0095-1137/89/061201-05$02.00/0
Copyright 1989, American Society forMicrobiology
Evaluation of the Titertek-NF System for Identification
of
Gram-Negative Nonfermentative and Oxidase-Positive
Fermentative Bacteria
PETER KAMPFER* AND WOLFGANG DOTT
FachgebietHygiene, Technische Univ'ersitatBerlin, Ainrurnerstrasse 32, D-1000Berlin 65, FederalRepublic of Germrany Received 28 July1988/Accepted 22 February 1989
TheTitertek-NF (TT-NF)system (Flow LaboratoriesGmbH, Meckenheim, Federal Republic of Germany)
was evaluated for the identification of 1,289 strains of gram-negative, nonfermentative bacteria and some
gram-negative, oxidase-positive bacteria. The oxidasetestwasalsoperformed. Identificationswereclassifiedas
correct, notidentified (twoor moretaxa possible and identification scoreof<80%; supplementary testsfor furthering the identication were not performed), and incorrect. Correct identification results were further
subdivided by thecorrectlevelof speciesorbiotype identificationas -98% (categorya),90to97% (category b), and 80to89% (categoryc). When compared withconventional identification results,the TT-NF system
correctly identified 90.3% ofstrains(1,164 of 1,289strains), with 72.5% (935 strains) belongingtocategorya,
14.7% of strains (189 strains) belongingto categoryb, and3.1% of strains (40 strains) belongingto category c. Among the remaining strains, 104 (8.1%) were notidentified and 14 (1.1%) were misidentified, and the system failed to generateidentificationresultsfor 7 strains (0.5%). Reactions within the TT-NFsystemwere
reproducible, withanestimated probability oferroneoustest results of 0.2%. Although gram-negative, nonfermentative bacteria
in-creasingly are becoming implicated in humandisease, iden-tification of these organisms has often been neglected in clinicalmicrobiology laboratories, especiallywhen antibiotic susceptibility data are available (2). A battery of
conven-tional biochemical tests is needed for their precise identifi-cation of the taxonomically heterogeneous group of
gram-negative, nonfermentative bacteria (13-15, 27). In recent
years, several commercial kits have been developed, some
of which areforthe identification ofmembersofthefamily
Enterobacteriaceae and commonly encountered nonfer-mentersandothers of whicharefor the specific identification
ofgram-negative nonfermentative bacteria (8).
Duringa3-year study,ascreeningwasperformedinorder to find miniaturized biochemical tests that would make possible an extensive species differentiation within these
bacterial genera (20, 21). As a result, in cooperation with
Flow Laboratories GmbH (Meckenheim, Federal Republic ofGermany), the Titertek-NF (TT-NF) system was
devel-oped; itprovides for thecomputer-aided identification of 57 different species. This study was undertaken to determine the accuracy, suitability, and reproducibility ofthe TT-NF systemforuse in clinical microbiology laboratories.
MATERIALS AND METHODS
Bacteria. A total of 1,289 strains of nonfermentative, gram-negative bacteria and fermentative, oxidase-positive organismsweretested. Of these,621wereobtained from the
National Collection ofType Cultures, Central PublicHealth Laboratory (Colindale, London, United Kingdom) (123
ref-erence strainsand 498 field strains). Theywereidentifiedby
the methods of the National Collection of Type Cultures (18). An additional 668strains were stock cultures obtained
from the Institute of Medical Microbiology, University of Heidelberg, Heidelberg, FederalRepublicofGermany; Cen-tre Hospitalier St. Nicolas, Sarrebourg, France; and the
*Correspondingauthor.
Culture Collection of the University Gôteborg, Gôteborg, Sweden, and were identified by conventional methods (13) or the API 20NE system (Analytab Products, Plainview, N.Y.) (12). Before testing,all isolates weregrown on sheep blood agar (Oxoid, Ltd., Hampshire, United Kingdom) for
24 hat30°C (strains ofthegeneraMoraxellaandPasteurella
and Bordetella bronchiseptica isolates weregrown for48 h at 30°C). Oxidase testing was carried out by using oxidase test strips (E. Merck AG, Darmstadt, Federal Republic of Germany).Toassesspurity,cultureswerechecked through-outthe study by colony morphology.
