Discussion

The above results are obtained after hot-block suppression. In order to do that, events with clusters that consist of just one LGD block are eliminated from the sample. Although we can see clear $\omega$ and $\phi$ production there are several problems that have to be resolved. First, the $\omega$ mass is close to its right value (782 MeV) for the low energy threshold, but it increases with increasing energy threshold in the LGD. For the last energy cut, it is higher than expected by $\approx$ 40 MeV. Second, the $\phi$ mass also increases with energy, but contrary to the $\omega$, the value from the fit is lower than expected (1020 MeV). The difference varies with energy from 90 to 50 Mev. Finally, the ratio of $\omega / \phi$ is less then expected: about 20 times as many $\omega$'s as $\phi$'s, independent of the energy cut. The production cross section is in favour of $\omega$ by a factor of $\approx$ 4. The ratio of branching ratios Br($\omega$ -$> \pi \gamma$)/Br($\phi$ -$> \eta \gamma$) is $\approx$ 6.5 and an additional factor of 2.5 comes from the ratio of $\pi$/$\eta$ decays to 2$\gamma$. Putting it all together, we expect $4*6.5*2.5 \approx$ 60 times more $\omega$ than $\phi$, but we see only 20.

Figure 2: Two Gaussian fit of $\pi$ $\gamma$ invariant mass in the range shown by solid line. The background is assumed to be one broad Gaussian (dashed line). The dot lines represent the same distributions for narrower $\pi ^0$ window.

Table 2: $\omega$ mass, width and yield as a function of deposited energy in the LGD.

	$E_{TOT} > [Gev]$		$M_{\omega} [Gev/c^2]$		$\sigma_{\omega} [Gev/c^2]$		$N_{\omega}$
	3.5		0.791		0.060		8.98 $10^6$
	4.0		0.798		0.065		8.30 $10^6$
	4.5		0.805		0.064		4.83 $10^6$
	5.0		0.819		0.061		1.95 $10^6$