Journal of General Microbiology (1989), 135, 2967-2978.
Printed in Great Britain
2967
A Reappraisal of the Terverticillate Penicillia Using Biochemical,
Physiological and Morphological Features
11. Identification
By P. D. B R I D G E , ' D . L. H A W K S W O R T H , l * Z . K O Z A K I E W I C Z , '
A. H . S . O N I O N S , ' R . R. M . P A T E R S O N ' A N D M . J . S A C K I N 2
' CAB International Mycological Institute, Ferry Lane, Kew, Surrey TW9 3AF, UK
Department of Microbiology, University of Leicester, PO Box 138, Medical Sciences Building,
University Road, Leicester LEI 9HN, UK
(Received 20 January I989 :revised 6 June I989 ;accepted I7 July 1989)
The data from an integrated numerical classification was used to construct identification
schemes for some fasciculate penicillia. The identification schemes were presented as a synoptic
key and a frequency matrix for computer-assisted identification. Statistical testing of the
frequency matrix showed that although character separation values were generally low, only
four pairs of taxa showed overlap greater than that expected for a rectangular distribution. The
identification schemes were tested practically with 52 previously studied strains and 51 further
cultures. A synoptic key based on 10 and 90% cutoff limits was used to correctly identify 44 of
the 51 additional strains, although this proved very sensitive to single test discrepancies. The
frequency matrix was used to correctly identify 45 of the additional strains with a Willcox
probability score and this was compared to identifications based on the modal likelihood
fraction.
INTRODUCTION
Schemes for the identification of filamentous microfungi almost exclusively involve
dichotomous keys derived from morphological features, although some physiological features
such as growth temperatures and growth rate may be included. Such keys are available for the
identification of terverticillate penicillia (e.g. Pitt, 1986),although little has been published as to
the reliability of the diagnostic characters. Similarly, there is no information available
concerning the performance of the different identification schemes. Pitt (1986) suggested that
70-80% of strains from commonly encountered sources could be readily identified, and that
identification of the remainder 'must increasingly rely on the skill and experience of the
taxonomist' (Pitt, 1986). In our opinion inability to identify two to three of every ten isolates is an
unacceptable level of uncertainty in a group of major applied importance, especially as
specialists often differ in subjective assessments. With improved species concepts based on
larger data sets, opportunities for improved levels of identification in the genus emerge.
The critical assessment of the performance of identification schemes is now a routine
procedure in bacteriology, where identification matrices are tested both statistically and
practically (e.g. Willcox et al., 1980; Feltham & Sneath, 1982; Williams et al., 1983). Numerical
taxonomy can often produce data suitable for constructing keys and identification matrices, and
the associated information on strain and test reproducibility can therefore be included.
This paper uses the results of the numerical classification of Bridge et al. (1989) (preceding
paper) to produce a synoptic key and a frequency matrix suitable for computer-assisted
identification.
Abbreviations: HMO, hypothetical median organism; MLF, modal likelihood fraction.
0001-5360 0 1989 SGM
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
2968
P . D . BRIDGE A N D OTHERS
METHODS
Strains. The 348 strains detailed in Bridge et al. (1989) were used. An additional 51 strains comprising fresh
isolates, further culture collection strains and further lines of the ex-type cultures of P . atramentosum, P . commune,
P . expansum, P . griseofulvum, P . psittacinum, P . uerrucosum and P.caseicola which, for various reasons had been
omitted from the previous numerical taxonomy were used.
Constructionof the identificationmatrix. The data set given for some fasciculate penicillia and a few other species
in Bridge et al. (1989) was used to produce an identification matrix. A full list of species names, verified synonyms
and details of unnamed clusters is given in Table 1;only names for which ex-type isolates have been studied are
included. All tests that could be considered diagnostic, i.e. that gave both clear positive and negative results for
clusters, were used to produce an initial frequency matrix. Characters that had consisted of quantitative
measurements in the numerical classification were plotted as frequency distributions for the clusters. Class
intervals were chosen, and each quantitative character was converted into several binary characters based on
these limits. For example, the character colony diameter was included in the matrix as two characters, colony
diameter I 2 0 mm and colony diameter > 30 mm. The computer programs CHARSEP and DIACHAR (Sneath, 19796,
1980a) were then used to select the most discriminatory characters from this matrix. The practicality of all
characters was also considered ; for example, although cellulose hydrolysis appeared a useful character, this test
involved a 3 week incubation period and so was not used in the final matrix. Characters concerning conidial size,
shape and ornamentation were originally derived from scanning electron microscopy (SEM). As a result, in order
to use all tests in the matrix to full advantage this equipment would be necessary, although conidial shape and size
can be determined by light microscopy; the effect of different techniques on such measurements has been
discussed separately (Oliver et al., 1987). These characters were retained in the matrix as although their use may
not be entirely practical, they were considered appropriate for a general reference identification scheme. The
characters in the identification matrix were ‘anticlustered’ in an attempt to identify complementary tests (P. H. A.
Sneath, unpublished). This procedure groups together the least similar tests into clusters (instead of the more usual
grouping of the most similar), and complementary tests such as those that were ‘mirror images’ were identified and
omitted.
Statistical evaluation of the matrix. The identification score attainable by the hypothetical median organism
(HMO) for each taxon was determined by the program MOSTTYP (Sneath, 1980b). Overlap between taxa in the
matrix was tested with the program OVERMAT, the cutoff selected being that expected for a rectangular distribution
(Sneath, 1980~).
Practical evaluation of the matrix. The identification matrix was used with a BASICidentification program (P. D.
Bridge, unpublished) similar to MATIDEN (Sneath, 1979a), which used the Willcox probability score (Willcox et al.,
1973) as the identification criterion, and also gave a value for the modal likelihood fraction (MLF; Dybowski &
Franklin, 1968), calculated as the highest absolute probability obtained over the highest absolute probability
possible for that group with the tests used.
The identification matrix was tested in two ways. First, test results for 52 strains studied in the original
numerical classification (Bridge et al., 1989) were used to ensure that the matrix could correctly identify strains
that had been used to construct the clusters. Secondly, the 51 additional fasciculate strains were tested.
Keys. Synoptic keys were constructed from the data represented in the identification matrix. Percent positive
values were converted to three plus/minus tables based on cutoffs of 1 and 99%, 10 and 90% and 20 and 80%
respectively. Synoptic keys were constructed by considering each character independently. For each character,
taxa that were above the positive cutoff were scored as positive, and taxa that were below the negative cutoff were
scored as negative. Taxa that were between the cutoff values were scored as both positive and negative. The keys
were tested with the 103 strains used for testing the identification matrix.