TT-NF system. The TT-NF system consists of 32 dehy-drated substrates in a standard microplate format (28 test and 4 control media). The system includes tests for indole production; esculin hydrolysis;urease;arginine dihydrolase; ornithine decarboxylase; fermentation of glucose and
su-crose; assimilation of D-glucose, mannose, maltose,
N-acetyl-D-glucosamine, mannitol, gluconate,
P-hydroxybu-tyrate, DL-lactate, adipate, suberate, malate, phenylacetate, and histidine; and hydrolysis tests for o-nitrophenyl-p-D-galactopyranoside, p-nitrophenylphosphorylcholine, 2-de-oxythymidine-5'-p-nitrophenylphosphate, L-proline-p-nitro-anilide, p-nitrophenyl-a-D-maltoside, p-nitrophenyl-N-acetyl-,-D-glucosaminide, and bis-p-nitrophenylphosphate. Each microplate contains three test sets for the identification of three different isolates. Inoculation of the test kit was asfollows. Each well was filled with 100 ,ul of a bacterial
suspensioninadefined minimalbasal medium withadensity
equivalent to a 0.1 McFarland standard. Test wells and
control wells ofesculin,esculincontrol, arginine dihydrolase, ornithinedecarboxylase, urease,glucoseandsucrose fermen-tation, and the fermentation controlwere overlaid with 50,ul
of sterile mineral oil. After inoculation, the microplate was
sealed withasterileplasticsealer.Followingincubationfor 24 hat30°C,50 plof thedimethylaminocinnamaldehyde reagent was added to the indole test wells, and then the test plates
were read with a photometer (Multiscan MCC 340; Flow Laboratories) connected to a personal computer
(Desk-1201
on April 11, 2020 by guest
http://jcm.asm.org/
pro286; Compaq, Munich, Federal Republic of Germany). Different tests were read at different wavelengths, and the resulting adsorbance values werecoded inbinary data. The
resultingtestprofileswere compared with aprobability
ma-trix, accordingtothe methodsofnumerical identification (39). This probability matrixwasconstructed by themethods given
in the literature (17, 24), but with upperand lower limits of
0.999 and 0.001, respectively. The evaluation results were
classifiedby the differentrangesof likelihoods(identification
scores), which are given as percentages throughout this
report.
The reproducibility of the TT-NFsystemwasdetermined
by testing 50 randomly chosen isolates on 3 differentdays.
Theresultingtestprofileswerecompared, andanestimate of
the probability ofan erroneoustestresultwascalculated by
methods described elsewhere (19).
RESULTS
The results of the evaluation of the TT-NF system are
summarized in Table 1. Of 1,289 strains tested, 935 (72.5%)
were correctly identified with an identification score of
-98%. For 189 strains (14.2%), the correct identification
scores werebetween 90 and 97%, while for 40 strains (3.1%),
thecorrectidentification scores were between 80 and 89%.
Foratotal of 104 strains (8.1%), the identification score was
below 80%, and hence, these strains were not identified correctly. To further the identification of 79 strains, oneto
fouradditionaltests were necessary; these appeared in the
formofpercentpositivetestresults onthecomputerscreen
incasesof low discrimination (additionaltestsrecommended by Flow Laboratories for the TT-NF system were nitrogen
formationfrom nitrate, nitrite formation from nitrate, motil-ity, growth on salmonella-shigella agar,growth in the pres-enceofcetrimide, growthon MacConkey agar, andgrowth
at42°C).
