Series Expansion Inconsistency

*To*: mathgroup at yoda.physics.unc.edu*Subject*: Series Expansion Inconsistency*From*: Cetin Cetinkaya <cetin at acm0.me.uiuc.edu>*Date*: Mon, 19 Apr 1993 20:19:27 -0500

Hello; While I am trying to do some perturbation analysis, I run into a peculiar behaviour. My function k looks like In[31]:= k 0. Second 1928.568 KByte 3 - n 3 - n (c1 eps) - (c20 + c21 eps) Out[31]= ------------------------------------------------------------- 2 - n 2 - n c1 eps (c20 + c21 eps) ((c1 eps) - (c20 + c21 eps) ) For n=19, I expand the function k with respect to eps up to eps^2 and collect the coefficients of eps: In[32]:= Collect[Simplify[Normal[Series[k/.n->19,{eps,0,2}]]] ,eps] 3.15 Second 1929.216 KByte 1 Out[32]= --- c20 This simply means that the expansion does not have eps or eps^2 orders terms. However, When I tried the following to see how the eps^3 order term looks like, I have got this: In[33]:= Collect[Simplify[Normal[Series[k/.n->19,{eps,0,3}]]] ,eps] 3.81667 Second 1930.156 KByte 2 2 3 3 1 c21 eps c21 eps c21 eps Out[33]= --- - ------- + --------- - --------- c20 2 3 4 c20 c20 c20 To my suprise, this expression has order eps and eps^2 terms. Could any one explain what is going on? The Mathematica verson I am running is 2.1. Cetin