Bromeliaceae diversity and conservation in Minas Gerais state, Brazil

Abstract

Field work and data from herbaria collections (2686 records) representing 283 taxa (265 species and 18 infraspecific taxa)
of Bromeliaceae occurring at Minas Gerais state, southeastern Brazil, were analyzed in order to obtain distribution and diversity
information, and to determine IUCN (The World Conservation Union) conservation status for each taxon. A map containing 1°×1°
grid cells was used to identify priority areas for new research collections, areas of high species diversity, and Bromeliaceae
conservation status. A clear decrease in Bromeliaceae diversity is observed between the eastern and the western portions of
Minas Gerais, and low floristic similarities were found between neighboring grid cells. The rocky mountains of Cadeia do Espinhaço
are considered the most important area for Bromeliaceae endemics. From the 283 taxa of Bromeliaceae that occur at Minas Gerais,
118 (42%) are considered threatened, and 124 taxa (44% of the total) do not occur inside any protected area. The region of
the Quadrilátero Ferrífero in the southern portion of the Cadeia do Espinhaço is the most threatened, and urgent strategies
for conservation of this rich Bromeliaceae flora are needed. Northeastern Minas Gerais, particularly the rocky outcrops or
inselbergs located in the Jequitinhonha and Mucuri rivers drainage basins need additional collection efforts and conservation
actions focused on these saxicolous taxa.

Full text

Biodivers Conserv (2007) 16:2989–3009
DOI 10.1007/s10531-007-9157-7
1 C
ORIGINAL PAPER
Bromeliaceae diversity and conservation in Minas
Gerais state, Brazil
Leonardo M. Versieux · Tânia Wendt
Received: 29 May 2006 / Accepted: 30 January 2007 / Published online: 7 March 2007
© Springer Science+Business Media B.V. 2007
Abstract Field work and data from herbaria collections (2686 records) representing
283 taxa (265 species and 18 infraspeciWc taxa) of Bromeliaceae occurring at Minas
Gerais state, southeastern Brazil, were analyzed in order to obtain distribution and
diversity information, and to determine IUCN (The World Conservation Union)
conservation status for each taxon. A map containing 1° £1° grid cells was used to
identify priority areas for new research collections, areas of high species diversity,
and Bromeliaceae conservation status. A clear decrease in Bromeliaceae diversity is
observed between the eastern and the western portions of Minas Gerais, and low
Xoristic similarities were found between neighboring grid cells. The rocky mountains
of Cadeia do Espinhaço are considered the most important area for Bromeliaceae
endemics. From the 283 taxa of Bromeliaceae that occur at Minas Gerais, 118 (42%)
are considered threatened, and 124 taxa (44% of the total) do not occur inside any
protected area. The region of the Quadrilátero Ferrífero in the southern portion
of the Cadeia do Espinhaço is the most threatened, and urgent strategies for
conservation of this rich Bromeliaceae Xora are needed. Northeastern Minas Gerais,
particularly the rocky outcrops or inselbergs located in the Jequitinhonha and
Mucuri rivers drainage basins need additional collection eVorts and conservation
actions focused on these saxicolous taxa.
Keywords Brazil · Bromeliaceae · Conservation · Cadeia do Espinhaço ·
Endemism · Epiphytes · Minas Gerais · Serra da Mantiqueira · Species richness
L. M. Versieux (&) · T. Wendt
Departamento de Botânica, Universidade Federal do Rio de Janeiro, CCS, IB,
Sala A1-92, Ilha do Fundão, Rio de Janeiro, RJ 21941-590, Brazil
e-mail: lversieux@yahoo.com.br
T. Wendt
e-mail: twendt@biologia.ufrj.br

2990 Biodivers Conserv (2007) 16:2989–3009
1 C
Introduction
The monocot family Bromeliaceae Juss. is among the most characteristic of the Neo-
tropical region. It is wholly American except for a single species of west tropical
Africa, and contains about 56 genera and 3000 species of generally herbaceous and
rhizomatozous plants, with lanceolate leaves that are often spirally arranged with
tightly overlapping basal sheathes that form free-water tanks, or phytotelma, and
both blades and sheathes often bear scale-like water-absorbing trichomes (Dahlgren
et al. 1985; Luther 2004). The distinct adaptations for a wide variety of terrestrial
and epiphytic life forms have enabled Bromeliaceae to colonize and diversify within
the Neotropics, and the family constitutes a noteworthy case of adaptive radiation
(Benzing 2000; Givnish et al. 2004). Being widely distributed and frequent in
Neotropical habitats, bromeliads act as an important ecological element in many
communities, contributing to structural complexity of the environment, which is
directly reXected on richness and diversity of associate fauna and Xora (Benzing 2000).
Bromeliaceae is also world renowned for its horticultural value. Over the last two
decades, the family has become more popular in Brazil, as home and garden orna-
mental plants, and this has promoted increased collecting pressures on natural popu-
lations. Although it is broadly accepted that Brazil is the richest country for
Bromeliaceae species, a precise and updated inventory for use in conservation
eVorts and to drive environmental agency decisions and actions, is lacking. The pres-
ent work is the Wrst to employ grid cell analysis with respect to diversity and conser-
vation data on Minas Gerais (MG) Bromeliaceae species. One degree grid cells have
been employed in research addressing phytogeography, species richness, and ende-
mism analyses for conservationist purposes (e.g. Kress et al. 1998; Serrato et al.
2004). Benzing (2000) notes that studies of Bromeliaceae phytogeography are rare,
usually restricted to areas continental in scale (e.g. Smith 1934), and that important
information on Bromeliaceae evolution would arise from analyses combining taxo-
nomic, Xoristic and life-form data.
Due to the great ecological importance and horticultural interest in the family,
versus the current, alarming status of human interference in natural bromeliad habi-
tats in MG, we present this work, which aims (1) to determine current knowledge on
Bromeliaceae distribution and their habitats within MG; (2) analyze the diversity
and collection eVorts for the family within 1° £1 ° grid cells; (3) establish the conser-
vation status of each taxon, and list those that are not protected; (4) indicate areas
where more scientiWc collections, or special conservation are needed.
Study sites
The state of MG is located in the southeastern region of Brazil (Fig. 1) contains
586,528 km2 of area, 853 municipalities, and a rich Bromeliaceae Xora, characterized
by high numbers of genera and species, and elevated levels of endemism (Versieux
and Wendt 2006). Unfortunately, the history of MG is rich with massive deforesta-
tion due to expansions of agricultural, cattle rising and urban areas, and also by min-
ing activities. Minas Gerais is characterized by a hilly relief with elevation ranging
from 79 to 2,890 m (Drummond et al. 2005). The state has a complex vegetation that
ranges from xerophytic dry forest to tropical evergreen forests, that can generally be
divided into three diVerent biomes: Atlantic Forest in the eastern and southern

Biodivers Conserv (2007) 16:2989–3009 2991
1 C
portions of the state; cerrado (savanna) in the central-western and northwestern
areas; and caatinga (low drought-decidous forest) in the extreme north of MG
(Drummond et al. 2005).
The highlands region formed by the Cadeia do Espinhaço and the Serra da
Mantiqueira have great importance as habitat for endemic Xora and fauna (Fig. 1).
The Cadeia do Espinhaço extends for ca. 1,100 km from MG (Ouro Preto munici-
pality) to Chapada Diamantina in Bahia state, its northernmost limit (Harley 1995).
A particular vegetation type known as campo rupestre (rocky Weld) is characteristic
of the more elevated areas along the Cadeia do Espinhaço. This grassland vegeta-
tion usually appears above 800 m where soils are shallow, and sandstones and
quartzite outcrops are very frequent, providing diVerent habitats for saxicolous
plants (Giulietti et al. 1987; Harley 1995). Robust vegetation with twisted low trees
appears wherever the soil is deeper and gallery forest and forest patches known as
capões occur along watercourses, ravines and hillsides, or even on hill tops (Pirani
et al. 1994). The extreme southern part of Cadeia do Espinhaço is called ‘Quadrilá-
tero Ferrífero’ and diVers from most of the remaining mountain range, due to
exposed iron oxide deposits which are known as ‘canga’ and provide habitat for
many saxicolous species (Rizzini 1997). Lying close to boundaries of Rio de Janeiro
and São Paulo states is another important mountain range, Serra da Mantiqueira. In
this range the forest resembles the most exuberant and rich vegetation that is found
along the southeastern Brazilian coast, called Atlantic Forest sensu stricto. As the
altitude increases, small areas of Araucaria forest, also called cloud forest, appear
and high humidity within the forest supports abundant epiphytic vegetation. Above
Fig. 1 Map of Minas Gerais showing the main mountain ranges and the biomes