Incorrect identifications were obtained for 14 strains
(1.1%) (Table 2). Forseven strains (0.5%) no identification
could be obtained. Investigations on the reproducibility of
the TT-NF systemprovided the following results. A total of 50 strains were chosen at random, and 28 tests were
per-formed oneach strain, in triplicate, foratotal of4,200test
results. Of these 4,200 test results, there were 7 errors (a single negative result in one of the triplicate tests was
classified as an error). The corrected error rate was
calcu-latedasfollows:
C = 1/2(1- V/1-T4E/N)
where E is the number of errors observed, and N is the
number oftest results, as described previously (19). The
correctederrorratefor theTT-NF systemwascalculatedto
be0.2%, witha95% confidence interval of 0.7to 1.4%. The reproducibility of individualtests is summarized inTable3.
DISCUSSION
The TT-NF system was easy to set up and posed no
problems of interpretation. Indole tests were sometimes
difficulttoreadphotometrically, because somestrains of the
Alcaligenes faecalis-A. odoransgroupand alsosome
Pseul-domonas acidovorans strains produced a cherry red color
after the wells werefilled with the indolereagent (4-dimeth-ylaminocinnamaldehyde). This caused higher adsorbance valuesand, hence, false-positive results. These nonspecific reactionswereconsidered in theprobability values of Alcali-genesfaecalis-A. odorans and Psezidomonas alcaligenes in the data base.
The reproducibility ofthekit tests was
excellent,
withanestimated probability of erroneous results of 0.2%. This value is low when compared with probabilities of 2 to 4% found in other studies (19),
probably
because thetests that wereperformedare well established in ourlaboratory.
Seventeen
species
of the genus Pseudomonas were in-cluded inour study. Most strains were identified correctlywith an identification score of
-90%.
For the 5 of 18 Pseudomonaspseudoalcaligenes
strains and the 8 of 24Pseudomonastestosteroni strains
tested, only
identification scores of <80% could be obtained. The differentiationamonginactive strains of the
species
Pseudomonas alcali-genes,Pseudomonaspseudoalcaligenes,
and Pseudomonas testosteroni was notsatisfactory,
as hasalready
beende-scribed for other test systems,
especially
the API 20NEsystem
(2, 23, 26).
Strains ofChryseomonas luteola
(CDC
group Ve-1)werecorrectly
identified,
whereas for two of the five strains ofFlavimonasoryzihabitans (CDC groupVe-2),
higher
identi-fication scoresof -80%were
prevented.
Within the genus
Alcaligenes,
Alcaligenes faecalis
andAlcaligenes
odorans weregrouped together
(Alcaligenes
odorans isalater synonymfor
Alcaligenes faecalis [22]).
Five taxa ofthe genus Moraxella were tested. The two
species
Moraxellanonliquefaciens
and Moraxella lacunataare not
easily separated
on routine biochemical tests and were therefore combined into asingle
taxon. Of the 14strains of these two
species,
13 achieved identification scores ofonly
68.7%. For 8 strains of the 12 Moraxellaphenylpyruvica
strains,
ascoreof<80%was obtained. Forthe 11strains of
Oligella
urethralis(Moraxella
urethralis),
10identification scores were only at a level of77.1%. Within the genera
Flavobacterium, Pasteurella,
andVibrio,
asforthe
species
Aeromonashydrophila,
Plesiomonasshigelloides,
and Chromobacteriumviolaceum,
identifica-tion in most cases were correct, with scores of >90%. Strains identified with alower accuracy (scores from80 to
90%)
inmost casesshowedatypical
testpatterns. ForVibriovulnificus, percentages
should beverified,
with the inclusion ofmorereferenceorganisms
in the database needed.Previous
investigations
onthe accuracy ofcommercially
available systems for the identification of nonfermenters produced various results (2, 8, 23, 26). Identification of gram-negative
fermenting
andnonfermenting
bacteriaby
probabilisticmethods has beendescribed in theliterature(6,
16, 18), and most
commercially
available test kits for theidentification ofvarious bacteriaarebasedonthese methods
(40). Most evaluation resultsof
commercially
available kitsreported
difficulties inidentifying
weak oxidativeorinactivestrains. The combination of carbon substrate assimilation
testswith enzymetestsin the API 20NEsystemhasoffered
the
highest
accuracysofar for identificationtothespecies
orbiogroup
level(2).
