Allegro

Figure 4.74 Breadboard Bode plots (Data provided courtesy of Allegro ).

John Allegro, a leading Dead Sea Scroll scholar and author of The Sacred Mushroom and the Cross, a daring volume of speculation about the origins of religion. 

The example we will consider in detail is a company that produces pressurized canisters of drug for asthma inhalers. Their factory has two machines that manufacture the canisters. One machine (Allegro model) is faster and produces 61 per cent of the factory s output. The slower machine (Andante model) produces the other 39 per cent. Occasional canisters are returned as faulty, and over the previous year, 126 of these have accumulated. From the batch numbers on the canisters we can determine that 52 were manufactured on the Allegro machine and 74 on the Andante. Since only a minority of canisters are made on the Andante machine, and yet it generates the majority of returns, we suspect that the output from this machine is less reliable. However, a formal statistical test is required. 

First we need a null hypothesis. This will state that the products of the two machines are equivalent and if we went on collecting dodgy canisters for long enough the proportions of returns from each machine would simply be proportional to the numbers they produced. More formally it would be among a large population of returned canisters 61 per cent would be from the Allegro machine and 39 per cent from the Andante . 

To perform the test we first calculate the so-called expected frequencies . These are an idealized set of frequencies calculated to match the null hypothesis exactly. So with our 126 returned canisters, the null hypothesis claims that 61 per cent of these should be from the Allegro and 39 per cent from the Andante machine. Hence, the expected frequencies are 126 x 0.61 = 76.86 (Allegro) and 126 x 0.39 = 49.14 (Andante). Notice that the term expected is being used in a technical sense. There is no... 

Allegro or Andante), so the critical y2 is 3.842. Our data set yielded a y2 of 20.62. As the value we achieved is way in excess of the critical value, our conclusion is very clearly significant. There is overwhelming evidence that the Andante machine is producing a disproportionately high proportion of the faulty canisters. 

Many GC detectors exist, but not all are suitable for phytochemicals. The thermal conductivity detector (TCD) is considered a universal detector and is appropriate for most analytes as long as the thermal conductivity of the carrier gas is different from that of the analytes. During the early development phase of GC, TCD was an easy choice because thermal conductivity measuring devices were already in use (Colon and Baird, 2004). Ionization detection arrived with its improved trace determinations and replaced TCD in many applications. While TCD is still used for some food applications (Allegro et ak, 1997 Sun et ak, 2007) and in the past was used for phenolic acids (Blakely, 1966), currently it is not generally used for phytochemicals. Rather, the flame ionization detector (FID) is better-suited due to its selectivity for organic compounds and superior measuring ability for trace measurements. 

Allegro C. Oliveira F.A.R. Pocas M.F. 1997. Influence of water activity and oxygen on keeping quality characteristics of ffeeze-dried strawberries. In Proceedings of the Third Main Meeting Process Optimisation and Minimal Processing of Foods, pp. 314-318. 

---