RESULTS A N D DISCUSSION
The final identification matrix comprised 57 characters and 37 species or species-complex
groups (Table 2). Part species groups such as P . granulatum var. globosum and coremial strains
recovered as single clusters were kept separate as in the numerical taxonomy (Bridge et al.,
1989). Four pairs of taxa showed overlap greater than that expected for a rectangular
distribution. However, these instances all involved comparisons between very large groups and
very small groups and so the significance of the apparent overlap could not be tested. The pairs
involved were P . viridicatum (cluster 28, two strains) and P . solitum var. solitum (45 strains), P.
viridicatum (cluster 28) and P . aurantiogriseum(44strains), P . aurantiogriseum and cluster 3 1 and
P . olivinoviride (two strains) and P . soliturn var. solitum.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
Penicillium identijcation
2969
Table 1. Species names and synonyms for taxa where ex-type cultures were studied that are
included in the identijcation matrix*
P . atramentosum Thom 1910
P . aurantiogriseum Dierckx 1901
= P . aurantiocandidum Dierckx 1901
= P . puberulum Bainier 1907
= P . cyclopium Westling 1911
= P . brunneoviolaceum Biourge 1923
= P . aurantiogriseum var. poznaniensis Zaleski
1927
= P . polonicum Zaleski 1927
= P . johannioli Zaleski 1927
= P . verrucosum var. cyclopiurn (Westling)
Samson, Stolk and Hadlok 1976
= P . cyclopium var. aurantiovirens (Biourge)
Fassatiova 1977
P . aurantiogriseum var. melanoconidium Frisvad
ined.0
P . aurantiogriseum var. neoechinulatum Frisvad,
Filtenborg and Wicklow 1987
P . brevicompactum Dierckx 1901
= P . stolonferum Thom 1910
= P . olsonii Bainier and Sartory 1912
P . camembertii Thorn 1906
= P . caseicola Bainier 1907
P . chrysogenurn Thorn 1910
= ?P.griseoroseum Dierckx 1901
P . citrinum Thom 1901t
P . clavforme Bainier 1905t
= ?P.vulpinum (Cooke and Massee) Seifert and
Samson 1986$
P . clavigerum Demelius 19237
P . corylophilum Dierckx 1901t
P . echinulatum Fassatiova 1977
= P . cyclopium var. echinulatum Raper and
Thom 1949
= P . palitans var. echinoconidium Abe 1956 nom.
inval.
P . expansum Link 1809
P . glandicola var. mononematosa Frisvad, Filtenborg
and Wicklow 1987
P . granulatum Bainier 1905
= ?P. glandicola (Cooke and Massee) Seifert and
Samson 1986$
P . granulatum var. globosum Bridge and others 1989
P . griseofulvum Dierckx 1901
= P . patulum Bainier 1906
= P . urticae Bainier 1907
= P . jexuosurn Dale in Biourge 1923
P . hirsutum Dierckx 1901
= P . corymbiferum Westling 1911
= P . verrucosurn var. corymbiferum (Westling)
Samson, Stolk and Hadlok 1976
P . hirsutum var. albocoremium Frisvad ined.0
P . hirsutum var. allii Frisvad ined.5
P . hordei Stolk 1969
P . olivinoviride Biourge 1923
P . raistrickii G. Smith 19337
P . roquefortii Thom 1906
= P . roquefortii var. viride Dattilo-Rubbo 1938
P . solitum Westling 1911
= P . palitans Westling 1911
= P . schneggii Boas 1914
= P . javoglaucum Biourge 1923
= P . gladioli McCulloch and Thom 1928
= P . lanosogriseum Thom 1930
= P . ochraceum Bainier apud Thom 1930
= P . carneolutescens G . Smith 1939
= P . roquefortii var. punctatum Abe 1956 nom.
inval.
= P . casei var. compactum Abe 1956 nom. inval.
= P . pseudocasei G . Smith 1963
= P . mali Novobranova 1974
= P . conservandii Novobranova 1972
= P . verrucosum var. melanochlorum Samson,
Stolk and Hadlok 1976
P . solitum var. crustosurn Bridge and others 1989
= P . crustosurn Thom 1930
P . verrucosum Dierckx 1901
P . viridicatum Westling 1911
* Unnamed taxa included in matrix.
Cluster 17. This cluster consisted of only two strains, received as the ex-type cultures of P . camembertii and P .
caseicola. The cluster, however, differed from other isolates of P . camembertii in production of citrinin, a reduced
level of physiological activity and in having a generally deteriorated morphology, one strain showing mainly onelevel branched penicilli.
Cluster 22. This cluster consisted of four strains received as two new varieties of P . glandicola, a possible synonym
of P . granulatum. These strains all produced the secondary metabolite 2-carboxy-3,5-dihydroxyphenylacetyl
carbinol, a relatively uncommon metabolite in this matrix. The taxonomic position of these strains is at present
unclear.
t These species do not have terverticillate branches but were included in the study as ‘marker’ species.
$ Earlier species epithets from the genus Coremium have been suggested for P . claviforme (P.vulpinum) and P .
granulatum (P.glandicola) by Seifert & Samson (1986).
5 These varietal names have yet to be published.
DIACHAR gave similar sums of scores for each group (16.1-24.7), the lower totals being found
for the large variable groups P . aurantiogriseum and P . solitum varieties. CHARSEP gave a wide
range of separation values for the characters (CSP indices from 0.174.76) although
anticlustering showed few tests to be linked. Only one character in the final matrix (production
of penitrem A) did not appear in the lists of the five most useful tests for each taxon. Characters
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
2970
P. D . B R I D G E AND O T H E R S
Table 2. Percent positive frequency matrix for 37 Penicillium taxa
Taxon
1. P . expansum (cluster 1)
2. P . expansum (cluster 2)
3. P. expansum (cluster 3)
4. P. aurantwgriseum (cluster 4)
5 . P. aurantwgriseum (cluster 6)
6. P. chrysogenum (cluster 7)
7. P. echinulatum
8. P. granulatum var. globosum
9. P . granulatum var. granulatum
10. P . hirsutum (cluster 12)
1 1. P . hordei (cluster 13)