Carbon source assimilation tests andcarbon substrate alkalinization testshave beenused earlier for the differentiation of
representatives
of the families Pseiidomonadaceae(3, 4, 25, 28, 30, 31, 33-35, 37,38)
andAlcaligenaceae
(29, 32) and the genus Flavobacterium(7).
Miniaturization of these methods (9-11, 20, 21) has made their
application possible
incommercially
availabletestkits.The
application
ofqualitative
enzyme testswithchromoge-nic substrates improved the accuracy of differentiation of
glucose-nonfermenting,
gram-negative,
andoxidase-posi-tive,
glucose-fermenting
bacteria(1,5).
The TT-NF system is a combination of miniaturized classical biochemical tests, carbon substrate assimilation tests, and growth-dependent
qualitative
enzyme tests withon April 11, 2020 by guest
http://jcm.asm.org/
TABLE 1. Results of the evaluation of the TT-NFsystem
No.of strainscorrectly identified No.of strains No.of No.of
No.of with an identification score of: with an strains strains
Species strainstested identification incorrectly not
>98% 90-97% 80-89% scoreof <80%" identified identified' Pseudomonas aeruginosa Pseudomonasfluorescens Pseudomonasputida Pseludomonaspseudomallei Pseudomonas cepacia Pseudomonas pickettii Pseudomonasstutzeri Pseudomonas mendocina Pseudomonasalcaligenes Pseudomonas pseudoalcaligenes Pseudomonas acidovorans Pseudomonas testosteroni Pseudomonas diminuta Pseudomonas vesicularis Pseudomonas maltophilia Pseudomonas paucimobilis Shewanellaputrefaciens
Chryseomonas luteola (CDCgroupVe-1) Flavimonas oryzihabitans(CDCgroupVe-2) Alcaligenesfaecalis-A. ordorans
Alcaligenesxylosoxidans subsp. denitrificans Alicaligens xylosoxidans subsp. xylosoxidans Ochrobactrumanthropi (Achromobacter
groupV-d)
Bordetellabronchiseptica Agrobacterium tumefaciens
Moraxellanonliquefaciens-M. lacunata Moraxellaosloensis
Moraxella phenylpyruvica Moraxellaatlantae
Oligella urethralis (Moraxellaurethralis) Acinetobacter calcoaceticus Acinetobacterlwoffii Flavobacterium meningosepticum Flavobacterium odoraltun Flavobacterium breve Flavobacterium multivorum Flavobacteriumthalpophilum Flavobacteriumspiritivorum Flavobacterium species group IIb
Weeksellazoohelcum (CDCgroup Ilj) Pasteurellaaerogenes
Pasteurella multocida
Pasteurellahaemolytica type A Pasteurella haemolytica type T Pasteurella pneumotropica Pasteurellaureae Aeromonashydrophila Plesiomnonas shigelloides Vibrioalginolyticus Vibrioparahaemolyticus Vibriovulnificus Vibriocholerae Vibriomimicus Vibriometschnikovii Vibriofurnissii Vibriofluvialis Chromnobacterium violaceuni 113 72 61 29 42 19 22 22 8 18 44 24 24 25 57 20 17 6 5 54 13 il 23 28 30 14 10 12 s il 26 25 25 21 18 42 7 31 28 14 1 34 2 2 3 4 28 23 22 18 10 23 s 12 10 6 10 106 58 42 27 39 15 20 4 38 21 20 47 16 17 6 3 50 2 10 19 6 6 6 3 20 7 9 3 15 1 3 7 1 4 25 26 9 4 4 12 il 20 3 25 15 3 9 7 il 22 s 29 22 1 14 32 2 25 17 20 14 2 19 2 12 7 3 8 i
s
s
2 1 1 1 i 1 1 2 2 1 2 6 1 1 3 8 i 2 1 2 1s
2 8 1 2 1 2 2 4 1 1 13 1 8 1 10 1 2 2 1 3 2 1 3 1 i 1 2 1 1 1 1 1 1 1 1 2 2 2 4 1 1 2 2 2 3 1 1 i i 2 2 i 1 1 4 1 1 2 1 1 1 1 1 1 11,289(100) 935(72.5) 189(14.7) 40(3.1) 104(8.1) 14(1.1) 7(0.5) Total (percent)
"Lowdiscriminationresults(identificationresults of strainsgroupedin this category showedtwoor more taxapossible [amongthemthecorrecttaxonl),but
theyfailed to reach the identification level (scores of.80%);acorrectidentificationcanbe madebyperformingthesuggestedsupplementarytests todifferentiate
between themostlikelytaxasuggested.
