1. The following table shows the distribution of predict (and noun prediction) and of forecast (verb and noun) in the theoretical research journal Nature (ca. 450,000 words) and the popular science journal New Scientist (ca. 2¼ million words):

   	predict			forecast
   	verb	noun	total	verb	noun	total
   Nature	176	33	209	0	0	0
   New Scientist	384	193	577	32	96	128

2. The following discussion of the meaning of prediction comes from a review of John Casti Searching for Certainty (Scribners) by Ian Stewart in New Scientist for 2nd May 1992:

   Can science foretell the future? People get terribly confused about the role of prediction in science. A lot of them have never understood that 'prediction' has two meanings. A theory of earthquakes, for example, can be 'predictive' without being able to predict earthquakes. A theory is predictive if it states that under certain conditions, certain consequences will follow. It is not obliged to state in advance when those conditions will hold.

3. Here are 16 more-or-less randomly selected KWIC citations for predict/prediction from Nature:

 1. volved in copper binding. Our findings predict that examples of selective editing of mitocho
 2. f the stratosphere. The present models predict that a cooling of the winter polar vortex by
 3. e analysis of this cDNA we are able to predict the complete amino-acid sequence of the polyp
 4. or this problem use the survey data to predict values on the vertical profile; by contrast,
 5. s the calcium-voltage hypothesis would predict an increase in release, locked in time to the
 6. igure 5 shows the discontinuity depths predicted by cross-correlation analysis, using three
 7. the velocity of the dendritic solution predicted for the same parameters, and so the former
 8. f the feedback is consistent with that predicted by climate models. This study demonstrates
 9. d an extraordinary 78% identity to the predicted amino-acid sequence of the recently reporte
10. equivocally placed. Figure 2 shows the predicted and observed structures of each of the hype
11. ults of this analysis were used in the prediction of the structure of the H2 region of the a
12. nciple, which would permit theoretical prediction of growth morphologies, must exist. The an
13. t least 10 55, in agreement with model predictions and ultraviolet observations 9-11. Photoi
14. backbones. The stereochemical analysis predicts that the stacked X structure should be right
15. the UAS of CYC7. The competition model predicts that such changes should increase expression
16. en major plates. Their analysis, which predicts that both east and west initial outer pseudo

4. And here are 16 more-or-less random citations for forecast from New Scientist

 1. alculations a second. The centre makes forecasts 10 days ahead for 18 national meteorologica
 2. e any action whose success hinges on a forecast being right. They might end up doing a lot l
 3. r stands up in the House of Commons to forecast Britain's economic performance for the next
 4. vice labour of its people. This gloomy forecast can be better understood by looking closely
 5. .  But three months earlier the secret forecast carried out by Treasury economists suggested
 6. f the equation, but also to failure to forecast correctly variables such as income and unemp
 7. er use of satellite data. This six-day forecast for Typhoon Yuri (below), which clipped the
 8. gnals at times." All users of economic forecasts, he adds, need to be aware of the wide marg
 9. s grossly underused because the demand forecast in the northeast has never materialised. An
10.  existing climate models would improve forecasts. `It's the things we don't know about that
11. s wrong: No one takes the government's forecasts of economic upturn seriously. The problem l
12. eorologists are still very cautious in forecasting stormier climates, stressing that the spa
13. the then health minister Roger Freeman forecast that `many hospitals will have to look to th
14. l by about 70 metres. This led them to forecast that modern ice sheets will grow and sea lev
15.  25 years ago. It was an experiment to forecast the result of the Greyhound Derby at White C
16. tion, and UK sales of sets in 1992 are forecast to be the lowest for 10 years, 25 per cent b

5. The following citations are taken from New Scientist:

   From the current values of each indicator, the values for the previous three quarters and the values for two years ago, the network predicts what will happen in two years' time. It combines these data to forecast another economic indicator calculated by the CSO called the Coincident Index, which is designed to follow the ups and downs of the gross domestic product but varies more smoothly.

   Palmer says that forecasts will soon change from predictions such as 'much of next week will be sunny' to 'there is a 90 per cent chance of sun next Monday and Friday'.