2992 Biodivers Conserv (2007) 16:2989–3009
1 C
1,200 m in Serra da Mantiqueira, the forest gives way to campo de altitude, or high
altitude grassland. This open habitat provides only a thin soil layer and vegetation is
scattered in small islands and adapted to high levels of humidity, winds, solar
exposure, and low winter temperatures (< 12°C) (Martinelli 1989; SaVord 1999).
Methods
Data were obtained from Weld work carried out from 2002–2005 and from collec-
tions located at 14 herbaria that were examined between years 2002–2004 (BHCB,
BHZB, CESJ, HB, HBR, MBM, R, RB, RFA, SEL, SP, SPF, US, VIC, acronyms
according to Holmgren et al. (1990) except BHZB = Fundação Zoobotânica de Belo
Horizonte). All herbarium specimens from MG were photographed and databased
using Brahms software (Botanical Research and Herbarium Management System,
version 5.55, Oxford University). The recent checklist of Minas Gerais Bromelia-
ceae (Versieux and Wendt 2006) was used as the main taxonomic framework. Data
on the type of vegetation in which Bromeliaceae taxa occurred were mostly taken
from specimen labels, but for some cases a vegetation map was consulted (Drum-
mond et al., 2005). Municipalities were grouped inside 74 grid cells of 1 per 1° and
those that were not totally conWned within a single grid cell were assigned as belong-
ing to the cell that hosts their administration center. The presence or absence of each
taxon of Bromeliaceae in each cell was recorded. The Jaccard’s measure of similarity
and the UPGMA clustering method were employed in order to identify Xoristic sim-
ilarities among grid cells, using BiodiversityPro software (ver. 2/1997). The IUCN
(2001) methodology was used to establish conservation status of each taxon, except
for the genus Encholirium that was previously deWned by Forzza et al. (2003). Bro-
meliaceae taxa occurrence within parks or natural reserves were veriWed with the list
of protected areas of MG (Camargos 2001).
Results and discussion
Phytogeographical analysis
Species richness and endemism
Bromeliaceae is represented within MG by 27 genera, 265 species, and 18 infraspe-
ciWc taxa (Fig. 2). The areas with highest species richness are concentrated along the
southern Cadeia do Espinhaço and in southeastern MG, and generally correspond
to well-known and historical collection localities (e.g. Diamantina, Serra da Piedade,
Ouro Preto, Serra de Ibitipoca). Some grid cells with numerous bromeliad taxa are
also important because they are among the Xoristically richest areas for other plant
groups. For example, cell F8 is the richest cell for Bromeliaceae and corresponds to a
very rich area for Xyridaceae and Mimosa (Wanderley 1992; Simon and Proença
2000). High generic-level diversity is related to the geographic position occupied by
MG, where many diVerent climatic types exist, allowing the development of distinct
vegetation forms and, consequently, distinct taxa of Bromeliaceae. Bromeliaceae
species diversity within MG is high, corresponding to almost 9% of the total number
of species for the entire family. Once again, this is related to unique sets of habitat

Biodivers Conserv (2007) 16:2989–3009 2993
1 C
attributes, local climates, and to the biogeographical aYnities among eastern
Brazilian areas, where Bromeliaceae has profoundly diversiWed (Versieux and
Wendt 2006). One genus (Andrea) and 98 species (37% of total) are endemic to MG.
Sixty two percent of all the endemic taxa are restricted to the Espinhaço, while 12%
are only known to occur in the Mantiqueira. Only one percent of Bromeliaceae
endemic species have general, broad distributions within MG.
The importance of the campo rupestre habitat of the Cadeia do Espinhaço for
Xora diversity and endemism has been observed by many authors while working
with diVerent plant families and is conWrmed here for MG Bromeliaceae (Fig. 2). As
observed by Pirani et al. (1994), many species from the Cadeia do Espinhaço rocky
Welds probably aroused in situ in response to speciWc environmental conditions due
to the isolation from nearby mountains. Morton (1972) explained the higher levels
of endemism in montane habitats of west African mountains by a rapid speciation
following extinctions of many grassland species resulting from repeated climatic
oscillations (i.e. either temperature and rain fall increase or decrease) that promoted
advances of the forest during climatic optima, when many of the savanna habitats
were reduced or eliminated. However, open habitat species rested conWned to suit-
able refuge such as cliVs and rock outcrops where hybridization provided a suYcient
genetic variation for the evolution of new species during the next expansion of grass-
lands (Morton 1972). Similar processes could have occurred in elevated areas of
MG, such as the Cadeia do Espinhaço and the Serra da Mantiqueira, allowing their
notable Xoristic diversity and endemism. Even endemic genera inhabit these moun-
tains, as the monotypic Andrea, from the Espinhaço. Considering only dispersal
capacity, higher endemism rates for the Bromeliaceae of the Espinhaço campo
Fig. 2 Species richness within 1° £1° grid cells for Bromeliaceae of Minas Gerais, Brazil

2994 Biodivers Conserv (2007) 16:2989–3009
1 C
rupestre may be attributed to lower dispersal ability for many saxicolous taxa (e.g.
Dyckia, Encholirium) and perhaps to an eYcient isolation from nearby mountains
by lower elevated forested valleys, where similar rock outcrops, or other environ-
mental factors required for seed germination and seedling growth are unavailable.
Collection eVorts
The most collected grid cell are located along the Cadeia do Espinhaço, particularly
in its southernmost portion, i.e. grid cells E8, F8, and G8 (Fig. 3) and three are in the
southeastern portion of MG (G9, H7, H8). Almost 26% of the Bromeliaceae listed
for MG are represented in herbaria by a single specimen (Table 1). Spiny and succu-
lent leaves, the large dimensions of many of its species, and their occurrence in loca-
tions diYcult to access seem to be the cause of these low numbers. A clear decrease
in collection numbers is observed toward the central-western and northwestern
portions of the 45° meridian. The 45° of longitude roughly corresponds to the west-
ern border of the Espinhaço, limiting the campo rupestre and the Atlantic forest to
the east, and the savanna and part of the caatinga at the western and northern
portions of MG, respectively. Thus this mountain chain seems to act as barrier for
the inland distribution of many Bromeliaceae species of the Atlantic domain, inXu-
encing diversity and sampling. Sampling also decreases toward northeastern (e.g.
B10, B11, C10, C11, D9, D11, E9, E10, F10) and the extreme southern (e.g. H6, I5,
I6) portions. Seventeen grid cells did not presented any voucher of Bromeliaceae.
Those undercollected areas should be considered priorities in future collections and
Weld work research.
Fig. 3 Total number of Bromeliaceae taxa collections within 1° £1° grid cells, in Minas Gerais,
Brazil