bStrains reached theidentificationlevel (scoresof.80%),buttothe incorrecttaxon.
' Noidentification result is shown(comparisonoftestprofileswith the data basegeneratedidentificationscoresthatweretoosmall).
on April 11, 2020 by guest
http://jcm.asm.org/
TABLE 2. Misidentifications with the TT-NF system Identification with the Conventional identification(no.) TTN
hyte
TT-NFsystern
Pseudoinonasdiminutl (1)...Moraxella atlantae
Pseudoinonas vesicularis(1)...Flat'obh(ateriun
ineningosepli-Pseudotnonaspaucimobilis (1)... Pseudomnonasdiniinuta
Agrobacteriurn rumefaciens(1) ....Flavobacterilm ,nultivorum
Moraxella
nonliquefaciens-M. lacunala (1)...Pasteurellahaeinolvtic-a typeA
Oligellaurethralis (1)...Moraxellaphenvlpvrulvica
Flavobacterinm brei'e(1)...Moraxella
nonliquefaciens-M.lacunata Flavobac-teriun species
groupIIb (1)... Flav'obwterinm eninengosepti-c(un
Aeroinonas hydrophila (1) ...Vibriofurnissii Plesioinonas shigelloides (1)... Pasteurellnllaocida Vibriov'ulnificus (1)...Flai'obal(teriiumi speciesgroup
lIb
Chroinobac'teriun v'iolaceun (1)....Flavobh(ateri-iiii inulti'oro7un AIcaligenes
fjecalis-A. odorans(2)... Akalaigenesxylosox.idans
subsp. denitrificansand
Mor-axelIa
phenvlpyruvica
chromogenic substrates. Hence, this system is designed to offer a relatively specific level of identification compared with that offered by other methods.
The TT-NF system allows, in most cases. the correct
identification (identification scores of >90%) ofcommonly
encountered nonfermenting bacteria and oxidase-positive, fermenting bacteria; and also, those groupsofbacteria that are rarely encountered in the clinical microbiology labora-tory were well separable. Disadvantages of the system included the lack of adequate resolution within the genus Moraxella, the lack of adequate differentiation among non-reactive Pseuidomonas testosteroni, Pseuidomonas alca( li-genes, and Pseudomonas pseudoahlcaligenes species and
Alcaligenes xylosoxidans subsp. denitrifi(ans. Combining these rarer species or genera into groups and identifying
themonthe genus or the group levelonly could improve the percent identification values, as has been suggested for the
PASCO MIC-ID system(36). For the majority of
laborato-ries, collective groupings as Moraxella sp. and
Pseuidomo-nas alcaligenes, Pseudoinonas pseutdoal(ctiligenes, and Pseudomonas testosteroni would be adequate. Vibrio
i'il-nificuiis
isolates also posed problemsfor the TT-NF system.Certain biotypes are not represented in the data base.
