discrepancy while running the eval funcion

The source code is below. R1 is just a multinomial in terms of T1 and T2. There is a discrepancy for the values m1 and m2. I am supposing them to be identical but they are not. Worse still, the third entries of them are quite different. Is there anything wrong when I used the eval function in the code? Does this discrepancy disappear if I used higher resolution (i.e., more digits) for the calculation? How? Thanks.

clc;clear all;

format long;

syms T1 T2;

R1=-(20*T1*T2 + 51*T1*T2^2 + 10*T2^2)*(146272*T1^2*T2^3 + 41342*T1^2*T2^2 + 1040*T1^2*T2 + 200*T1^2 + 27099*T1*T2^3 + 2040*T1*T2^2 + 400*T1*T2 + 1690*T2^3 + 200*T2^2)*(- 250710208*T1^3*T2^5 + 41873972*T1^3*T2^4 + 14400044*T1^3*T2^3 - 1797278*T1^3*T2^2 + 869840*T1^3*T2 - 9800*T1^3 + 54996498*T1^2*T2^5 + 75095644*T1^2*T2^4 + 8271756*T1^2*T2^3 + 1257020*T1^2*T2^2 + 10800*T1^2*T2 + 2000*T1^2 + 16185776*T1*T2^5 + 7516322*T1*T2^4 + 819720*T1*T2^3 + 78200*T1*T2^2 + 4000*T1*T2 + 1209560*T2^5 + 427020*T2^4 + 57600*T2^3 + 2000*T2^2)*(2002492630154240*T1^5*T2^6 - 1241222117102912*T1^5*T2^5 - 191601385814160*T1^5*T2^4 + 32207386732184*T1^5*T2^3 - 11523881039144*T1^5*T2^2 + 1400056482720*T1^5*T2 - 17878754400*T1^5 - 3809563673058304*T1^4*T2^7 + 2827038395565216*T1^4*T2^6 + 140073437800760*T1^4*T2^5 - 2811507824820*T1^4*T2^4 + 21912078533196*T1^4*T2^3 - 42466040126*T1^4*T2^2 - 171398217520*T1^4*T2 + 6002953400*T1^4 + 205333666033664*T1^3*T2^8 - 1324036748127328*T1^3*T2^7 + 654778663392480*T1^3*T2^6 - 98796906066260*T1^3*T2^5 - 13089768910652*T1^3*T2^4 - 3618893057244*T1^3*T2^3 - 59609267456*T1^3*T2^2 + 2257369680*T1^3*T2 - 556801600*T1^3 - 45042571833984*T1^2*T2^8 - 165802236565568*T1^2*T2^7 + 105699179029248*T1^2*T2^6 + 899931973192*T1^2*T2^5 + 313813414778*T1^2*T2^4 - 12802978208*T1^2*T2^3 - 16720825360*T1^2*T2^2 - 644966400*T1^2*T2 + 11204000*T1^2 - 13256280030208*T1*T2^8 - 3828427453056*T1*T2^7 + 7580810037504*T1*T2^6 + 1081612069392*T1*T2^5 - 12485342896*T1*T2^4 - 191758960*T1*T2^3 + 532846400*T1*T2^2 + 22408000*T1*T2 - 990639316480*T2^8 + 227077392640*T2^7 + 379437840640*T2^6 + 113723069920*T2^5 + 12752559640*T2^4 + 621011200*T2^3 + 11204000*T2^2);

T1=-0.4;

R11=eval(R1);

rt=eval(solve(R11));

rt_2=rt(find(rt>-0.9 & rt<-0.4));

m1=zeros(1,length(rt_2));

m2=zeros(1,length(rt_2));

     for ii=1:length(rt_2)

        T2=rt_2(ii);

        m1(ii)=eval(R11);

        m2(ii)=eval(R1);

     end

m1

m2