Vegetative growth variables

Total plant height (AP) was measured from the base of the pseudobulb up to the apex of the largest leaf; the total number of leaves (NTH), pseudobulbs (NTPsb), and shoots (NTBr); as well as diameter (DPsb) and length of the pseudobulb (LPsb) and leaf length (LH) of the shoot induced and not induced to flowering. Variables were weekly measured.

Flowering variables

1) Days to anthesis of the first flower (DA); 2) flower life (VF); 3) number of flowers per raceme (NF); 4) flower stalk length (LTF), from the first node of the stalk, next to the apex of the pseudobulb, up to the start of the raceme; 5) flower stalk diameter (DTF) measured in the first node, next to the apex of the pseudobulb; 6) flower size, which included the length from the apex of the superior sepal to the apex of the inferior sepal (LS), as well as the width, from the apex of the left petal to the apex of the right petal (LP); 7) labellum size, which include its length (LLab) and width (ALab).

Supplementary variables

1) Probability model of the flower induction, according to the effect of light intensity (in both controlled environments) and the CFA (in the three environments), by number of induced plants in each treatment; 2) percentage of flower induction; 3) vegetative and reproductive growth variables correlation.

Data analysis

To evaluate the relationships among the response variables for vegetative growth and flowering, we performed an ANOVA on the data and compared their means using Tukey test (p≤0.05). For which, SAS Enterprise Guide was used (^{SAS Institute, Inc., 2009}). The blooming induction probability in the three environments was modeled with logistic regression. We used the Pearson correlation coefficient to estimate the relation between the variables of vegetative growth vs. flowering variables, and the correlation between the flowering variables.

Plant height

The plants height of induced and non-induced shoots during the experiment was significantly different between the controlled environment and the greenhouse: CFA plants in greenhouse were higher, in relation to the AInv+SFA and AC2+SFA (induced plants), AC1+CFA and AC2+CFA (non induced plants) treatments (p≤0.05; Figure 1).

Mean values with different letters on the bars indicate significant differences between treatments (Tukey, p≤0.05).

AInv: greenhouse environment (natural light with 90 μmol m^-2 s^-1 average light intensity), AC1: controlled environment 1 (HPS + MH + FL with 80 μmol m^-2 s^-1 light intensity), AC2: controlled environment 2 (LED with 64 μmol m^-2 s^-1 light intensity), CFA: plants with previous flowering, and SFA: plants without previous flowering.

Figure 1 Plant height of L. anceps subsp. anceps with vegetative shoots induced and not induced to flowering, in different cultivation environments and with different flowering background, at the end of the experimental cycle (60 d). 

These results show that in the analyzed controlled environments, plants that bloomed the previous year exhibited lower height, which can be one of the causes of the blooming non-induction in our research. Also, the height of the L. anceps subsp. anceps plants induced or not to flowering in a controlled environment (LED light or combination of HPS + MH + FL) was equal to or less than the plants in the greenhouse, where the CFA factor is an important supplement to the obtained results from the same lighting source. Our results contrasted with those reported by ^{Terfa et al. (2012)}, who analyzed quality parameters of potted roses (Rose×hybrid ‘Toril’) grown in 100 μmol m^-2 s^-1 light intensity with different light qualities provided by LEDs (80 % red and 20 % blue) and HPS lamps, in closed growth chambers and in greenhouse. The length of the stem and the pedicel were significantly shorter in plants that grew under LED light compared to those on HPS lighting, although the total production period did not change. ^{Yamada et al. (2008)} used different light sources: red (660±30 nm) and far red (730±30 nm), all with 5 μmol m^-2 s^-1 FFF, in Eustoma grandiflorum ‘Neilpeach Neo’. They found that the stem elongation rate was higher in plants grown under far red fluorescent lamps, it was similar under fluorescent and incandescent lamps and lower under daylight like fluorescent lamps compared with the control (without additional lighting). But in our study there was greater FFF (64 μmol m^-2 s^-1) with LED light, with a red light component, and the height of the plants was reduced compared to the greenhouse plants (without supplementary red light). The nature of the studied species promotes different responses.

^{Heo et al. (2006)} used different mixes of LED type (with FFF=90 ± 10 μmol m^-2 s^-1) radiation in the growth of agerato (Ageratum houstonianum Mill. cv. Blue Field), marigold (Tagetes erecta L. cv. Orange Boy) and sage (Salvia splendens E Sello ex Ruem & Schult. cv. Red Vista). They found that the stem length was greater (257, 147 and 140 %, respectively) with a far red light component combined with blue or red light, compared to the fluorescent light (control). Our results differed because the plant height growth was lower when using LED lighting (red and blue light combination).