Biodivers Conserv (2007) 16:2989–3009 2995
1 C
T
a
bl
e
1Li
st o
f
B
rome
li
aceae o
f
Mi
nas
G
era
i
s state,
B
raz
il
,
f
o
ll
owe
d
b
y tota
l
num
b
er o
f
recor
d
s,
occurrence within protected areas, and IUCN conservation status
Taxon No. of
records Occurrence
within parks IUCN Category
1. Acanthostachys strobilacea (Schult. &
Schult. f.) Klotzsch 44 y Least Concern (LC)
2. Aechmea alba Mez 3 n Vulnerable (VU)
3. A. alopecurus Mez 1 n Endangered (EN)
4. A. bambusoides L.B. Sm. & Reitz 6 n Vulnerable (VU)
5.1. A. bromeliifolia (Rudge) Baker var.
bromeliifolia 104 y Least Concern (LC)
5.2. A. bromeliifolia var. albobracteata
Philcox 1 y Least Concern (LC)
6. A. brueggeri Leme 5 n Critically Endangered (CR)
7. A. burle-marxii E. Pereira 1 n Data DeWcient (DD)
8.1. A. distichantha Lem. var. distichanta 22 y Least Concern (LC)
8.2. A. distichantha var. glaziovii (Baker)
L.B. Sm. 3 n Least Concern (LC)
8.3. A. distichantha var. schlumbergeri
E. Morren ex Mez 7 y Least Concern (LC)
9. A. gamosepala Wittm. 1 n Not Evaluated (NE)
10. A. lamarchei Mez 40 y Least Concern (LC)
11. A. maculata L.B. Sm. 3 y Vulnerable (VU)
12.1. A. nudicaulis (L.) Griseb. var. nudicaulis 59 y Least Concern (LC)
12.2. A. nudicaulis var. aureorosea
(Antoine) L.B. Sm. 13 y Least Concern (LC)
12.3. A. nudicaulis var. cuspidata Baker 9 y Least Concern (LC)
13. A. organensis Wawra 3 y Least Concern (LC)
14. A. phanerophlebia Baker 42 y Least Concern (LC)
15. A. pineliana (Brongn. ex Planch.) Baker 4 y Least Concern (LC)
16. A. purpureorosea (Hook. f.) Wawra 1 n Endangered (EN)
17. A. ramosa Mart. ex Schult. & Schult. f. 22 y Least Concern (LC)
18. A. vanhoutteana (Van Houtte) Mez 4 y Least Concern (LC)
19. A. weilbachii Didr. 1 y* Vulnerable (VU)
20. Alcantarea burle-marxii(Leme)
J.R. Grant 3 n Endangered (EN)
21. A. duarteana (L.B. Sm.) J.R. Grant 9 n Endangered (EN)
22. A. extensa (L.B. Sm.) J.R. Grant 8 y Least Concern (LC)
23. A. hatschbachii (L.B. Sm. & Read) Leme 2 n Critically Endangered (CR)
24. A. imperialis (Carrière) Harms 6 y Vulnerable (VU)
25. A. odorata (Leme) J.R. Grant 1 n Data DeWcient (DD)
26. A. turgida Versieux & Wand. 2 y Not Evaluated (NE)
27. A. sp. 1. 2 n Not Evaluated (NE)
28. Ananas ananassoides (Baker) L.B. Sm. 46 y Least Concern (LC)
29. A. bracteatus (Lindl.) Schult. & Schult. f. 5 y Least Concern (LC)
30. A. comosus (L.) Merr. 1 y Least Concern (LC)
31. A. nanus (L.B. Sm.) L.B. Sm. 2 y Vulnerable (VU)
32. Andrea selloana Baker 11 y Endangered (EN)
33.1. Billbergia amoena (G. Lodd.)
Lindl. var. amoena 46 y Least Concern (LC)
33.2. B. amoena var. carnea E. Pereira 3 n Vulnerable (VU)
33.3. B. amoena var. minor
(Antoine & Beer) L.B. Sm. 4 y* Endangered (EN)
34. B. distachia (Vell.) Mez 26 y Least Concern (LC)
35. B. elegans Mart. ex Schult. & Schult. f. 54 y Least Concern (LC)

2996 Biodivers Conserv (2007) 16:2989–3009
1 C
T
a
bl
e
1
cont
i
nue
d
Taxon No. of
records Occurrence
within parks IUCN Category
36. B. euphemiae E. Morren 15 y Least Concern (LC)
37. B. horrida Regel 13 y Least Concern (LC)
38. B. iridifolia (Nees & Mart.) Lindl. 12 y Least Concern (LC)
39. B. leptopoda L.B. Sm. 5 y* Endangered (EN)
40. B. lymanii E. Pereira & Leme 7 y Least Concern (LC)
41. B. macrocalyx Hook. f. 1 n Not Evaluated (NE)
42. B. meyeri Mez 3 n Vulnerable (VU)
43. B. minarum L.B. Sm. 3 y Not Evaluated (NE)
44. B. nutans H. Wendl. ex Regel 3 y Least Concern (LC)
45. B. pohliana Mez 1 n Data DeWcient (DD)
46. B. porteana Brongn. ex Beer 9 y Least Concern (LC)
47. B. reichardtii Wawra 4 n Endangered (EN)
48. B. sanderiana E. Morren 1 n Data DeWcient (DD)
49. B. saundersii W. Bull 1 n Data DeWcient (DD)
50. B. tweedieana Baker 3 n Vulnerable (VU)
51. B. vittata Brongn. 50 y Least Concern (LC)
52. B. zebrina (Herb.) Lindl. 18 y Least Concern (LC)
53. Bromelia antiacantha Bertol. 4 y Least Concern (LC)
54. B. balansae Mez 20 y Least Concern (LC)
55. B. glaziovii Mez 2 y Vulnerable (VU)
56. B. interior L.B. Sm. 4 n Least Concern (LC)
57. B. regnellii Mez 3 y Least Concern (LC)
58. B. serra Griseb. 5 n Vulnerable (VU)
59. B. villosa Mez 1 n Data DeWcient (DD)
60. Canistrum auratum Leme 1 n Vulnerable (VU)
61. Cryptanthus. caracensis
Leme & E. Gross 9 y* Critically Endangered (CR)
62. C. glaziovii Mez 4 y* Critically Endangered (CR)
63. C. leopoldo-horstii Rauh 6 n Critically Endangered (CR)
64. C. minarum L.B. Sm. 2 n Critically Endangered (CR)
65. C. schwackeanus Mez 47 y Vulnerable (VU)
66. C. warasii E. Pereira 1 n Critically Endangered (CR)
67. Dyckia argentea Mez 2 n Vulnerable (VU)
68. D. brachyphylla L.B. Sm. 14 n Vulnerable (VU)
69. D. bracteata (Wittm.) Mez 21 y Vulnerable (VU)
70. D. brevifolia Baker 2 n Not Evaluated (NE)
71. D. cinerea Mez 12 y Vulnerable (VU)
72. D. consimilis Mez 8 n Endangered (EN)
73. D. densiXora Schult. & Schult. f. 4 y Vulnerable (VU)
74. D. dissitiXora Schult. & Schult. f. 5 n Least Concern (LC)
75. D. elata Mez 1 n Data DeWcient (DD)
76. D. glandulosa L.B. Sm. & Reitz 1 y Data DeWcient (DD)
77. D. goehringii Rauh 1 n Data DeWcient (DD)
78. D. granmogulensis Rauh 2 n Data DeWcient (DD)
79. D. lagoensis Mez 11 y Least Concern (LC)
80. D. leptostachya Baker 3 n Least Concern (LC)
81. D. linearifolia Baker 2 n Data DeWcient (DD)
82. D. macedoi L.B. Sm. 14 y Vulnerable (VU)
83. D. macropoda L.B. Sm. 1 n Data DeWcient (DD)
84. D. marnier-lapostollei L.B. Sm. 2 n Critically Endangered (CR)
85. D. mello-barretoi L.B. Sm. 3 y Vulnerable (VU)
86. D. minarum Mez 27 y Least Concern (LC)

Biodivers Conserv (2007) 16:2989–3009 2997
1 C
T
a
bl
e
1
cont
i
nue
d
Taxon No. of
records Occurrence
within parks IUCN Category
87. D. monticola L.B. Sm. & Reitz 1 n Not Evaluated (NE)
88. D. orobanchoides Mez 1 n Data DeWcient (DD)
89. D. pectinata L.B. Sm. & Reitz 2 n Critically Endangered (CR)
90. D. princeps Lem. 1 n Data DeWcient (DD)
91. D. rariXora Schultes f. 9 y Vulnerable (VU)
92. D. remotiXora Otto & A. Dietr. 3 n Vulnerable (VU)
93. D. saxatilis Mez 52 y Least Concern (LC)
94. D. schwackeana Mez 3 n Endangered (EN)
95. D. simulans L.B. Sm. 4 y Endangered (EN)
96. D. sordida Baker 28 y Endangered (EN)
97. D. sp. 1. 17 n Not Evaluated (NE)
98. D. sp. 2. 2 y Not Evaluated (NE)
99. D. sp. 3. 2 y Not Evaluated (NE)
100. D. sp. 4. 1 n Not Evaluated (NE)
101. D. sp. 5. 29 n Not Evaluated (NE)
102. D. spinulosa L.B. Sm. & Reitz 1 n Data DeWcient (DD)
103. D. tenebrosa Leme & H. Luther 2 n Data DeWcient (DD)
104. D. trichostachya Baker 5 y Endangered (EN)
105. D. tuberosa (Vell.) Beer 4 n Not Evaluated (NE)
106. D. ursina L.B. Sm. 13 y Critically Endangered (CR)
107. D. warmingii Mez 1 n Data DeWcient (DD)
108. D. weddelliana Baker 2 n Data DeWcient (DD)
109. Edmundoa lindenii var. rosea
(E. Morren) Leme
2 n Vulnerable (VU)
110. Encholirium belemii L.B. Sm. &Read 1 n Data DeWcient (DD)
111. E. biXorum (Mez) Forzza 4 n Endangered (EN)
112. E. bradeanum L.B. Sm. 1 n Data DeWcient (DD)
113. E. gracile L.B. Sm. 1 n Critically Endangered (CR)
114. E. heloisae (L.B. Sm.) Forzza & Wand. 39 y Vulnerable (VU)
115. E. horridum L.B. Sm. 1 n Critically Endangered (CR)
116. E. irwinii L.B. Sm. 10 y Vulnerable (VU)
117. E. longiXorum Leme 2 n Critically Endangered (CR)
118. E. luxor L.B. Sm. & Read 8 n Critically Endangered (CR)
119. E. magalhaesii L.B. Sm. 25 y Vulnerable (VU)
120. E. pedicellatum (Mez) Rauh 4 n Critically Endangered (CR)
121. E. reXexum Forzza & Wand. 4 n Vulnerable (VU)
122. E. scrutor (L.B. Sm.) Rauh 7 n Endangered (EN)
123. E. subsecundum (Baker) Mez 69 y Vulnerable (VU)
124. E. vogelii Rauh 5 y Vulnerable (VU)
125. Fernseea itatiaiae (Wawra) Baker 4 y Critically Endangered (CR)
126. Hohenbergia augusta (Vell.) E. Morren 1 n Data DeWcient (DD)
127. H. catingae Ule 2 n Vulnerable (VU)
128. H. pabstii L.B. Sm. & Read 1 n Vulnerable (VU)
129. H. ramageana Mez 2 n Vulnerable (VU)
130. Neoglaziovia variegata (Arruda) Mez 16 y Least Concern (LC)
131. Neoregelia bahiana (Ule) L.B. Sm. 40 y Least Concern (LC)
132. N. brigadeirensis Paula & Leme 2 y* Vulnerable (VU)
133. N. brownii Leme 4 y Vulnerable (VU)
134. N. chlorosticta (Baker) L.B. Sm. 1 y Data DeWcient (DD)
135. N. cyanea (Beer) L.B. Sm. 2 n Data DeWcient (DD)
136. N. farinosa (Ule) L.B. Sm. 1 y Data DeWcient (DD)
137. N. ibitipocensis (Leme) Leme 3 y Vulnerable (VU)
138. N. leprosa L.B. Sm. 1 n Data DeWcient (DD)
139. N. lymaniana R. Braga & Sucre 4 y Vulnerable (VU)

