I have a problem with wblfit and wblpdf.

My fitting result are shown on the following image.

Full code is almost same to the Matlab tutorial code except binWidth, binCtrs, and life.

      paramEsts = wblfit(life); % 0.1663, 0.6049
      n = length(life); %50
      binWidth = 0.032; 
      binCtrs = 0.032:binWidth:5;
      counts = hist(life, binCtrs);
      prob = counts / (n * binWidth);
      bar(binCtrs,prob,'hist');
      xlabel('Time'); ylabel('Probability Density'); ylim([0 4]);
      xgrid = linspace(0,5,100);
      pdfEst = wblpdf(xgrid,paramEsts(1),paramEsts(2));
      line(xgrid,pdfEst)

Life is: 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0320 0.0640 0.0640 0.0640 0.0640 0.0640 0.0640 0.0640 0.0640 0.0960 0.0960 0.1280 0.1280 0.1280 0.3520 0.3520 0.4160 0.6400 0.6720 0.8960 0.9280 1.1520 3.1680 4.4480

My question here is why the histgram is not integrated to 1? If I use

prob = counts / sum(counts);

isstead of

prob = counts / (n * binWidth);

The result is like the following:

Fitting is same and the area of histogram was integrated to 1. What's wrong here? And I found that even tutoral histogram was integrated to 0.5, not to 1.