TABLE 3. Reproducibility ofindividual tests
Test' No. of Corrected
errors" error rate(%)' Glucosefermentation 2 1.4(0-2.8)"'
Gluconate assimilation i Histidine assimilation 1
Mannoseassimilation 1 0.7(0-2(.0)
Phenylacetate assimilation 1
Urease 1
Othertests
gave
Oerrors
inthe
150
teststhat
were
performed.b Asinglenegativeresult in oneofthetriplicatetestswasclassified
als
anerror.
Thecorrected error rate isgivenbyC= 1/2 (1 - N. as
de-scribedin the text.
dThe95%confidence intervals aregivenin parentheses.
Despitetheseproblems,the TT-NFsystemprovidesahighly
differentiating approach for the identification of
gram-nega-tive, nonfermenting and gram-negative, oxidase-positive, fermenting bacteria,and thusoffersapracticalalternativeto
already existing identification systems for this
heteroge-neousgroup of
organisms.
ACKNOWLEDGMENTS This studywas supported byFlow Laboratories.
Wethank H. Backes(FlowLaboratories)for invaluablehelpand discussions. Furthermore, we thank Katrin Noack for excellent technical assistanceand H. K. Geiss(Instituteof Medical
Microbi-ology, UniversityofHeidelberg)andG. Barth(CentreHospitalier
St. Nicolas, Sarrebourg, France) for the contributions of test
organisms.
LITERATURE CITED
1. Algorta,G., V. Monteil, and M. Y. Popoif. 1987. Profils d'ac-tivitésenenzymatiquesdes genresAgrobacterium,Aicaligenes, Altero<lnonas,FIai'obacteriumetPseudoinonas.Ann.Inst. Pas-teur/Microbiol. 138:297-302.
2. Appelbaum,P. C.,and D.J.Leathers. 1984. Evaluation of the
RapidNFT systemfor identification ofgram-negative, nonfer-mentingrods. J. Clin. Microbiol. 20:730-734.
3. Ballard, R. W.,M. Doudoroif, R. Y. Stanier, and M. Mandel. 1968. Taxonomy of the aerobic pseudomonads: Pseudoinonas
dinintwa and P. tesiculare. J.Gen. Microbiol.53:349-361. 4. Ballard,R.W.,N.J. Palleroni,M.Doudoroff,R.Y.Stanier,and
M. Mandel. 1970. Taxonomy of the aerobic pseudomonads: Pseudonîonascepalcia,P.inarginalta,P.allîicola,and P.
carx'o-phvlli. J. Gen. Microbiol. 60:199-214.
5. Bascomb, S. 1985. Growth independent rapid automated en-zyme identification systems, p. 377-389. hI K.-O. Habermehl (ed.), Rapid methods and automation in microbiology and
immunology. Springer-Verlag, Heidelberg,FederalRepublicof
Germany.
6. Bascomb, S., S. P. Lapage, M. A. Curtis, and W. R. Willcox. 1973. Identification of bacteria by computer: identification of reference strains. J. Gen. Microbiol. 77:291-315.
7. Bruun, B. 1983. Studieson acollection ofstrains of the genus Flarobacteriuin 2. Nutritional studies. Acta Pathol. Microbiol. Immunol. Scand. 91:35-41.
8. D'Amato, R. F., J. C. McLaughlin, andM. J. Ferraro. 1985. Rapid manual and mechanized/automated methods for the de-tection and identification ofbacteria and yeasts, p. 52-65. In E. H. Lennette. A. Balows, W. J. Hausler, Jr., and H. J.
Shadomy (ed.). Manual ofclinicalmicrobiology,4th ed. Amer-ican SocietyforMicrobiology, Washington,D.C.
9. Freney, J., J. Fleurette,P.Laban, J.-P.Gayral,andM.
Desmon-ceaux.1983. Systeme automatiqued'identificationdesbacillesà
gram négatif. Déscription et evaluation d'un prototype, p. 245-251. Io H. Leclerc (ed.), Les bacilles à gram négatif
d'interêt médicalet ensantépublique: taxonomie-identification-applications. Les Editions Institut National de la Santéetde la Recherche Médicale, Paris.