2998 Biodivers Conserv (2007) 16:2989–3009
1 C
Table 1 cont
i
nue
d
Taxon No. of
records Occurrence
within parks IUCN Category
140. N. oligantha L.B. Sm. 1 y* Data DeWcient (DD)
141. N. sarmentosa (Regel) L.B. Sm. 5 y Least Concern (LC)
142. N. simulans L.B. Sm. 1 y Vulnerable (VU)
143. Nidularium antoineanum Wawra 8 y Least Concern (LC)
144. N. azureum (L.B. Sm.) Leme 2 n Critically Endangered (CR)
145. N. bicolor (E. Pereira) Leme 9 y Vulnerable (VU)
146. N. ferdinando-coburgii Wawra 4 y Vulnerable (VU)
147. N. linehamii Leme 1 y Critically Endangered (CR)
148. N. longiXorum Ule 3 y Vulnerable (VU)
149. N. marigoi Leme 9 y Least Concern (LC)
150. N. purpureum Beer 1 n Data DeWcient (DD)
151. N. meeanum Leme 1 n Data DeWcient (DD)
152. N. rutilans E. Morren 2 y Endangered (EN)
153. Orthophytum benzingii
Leme & H. Luther
1 n Endangered (EN)
154. O. compactum L.B. Sm. 4 n Vulnerable (VU)
155.1. O. disjunctum L.B. Sm. var. disjunctum 3 n Not Evaluated (NE)
155.2. O. disjunctum var.
angustobracteatum Rauh 1 n Data DeWcient (DD)
155.3. O. disjunctum var. variegatum Rauh 1 n Data DeWcient (DD)
155.4. O. disjunctum var. viridiXorum Rauh 1 n Data DeWcient (DD)
156. O. duartei L.B. Sm. 1 n Data DeWcient (DD)
157. O. eddie-estevesii Leme 1 n Endangered (EN)
158. O. estevesii (Rauh) Leme 1 n Data DeWcient (DD)
159. O. foliosum L.B. Sm. 1 n Vulnerable (VU)
160. O. glabrum (Mez) Mez 7 n Vulnerable (VU)
161. O. grossiorum Leme & Paula 1 n Vulnerable (VU)
162. O. gurkenii Hutchison 1 n Critically Endangered (CR)
163. O. horridum Leme 1 n Endangered (EN)
164. O. humile L.B. Sm. 3 y Vulnerable (VU)
165. O. itambense Versieux & Leme 1 y Critically Endangered (CR)
166. O. leprosum (Mez) Mez 10 n Vulnerable (VU)
167. O. lucidum Leme & H. Luther 2 n Endangered (EN)
168. O. magalhaesii L.B. Sm. 4 n Vulnerable (VU)
169. O. maracasense L.B. Sm. 3 n Vulnerable (VU)
170. O. mello-barretoi L.B. Sm. 32 y Vulnerable (VU)
171. O. supthutii E. Gross & Barthlott 3 n Critically Endangered (CR)
172. Pepinia bradei (Markgr.) G.S. Varad.
& Gilmartin 3 y Vulnerable (VU)
173. Pitcairnia carinataMez 9 y Least Concern (LC)
174. P. curvidens L.B. Sm. & Read 5 y Vulnerable (VU)
175. P. decidua L.B. Sm. 9 y Vulnerable (VU)
176.1. P. Xammea Lindl. var. Xammea 25 y Least Concern (LC)
176.2. P. Xammea var. Xoccosa L. B. Sm. 6 y Least Concern (LC)
176.3. P. Xammea var. glabrior L.B. Sm. 4 y Least Concern (LC)
176.4. P. Xammea var. macropoda L.B. Sm.
& Reitz 4 n Endangered (EN)
177. P. lanuginosa Ruiz & Pav. 13 y Least Concern (LC)
178.1. Portea petropolitana (Wawra)
Mez var.petropolitana 7 y Least Concern (LC)
178.2. P. petropolitana var. noettigii
(Wawra) L.B. Sm. 4 y Vulnerable (VU)
179. P. silveirae Mez 8 y Least Concern (LC)

Biodivers Conserv (2007) 16:2989–3009 2999
1 C
Table 1 cont
i
nue
d
Taxon No. of
records Occurrence
within parks IUCN Category
180. Pseudananas sagenarius (Arruda)
Camargo 13 y Least Concern (LC)
181. Quesnelia arvensis (Vell.) Mez 1 y Data DeWcient (DD)
182. Q. augusto-coburgii Wawra 4 y Vulnerable (VU)
183. Q. indecora Mez 18 y Least Concern (LC)
184. Q. kautskyi C.M. Vieira 7 y Vulnerable (VU)
185. Q. liboniana (De Jonghe) Mez 3 y Least Concern (LC)
186. Q. quesneliana (Brongn.) L.B. Sm. 3 n Not Evaluated (NE)
187. Q. strobilispica Wawra 7 y Least Concern (LC)
188. Racinaea aerisincola (Mez) M.A.
Spencer & L.B. Sm. 6 y Least Concern (LC)
189. Tillandsia arhiza Mez 1 y Least Concern (LC)
190. T. chapeuensis Rauh 1 y Not Evaluated (NE)
191. T. copynii Gouda 2 n Data DeWcient (DD)
192. T. gardneri Lindl. 46 y Least Concern (LC)
193. T. geminiXora Brongn. 42 y Least Concern (LC)
194. T. globosa Wawra 6 n Least Concern (LC)
195. T. graomogolensis Silveira 7 y Least Concern (LC)
196. T. horstii Rauh 1 n Data DeWcient (DD)
197. T. leonamiana E. Pereira 1 n Data DeWcient (DD)
198. T. loliacea Mart. ex Schult. & Schult. f. 20 y Least Concern (LC)
199. T. mallemontii Glaziou ex Mez 1 n Least Concern (LC)
200. T. parvispica Baker 5 y Least Concern (LC)
201. T. pohliana Mez 19 y Least Concern (LC)
202. T. polystachia (L.) L. 12 n Least Concern (LC)
203. T. pruinosa Sw. 3 n Data DeWcient (DD)
204. T. recurvata (L.) L. 82 y Least Concern (LC)
205. T. sp. 1. 1 n Not Evaluated (NE)
206. T. aV. sprengeliana Klotzsch ex Mez 1 n Not Evaluated (NE)
207. T. streptocarpa Baker 45 y Least Concern (LC)
208. T. stricta Sol. 115 y Least Concern (LC)
209.1. T. tenuifolia L. var. tenuifolia 38 y Least Concern (LC)
209.2. T. tenuifolia var. surinamensis
(Mez) L.B. Sm.
7 y Least Concern (LC)
209.3. T. tenuifolia var. vaginata (Wawra)
L. B. Sm. 18 y Least Concern (LC)
210. T. tricholepis Baker 8 y Least Concern (LC)
211. T. usneoides (L.) L. 31 y Least Concern (LC)
212. Vriesea arachnoidea And. Costa 2 y Endangered (EN)
213. V. atropurpurea Silveira 4 y Critically Endangered (CR)
214. V. billbergioides E. Morren ex Mez 6 y Least Concern (LC)
215. V. bituminosa Wawra 8 y Least Concern (LC)
216. V. cacuminis L.B. Sm. 6 y* Vulnerable (VU)
217. V. carinata Wawra 3 y Least Concern (LC)
218. V. clausseniana (Baker) Mez 21 y Vulnerable (VU)
219. V. crassa Mez 14 y Vulnerable (VU)
220. V. densiXora Mez 3 y Endangered (EN)
221. V. diamantinensis Leme 6 n Vulnerable (VU)
222. V. ensiformis (Vell.) Beer 14 y Least Concern (LC)
223.1. V. friburgensis Mez var. friburgensis 22 y Least Concern (LC)
223.2. V. friburgensis var. tucumanensis
(Mez) L.B. Sm. 1 n Data DeWcient (DD)
224. V. gigantea Gaudich. 1 y Vulnerable (VU)

