diff --git a/README.md b/README.md
index 3e823cf..ed4e45b 100644
--- a/README.md
+++ b/README.md
@@ -6,6 +6,8 @@
 
 RAIL interface to BPZ algorithms via the [DESC_BPZ](https://github.com/LSSTDESC/DESC_BPZ) package implementation (also available via PyPI with `pip install desc-bpz`).  Anyone using BPZ via either rail_bpz or DESC_BPZ should cite both [Benitez (2000)](https://ui.adsabs.harvard.edu/abs/2000ApJ...536..571B/abstract) and [Coe et al. (2006)](https://ui.adsabs.harvard.edu/abs/2006AJ....132..926C/abstract).
 
+In addition to the "default" BPZ CWWSB SED template set, `rail_base` now also includes an additional set of 31 template SEDs, a set that was used to compute the COSMOS 30-band photo-z's using LePhare.  The template set consists of empirical templates from [Poletta et al 2007](https://arxiv.org/pdf/astro-ph/0703255) supplemented with blue [BC03](https://ui.adsabs.harvard.edu/abs/2003MNRAS.344.1000B/abstract) SEDs, see [Ilbert et al 2009](https://ui.adsabs.harvard.edu/abs/2009ApJ...690.1236I/abstract) and [Dahlen et al 2013](https://ui.adsabs.harvard.edu/abs/2013ApJ...775...93D/abstract) for details.  The demo notebook [BPZ_lite_demo.ipynb](https://github.com/LSSTDESC/rail/blob/main/examples/estimation_examples/BPZ_lite_demo.ipynb) in the rail examples repository shows an example of how to use this alternate template set.
+
 As the "lite" name implies, not all features of BPZ are implemented, the main product is the marginalized redshift PDF, which is output for a sample as a `qp` ensemble.  However, several other quantities are computed and stored as "ancillary" data and stored with the ensemble, these are:
 - zmode (float): the mode of the marginalized posterior redshift PDF distribution.
 - zmean (float): the mean of the marginalized posterior redshift PDF distribution.