10. Freney,J.,P.Laban,M.Desmonceaux,H.Alexandre,B. Poggi,
J.-P.Gayral,andJ.Fleurette.1985. Anautomated micromethod for the identification of gram-negative bacilli by carbon
sub-strateassimilation tests. p. 377-389.J) K.-O.Habermehl(ed.),
Rapid methods and automation in microbiology and immunol-ogy. Springer-Verlag, Heidelberg, Federal Republic of Ger-many.
11. Freney, J., P. Laban, M. Desmonceaux, J.-P. Gayral, and J. Fleurette. 1984.Differentiation ofgram-negativerods other than
Enlerobact1eriaceae and Vibrionaceae by a micromethod for determination of carbon substrate assimilation. Zentralbl. Bak-teriol. Hyg. Abt. 1Orig. ReiheA258:198-214.
12. Geiss, H.K., V. Hingst, and H. D. Piotrowski. 1984. Untersu-chung zur Identifizierung von Nonfermentern. Hyg. Med. 5: 171-173.
13. Gilardi, G. L. (ed.). 1978. Glucose nonfermenting rods in
on April 11, 2020 by guest
http://jcm.asm.org/
clinical microbiology. CRCPress,Inc., Boca Raton, Fla. 14. Gilardi, G. L.(ed.). 1985.Nonfermentative gram-negative rods.
Laboratory identificationandclinical aspects. Marcel Dekker, Inc., New York.
15. Gilardi, G. L. 1988. Identification of glucose-nonfermenting gram-negativerods. NorthGeneral Hospital, New York. 16. Holmes, B., C. A. Dawson, and C. A.Pinning. 1986. A revised
probabilitymatrixfortheidentification ofgram-negative, aero-bic, rod-shaped, fermentative bacteria. J.Gen. Microbiol. 132: 3113-3135.
17. Holmes, B., and L. R. Hill. 1985. Computers in diagnostic microbiology, includingidentification, p. 265-287. In M. Good-fellow, D. Jones, and F. G. Priest. (ed.), Computer assisted bacterial systematics. Special publication no. 15. Society for General Microbiology. Academic Press, Inc. (London), Ltd., London.
18. Holmes, B., C. A. Pinning, and C. A. Dawson. 1986. A proba-bility matrix for the identification of gram-negative, aerobic, non-fermentative bacteria that grow on nutrient agar. J. Gen. Microbiol. 132:1827-1842.
19. Holmes, B., W. R. Willcox, S. P. Lapage, and H. Malnick. 1977. Testreproducibility ofthe API (20E), Enterotube, and Pathotec systems. J. Clin. Pathol. 30:381-387.
20. Kampfer, P., W. Bette, and W. Dott. 1987. Automatisierte MikromethodezurBestimmungderVerwertbarkeitorganischer Kohlenstoffquellen durch klinisch bedeutsame Pseudomonas Arten. Zentralbl. Bakteriol. Hyg. Abt. 1 Orig. Reihe A 265: 62-73.
21. Kampfer,P., W. Bette, and W. Dott.1987.Phenotypic differen-tiation of members of the family Vibrionaceae using miniatur-ized biochemical tests. Zentralbl. Bakteriol. Hyg. Abt. 1 Orig. ReiheA 266:425-437.
22. Kersters, K., and J. De Ley. 1984.Alcaligenes, p. 361-373. In N. R.Krieg, and S.J.Holt(ed.), Bergey'smanualofsystematic bacteriology, vol. 1.The Williams & WilkinsCo., Baltimore. 23. Lampe, A.S., and T. J. K. van der Reyden. 1984. Evaluation of
commercial systems for the identification of nonfermenters. Eur. J.Clin. Microbiol.3:301-305.
