diff --git a/README.md b/README.md
index 4fb4fc766ee1c59bd591fbe712b8d47e5c0a8ae6..c053bf97e57539b2646b044ad694dd1f2a1ccfe3 100644
--- a/README.md
+++ b/README.md
@@ -208,6 +208,14 @@ I put together a load cell "dev kit" package to send to Mike at Moog:
 <img src='img/package.jpg' height=300px>
 
 
+### Noncontact, bipolar, nonlinear loadcells
+
+I've become interested in nonlinear load cells -- those whose transfer function changes slope in order to be very sensitive at small loads and less sensitive at larger loads (in order to maximize range).  One way to think about this transfer function is as a soft transition into a protective stop.
+
+Nonlinear load cells have been made by geometric design of flexures, as in this <a href="https://dspace.mit.edu/handle/1721.1/119789">thesis</a>. This is very cool, but seems it might be sensitive to contamination and the contact forces might exhibit some hysteresis.  It also doesn't seem to lend itself to miniaturization.
+
+Another approach is to use a flexure that is more linear but to measure displacement using a physical principle with a nonlinear transfer function.  One example of such a principle is frustrated total internal reflection, as the reflectivity is a near exponential function of gap distance.  <a href='https://link.springer.com/article/10.1140/epje/i2017-11542-4'>This paper</a> demonstrates the measurement of film thickness using this principle (and as a bonus includes some force measurement based on Hertzian contact between planar and spherical lenses!).  
+