3000 Biodivers Conserv (2007) 16:2989–3009
1 C
Floristic inventories concentrated in small areas have produced signiWcant contri-
butions to the knowledge of the Brazilian Xora (Prance 2001; Giulietti et al. 2005).
We observed that the existence of protected areas or reserves raise many grid cell
collection totals, and these numbers are higher if such areas have undergoing
detailed Xoristic work (e.g. Flora da Serra do Cipó, grid cell F8). Decreased
Table 1 cont
i
nue
d
DiVerent genera are separated by bold face. y = occur in protected area(s), y* = restricted and only
known to occur in one protected area, n = do not occur in any protected area
Taxon No. of
records Occurrence
within parks IUCN Category
225. V. gradata (Baker) Mez 9 y Least Concern (LC)
226. V. guttata Linden & André 6 y Least Concern (LC)
227. V. heterostachys (Baker) L.B. Sm. 11 y Least Concern (LC)
228. V. aV. hieroglyphica (Carrière)
E. Morren 1 n Not Evaluated (NE)
229. V. hoehneana L.B. Sm. 2 y Data DeWcient (DD)
230. V. itatiaiae Wawra 1 n Endangered (EN)
231. V. jonghei (K. Koch) E. Morren 1 n Not Evaluated (NE)
232. V. longicaulis (Baker) Mez 12 y Least Concern (LC)
233. V. longistaminea Paula & Leme 2 n Vulnerable (VU)
234. V. lubbersii (Baker) E. Morren ex Mez 2 y Least Concern (LC)
235. V. minarum L.B. Smith 36 y Endangered (EN)
236. V. minor (L.B. Sm.) Leme 25 y Least Concern (LC)
237. V. modesta Mez 1 n Data DeWcient (DD)
238. V. monacorum L.B. Sm. 4 y Endangered (EN)
239. V. morrenii Wawra 2 y Endangered (EN)
240. V. nanuzae Leme 1 n Endangered (EN)
241. V. neoglutinosa Mez 1 n Data DeWcient (DD)
242. V. oligantha (Baker) Mez 55 y Least Concern (LC)
243.1. V. paraibica Wawra var. paraibica 3 n Data DeWcient (DD)
243.2. V. paraibica var. interrogatoria
(L.B. Sm.) And. Costa 3 n Vulnerable (VU)
244. V. pauperrima E. Pereira 3 y Least Concern (LC)
245. V. pardalina Mez 6 y Least Concern (LC)
246. V. penduliXora L.B. Sm. 1 y Vulnerable (VU)
247. V. procera (Mart. ex Schult. & Schult. f.)
Wittm. 5 y Least Concern (LC)
248. V. racinae L.B. Sm. 1 y Data DeWcient (DD)
249. V. rafaelii Leme 2 n Critically Endangered (CR)
250. V. regnellii Mez 1 n Data DeWcient (DD)
251. V. ruschii subsp. leonii Leme 6 y Least Concern (LC)
252. V. sazimae Leme 1 n Data DeWcient (DD)
253. V. saxicola L.B. Sm. 3 n Endangered (EN)
254. V. scalaris E. Morren 3 y Least Concern (LC)
255. V. sceptrum Mez 5 n Least Concern (LC)
256. V. schwackeana Mez 4 y Vulnerable (VU)
257. V. segadas-viannae L.B. Sm. 2 n Data DeWcient (DD)
258. V. simulans Leme 1 n Vulnerable (VU)
259. V. sp. 1. 3 y Not Evaluated (NE)
260. V. stricta L.B. Sm. 11 y Vulnerable (VU)
261. V. vagans (L.B. Sm.) L.B. Sm. 5 y Least Concern (LC)
262. Wittrockia cyathiformis (Vell.) Leme 4 n Least Concern (LC)
263. W. gigantea (Baker) Leme 10 y Least Concern (LC)
264. W. tenuisepala (Leme) Leme 2 n Data DeWcient (DD)
265. W. sp. 1. 1 y Not Evaluated (NE)

Biodivers Conserv (2007) 16:2989–3009 3001
1 C
collections toward northeastern MG (e.g. cells E9, E10, B10, B11), may be related to
the lower number of protected areas, or to the highly fragmented original vegetation
(cf. IEF-MG 1994), rather than to unsuitable natural environmental conditions for
Bromeliaceae occurrence. Greater collecting eVorts within parks/reserves were also
observed by Calvente et al. 2005 while working with the Cactaceae of Rio de Janeiro.
Thus we conclude that Brazilian botanists are becoming more highly dependent on
oYcially protected areas due to the lack of original vegetation outside of these areas,
or due to poor access to private lands. Greater eVorts are needed to establish more
preserved areas, and to develop a more eYcient system of permitting to allow for
needed scientiWc collecting and study in these parks and reserves.
Uneven sampling can bias or inXuence Xoristic analyses, but the data presented
here and our personal observations indicate that within the savanna and caatinga the
Bromeliaceae tend to be rarer and have lower species diversity than in the Atlantic
forest. Rizzini (1997) observed that the epiphytic life form is uncommon for plants in
the savanna, and that is certainly reXected in Bromeliaceae diversity in western MG.
The greater richness of epiphytic Bromeliaceae in eastern MG Atlantic forest is con-
cordant with Gentry and Dodson (1987) observation that epiphytes are most diverse
in wet, middle elevation, rich-soil, tropical American forests. Reitz (1983) observed
in Santa Catarina state a drastic decrease in the total number of Bromeliaceae spe-
cies while going from east to west and attributed this pattern to the decreasing inland
temperature. Lower temperatures, particularly frost, were also recognized as a fac-
tor inXuencing Bromeliaceae species richness in Bolivia (Kessler 2002). In MG other
climatic factors, such as the distribution of rain fall, seems to be more important than
temperature in inXuencing Bromeliaceae taxa occurrence in the eastern and western
sides of the state, and also toward the north portion of the Espinhaço range. As
shown by Harley (1995) rainfall regime changes in the Espinhaço, becoming pro-
gressively sparser, the dry season is longer, and temperatures are generally greater
northward. Bromeliaceae diversity is known to peak in humid montane forests
(Ibisch 1996 apud Kessler 2002). In MG, reduced humidity could aVect Bromelia-
ceae distribution patterns since in the cerrado and caatinga few genera, usually ter-
restrial, are observed (e.g. Aechmea, Ananas, Bromelia, Dyckia, Neoglaziovia and
Pseudananas).
We identify as priorities for future scientiWc research the southern portion of the
state, as well as the Jequitinhonha, Mucuri, and Doce rivers drainage basin (e.g. B10,
B11, C11, D9, D11, E9, E10, F10) all being under collected areas that should poten-
tially present much higher Bromeliaceae diversity. The northwestern portion
(A5, B4, B5, B6, C4, C5, C6, C7, D4, D5, D6) and the Triângulo Mineiro region
(E1, E2, E3, F1, F2, F3) are also relevant for future Xoristic surveys with the family.
Bromeliaceae habitat and life form
Of the total number (283) of Bromeliaceae species known from MG, 97 (34%) are
exclusive of the Atlantic forest, 65 (23%) of the campo rupestre, and 32 (11%) can
be found in these both habitats. Thirteen taxa (4.6%) are exclusive of the cerrado,
and twelve (4.2%) inhabit both cerrado and campo rupestre (Fig. 4). Habitats
combinations involving the caatinga, Atlantic forest, cerrado, high altitude grass-
lands, and campo rupestre are less frequent and vary from one to Wve taxa (Fig. 4).
The more representative occurrence of Bromeliaceae in the Atlantic forest is con-
cordant with the extensive diversiWcation of the family in eastern Brazil, especially

