BIOLOGICAL CONTROL IN TUMOR GROWTH
                                    
               Seth Michelson, Roche Bioscience, MS S1-137
                 3401 Hillview Ave, Palo Alto, CA  94303
                    e-mail: SETH.MICHELSON@SYNTEX.COM
                                    
                                    
     We believe that real science is the unification of
hypothesis, synthesis, and testing.  We think that science
requires the design, implementation, and execution of good
experiments within the context of a firm theoretical structure. 
We believe that data is not information; that data requires a
context for interpretation, and experimentation.  We believe that
theory without a firm grasp of the experiment, its implications,
and its limits is kinderspiel;  elegant theory, while intellectu-
ally exciting,  is of no purpose if it fails to predict observa-
tion.  These are fairly strong comments.  But we think they are,
for the most part, valid.  
     We have In modeled as many of tumor growth phenomena as
possible to gain as much insight from them as possible and deter-
mined holes in the evidence exist.  The analyses of our models,
using both constant coefficients and allowing the parameters to
vary with signal processing capacity, yielded a model of tumor
dormancy and recurrence, and one of tumor growth in the presence
of other growing tissues (concomitant resistance and growth in
the presence of a partial hepatectomy).  Our models defined two
forms of dormancy, one in which all the cells are quiescent (TYPE
I) and one in which the tumor was at a dynamic equilibrium (TYPE
II).  Recurrence as a phenomenon then depends explicitly on the
type of dormancy the tumor is in.