12. P . chrysogenum (cluster 14)
13. P. glandicola var. mononematosa
14. P. camembertii (cluster 16)
15. Penicillium sp. (cluster 17)
16. P . verrucosum
17. P. atramentosum (cluster 19)
18. P. atramentosum (cluster 20)
19. P. atramentosum (cluster 21)
20. Penicillium sp. (cluster 22)
21. P. citrinum
22. P. brevicompactum
23. P. raistrickii
24. P . aurantwgriseum var. neoechinulatum
25. P. viridicatum (cluster 27)
26. P. viridicatum (cluster 28)
27. P. griseofulvum
28. P . hirsutum var. allii
29. P. aurantiogriseum var.
melanoconidium
30. P. clavigerum
31. P . roquefortii
32. P. corylophilum
33. P. claviforme
34. P. solitum var. crustosum
35. P. olivinoviride
36. Coremial isolates
37. P. solitum var. solitum
1 1 9 9 16799 189 1638999 1 1 1 1 1 1 6 7 9 9
99 1 9 9 1 6 7 9 9 1 9 9 1 9 9 9 9 9 9 1 1 1 1 1 1 1 9 9
1 1 99 99 99 99 1 99 1 99 99 99 1 50 1 1 1 1 1 99
1 1 80 96 89 99 1 99 2 25 82 99 1 2 1 1 2 1 80 82
1 1 99 80 99 99 1 99 40 80 99 99 1 1 60 1 20 1 20 99
33 1 99 99 99 99 1 67 99 50 33 99 1 1 99 1 1 1 67 33
1 1 83 99 1 99 1 99 33 99 99 99 1 1 1 1 1 1 33 67
1 1 9 9 6 7 199 1996799999933 1 1 1 1 19999
33 1 99 33 1 99 1 99 67 1 99 99 33 1 1 67 1 1 67 99
1 199999999 199 1 19999 1 150 1 1 1 1 1
1 199999999 164 1 9 6 4 9 9 9 1 9 9 1 118 16418
50 1 63 13 63 99 13 99 25 1 88 99 1 1 63 1 1 1 88 1
99 1 99 60 99 99 1 99 80 1 99 99 1 1 20 1 1 1 80 99
1 1809950996099 1999999 1 1 1 1 199 1 1
1 1 1 1 1999999 1999999 1 1 1 1 199 1 1
8 1 17 67 1 99 99 92 92 1 99 92 1 1 1 1 8 1 33 33
1 1 1 57 14 99 99 99 99 29 29 99 1 1 57 1 1 1 43 99
1 1 1 88 13 99 99 99 99 38 88 99 1 1 63 1 1 1 63 88
1 1 1 1 199999999 1 1 1 1 150 1 1 150 1
25 1 50 50 1 99 99 99 25 75 99 99 1 1 1 1 1 1 1 99
99 1 99 1 33 99 99 1 67 1 99 99 1 1 99 1 1 1 33 99
8 1 99 1 85 99 99 23 39 1 99 92 1 1 99 1 1 1 8 92
1 150 15099 1 1 1 15099 1 199 1 1 15099
50 1 50 99 75 99 1 25 1 1 25 99 75 50 25 1 1 1 99 75
11 1 9 9 7 8 8 9 9 9 1 7 8 1 1 1 1 9 9 8 9 1 1 1 9 9 1 1 1
1 1 9 9 9 9 9 9 9 9 199 1 1 5 0 9 9 9 9 1 1 150 1 1 1
20 1 2 0 9 9 4 0 9 9 4 0 1 3 0 1 8 0 9 9 1 1 1 1 1 1 1 9 9
1 150 1 1 1 1 1 1 15099 199 1 150 15050
50 1 9 9 9 9 5 0 5 0 1 9 9 1 1 9 9 9 9 1 1 1 1 5 0 1 1 1
1
1
1
1
1
1
1
2
199 1
99 86 43
1 99 33
1 1 1
1 99 99
1 199
1 40 20
1 76 92
19999 1 1 15050
29 99 99 99 99 14 43 14
99 99 99 1 99 1 33 1
1 33 50 99 99 33 67 99
99 99 1 99 7 99 89 99
199 19999995099
1 99 40 20 1 20 40 99
76 99 4 98 34 68 72 99
1
1
1
1
1
1
1
1
1
1
1
1
1
1
80
1
199
99 1
99 1
1 99
4 1
1 1
1 40
1 1
1
14
1
1
1
1
1
6
1
1
1
1
4
1
1
1
199
43 43
67 99
83 67
26 82
199
80 80
68 82
which were either positive or negative in only a few taxa gave low values in the CHARSEP
program. However, these characters were particularly useful in separating certain taxa and so
were retained in the matrix. The CSP values in this matrix were generally lower than found in
bacteriological data (e.g. Sneath, 1979b; Priest & Alexander, 1988),where 0.5 was considered an
acceptable cutoff. However, excluding characters where the CSP value was less than 0.5
significantly decreased the number of successful identifications in this case.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
297 1
Pen icillium ident $cat ion
Table 2 (continued)
::
El
G
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
67 1 3 3 4 4 3 3 1 6 7 9 9 1 1 9 9 1 1 1 7 8 5 6 1 8 9 6 7 1 1 2 2
99 1 9 9 9 9 9 9 1 1 9 9 1 1 9 9 1 1 1 6 7 6 7 1 6 7 1 1 1
99 1 9 9 1 1 1 1 9 9 1 1 1 1 9 9 1 9 9 1 5 0 9 9 1 1 1
66 2 55 52 86 1 43 99 7 1 89 2 2 7 27 77 11 7 2 23 1
9940 1 1 4 0 2 0 8 0 8 0 1 1 9 9 1 1 1 8 0 6 0 1 4 0 2 0 2 0 2 0
99 1 3 3 1 1 1 9 9 1 1 1 9 9 1 1 1 1 9 9 1 1 1 1 1
679917 199 1 1 9 9 1 1 199 150 150 1 1 1 1 1
333399 199 1 199 1 199 1 1 9 9 1 1 1 1 3 3 6 7 6 7
67 1 9 9 1 9 9 1 1 9 9 1 1 9 9 1 1 1 9 9 1 1 1 6 7 1 1
99 1 5 0 1 9 9 1 1 9 9 9 9 1 1 1 1 1 1 9 9 1 1 1 9 9 9 9
99 1 9 1 9 6 4 1 1 9 9 1 9 1 9 1 1 1 8 1 8 2 9 1 8 7 3 5 4 9
1 188 1 1 1 9 9 9 9 1 3 188 1 1 2 5 1 3 7 5 1 3 1 1 1 1
606060 120 18099 1 199 1 1 1 199 1 1 1 1 1
1 1 1 1 9 9 1 1 9 9 9 9 1 1 1 1 15099 1 1 1 1 1
1 1 1 1 9 9 5 0 9 9 9 9 1 1 5 0 15050 150 1 1 9 9 1 1
75 1 1 1 9 2 1 1 7 9 9 1 7 1 7 5 8 1 8 1 1 9 9 1 8 1 1 1
57 29 1 14 29 14 86 99 14 1 14 1 71 99 1 1 14 14 14 14 1
88 ‘13 13 1 38 1 13 88 13 1 50 1 38 1 99 1 1 1 1 1 1
5099 1 5 0 1 1 9 9 9 9 1 1 1 1 9 9 1 9 9 1 1 1 1 1 1
99 1 5 0 1 2 5 1 1 9 9 1 1 1 1 9 9 1 5 0 5 0 1 1 2 5 1 1
67 1 6 7 1 1 6 7 9 9 9 9 9 9 1 1 1 1 3 3 1 6 7 1 1 9 9 1 1
9215 1 1 1 5 1 8 9 9 9 2 8 1 1 1 8 8 9 2 2 3 1 8 1 8
995050 19999 199 1 1 1 1 9 9 1 1 9 9 5 0 1 1 1 1
2525997550 12599 1 1 199 150 150 1 1 1 1 1
1 33 89 56 99 1 1 99 78 1 1 22 1 33 1 78 99 33 22 1 22
99 1 1 9 9 9 9 1 1 9 9 5 0 1 5 0 1 1 5 0 1 5 0 1 1 1 1 1
20 1 1 6 0 1 1 9 0 3 0 1 1 9 9 1 1 1 8 0 6 0 1 3 0 1 1 1 0
99 1 5 0 9 9 9 9 1 5 0 9 9 1 1 9 9 1 1 1 1 9 9 1 1 1 1 1
509999 150 1 1 9 9 1 1 9 9 1 1 199 1 1 1 1 1 1
30.