3002 Biodivers Conserv (2007) 16:2989–3009
1 C
for subfamily Bromelioideae (Smith 1934; Smith and Downs 1974). Many genera are
endemic of this biome (Benzing 2000) and will occur in eastern MG reXecting the
biogeographycal continuity. In fact, 19 of the 27 genera occurring in MG belong to
subfamily Bromelioideae. Although seemingly contradictory, the Atlantic forest
shares taxa with caatinga (e.g. Alcantarea, Orthophytum), but these are lithophytic
taxa, that even while growing inside the Atlantic forest domain are exposed to
totally diVerent edaphic and microclimatic conditions.
Traditionally, the campo rupestre vegetation has been closely associated with the
cerrado domain. However the data presented here shows that a greater number of
Bromeliaceae taxa are shared between Atlantic forest and campo rupestre than
between campo rupestre and cerrado (Fig. 4). It should be emphasized, however,
that an usual view of the campo rupestre vegetation of the Espinhaço includes, as
part of this habitat, the gallery forests and forest “islands” known as capões, which
occur scattered among the grassland, and this seems to be the key to the Xoristic con-
nection between the Bromeliaceae Xora of the Atlantic forest and campo rupestre of
the southern portion of the Espinhaço.
Most of the taxa (55%) in MG exhibit terrestrial and/or saxicolous life form. Next
are those that are either ephyphytic, saxicolous and/or terrestrial (22%), followed by
obligatory epiphytic (20%). The remaining 3% have unknown life forms. The pre-
dominant terrestrial/saxicolous life form reXects the existence of entire genera in
each of the three subfamilies that grow only in these circumstances (e.g. Dyckia,
Alcantarea, Orthophytum). It is notable that the open areas, characteristic of campo
rupestre and cerrado are probably more conducive for the establishment of
Fig. 4 Total number of Minas Gerais Bromeliaceae taxa in diVerent habitats
97
65
32
23
13
12
11
7
6
5
5
2
2
1
1
1
1200 20 40 60 80 100
Atlantic forest
campo rup estre
Atlantic forest, campo rup estre
Atlantic forest, campo rup estre, cerrado
cerrado
campo rupestre, cerrado
unknown
Atlantic forest, caatinga
Atlantic forest, high altitude grassl ands
caatinga, cerrado
high altitude grasslands, campo rupestre
Atlantic forest, caatinga, campo rupes tre, cerrado
Atlantic forest, cerrado
caatinga
caatinga, campo rup estre, cerrado
high altitude grasslands
Habitat
No. of Taxa

Biodivers Conserv (2007) 16:2989–3009 3003
1 C
terrestrial/saxicolous bromeliad species due to the high degree of outcropping.
However, even within the Atlantic forest domain there are genera (i.e. Cryptanthus,
Orthophytum) totally restricted to rock outcrops or to the understory soils.
Grid cells clustering
To reduce uneven sampling, we used only those grid cells that presented at least 10
Bromeliaceae vouchers. Figure 5 presents a dendrogram based on 254 Bromeliaceae
taxa, distributed among 28 of the 74 grid cells. The dendrogram allowed us to iden-
tify two sets of areas, subdivided into seven smaller subsets. The Wrst set is the larg-
est, and is subdivided into Wve subsets, characterizing an area of similar climatic and
ecological features mainly composed by grid cells located along Atlantic forest,
campo rupestre, and cerrado. The second subset includes areas of the northern por-
tion of MG and includes subsets number six and seven, mainly covered by caatinga,
campo rupestre and cerrado. We observed that the Bromeliaceae Xora of the Cadeia
do Espinhaço (partially represented within subset 3) is more related to the Xora of
southeastern MG (subset 2), which is covered by Atlantic forest. The southernmost
area of the Espinhaço also corresponds to the inland western distribution limit for
many characteristically Atlantic forest taxa (e.g. Aechmea lamarchei, Quesnelia
strobilispica, Vriesea pardalina), that usually use the gallery forests as corridors for
dispersion, enabling them to reach isolated forested areas inside the campo rupestre.
On the other hand, the northernmost portion of the Espinhaço (subsets 5, 6, 7) is
more closely related to cerrado or caatinga areas. Climatic factor seems to be the
key to understand those connections, and increased sampling would likely reveal
important, new information regarding this. Cell C10 appears isolated from subsets
number 6 and 7. Indeed C10 presents a strong vegetation transition, hosting a pecu-
liar mix of taxa of Atlantic forest, caatinga, and cerrado biomes. Our results show a
low similarity among all the grid cells, which varied from 2 to 40%. Probably the
high number of endemic species and those with very narrow ranges contribute to
these low values. Simon and Proença (2000) obtained similar results while studying
the genus Mimosa (Leguminosae). They observed that grid cells with higher number
of collections were frequently close to each other, as neighbors, and occurred in
mountainous areas, where endemic and narrowly distributed species were frequent,
generating a low similarity among areas, regardless of their geographical proximity.
This pattern repeats here for the Bromeliaceae. As an example, a similarity of only
31% was found between cells F8 and E8, in Cadeia do Espinhaço.
Conservation of Minas Gerais Bromeliaceae taxa
Of the 283 taxa of Bromeliaceae that occur in MG, 118 are considered threatened
(critically endangered, endangered or vulnerable). One hundred taxa have a conser-
vation status of least concern, 44 lack suYcient data, and 21 were not evaluated due
to their doubtful taxonomic status (Table 2). Taxa that were only recently recorded
for MG, with probably partially known geographic distribution or represented by
few specimens from relatively under collected areas, are among those 44 taxa with
deWcient data (Table 2).
Half of the threatened taxa (59) belong to the subfamily Bromelioideae (12 criti-
cally endangered, 11 endangered, 36 vulnerable). Bromelioideae was followed by

3004 Biodivers Conserv (2007) 16:2989–3009
1 C
Pitcairnioideae subfamily, with 34 taxa threatened, and by Tillandsioideae with 25
threatened taxa (Table 2). In Bromelioideae, the genusOrthophytum has nine
vulnerable species, most of them endemic of inselbergs in northeastern MG.
Fig. 5 Floristic similarities (Jaccard) among 28 grid cells based on Bromeliaceae taxa

Biodivers Conserv (2007) 16:2989–3009 3005
1 C
A similar situation is observed for many Aechmea, Hohenbergia, and Neoregelia
species characteristic of the Atlantic forest, with taxa occurring within few forest frag-
ments or, in some cases, restricted to only one protected area. The situation for Cryp-
tanthus species is the most critical with Wve of its six species critically endangered,
mainly because they form very small populations with reduced ranges in rocky Weld.
For example, Cryptanthus warasii is only known from the type collection, whereas
C. leopoldo-horstii, only from two populations, both at road side localities.
The subfamily Pitcairnioideae has eight taxa critically endangered, eight endan-
gered, and 18 vulnerable (Table 2). Many Pitcairnioideae species presents small
appendaged seeds, unable to disperse long distances, thus favoring narrowly
endemic species (Holst 1994). The genus Dyckia deserves special attention because
of the elevate number of species, many of them poorly known, and usually with
overlapping diagnostic features, and frequently represented in herbaria by only a
single leaf and part of the inXorescence. In addition, species from the Quadrilátero
Ferrífero (e.g. Dyckia densiXora, D. elata, D. schwackeana, D. simulans) have very
restricted distributions that are currently exposed to habitat destruction by gold and
iron ore mining.
The subfamily Tillandsioideae contains three critically endangered, 10 endan-
gered, 12 vulnerable taxa. This subfamily holds the fewest critically endangered spe-
cies, and the highest percentage of least concern taxa (Table 2). Plumose appendaged
seeds allow a broader dispersion for many species of Vriesea. However, some species
like Alcantarea hatschbachii and Vriesea segadas-viannae, both endemic to the Espi-
nhaço and only known from two collections, exemplify that even inside well collected
areas there are overlooked species that deserve attention regarding conservation.
Special attention should be given to those species that form large/colorful rosettes,
such as Vriesea atropurpurea, V. bituminosa, V. crassa, V. minor and Alcantarea spp.,
because these are the ones preferred by landscape designers for use in private gar-
dens and are frequently extracted from the wild for this type of commercial use. Dry
inXorescences and fruits of Vriesea diamantinensis, V. nanuzae and V. simulans are
also locally sold in dry Xowers bouquets, compromising seed production, and thus
dispersion. Study is needed to evaluate the eVects of this kind of extractive activity on
population structure. For Tillandsia species the conservation statuses are less critical,
because many of its species have a broad distribution within the state and, in many
cases, are indigenous to other Neotropical countries.
The published red list for the Xora of MG (Mendonça and Lins 2000) lists 27
threatened Bromeliaceae species (19 endangered and 8 vulnerable). The numbers
reported in this study are greater due to the current, better knowledge about Brome-
liaceae distribution. However, as more data are generated, especially for poorly
Table 2 IUCN categor
i
es tota
l
s for M
i
nas Gera
i
s Brome
li
aceae taxa
i
n eac
h
su
b
fam
il
y
Subfamily Threatened Least
Concern Data
DeWcient Not
Evaluated
Critically
Endangered Endangered Vulnerable
Bromelioideae 12 11 36 51 16 5
Pitcairnioideae 8 8 18 10 14 8
Tillandsioideae 3 10 12 39 14 8
Total2329661004421