Number of growth organs

We report our results at 60 d because weekly samplings did not show significant differences. The orchid growth is slow and the experimental period here analyzed was short to appreciate differences before the evaluated period.

The NTH, NTPsb and NTBr in plants induced to flowering showed no significant differences (p>0.05) among the environments (Figure 2A). On plants not induced to flowering, the AC2 + CFA treatment had a negative effect on the number of leaves, pseudobulbs and shoots (p≤0.01) compared with other treatments (Figure 2B). As for the total plant height, the CFA plants cultivated in the AC2 (with LED light) decreased the growth organs production, which resulted in the absence of flower induction. This may indicate that for L. anceps subsp. anceps the light intensity threshold for growth promoting induction could be of more than 60 μmol m^-2 s^-1, but less than 120 μmol m^-2 s^-1, because the plant leaves turn yellow, wither and fall when grown in a greenhouse with higher light intensity.

Mean values with different letters in bars show significant differences between treatments (Tukey, p≤0.05). AInv: environment greenhouse (natural light with 90 μmol m^-2 s^-1 average light intensity), AC1: controlled environment 1 (HPS + MH + FL with 80 μmol m^-2 s^-1 light intensity), AC2: controlled environment 2 (LED with 64 μmol m^-2 s^-1 light intensity), CFA: plants with previous flowering, and SFA: plants without previous flowering.

Figure 2 Vegetative growth of L. anceps subesp. anceps with induced shoots (A) and not induced shoots (B) to flowering, in different cultivation environments and with different flowering background at the end of the experimental cycle (60 d). 

Flower induction

The regression analysis used to model the flower induction showed that in the controlled environments and in the greenhouse the flowering background significantly affected the induction (Table 2). This result is consistent with the percentages of induction in different environments because the highest corresponded to the treatments where the plants had previous flowering, compared to those that did not (Figure 3). In controlled environments (AC1+CFA and AC2+CFA with 50 and 35 % flower induction, and AC+SFA and AC2+SFA, with 15 % induction) the percentages were inferior compared to the greenhouse percentages. Even so, plants in the greenhouse did not show a 100 % of flower induction (90 % in CFA plants and 35 % in SFA plants).

AInv: greenhouse environment (natural light with 90 μmol m^-2 s^-1 average light intensity), AC1: controlled environment 1 (HPS + MH + FL with 80 μmol m^-2 s^-1 light intensity), AC2: controlled environment 2 (LED with 64 μmol m^-2 s^-1 light intensity), CFA: plants with previous flowering, and SFA: plants without previous flowering.

Figure 3 Percentage of induced plants to flowering of L. anceps subsp. anceps in different cultivation environments and with different flowering background. 

The growth results showed above indicated that the non-induced plants in AC2, in combination with CFA plants, mainly, reduced the vegetative growth of the orchid under study. The results show that induced plants and with previous flowering have higher probability to induce flowering the next year (Table 2). In such situation, although the light intensity was not a significant factor in the probability of the flower induction, each environment (with different light intensities) in combination with the previous flowering suggests a directly proportional effect on the induction percentage (to higher light intensity in the environment, higher induction percentage) (Figure 3). The above shows that SFA plants in controlled environments might have a reproduction cost on the current induction associated with the light intensity.

Based on the growth decrease in plants of AC2 treatment (lower light intensity), it suggests that the plant size could be another factor that participated in the flower induction. This idea is substantiated by ^{Jacquemyn et al. (2010)} study; they pointed out that Orchis purpurea plants in both shaded and with light environments need to have a certain size to start flowering. Nevertheless, this size threshold was almost three times larger in shaded environments than in the light.