31.
32.
33.
34.
35.
36.
37.
99 1 1 5 0 1 1 1 9 9 1 1 1
9943 1 1 9 9 1 8 6 5 7 1 1 4 8 6
9933 1 1 1 9 9 9 9 1 1 1 3 3
17 1 1 1 1 1 8 3 1 7 1 1 8 3
96 1 6 3 1 9 9 1 5 9 9 9 9 9 1 1
5050 1 1 9 9 1 1 9 9 9 9 1 1
12080 180 1209999 1 1
661414 290 2409088 1 1
1 9 9 199 1995050 1
1 186 114 1 1 1 1
167 1 1 9 9 1 1 1 1
117 199 1 1 1 1 1
1 126 178 1 1 119
1 1 1 1 9 9 1 1 5 0 1
1 1 1 199 1 120 1
6 228 464 6 4 2 4
1
1
1
1
1
1
1
4
1
1
1
2
1
1
1
1
33
1
1
1
1
1
1
8
1
13
1
9 9
1
1
1
1
33
1
1
1
1
1
1
1
1
1
1
1
1
1 1 1336711
167 1 1 1 1
150 1 1 1 1
7 7 14 2 14 34
20 1 1 1 1 1
1 1 1 1 9 9 1
183 117 117
33 1 6 7 6 7 1 1
6733 1 1 3 3 1
5050 1 1 1 9 9
1 1 1 11882
25 1 1 1 1 3 1
14020 1 1 1
1 1 1 1 1 1
1 1 1 1 1 1
1 1 8 8 1 1
29 1 14 1 1 1
13 50 1 1 1 1
50 1 1 5 0 1 1
757550 150 1
1 1 1 1 1 1
87762 8 8 8
199 1 1 1 1
125 1 1 5 0 1
1 1 1 44 22 1
1 1 150 1 1
190 1 1 5 0 1
995099 1 9 9 1
99 1 1 1 1 1
1 1
1 1
1 1
33 1
1 7
199
2020
1 8
1 1 1 1
1 129 1
1 1 1 1
1 16717
4597082
1 1 1 1
1 14060
6 1 6 8
The HMO for each taxon identified to the appropriate taxon with Willcox scores in excess of
0.9999 (that is a less than 1 in 10000 chance that the HMO did not belong to that taxon, assuming
that it belongs exclusively to one taxon in the matrix). The only exception to this was for P .
solitum var. solitum, where the score for the HMO was 0.989 for that taxon and 0.0102 for P .
solitum var. crustosum. In practice, a number of strains gave identifications to the correct taxa
that were higher than those in MOSTTYP.This is most probably due to the normalization of the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
2972
P. D . B R I D G E A N D O T H E R S
Table 2 (continued)
Taxon
1. P. expansum (cluster 1)
2. P. expansum (cluster 2)
3. P. expansum (cluster 3)
4. P . aurantiogriseum (cluster 4)
5 . P . aurantiogriseum (cluster 6)
6. P . chrysogenum (cluster 7)
7. P. echinulatum
8. P. granulatum var. globosum
9. P. granulatum var. granulatum
10. P. hirsutum (cluster 12)
11. P. hordei (cluster 13)
12. P . chrysogenum (cluster 14)
13. P. glandicola var. mononematosa
14. P. camembertii (cluster 16)
15. Penicillium sp. (cluster 17)
16. P. verrucosum
17. P. atramentosum (cluster 19)
18. P . atramentosum (cluster 20)
19. P. atramentosum (cluster 21)
20. Cluster 22
21. P. citrinum
22. P. brevicompactum
23. P. raistrickii
24. P. aurantiogriseum var. neoechinulatum
25. P. viridicatum (cluster 27)
26. P. viridicatum (cluster 28)
27. P . griseofulvum
28. P. hirsutum var. allii
29. P . aurantiogriseum var.
melanoconidium
30. P. clavigerum
3 1. P. roquefortii
32. P. corylophilum
33. P. claviforme
34. P. solitum var. crustosum
35. P. olivinoviride
36. Coremial isolates
37. P. solitum var. solitum
22 1 67 1 99 1 44 78
99 1 1 1 99 33 1 99
1 1 9 9 199 1 1 5 0
9 3 18 3 85 36 5 22
20 1 60 1 60 1 20 60
1 33 33 1 1 33 67 33
17 1 1 1 99 1 1 99
199 1 199 1 1 1
1 67 33 1 33 67 1 33
1 1 119950 1 1
18 1 36 91 99 27 18 1
1 13 25 1 1 43 13 1
20 1 1 1 80 80 1 20
1 50 1 50 25 25 25 67
199 199 1 199 1
1 99 1 75 83 33 1 17
1 43 1 1 86 28 28 14
38 75 1 1 99 63 1 75
150 1 15050 1 1
75 50 1 25 99 99 1 99
199 1 1 1 1 1 1
170 1 1 161 116
1 1 1 1 150 1 1
1 1 1 50 99 25 25 50
1 1 22 1 99 67 1 22
1 1 1 199 1 1 1
70 30 1 1 93 10 60 10
1 50 1 50 99 99 1 50
1 50 1 1 99 99 1 99
50 1
1 1
133
67 17
30 1
50 1
20 20
2 4
1
71
1
1
67
1
20
28
1
1
1
1
1
1
20
12
1
1
1
1
99
50
60
85
11
1
1
1
20
1
99
1
1
1
1
1
1
67
1
1
1
1
1
1
1
1
1
50
1
1
1
1
1
50 50 50 1
1 71 86 71
16733 1
1 99 99 1
41 11 44 27
1 50 99 50
80 20 80 20
12 14 78 46
probabilities involved in calculating the Willcox score, so that although such a strain is less
typical than the HMO, it is even less likely to belong to any other group.