24. Lapage, S. P., S. Bascomb, W. R. Willcox, and M. A. Curtis. 1973. Identification of bacteria by computer: general aspects andperspectives. J. Gen. Microbiol.77:273-290.
25. Martin, R., P. S. Riley, D. G. Hollis, R. E. Weaver, and M. J. Krichevsky. 1981. Characterization ofsome groups of gram-negative nonfermentative bacteria by the carbonsource alkalin-izationtechnique. J. Clin. Microbiol. 14:39-47.
26. Martin, R., F.Siavoshi,and D. L.McDougal. 1986.Comparison of theRapidNFT systemandconventional methods for
identi-fication of nonsaccharolytic gram-negative bacteria. J. Clin. Microbiol. 24:1089-1092.
27. Oberhofer, T. R. 1985.Manualof nonfermentinggramnegative bacteria.JohnWiley & Sons, Inc., NewYork..
28. Palleroni, N. J., M. Doudoroff, R. Y. Stanier, R. E. Solanes, and M. Mandel.1970. Taxonomyoftheaerobic pseudomonads: the properties ofthePseudomonas stutzerigroup. J. Gen. Micro-biol. 60:215-231.
29. Pichinoty, F., M. Véron, M. Mandel, M. Durand, C. Job, and J.-L. Garcia. 1978. Étude physiologique et taxonomique du genreAlcaligenes: A. denitrificans, A. odorans et A.faecalis. Can.J.Microbiol. 24:743-753.
30. Pickett, M. J., and J. R. Greenwood. 1986. Identification of oxidase-positive, glucose-negative, motile species of nonfer-mentative bacilli. J. Clin. Microbiol.23:920-923.
31. Pickett, M. J., and J. R. Greenwood. 1986. Pseudomonas alcaligenes and Pseudomonas testosteroni: characterization andidentification. Curr. Microbiol. 13:197-201.
32. Pinter, M., and M. Kantor. 1974. Comparison of Alcaligenes faecalis and Alcaligenes odoransvar.viridansbycarbonsource utilizationtests. ActaMicrobiol.Acad.Sci. Hung.21:293-295. 33. Ralston, E., N. J. Palleroni, and M. Doudoroff. 1973. Pseudo-monas pickettii, a new species of clinical origin related to Pseudomonas solanacearum.Int.J. Syst. Bacteriol. 23:15-19. 34. Ralston-Barrett, E., N. J. Palleroni, and M. Doudoroif. 1976. Phenotypic characterization anddeoxyribonucleicacid homol-ogies ofthe "Pseudomonas alcaligenes" group. Int. J. Syst. Bacteriol. 26:421-425.
35. Rarick, H. R., P. S. Riley, and R. Martin. 1978. Carbon substrate utilization studies of some cultures of Alcaligenes
denitrificans, Alcaligenes faecalis, and Alcaligenes odorans isolatedfrom clinicalspecimens.J. Clin. Microbiol. 8:313-319. 36. Rhoden, D. L., B.Schable,and P. B.Smith. 1987.Evaluationof PASCO MIC-IDsystemforidentifying gram-negative bacilli.J. Clin. Microbiol. 25:2363-2366.
37. Snell,J. J.S.,andS. P.Lapage.1973.Carbonsourceutilization tests as an aid to the classification of nonfermenting gram-negativebacteria. J.Gen.Microbiol. 74:9-20.
38. Stanier, R. Y., N. J. Palleroni, and M. Doudoroif. 1966. The aerobicpseudomonads:ataxonomicstudy. J.Gen. Microbiol. 43:159-271.
39. Willcox, W.R., S. P. Lapage, S. Bascomb, and M. A.Curtis. 1973. Identification ofbacteria by computer: theory and pro-gramming. J.Gen. Microbiol. 77:317-330.
40. Willcox,W.R.,S. P.Lapage,and B.Holmes.1980. Areviewof numerical methods in bacterial identification. Antonie van Leeuwenhoek J. Microbiol. Serol.46:233-299.