FFT'ed two gaussian functions then multiplied them together then ifft'ed the result, but I got something that is completely out of whack... Why?

The following code is what I ran. Essentially I created 2 gaussian functions then moved them both into frequency domain via fft and then multiplied them together in frequency domain. I then ifft the result hoping to see something that looks like a "blurred" Gaussian. However, instead what I get is the monstrosity in the blue curve in figure(1)... I don't understand why is the blue curve anything but what I expected. Any help is much appreciated.

if true
close all;
clear all;
clc;
col=1000;
sx=(col+1)/2;
k=50;
a=1000/k;
x=1:1:1000;
t=exp(-5*(x-sx).^2/a^2);
t=t/sum(sum(t));
T=fft(t);
b=100;
f=exp(-5*(x-sx).^2/b^2);
F=fft(f);
O=F.*T;
tt=ifft(T);
ft=ifft(F);
ot=ifft(O);
figure(1)
plot(x,tt,'r',x,ft,'g',x,ot,'b');
figure(2)
plot(x,abs(T),'r',x,abs(F),'g',x,abs(O),'b');
figure(3)
plot(x,angle(T),'r',x,angle(F),'g',x,angle(O),'b');
end

Answer 1

[Brad Liu]

Problem Resolved... Essentially, the length of the results need padding. Cannot get away with the increased length of vectors. Have to truncate it manually.