The lack of overlap and the good identification scores were unexpected considering the low
values of sums of scores and CSP indices. One possible explanation is that many of the tests
included in the matrix separated only a few different taxa. This is supported by the results from
the anticlustering which suggested that all the tests were very different, most being recovered in
one very large loosely formed cluster.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
Penicillium identijicat ion
2973
Of the 52 previously studied strains, 50 identified to their correct groups with Willcox
probability scores of better than 0-9,with between 0 and 7 atypical test results (47 scoring above
0.99 and 43 scoring above 0-999)when using original data from the numerical taxonomy. MLF
scores were in general low, only 11 strains giving values greater than 0.1. Despite the low scores
all strains gave a scorelarger than
Of the two strains that did not identify correctly at the
0.9 probability level, one identified to the correct group with a score of 0.718, although the MLF
was relatively high at
and one was a representative of P . solitum var. crustosum which
identified at 0.98 (MLF
to P . solitum.
Of the 5 1 additional fasciculate penicillia identified, 45 with Willcox probabilities of better
than 0.9 (36 scoring above 0.99). Of the additional authentic ex-type cultures, five identified to
their respective taxa (P. atramentosum, P . expansum, P . griseofulvum, P . vetrucosum and P .
camembertii). The ex-type culture of P . psittacinum identified as P . soliturn; this was a different
line of the same strain that was found loosely linked to this taxon in the numerical classification.
The ex-type culture of P . commune identified as P . aurantiogriseum. Pitt (1980) considered P .
commune a synonym of P. puberulum, which in turn has been considered a synonym of P .
uurantiogriseum (Cruickshank & Pitt, 1987). However, more recent work has suggested that P .
commune is a distinct taxon (Cruickshank & Pitt, 1987; Samson & van Reenen-Hoekstra, 1988).
However, in the numerical classificationof Bridge et al. (1989), strains of P . commune clustered in
either P . aurantiogriseum or P . solitum. Again, MLF scores were low, 40 of the 45 strains giving
values of greater than
Of the six remaining strains, five gave MLF scores of greater than
to the correct taxa, although the Willcox scores were between 0.555 and 0.818. One strain
remained unidentified to any group, the best scores being 0.617 and 1 x 10-lo to P . roquefortii.
This strain had originally been identified as P . mali, although it showed SEM conidial features
similar to P . roquefortii.
While the matrix was in most cases tested with at least 52 of the 57 test results, identifications
based on a smaller number of tests were also assessed. This involved considering sets of tests
from particular techniques as missing in the unknowns, such as omitting results from SEM, or
secondary metabolite tests, giving character sets varying from the original 57 tests to a minimum
of 16 tests. The omission of characters relating to conidial colour and secondary metabolite
production did not significantly lower identification scores. Detailed results from this procedure
are given by Bridge (1989).
Based on the results for the known strains, successful identification was determined as a
or greater. Situations where only one of the
Willcox score of 0.9 or greater and a MLF of
above criteria was met were less clear and the number and type of test results against
identification needed to be considered. The Willcox scores obtained in this investigation were of
a similar magnitude to those found in studies with bacteria (e.g. Williams et al., 1983). The MLF
scores were low compared to those given in Dybowski & Franklin (1968), where scores of
between
and 1 were obtained with a small set of tests. In this case, where there were a large
number of tests, absolute probabilities were very small, and so MLF values could be expected to
be a little reduced. MLF values were also depressed by the relatively large proportion of results
in the matrix that were between 10 and 90% positive. In practice, the negative power of 10 of the
MLF score was often within one digit of the number of tests against identification, when there
were four or fewer tests against. While the tests in the identification matrix were selected on
their diagnostic and practical capabilities from the previous study, they all contained a
significant number of variable results and these would serve to depress the absolute probability.
Identijication keys
The three synoptic keys constructed with all the characters in the matrix performed quite
differently from each other. When 1 and 99% were used as the cutoff values for minus and plus,
strains were usually not identified or else were identified to two or three groups that could not be
separated. When 20 and 80% were used, strains were often misidentified, often after only a small
number of tests. When cutoffs of 10 and 90% were used, 50 of the 52 previously studied strains
were correctly identified. With these criteria 44of the 51 additional strains were also successfully
identified. Strains which belonged to either the P . solitum var. solitum or the P . solitum var.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
2974
P . D . BRIDGE A N D OTHERS
Table 3 . Synoptic key based on 2.5 characters giving highest CSP scores from identiJication
matrix (+ue, > 10%; - ue, < 90%)
Numbers for taxa are as in Table 2. Further characteristics may be found in Table 2 and in the
preceding paper (Bridge et al., 1989).
Taxa
Character
Morphological
Rough stipes
Pink diffusing pigment
Yellow colony reverse
Yellow mycelium
+
Production of coremia
Velvet colony texture
Fasciculate colonies
Blue-green conidia
Grey-green
First-level branches divergent
Ellipitical conidia
Subgloese conidia
Conidia length 5 2.8 pm
Colony diameter 5 20 mm
Physiological and biochemical
Growth on N O 2 agar
Growth on creatin
+ve 1,2,4,5,7,8,9,10,11,13,14,15,16,17,18,20,22,23,24,25,26,28,29,31,
34,35,36,37
-ve 1,3,4,5,6,11,12,13,17,18,19,20,21,22,24,27,29,30,32,33,36
+ve 1,2,4,17,19,24,25,26,27,28,30
-ve 1,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,
27,29,30,3 1,32,33,34,35,36,37
+ve 1,2,3,4,6,7,8,9,10,11,12,13,18,20,21,23,24,25,28,29,34,36,37
-ve 1,4,5,6,7,10,12,13,14,15,16,17,18,19,20,21,22,23,25,26,27,28,30,
31,32,33,34,35,36,37
+ve 3,11,24,28,36
-ve 1,2,3,4,5,6,7,8,9,10,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,