3006 Biodivers Conserv (2007) 16:2989–3009
1 C
known taxa, and habitat disturbance broadens, it is expected that the frequency-
level in status categories presented here will change. According to Martinelli (2000),
74% of Bromeliaceae species within the Brazilian Atlantic forest are endemic. This
vegetation is considered the diversity center of subfamily Bromelioideae (Smith and
Downs 1974; Benzing 2000) and also for many other plant groups (Mori et al. 1981).
To ensure the preservation of this biodiversity, creation of more protected areas
within the Atlantic forest should be encouraged, as this biome has been reduced to 4
% of its original coverage area within MG (Costa et al. 1998). Similar actions should
apply to the cerrado, also considered to be one of the richest Xoristic regions in the
world, and treated as the second Brazilian biodiversity hotspot with high levels of
endemism (Mittermeier et al. 2000). Currently, the cerrado is reduced to 25% of its
original area in MG (Mendonça and Lins 2000) and is suVering from increasing
human pressure, mainly from the expansion of soy bean plantations. This destruc-
tion of cerrado vegetation will likely include loss of the few recorded populations of
Billbergia meyeri and Bromelia glaziovii.
The campo rupestre, which has long provided natural grazing areas for livestock,
has special relevancy for endemic lithophytic taxa, such as the genera Cryptanthus,
Dyckia, Encholirium, and Orthophytum (Versieux and Wendt 2006). The long-held
belief that the Wre does not compromise grassland vegetation because many of its
species are adapted to this disturbance, does not seem to be true for many rocky
Weld bromeliads. During our Weld work we have returned to several places where
Wres were so intense that entire populations disappeared within a few years, regard-
less of their saxicolous life form.
Rapini et al. (2002) observed that many taxa endemic to the Espinhaço are rare
in herbaria collections due to limited blooming periods for many species, and
uneven collecting eVorts. These authors considered that few taxa would be severely
threatened, but many would be considered vulnerable, due to the destruction of
their limited habitat. Our results seem to be concordant with those of Rapini et al.
(2002), since most Bromeliaceae considered threatened here are grouped in the
vulnerable category (Table 2). Under collecting can bias results on species distribu-
tions. However, we consider ongoing Wre disturbance associated with urban devel-
opment, road construction, and mining activities to be major hazards for small
populations, and also an impediment in gaining improved knowledge of Bromelia-
ceae distributions in the Espinhaço. In regards to the Espinhaço, special attention
should be directed to taxa endemic of the Quadrilátero Ferrífero because of the
accelerated loss of habitats by urban growth and mining. Many mining companies
create private reserves as a compensatory measure for the destructive mining activ-
ities. This practice, while commendable, should be carefully managed and moni-
tored by environmental agencies so that these reserve plots are selected to better
reXect the size and vegetation type that is being consumed by mine activity. For
example, selecting an area of semi-deciduous forest for protection in compensation
for canga destruction has no equivalent value because their biotic compositions are
distinct. A conservation procedure that could be employed sometimes is to remove
individuals of Bromeliaceae and transplant them from the mining area to a similar
and protected place in the nearby or to botanic gardens. As shown by Cavallari
et al. (2006), great part of the genetic diVerence of rare rupiculous species of
Encholirium is found among individuals. Thus, every eVort to preserve individuals
and their populations are signiWcant to conserve the whole genetic variation for
some species.

Biodivers Conserv (2007) 16:2989–3009 3007
1 C
Forty four percent (124 spp.) of Bromeliaceae taxa of MG do not occur inside any
protected area (Table 1). Among these, 56 that are threatened (16 critically endan-
gered, 17 endangered, 23 vulnerable), some are only known from roadside localities
(e.g. Aechmea bambusoides) and others from areas adjacent to hydroelectricpower
plant impoundments (e.g. Aechmea bruegerii, A. purpureorosea). Nine taxa are
restricted to a single protected area (Table 1), not occurring outside their limits and
requiring special measurements (i.e., environmental education, isolation of exposed
populations) for their eVective preservation. Camargos (2001) provided a map show-
ing the distribution of all protected areas within MG that clearly indicates that the
northeastern, northwestern and the extreme western (Triângulo Mineiro) regions
are less protected. The northeastern portion of MG, particularly the Jequitinhonha
river basin, is rich in Cactaceae (Taylor and Zappi 1991), new occurrences of Brome-
liaceae (Versieux and Wendt 2006), and other saxicolous taxa characteristic of insel-
berg formations (Porembski et al. 1998), yet it is still poorly protected despite the
advanced fragmentation process (Camargos 2001; IEF-MG 1994) and the threat to
saxicolous taxa by the granite mining companies (Forzza et al. 2003). Although set-
ting targets for protected areas is not a trivial task because biodiversity represents a
continuum of ecological organization that can not be encapsulated in a single
variable (Brooks et al. 2004), we considered that these areas in northeastern MG
(i.e. Jequitinhonha and Mucuri rivers basins) should be given priority-status for
Bromeliaceae conservation. For MG, Drummond et al. (2005) presented an atlas
that mapped areas for biodiversity preservation, taking many groups of organisms
(e.g. invertebrates, mammals) into account. Nevertheless, we reiterate here that
strategies for Bromeliaceae conservation in grid cell F8 are urgent, due to the high level
of endemism and species richness that are threatened by mining and deforestation.
Conclusions
This paper provides an updated analysis of Bromeliaceae distribution within MG
state, using maps containing 1° £ 1° grid cells. A clear decrease in Bromeliaceae
diversity is observed when going from the eastern to the western sides of the state.
Within MG, we observed that areas covered by the Atlantic forest present greater
bromeliad diversity than areas of caatinga, and cerrado vegetation. Also, the Atlantic
forest vegetation shares species with the campo rupestre of the southernmost por-
tion of the Cadeia do Espinhaço. The most diverse grid cells are located in elevated
areas in the southern Espinhaço, or in the Serra da Mantiqueira, in the southeastern
portion of the state. Low Xoristic similarities, favored by the very narrowly distrib-
uted and endemic taxa, were found between neighboring grid cells, even for areas
located along the same mountain chain. Strategic taxonomic surveys and inventories
are required for many grid cells that did not present any Bromeliaceae record, or
that showed low values, particularly those located along the northeastern and
extreme southern portions, where new records for the state are expected, since these
areas are within the Atlantic forest domain. The Cadeia do Espinhaço is considered
the most important area of endemism for Bromeliaceae, sheltering 62% of all the
endemic taxa. From the 283 taxa of Bromeliaceae that occur within MG, 118 were
considered threatened and 124 taxa do not occur inside any protected area. Many
taxa are still poorly known and therefore categorized as data deWcient or were not
evaluated. Changes are expected to occur in this classiWcation, as new population

