### Table 3: We nd little or no evidence that these E. coli pathways occur in H. in uenzae.

1996

"... In PAGE 9: ...iosynthesis pathways appear to be missing from H. in uenzae. Also missing are several catabolic pathways for carbohydrates, which could be linked to the growth-medium versatility of this organism. Table3 should be considered the least solid prediction; because we used more conservative evidence in making these predictions, a lack of evidence for a given pathway should not be considered strong evidence that the pathway does not occur. That is, were we performing a more liberal prediction, some evidence would no doubt appear for some of these pathways.... ..."

Cited by 20

### Table 3 reports the results. Examining the returns across deciles reveals little evidence

### TABLE 1 Effect of incubation temperatuires c)f 24.5, 37, and 48 C on7 the percentage of dissociants of Lactobacilllls builgaricuis in carrot liver broth DISSOCIANTS*

1945

### Table 4: Ordinary standard deviations and correlations estimated on a 500 days sample (1994-95) The MSW estimation results shown below are based on a two state model (k = 2) using the speci cation indicated in section 2.3 restricting means to zero. Results for a three state model are omitted because there was little evidence found for a third state. See for example the results of the score tests presented below (cf. table 9). 11

### Table 1: Correlations between Random Initial Conditions methodologies in di erent Methodologies based on di erent initial conditions. Each entry in the table corresponds to the correlation between two methodologies (two di erent random seeds), where each metholology consists of 9 di erent ver- sions, or training sets. As is apparent from the table, there is little evidence here of any resulting diversity. The failures between the methodologies are all highly correlated. Figure 3 shows a plot of one of these comparisons (between Random Seed 1, and Random seed 2). As should be apparent from the gure, as the probability of failing on one methodology increases, so does the probability of failing on the other.

1995

"... In PAGE 6: ... Using the statistical methods of Lit- tlewood and Miller (1989), it was then possible to calculate the correlation between the failures of pairs of methodologies, testing across several ver- sions. In Table1... In PAGE 10: ...ion of 0.5957. Thus, the correlations achieved are lower than those found as a result of the other two methods. The lowest correlation achieved as a result of varying the initial conditions ( Table1 ) was 0.9273, and as a result of varying the training sets (Table 2) was 0.... ..."

Cited by 3

### Table 4.1: Timings in minutes:seconds for GKA (Phase One, 1st descent). prime inputs of 201 digits and one of 401 digits. For the 201 digit triple, we observe obvious bene ts in increasing the number of subtasks from 2 to 8. Evidently, for this size input, there is little to be gained from having more than 8 subtasks.

1992

### Table 1 shows very clearly that 1 belongs to ER(f) while Tables 3 and 4 show that f0:5; 0g\ER(f)=;. Table 2 is more interesting: the value 8=9=0:8889 = 2 ER(f), but the interval (8=9 ; 8=9+ )=(0:7889; 0:9889) is very close to 1 and mfx 2 [ ; ]:f(x)=1g = . This little pathology is evident since for n =16; 32 it seems that 8=9 2ER(f). However it is enough to inspect Table 2 for n 64 to recover the correct negative information.

"... In PAGE 19: ...1551 Table1 . (Example 1): = 1, Un = Fn n 16 32 64 128 256 512 Z 7 15 31 63 127 255 R * 2.... ..."

### Table 2: The Effects of Capital Account Liberalization on Growth, Investment and Inflation: Cross-Sectional Evidence

1998

"... In PAGE 10: ... I estimate these regressions using ordinary least squares (OLS), and also instrumenting for financial openness using its own average value over the previous ten years 1975-1984 as an instrument (IV).13 The results are presented in Table2 . Throughout the paper, I adopt the following convention for reporting results.... In PAGE 11: ... The results in Table2 are generally consistent with those existing in the literature. There is very little evidence that growth or investment are higher in more financially open economies.... ..."

Cited by 4