27,29,30,31,32,33,34,35,36,37
ve 9,30,33,36
-ve 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,
26,27,28,29,31,32,34,35,36,37
+ve 5,6,10,12,13,17,18,19,21,22,23,24,31,32
-ve 1,2,3,4,5,7,8,9,10,11,12,13,14,15,16,17,18,19,20,24,25,26,27,28,
29,30,33,34,35,36,37
+ve 1,2,3,4,5,7,8,9,10,11,13,14,16,17,18,19,20,24,25,26,27,28,29,34,
35,36,37
-ve 4,5,6,9,12,13,14,15,16,17,19,21,22,23,30,31,32,33,35,36,37
+ve 1,4,5,6,7,8,9,11,12,13,16,17,18,19,21,23,24,28,31,32,33,34,36,37
-ve 1,2,3,4,5,6,7,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,25,26,
27,28,29,30,3 1,32,33,34,35,36,37
+ve 1,2,3,4,5,6,7,8,9,11,13,16,17,18,20,21,22,23,24,27,28,30,31,32,33,
34,35,36,37
-ve 4,6,7,10,11,12,14,15,16,18,19,24,25,26,28,29,31,33,34,36,37
+ ve 1,3,4,5,12,13,15,16,17,18,19,21,24,27,28,3 1,32,33,34,36,37
-ve 1,2,3,4,5,7,8,9,10,11,13,14,16,17,18,20,22,23,24,25,26,28,29,30,
31,33,34,35,36,37
+ ve 1,2,3,4,5,9,12,14,18,19,20,27,29,30,33
-ve 1,2,4,5,6,7,8,10,11,12,13,14,15,16,17,20,21,22,23,24,25,26,27,28,
31,32,34,35,36,37
+ve 1,2,4,5,6,7,10,11,12,13,14,15,16,20,21,22,23,24,25,26,27,28,31,32,
34,3536.37
- ve 1 ~2,3,4,5,7,8,9,11,12,15,17,18,19,20,21,24,25,26,27,29,30,3 1,33,
34,37
+ve 1,2,3,4,5,6,7,9,13,14,16,17,18,20,22,24,25,28,29,30,31,32,33,34,
35,36,37
-ve 1,3,4,5,6,8,9,10,11,12,13,14,15,16,17,18,19,21,22,23,24,25,26,27,
28,30,3 1,32,34,36,37
+ve 6,8,9,12,14,15,16,17,18,19,20,21,22,27,28,29,32,33,36
- ve 1,2,3,4,5,6,7,9,10,11,12,13,14,17,18,19,20,22,23,24,25,26,27,28,
29,30,31,32,33,34,35,36,37
+ve 12,14,15,16,17,18,19,20,21,22,27,30,31,32,33,36
-ve 1,2,3,4,5,6,7,8,9,10,11,12,13,14,23,24,25,26,27,28,29,33,34,35,36,
37
+ve 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,22,24,25,26,29,
31,33,34,35,36,37
- ve 1,6,11,21,22,23,24,25,27,28,30,32,36
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
Penicillium identiJication
Character
Physiological and biochemical
(continued)
Base from creatin
Table 3 (continued)
2975
Taxa
+
ve 1,2,3,4,5,6,7,8,14,15,17,18,20,25,3 1,33,34,35,36,37
-ve 1,4,5,6,9,10,11,12,13,16,17,18,19,20,21,22,23,24,25,26,27,28,29,
30,31,32,33,36,37
Base from lactic acid
+ve 3,4,5,6,7,8,9,10,11,12,13,14,16,17,18,20,24,25,26,27,29,31,32,34,
35,36,37
- ve 1,2,5,8,9,12,13,15,16,17,18,19,20,21,22,23,25,28,30,3 1,32,33,36
Base from citric acid
ve 1,2,3,4,5,6,10,11,12,13,14,17,18,21,22,23,24,25,26,27,29,31,32,34,
37
-ve 1,2,4,7,8,9,12,14,15,16,17,18,19,20,21,22,23,24,25,27,28,29,30,31,
33,35,36,37
Base from urea
+ve 5,6,7,8,9,12,13,16,17,18,19,20,21,22,27,31,32,33,35,37
-ve 1,2,3,4,5,7,8,9,1O,1l,l2,13,14,15,2O,2ly22,23,24,25,26y27,28,29,
30,34,36,37
+ve 1,2,3,4,5,6,7,8,9,10,11,12,13,14,16,20,21,22,23,24,25,26,27,28,29,
Growth on tannic acid
30,31,32,34,36,37
-ve 4,7,12,14,15,16,17,18,19,20,23,24,27,28,31,33,35,36,37
Griseofulvin production
+ve 2,3,7,9,10,13,18,20,22,23,24,27,28,35,36
-ve 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,24,25,26,
28,29,30,31,32,33,34,36,37
Additional characters where electron microscopy is necessary
Lobate conidia
+ve 10,12,14,16,17,18,21,22,25,26,34,35,36,37
-ve 1,2,3,4,5,6,7,8,9,11,12,13,15,16,17,18,19,20,23,24,25,26,27,28,29,
30,31,32,33,37
Microtuberculate conidia
+ve 1,2,4,5,6,8,9,12,13,15,16,17,18,26,27,28,29,31,32,33
-ve 3,4,7,10,11,12,14,15,16,17,18,19,20,21,22,23,24,25,26,30,31,32,
33,34,35,36,31
Microverrucate conidia
ve 3,15,17,18,19,20,23,30,32,33
-ve 1,2,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,21,22,24,25,26,27,28,
29,31,32,33,34,35,36,37
+
+
crustosum groups could not be separated by the key. The key performed better with some
characters than with others, and the use of secondary metabolite and conidial colour characters
often resulted in misidentifications. As an example, the synoptic key containing the 25 tests that
gave the highest CSP values (> 0.46) is given in Table 3.
Comparison of the performance of the identification matrix with the synoptic key was
difficult, as results from the key depended to an extent upon which characters were used. In
general, characters were used in the key in the order gross morphology, micromorphology,
physiology, SEM and secondary metabolites, strains being identified after seven to fourteen
characters. Identification by key, however, is very sensitive to minor test discrepancies. One
example of this was seen in testing the matrix and the key, where one strain identified as a P .
echinulatum with the matrix, the only discrepancy being the production of lighter coloured
conidia than usual. This strain would misidentify in a key if conidial colour was one of the
characters selected. The subsequent omission of conidial colours and secondary metabolite
characters did not have a significant effect on the performance of the matrix, and in most cases
improved the performance of the key. There are, however, exceptions; strains of P . camembertii
(cluster 16) cannot always be separated from P . solitum without the white conidia character in
the key but this did not prevent identification by the matrix. In the numerical classification
(Bridge et a f . ,1989), a white-spored mutant of P . solitum clustered as P . camembertii (cluster 16),
and so this result was perhaps not surprising. Characters based on conidial colour can be
subjective (Onions et al., 1984) and this may explain the poor performance of these characters in
these identification schemes. Problems associated with secondary metabolite characters are less
clear. Some authors have found these to provide reliable identification criteria (e.g. Frisvad,
1981), although this has not always been the experience of other workers (Land & Hult, 1987;
Bridge et al., 1989). Although more genes can be involved in secondary metabolism than
primary metabolism (Bu’hck, 1980), they are under less-precise genetic control (Bennett,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
2976
P. D. BRIDGE A N D OTHERS
1985). In some parts of the genus they may be useful, e.g. penitrem A production in P. solitum
var. crustosum. In the numerical classification, secondary metabolite characters showed poor
reproducibility (Bridge et al., 1989) and it seems likely that further work is needed on both their
induction and detection before they can be routinely used as reliable identification criteria ;
certainly while the presence of particular metabolites may be useful characters, little can be
inferred from their absence. An example of this is the so-called chemosyndromic variation seen
in lichens, where presumed qualitative differences were subsequently found to be quantitative
(Culberson & Culberson, 1977).