3008 Biodivers Conserv (2007) 16:2989–3009
1 C
and distribution data become available because taxa considered to be narrowly
endemic can present broader ranges. Immediate actions to protect the Bromeliaceae
Xora of the Quadrilátero Ferrífero in the southern portion of the Cadeia do
Espinhaço are needed, since the area suVers with accelerated urban growth and
mining activities.
Acknowledgements This paper represents a part of the M.Sc. dissertation of LMV undertaken at
the Graduate Program in Botany of the Universidade Federal do Rio de Janeiro/Museu Nacional.
We acknowledge support from CAPES, CNPq, NSF (DEB 0129446), Marie Selby Botanical Gar-
dens, and Smithsonian Women Committee. We gratefully acknowledge the Instituto Estadual de
Florestas de Minas Gerais and IBAMA for permission to collect, Alice Calvente and Dr. Gregory
Brown for review of the English and suggestions to the manuscript, two anonymous reviewers for
their constructive criticism and suggestions and the curators, keepers, and plant collectors of the cited
herbaria for access to the specimens and data.
References
Benzing DH (2000) Bromeliaceae: proWle of an adaptive radiation. Cambridge University Press,
Cambridge
Brooks TM, Fonseca GAB, Rodrigues ASL (2004) Protecte d areas and species. Conserv Biol 18:616–
618
Calvente AM, Freitas MF, Andreata RHP (2005) Listagem, distribuição geográWca e conserva ção das
espécies de Cactaceae no estado do Rio de Janeiro. Rodriguésia 56:141–162
Camargos RMF (2001) Unidades de conservação em Minas Gerais: levantamento e discussão.
Fundação Biodiversitas, Belo Horizonte
Cavallari MM, Forzza RC, Veasey EA, Zucchi MI, Oliveira GCX (2006) Genetic variation in three
endangered species ofEncholirium (Bromeliaceae) from Cadeia do Espinhaço, Brazil, detected
using RAPD markers. Biodivers Conserv 15:4357–4373
Costa CMR, Herrmann G, Martins CS, Lins LV, Lamas IR (orgs.). (1998) Biodiversidade em Minas
Gerais: um atlas para sua conservação, 1a ed. Fundação Biodiversitas, Belo Horizonte
Dahlgren RMT, CliVord HT, Yeo PF (1985) The families of the Monocotyledons: structure, evolu-
tion and taxonomy. Springer-Verlag, Berlin
Drummond GM, Martins CS, Machado ABM, Sebaio FA, Antonini Y (orgs). (2005) Biodiversidade
em Minas Gerais: um atlas para sua conservação, 2a ed. Fundação Biodiversitas, Belo Horizonte
Forzza RC, Christianini AV, Wanderley MGL, Buzato S (2003) Encholirium (Pitcairnioideae –
Bromeliaceae): conhecimento atual e sugestões para conservação. Vidalia 1:7–20
Gentry AH, Dodson CH (1987) Diversity and biogeography of Neotropical vascular epiphytes. Ann
Mo Bot Gard 74:205–233
Giulietti AM, Menezes NL, Pirani JR, Meguro M, Wanderley MGL (1987) Flora da Serra do Cipó,
Minas Gerais: caracterização e lista das espécies. Boletim de Botânica da Universidade de São
Paulo 9:1–151
Giulietti AM, Harley RM, Queiroz LP, Wanderley MGL, Van Den Berg C (2005) Biodiversity and
conservation of plants in Brazil. Conserv Biol 19:632–639
Givnish TJ, Millam KC, Evans TM, Hall JC, Pires JC, Berrie PE, Sytsma KJ (2004) Ancient vicariance
or recent long-distance dispersal? Inferences about phylogeny and South American-African
disjunctions in Rapateaceae and Bromeliaceae based on ndhF sequence data. Int J Plant Sci 165
(4 Suppl.):S35-S54
Harley RM (1995) Introduction. In: Stannard BL (ed), Flora of Pico das Almas, Chapada Diamanti-
na, Bahia, Brazil. Royal Botanic Gardens, Kew, pp 1–40
Holmgren PK, Holmgren NK, Barnett LC (1990) Index Herbariorum Part 1 The Herbaria of the
World, 8th edn. New York Botanical Garden Bronx, New York
Holst BK (1994) Checklist of Venezuelan Bromeliaceae with notes on species distribution by state
and levels of endemism. Selbyana 15:132–149
IEF-MG - Instituto Estadual de Florestas de Minas Gerais (1994) Mapa de cobertura vegetal e uso
do solo do Estado de Minas Gerais. IEF-MG, Belo Horizonte, Brasil
IUCN (2001) IUCN Red list categories and criteria: Version 3.1 IUCN Species Survival Commission.
IUCN, Gland Switzerland and Cambridge, UK

Biodivers Conserv (2007) 16:2989–3009 3009
1 C
Kessler M (2002) Species richness and ecophysiological types among Bolivian bromeliad communi-
ties. Biodivers Conserv 11:987–1010
Kress WJ, Heyer WR, Acevedo P, Coddington J, Cole D, Erwin TL, Meggers BJ, Pogue M,
Thorington RW, Vari RP, Weitzman MJ, Weitzman SH (1998) Amazonian biodiversity: assessing
conservation priorities with taxonomic data. Biodivers Conserv 7:1577–1587
Luther HE (2004) An alphabetical list of bromeliad binomials, 9th edn. The Bromeliad Society Inter-
national, Inc., Orlando, Florida, USA
Martinelli G (1989) Campos de Altitude. Editora Index, Rio de Janeiro
Martinelli G (2000) The bromeliads of the Atlantic forest. ScientiWc American, March:86–93
Mendonça MP, Lins LV (orgs) (2000) Lista vermelha das espécies ameaçadas de extinção da Xora de
Minas Gerais. Fundação Biodiversitas e Fundação Zoo-Botânica de Belo Horizonte, Belo Hori-
zonte
Mittermeier RA, Gil PR, Mittermeier CG (2000) Hotspots: earth’s biologically richest and most
endangered terrestrial ecorregions. Cemex, México
Mori SA, Boom BM, Prance GT (1981) Distribution patterns and conservation of eastern Brazilian
coastal forest tree species. Brittonia 33:233–245
Morton JK (1972) Phytogeography of the West African mountains. In: Valentine DH (ed) Taxon-
omy, phytogeography and evolution, Academic Press, London, pp 221–236
Pirani JR, Giulietti AM, Mello-Silva R, Meguro M (1994) Checklist and patterns of geographic
distribution of the vegetation of serra do Ambrósio, Minas Gerais, Brazil. Rev Bras Bot 17:133–147
Porembski S, Martinelli G, Ohlemüller R, Barthlott W (1998) Diversity and ecology of saxicolous
vegetation mats on inselbergs in the brazilian atlantic rainforest. Divers Distrib 4:107–119
Prance GT (2001) Discovering the plant world. Taxon 50: 345–359
Rapini A, Mello-Silva R, Kawasaki ML (2002 ) Richness and endemism in Asclepiadoideae (Apocyn-
aceae) from the Espinhaço range of Minas Gerais, Brazil – a consevationist view. Biodivers
Conserv 11:1733–1746
Reitz R (1983) Bromeliáceas e a malária-bromélia endêmica, Herbário Barbosa Rodrigues, Itajaí,
Santa Catarina, Brasil
Rizzini CT (1997) Tratado de WtogeograWa do Brasil. Âmbito Cultural Edições, Rio de Janeiro
SaVord HD (1999) Brazilian Páramos I. An introduction to the physical environment and vegetation
of the campos de altitude. J Biogeogr 26:693–712
Serrato A, Ibarra-Manríquez G, Oyama K (2004) Biogeography and conservation of the genus Ficus
(Moraceae) in Mexico. J Biogeogr 31:475–485
Simon MF, Proença C (2000) Phytogeographyc patterns of Mimosa (Mimosoideae, Leguminosae) in
the savanna biome of Brazil: an indicator genus of high-altitude centers of endemism? Biol
Conserv 96:279–296
Smith LB (1934) Geographical evidence on the lines of evolution in the Bromeliaceae. Botanische
Jahrbücher für Systematic PXanzengeschichte und PXanzengeographie 66:446–468
Smith LB, Downs RJ (1974) Pitcairnioideae (Bromeliaceae) Flora Neotropica Monograph No 14
Part 1. Hafner Press, New York
Taylor NP, Zappi DC (1991) Cactaceae do vale do rio Jequitinhonha. Acta Bot Brasilica 5:63–69
Versieux LM, Wendt T (2006) Checklist of Bromeliaceae of Minas Gerais, Brazil, with notes on
taxonomy and endemism. Selbyana 27:107–146
Wanderley MGL (1992) Estudos taxonômicos no gênero Xyris L. (Xyridaceae) da Serra do Cipó,
Minas Gerais, Brasil. Ph.D thesis, Universidade de São Paulo, São Paulo