Laserfiche WebLink
<br />BIOLOGICAL METHODS <br /> <br />TABLE 3. FREQUENCY TABLE FOR DATA <br />IN TABLE I GROUPED AT AN INTERVAL <br />WIDTH OF 20,000 CELLS/ML <br /> <br />Interval <br /> <br />Interval <br /> <br />Frequency <br /> <br />Frequency <br /> <br />0- 20 <br />20 - 40 <br />40 - 60 <br />60 - 80 <br />80 - 100 <br />100 - 120 <br />120 - 140 <br />140 - 160 <br />160 -180 <br />180 - 200 <br /> <br />13 <br />11 <br />11 <br />2 <br />1 <br />2 <br />2 <br />2 <br />1 <br />o <br /> <br />200 - 220 <br />220 c 240 <br />240 - 260 <br />260 - 280 <br />280 - 300 <br />300 - 320 <br />320 - 340 <br />340 - 360 <br />360 - 380 <br />380 - 400 <br /> <br />o <br />1 <br />o <br />o <br />1 <br />1 <br />o <br />o <br />o <br />o <br /> <br />Closely related to the cumulative frequency <br />histogram is the cumulative frequency distribu- <br />tion graph, a graph of relative frequencies. To <br />obtain the cumulative graph, merely change the <br />scale of the frequency axis on the cumulative <br />frequency histogram. The scale change is made <br />by dividing all values on the scale by the highest <br />value on the scale (in this case the number of <br />observations or 48). <br />The value of the cumulative frequency distri- <br />bution graph is to allow relative frequency to be <br />read, i.e., the fraction of observations less than <br />or equal to some chosen value. Exercise caution <br />in extrapolating from a cumulative frequency <br />distribution to other situations. Always bear in <br />mind that in spite of a planned lack of bias, each <br />sample, or restricted set of san'F~Cs, is subject to <br />influences not accounted for and is therefore <br />unique. This caution is all the more pertinent for <br />cumulative frequency plots because they tend to <br /> <br />14 <br /> <br /> <br />10 <br /> <br />>- <br />"" <br />z: <br />.... <br />=> <br />e 6 <br />co:: <br />.... <br /> <br />40 80 120 160 200 240 280 320 <br />ALGAL CELLS/ML, THOUSANOS <br /> <br />Figure 2. Frequency histogram; interval width is <br />20,000 cells/mt. <br /> <br />. <br /> <br />smooth out some of the variation noticed in the <br />frequency histogram. In addition, the phrase <br />"fraction of observations less than or equal to <br />some chosen value" can easily be read "fraction <br />of time the observation is less than or equal to <br />some chosen value." It is tempting to generalize <br />from this reading and exten.d tht,se results <br />beyond their range of applicability. <br /> <br />14 <br /> <br /> <br />10 <br /> <br />>- <br />"" <br />z: <br />.... <br />=> <br />e 6 <br />e: <br /> <br />40 80 120 160 200 240 280 320 <br />ALGAL CELLS/ML, THOUSANDS <br /> <br />Figure 3. Frequency polygon; interval width is <br />20,000 cells/mt. <br /> <br />50 <br /> <br />40 <br /> <br /> <br />>- <br />.., <br />z: <br />~ 30 <br />~ <br />> <br />~ <br />~ 20 <br />~ <br /> <br />10 <br /> <br />o <br />o 40 80 120 160 200 240 280 320 360 400 <br />ALGAL CEllS/Ml. THOUSANDS <br /> <br />Figure 4. Cumulative frequency histogram; in- <br />terval width is 10,000 cells/ml. <br /> <br />3.5 Two-dimensional Graphs <br /> <br />Often data are taken where the observations <br />are recorded as a pair (cell count and time), <br />(biomass and nutrient concentration). Here a <br />quick plot of the set of pairs will usually be of <br />value. Figure 5 is such a graph of data taken <br />from Table 1. Each point is plotted at a height <br /> <br />. <br /> <br />8 <br />