Conclusions
Numerical classification (Bridge et al., 1989) and other studies on serology (Polonelli et al.,
1986), protoplast fusion (Ann6 & Peberdy, 1981), strain variation (Bridge et al., 1986, 1987),
isozyme studies (following paper: Paterson et al., 1989) and DNA melting curves (Paterson et
al., 1990) have shown that the terverticillate penicillia consist of a large number of closely
related and sometimes variable species. This study has shown that these organisms can be
identified objectively to discrete groups with an acceptable degree of confidence, although some
of these groups may be part of a larger continuum (Bridge et al., 1989). However, test selection
can also be an important criterion, and those characters that were shown to give good
reproducibility, such as conidial shape and ornamentation (see Bridge et al., 1989) should be
favoured if synoptic or dichotomous identification keys are to be adopted. With the data
available a considerable amount of time and specialist equipment, such as a scanning electron
microscope for conidial ornamentation, would be needed to use all the characters for an
unknown. While this would not be acceptable for routine identification purposes, it is feasible
for particularly difficult strains at reference laboratories. However, initial testing of this matrix
for unknowns where only some characters are known suggests that acceptable results can often
be obtained where traditional morphological characters are used in conjunction with results
from one or more additional character sets (Bridge, 1989).
This study and the associated numerical classification (Bridge et al., 1989) have mainly
confirmed the validity of the revised species concepts adopted by Pitt (1980, 1986) by
considering additional types of characters. These concepts can now be adopted with confidence
on the basis of the larger data sets considered here. Diagnostic morphological features can be
reinforced with physiological and SEM observations, and these approaches can prove helpful
where morphological criteria are ambiguous or may be subject to differences in interpretation
between workers. Characters such as conidial shape and ornamentation and growth on nitrite
and creatine agars can be particularly useful.
We would like to thank Elizabeth Oliver, Penny Fame11 and Devota Kavishe for their excellent technical
support, Professor P. H. A. Sneath for his many helpful suggestions and discussions and Professor E. A. Bell and
Dr L. E. Fellows of the Royal Botanic Gardens, Kew for providing biochemical facilities. This work was
supported by the Science and Engineering Research Council contract Sol17/84, Systematics of Microfungi of
Biotechnological and Industrial Importance.
REFERENCES
ABE, S. (1956). Studies on the classification of the
Penicillia. Journal of General and Applied Microbiology 2, 1-344.
ANNE, J. & PEBERDY,
J. F. (1981). Characterisation
of inter-species hybrids of Penicillium chrysogenum
and Penicillium roquefortii by isoenzyme analysis.
Transactions of the British Mycological Society 77,
401-408.
BAINIER,
G. (1905). Mycothbque de l’icole de pharmacie. Sur deux Penicillium. Bulletin trimestriel de la
Socihth mycologique de France 21, 126-130.
BAINIER,
G. (1906). Mycothbque de l’icole de pharmacie. V.Bulletin trimestriel de la SociPth mycologique de
France 22, 205-208.
BAINIER,
G . (1907). Mycothbque de l’kcole de pharmacie. IX. Bulletin trimestriel de la Socihth mycologique
de France 23, 11-22.
BAINIER,G. & SARTORY,
A. (1912). Etude d’un
Penicillium nouveau, P . olsonii (sp. nov.). Annales
mycologici 10, 398-399.
BENNETT,
J. W. (1985). Molds, manufacturing and
molecular genetics. In Molecular Genetics of Filamentous Fungi, pp. 345-367. Edited by W. E. Timberlake. New York: Alan R. Liss.
BIOURGE,
P. (1923). Les moisissures du groupe Penicillium Link. Cellule 33, 7-331.
BOAS,F. (1914). Uber ein neues Coremium-bildendes
Penicillium. Mykologisches Zentralblatt 5, 73-83.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
Pen icillium ident$cat ion
2977
L. & THOM,C. (1928). A rot of Gladiolus
BRIDGE,P. D. (1989). Identification of terverticillate MCCULLOCH,
corms caused by a Penicillium. Science 67, 217penicillia from a matrix of percent positive test
218.
results. In Modern Concepts in Penicillium and
NOVOBRANOVA,
T. I. (1972). Species composition of
Aspergillus Classification. Edited by R. A. Samson &
grape mycoflora under storage conditions in the
J. I. Pitt. New York: Plenum Press.
BRIDGE,P. D., HAWKSWORTH,
D. L., KOZAKIEWICZ, Alma-Ata region. Nauchnye doklady vysshei shkoly
biologicheskie nauki 10, 103-108 (in Russian).
Z., ONIONS,A. H. S. & PATERSON,
R. R. M. (1986).
T. I. (1974). Species novae penicillii in
Morphological and biochemical variation in single NOVOBRANOVA,
fructibus mali et vitis inventae e regione Almaisolates of Penicillium. Transactions of the British
Ataense. Novitates systematicae plantarum non vascuMycological Society 87, 389-396.
larium 11, 225-234.
BRIDGE,P. D., HUDSON,L., KOZAKIEWICZ,
Z.,
D. L.
ONIONS,A. H. S. & PATERSON,
R. R. M. (1987). OLIVER,E. M., BRIDGE,P. D. & HAWKSWORTH,
Investigation of variation in phenotype and DNA
(1987). The effect of selected mountants and stains
on the measurement of fungal spores. Mycopathocontent between single-conidium isolates of single
logia 97, 165-172.
Penicillium strains. Journal of General Microbiology
ONIONS,
A. H. S., BRIDGE,P. D. & PATERSON,
R. R. M.
133, 995-1004.
BRIDGE,P. D., HAWKSWORTH,
D. L., KOZAKIEWICZ, (1984). Problems and prospects for the taxonomy of
Penicillium. Microbiological Sciences 1, 185-1 89.
Z., ONIONS,
A. H. S., PATERSON,
R. R. M., SACKIN,
M. J.
R. R. M., BRIDGE,P.D., CROSSWAITE,
M. J. & SNEATH,
P. H. A. (1989). A reappraisal of PATERSON,
& HAWKSWORTH,
D. L. (1989). A reappraisal of
terverticillate penicillia using biochemical, physioloterverticillate penicillia using biochemical, physiological and morphological features. I. Numerical
taxonomy. Journal of General Microbiology 135,
gical and morphological features. 111. An evaluation
of pectinase and amylase isenozymes for species
2941-2966.
BU’LOCK,J. D. (1980). Mycotoxins as secondary
characterization. Journal of General Microbiology
metabolites. In The Biosynthesis of Mycotoxins. A
135, 2979-299 1.
Study of Secondary Metabolism, pp. 1-16. Edited by
PATERSON,
R. R. M., KING,G. & BRIDGE,P. D. (1990).
P. S. Steyn. London: Academic Press.
High resolution melting curves on DNA from 14
CRUICKSHANK,
R. H. & PITT,J. I. (1987). Identification
Penicillium strains. Mycological Research (in the
of species in Penicillium subgenus Penicillium by
Press).