Flowering variables

Differences were only in the days to anthesis (DA) of the first flower when the treatments AInv+CFA and AC2+CFA promoted longer times to anthesis. These had an average of 47 and 17 d more than the other treatments. However, only AInv+CFA was statistically different (p ≤ 0.01) to the AInv+SFA, AC1+CFA and AC1+SFA treatments (Table 3). The days to anthesis were shortened in AC1 by 45 d, with and without previous flowering, compared to those of the greenhouse. Also, AC2+CFA promoted less time to anthesis (30 d) in relation to the AInv treatment. ^{Paradiso and De Pascale (2014)} mention that greater light intensity (350 μmol m^-2 s^-1, where 150 μmol m^-2 s^-1 were added light) favored the inductive phase in Phalaenopsis hybrids which reduced from 51 to 41 d the time for the floral stem emergence in all plants; also, it anticipated the first flower anthesis by 6 d, compared to the reference light (200 μmol m^-2 s^-1). According to ^{Runkle (2010)}, once an inflorescence has begun, the time to anthesis of the first flower depends on its variety and the average temperature. In our study, it is important to highlight that, most likely, a constant temperature or constant inductive photoperiods in controlled environments (9/15 h photoperiod and 25/13 °C day/night temperature) also promoted a faster time to anthesis. After induction, all plants developed in the same environment.

The other analyzed flower development variables there were no treatment effect. There was induction in the AC2 + SFA treatment, but flower stalks aborted shortly after exceeding 5 cm long (size considered as the inductive threshold in our study) (Table 3).

Some of the most outstanding features of the flower development were observed in individual flowers of each treatment (Figure 4).

Figure 4 Flowers of L. anceps subsp. anceps in different cultivation environments and with different flowering backgrounds. A) AC1 + CFA, B) AC2 + CFA, C) AInv + CFA, D) AC1 + SFA and E) AInv + CFA. 

Correlation coefficient

The vegetative variables correlated with the blooming variables of the plants in a controlled environment were different from the correlations of plants in the greenhouse (Table 4). Negative correlations between some vegetative and flower growth characteristics suggest that under experimental conditions, if the plant invests more energy in vegetative growth, decreases flower growth and the number of flowers. On the contrary, the higher the vegetative shoot growth is, which is responsible for current year blooming, it promotes greater flower stem growth (positive correlation) (Table 4). In the greenhouse conditions, more height growth in the plant promotes greater time to anthesis (positive correlation) (Table 4), as shown in Table 3. ^{Amin et al. (2004)} analyzed the correlation coefficients of nine characters for six species of native epiphytic orchids from Bangladesh (Aerides odoratum, A. multiflorum, A. jaintapuri, Rhynchostyles retusa alba, R. retusa pink and Vanda teres). They found positive correlations in 20 out of 36 associations. Some of our data have partial similarity with that study: they found significant negative correlation between the number of leaves and the number of flowers. Also, non-significant correlation between the height of the plant and the length of the inflorescence were found (but not with the anthesis like in our research).

The largest number of significantly correlated blooming characters is in greenhouse-induced plants, compared to the ones in controlled environments. With the exception of one correlation (LP with ALab), the other plants characters from the evaluated environments, such as flower size (LS and LP) and the size of labellum (LLab and ALab) had positive correlations among themselves, and suggests that this is due to the bilateral symmetry characteristic (zigomorphic) of orchids (^{Davenport and Kohanzadeh, 1982}). These four characters were also positively correlated with the number of flowers (NF) in greenhouse-induced plants, but not in the controlled environments (Table 5).

In addition, plants induced in the controlled environment showed significant positive correlations between the length of the flower stalk (LTF) with its own diameter (DTF) and the days to anthesis (DA) (Table 5). Based on above and the flowering of the species under study during the experiment, it was considered that LTF could be correlated with DA in induced greenhouse plants, because at greater length, greater the time to anthesis is. However, there was no significant correlation between these characters, probably due to the variation in the stems size in these plants in the greenhouse. Regard to no significant values, the data is supported in Table 3 in relation to DA. These data are partially inconsistent with the data obtained by ^{Moniruzzaman et al. (2012)} for Dendrobium orchids. The length of the flower stem and the length of the inflorescence correlate with the number and life of flowers. Our research suggests that the stem length characteristic can be standardized with the influence of the controlled environment in the induced plants.

Throughout the four years of observations of the L. anceps subsp. anceps population, under experimental conditions in greenhouse, there are flowering variations (flower induction, flower stem size, flower size, days to anthesis), as its management (depending on experimental conditions) or the environmental factors changed (high or low temperature; highly rainy years which promoted overcast cloudiness and low light intensity), although, the same prevailing photoperiod. Despite multiple factors, our research shows important findings for the induction and flower development of the subjects under study, such as their reduction in the days to anthesis. This also highlights the development the flowers without abnormalities, after their removal from the chambers at a specific time once the flower induction had happened.