new lens on how the first galaxies assembled.