PITT,J. I. (1980) (dated 1979).* The Genus Penicillium
enzyme electrophoresis. Mycologia 79, 6 14-620.
and its Teleomorphic States Eupenicillium and TalaroCULBERSON,
C. F. & CULBERSON,
W. L. (1977).
myces. London : Academic Press.
Chemosyndromic variation in lichens. Systematic
PITT, J. I. (1986). A Laboratory Guide to Common
Botany 1, 325-339.
Penicillium Species, 2nd edn. North Ryde, Australia :
DATTILO-RUBBO,
S. (1938). The taxonomy of fungi of
CSIRO.
blue-veined cheese. Transactionsof the British MycoPOLONELLI,L., CASTAGNOLA,
M., D ’ U w , C. &
logical Society 22, 174- 18 1.
MORACE,G. (1986) (dated 1985).t Serological apDEMELIUS,
P. (1923). Beitrag zur Kenntnis der Hyphoproaches for identification of Aspergillus and Penicilmyceten Niederosterreichs. Verhandlungen der Zoolium species. In Advances in Penicillium and Aspergillogisch-botanischen Gesellschaft in Wien 72, 74.
DIERCKX,
R. P. (1901). Un essai de revision du genre
lus Systematics, pp. 267-280. Edited by R. A.
Samson & J. I. Pitt. New York: Plenum Press.
Penicillium Link. Annales de la Sociktk Scient$que de
Bruxelles 25, 83-88.
B. (1988). A frequency
PRIEST,F. G. & ALEXANDER,
DYBOWSKI,
W. & FRANKLIN,
matrix for probabilistic identification of some
D. A. (1968). Conditional
bacilli. Journal of General Microbiology 134, 301 1probability and the identification of bacteria: a pilot
3018.
study. Journal of General Microbiology 54, 215-229.
FASSATIOVA,
0. (1977). A taxonomic study of Penicil- RAPER,K. B. & THOM,C. (1949). A Manual of the
Penicillia. Baltimore : Williams & Wilkins.
lium series Expansa Thom emend. Fassatiova. Acta
SAMSON,R. A., STOLK,A. C. & HADLOK,R. (1976).
Universitatis Carolinae-Biologica 1974, 283-335.
Revision of the subsection fasciculata of Penicillium
FELTHAM,
R. K. A. & SNEATH,P. H. A. (1982).
and some allied species. Studies in Mycology Baarn
Construction of matrices for computer-assisted iden11, 1-47.
tification of aerobic Gram-positive cocci. Journal of
SAMSON,
R. A. & VAN REENEN-HOEKSTRA,
E. S. (1988).
General Microbiology 128. 713-720.
Introduction to Food-borne Fungi. Baarn : CentraalFRISVAD,J. C. (198 1). Physiological criteria and
bureau voor Schimmelcultures.
mycotoxin production as aids in identification of
K. A. & SAMSON,R. A. (1986) (dated 1985).t
common asymmetric penicillia. Applied and Environ- SEIFERT,
The genus Coremiumand the synnematous penicillia.
mental Microbiology 41, 568-579.
In Advances in Penicillium and Aspergillus SystemFRISVAD,J. C., FILTENBORG,0. & WICKLOW,
atics, pp. 143-161. Edited by R. A. Samson & J. I.
D. T. (1987). Terverticillate penicillia isolated from
Pitt. New York: Plenum Press.
underground seed caches and cheek pouches of
banner-tailed kangaroo rats (Dipodomys spectabilis). SMITH,G. (1933). Some new species of Penicillium.
Transactionsof the British Mycological Society 18,88Canadian Journal of Botany 65, 765-773.
91.
LAND,C. J. & HULT,K. (1987). Mycotoxin production
by some wood-associated Penicillium spp. LRtters in SMITH,G. (1939). Some new species of mould fungi.
Transactions of the British Mycological Society 22,
Applied Microbiology 4, 4 1-44.
252-256.
LINK, H. F. (1809). Observationes in ordines plantarum naturales. Magazin Gesellschaft Naturfors- SMITH,G . (1963). Some new species of Penicillium and
some observations on the taxonomy of the genus.
chender Freunde zu Berlin 3, 1-42.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37
2978
P . D . B R I D G E AND O T H E R S
Transactions of the British Mycological Society 46,
331-337.
SNEATH,P. H. A. (1979a). BASIC program for the
identification of an unknown with presence-absence
data against an identification matrix of percent
positive characters. Computers and Geosciences 5,
195-2 13.
SNEATH,
P. H. A. (1979b). BASIC program for character
separation indices from an identification matrix of
percent positive characters. Computers and Geosciences 5, 349-357.
SNEATH,
P. H. A. (1980a). BASIC program for the most
diagnostic properties of groups from an identification matrix of percent positive characters. Computers
and Geosciences 6, 21-26.
SNEATH,
P. H. A. (1980b). BASIC program for determining the best identification scores possible for the
most typical example when compared with an
identification matrix of percent positive characters.
Computers and Geosciences 6, 27-34.
SNEATH,
P. H. A. (1980~).BASIC program for determining overlap between groups in an identification
matrix of percent positive characters. Computers and
Geosciences 6, 267-278.
STOLK,A. C. (1969). Four new species of Penicillium.
Antonie van Leeuwenhoek 35, 261-274.
THOM,C. (1906). Fungi in cheese ripening : Camembert
and roquefort. Bulletin of the Bureau of Animal
Industry (US Department of Agriculture) 82, 1-39.
THOM, C. (1910). Cultural studies of species of
Penicillium. Bulletin of the Bureau of Animal Industry
(US Department of Agriculture) 118, 5-109.
THOM,C. (1930). The Penicillia. London: Baillitre,
Tindall & Cox.
WESTLING,R. (1911). Uber die griinen Spezies der
Gattung Penicillium. Archiv fur Botanik 11, 1-156.
WILLCOX,W. R., LAPAGE,S. P., BASCOMB,
S. &
CURTIS,M. A. (1973). Identification of bacteria by
computer: theory and programming. Journal of
General Microbiology 77, 3 17-370.
WILLCOX,
W. R., LAPAGE,
S. P. & HOLMES,
B.(1980). A
review of numerical methods in bacterial identification. Antonie van Leeuwenhoek 46, 233-299.
WILLIAMS,
S. T., GOODFELLOW,
M., WELLINGTON,
E. M. H., VICKERS, J. C., ALDERSON,
G., SNEATH,
P. H. A., SACKIN,M. J. & MORTIMER,A. M.
(1983). A probability matrix for identification of
some streptomycetes. Journal of General Microbiology 129, 1815-1830.
ZALESKI,A. (1927). Uber die in Polen gefundenen
Arten der Gruppe Penicillium Link I, I1 and 111.
Bulletin international de I'Acadimie polonaise des
sciences, series B 41 7-563.
* This book, although dated
1979 was not available until 1980.
This book, although dated 1985 was not available until 1986.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 15 Jun 2017 01:22:37