4.2. Extremely Red Objects (EROs)
The development of large format near-IR detectors has enabled relatively deep, wide-field IR surveys, and lead to the discovery of a class of faint EROs (galaxies with colors in the range R - K > 5.5 - 6), supplementing traditional low-mass stellar EROs (Lockwood, 1970). At first EROs were found one by one, (Hu and Ridgeway, 1994; Graham and Dey, 1996), but statistical samples are now detected in K < 20 near-IR surveys (Thompson et al., 1999; Yan et al., 2000; Daddi et al., 2000; Totani et al., 2001), in parallel to their identification in multiwaveband surveys (Smail et al., 1999; Pierre et al., 2001; Smith et al., 2001; Gear et al., 2000; Lutz et al., 2001). The clustering of relatively bright K < 19.2 EROs (Daddi et al., 2000) is observed to be very strong, fueling speculation that EROs are associated with the deepest potential wells that have the greatest density contrast at any epoch. This has been used as an argument in favor of their association with submm galaxies, which could be good candidates for massive elliptical galaxies in formation (Eales et al., 1999; Lilly et al., 1999; Dunlop, 2001).
There are two obvious categories of extragalactic EROs: very evolved galaxies, containing only cool low-mass stars, and strongly reddened galaxies, with large amounts of dust absorption, but which potentially have a very blue underlying SED. Only the second are good candidates for identification with submm-luminous galaxies. Detections of faint radio emission associated with young supernova remnants in EROs, and determinations of signatures of ongoing star-formation in their rest-frame UV colors should allow these cases to be distinguished. Radio and submm follow-up observations of EROs have tended to show that most are passively evolved non-star-forming galaxies without detectable radio emission (Mohan et al., 2002), with at most about 10-20% being candidates for dust-enshrouded starbursts/AGN. It is important to remember that few EROs selected from wide-field near-IR surveys, which reach limits of K 20, are actively star-forming submm galaxies (see the summary of results in Smith et al., 2001).
A small but significant fraction of submm galaxies appear to be associated with EROs at bright magnitudes K < 20 (Smail et al., 1999, 2002). Many more submm galaxies probably fit the ERO color criterion, but at much fainter magnitudes; see for example the K = 22.5 SCUBA galaxy shown in Fig. 15 (Frayer et al., 2000). It is certainly possible that future fainter ERO samples with K > 22 could contain a greater fraction of submm-luminous galaxies and fewer passive ellipticals than the K 20 samples. Note also that EROs have unfavorable K corrections for detection at high redshifts (Dey et al., 1999; Gear et al., 2000): beyond z 2.5 any ERO would be extremely faint at even near-IR wavelengths. This could account, in part or in whole, for the extreme faintness of counterparts to submm galaxies, if many do have extremely red intrinsic colors.