A SEARCH FOR CARBON-CHAIN-RICH CORES IN DARK CLOUDS

For all of the detected lines, peak antenna temperature, line width, and LSR velocity are derived by fitting the Gaussian profile to each spectrum, which are summarized in Table 4. For the NH₃ lines toward L429-1 and L483, we derived the LSR velocities and the line widths assuming that they are common for all the hyperfine components. Because of the insufficient spectral resolution of 37 kHz, which is larger than most of the hyperfine splittings due to the H nuclei in the NH₃ (1,1) line, we fitted only the five major spectral components as labeled in Table 1 (Ungerechts et al. 1980). The line widths and LSR velocities are slightly different from molecule to molecule for some sources. This is partly due to the hyperfine structure of the HC₃N and NH₃ lines, whose line widths tend to be broader than those of the other molecules.

Using these line parameters, column densities of CCS, HC₃N, HC₅N, and NH₃ were calculated by the similar method described by Suzuki et al. (1992). In addition, we calculated column densities of NH₃ for cores that we did not observe, using the results of Benson & Myers (1989) if they are available. The dipole moments of CCS, HC₃N, HC₅N, and NH₃ were assumed to be 2.81, 3.72, 4.33, and 1.46 Debye, respectively (Murakami 1990; Lafferty & Lovas 1978; Alexander et al. 1976; Cohen & Poynter 1974). The LTE condition was assumed, and the excitation temperatures were fixed to 5 K for CCS and 6.5 K for HC₃N, HC₅N, and NH₃. In this paper, we do not use the results of J_N = 2₁–1₀ lines of CCS, because the beam size at the 22 GHz band, 74'', is almost comparable to the typical size of the observed dark cloud cores as shown in Figure 4, and the derived column densities may be affected due to the unknown source coupling efficiency.

Zoom In Zoom Out Reset image size

The derived column densities are summarized in Table 5. For L492, line parameters and column densities are taken from Hirota & Yamamoto (2006). Calculated optical depths, τ, of the observed lines were 0.17–3.4, 0.07–8.8, and 0.15–0.65 for the CCS, HC₃N, and HC₅N lines, respectively. The CCS and HC₅N lines were found to be optically thin (τ < 1), except for the CCS line toward L492 (τ = 3.4; Hirota & Yamamoto 2006). On the other hand, the HC₃N lines are optically thick toward L1527 (τ = 2.4), L1512 (τ = 1.5), B68 (τ = 1.0), L492 (τ = 8.8; Hirota & Yamamoto 2006), and L483 (τ = 1.5). For L492, the optical depth and excitation temperatures were derived from the intensity ratio of the hyperfine components of the HC₃N line (Hirota & Yamamoto 2006).

In order to estimate uncertainties in the derived column densities, we calculated the column densities of CCS, HC₃N, and HC₅N by varying the assumed excitation temperature of ±1 K. We found that the uncertainties in the derived column densities were typically 10%–30%. In addition, we compared column densities of some of the sources where the CCS lines were observed previously. We found that the column densities of CCS toward these sources differ by a factor of 0.4–3 mostly due to the difference in the observed positions, as in the case of L1512 and B68, for instance. Considering these results, we estimate that the uncertainties in the derived molecular column densities are typically about 30%–40%, while it could be as high as a factor of 3 in the worst case.

For several sources where the CCS and HC₃N lines are intense, we obtained integrated intensity maps as shown in Figure 4. Detailed descriptions for individual sources are summarized in Appendix A. The CCS and HC₃N maps show similar morphology (Hirahara et al. 1992; Tafalla et al. 2006), although some of them show slight differences in their sizes and peak positions. It is well known that CCS and some carbon-chain molecules are depleted at the central part of a starless dense core, and their distributions show a central hole (Velusamy et al. 1995; Kuiper et al. 1996; Ohashi et al. 1999; Lai & Crutcher 2000; Lai et al. 2003). We can see such an apparent hole only in the CCS map of L1172D, possibly due to the lack of significant depletion of CCS and HC₃N in other mapped sources. However, it is also likely that we could not trace such a small structure in the cores because of the large beam size and the coarse grid spacing of our mapping observations, 40'', which are comparable to or slightly larger than the expected size of the hole. In fact, the central hole of CCS in B335 was not detected by the single dish observation, but was first imaged by the interferometric observation using the VLA (Velusamy et al. 1995) with the spatial resolution of 12''. Further high-resolution observations would settle this issue.

In the present survey, we observed the CCS, HC₃N, and HC₅N lines toward 40 dark cloud cores, and found several dense cores with high column densities of carbon-chain molecules and/or extremely low NH₃/CCS ratios. Here we tentatively define the CCPR as the core with the NH₃/CCS ratio of less than 10, which can easily be distinguished from other cores in Table 5 and Figure 5 as discussed later. Based on this criterion, we can identify five new candidates for CCPRs, L492, L1517D, L530D, L1147, and the northeast clump of L1172B, as well as four known CCPRs, L1495B, L1521B, L1521E, and the cyanopolyyne peak of TMC-1 (Suzuki et al. 1992). As briefly described in Appendix A, the chemical compositions of the newly found sources show a mutual variation, and are different from originally identified CCPRs. Here we classify the newly found candidates into two groups and summarize their properties below.

Zoom In Zoom Out Reset image size

One of the most outstanding source is L492, where the carbon-chain molecules have high column densities comparable to the known CCPRs. The high abundances of carbon-chain molecules, the lack of significant depletion, and low deuterium fractionation all suggest chemically less evolved nature of L492 (Hirota & Yamamoto 2006; Appendix A.8). Contrary to the other CCPRs, the NH₃ lines are also detected in L492 with moderate intensities. Furthermore, L492 is thought to be dynamically evolved according to a strong infalling signature found in the spectral lines (Lee et al. 2001). Detailed discussions in Hirota & Yamamoto (2006) concluded that L492 is chemically and dynamically evolved than other CCPRs, while younger than prestellar cores such as L1498 and L1544 (Kuiper et al. 1996; Ohashi et al. 1999). This order does not always mean the absolute age, because timescales of contraction, gas-phase chemistry, and depletion would be different from source to source.

The other group of the sources with remarkably low NH₃/CCS ratios are the starless cores L1517D, L530D, and L1147, as well as the northeast clump of L1172B. In these sources, the NH₃/CCS ratios are comparable to those of other known CCPRs, although column densities of both CCS and NH₃ are generally lower than the known CCPRs as can be seen in Table 5. Therefore, they are possible candidates for CCPRs. In the case of L1147, the weak NH₃ line is detected toward the dust continuum peak, and the CCS peak coincides with the same position (see Appendix A.12), which results in the low NH₃/CCS ratio toward the central part of the dense core as in the case of L492. However, the column densities of both CCS and NH₃ in L1147 are one order of magnitude lower than those of L492. Because the dust continuum emission of L1147 is weaker than the other cores (Kirk et al. 2005), it is also likely that the column density of H₂ is also less than L492. On the other hand, in L1517D and L530D, the peak positions of CCS are located at the edge of dense cores, which would be a reason for the lower column densities of the observed molecules (see Appendices A.4 and A.11). Therefore, it cannot be ruled out the possibility that the CCS line in L1517D and L530D would trace the outer part of the dense core, as can be seen in L1498 and L1544 (Kuiper et al. 1996; Ohashi et al. 1999). For L1172B, dust continuum maps are not available in our knowledge, and hence, we will not discuss further the relationships between the CCS, HC₃N, NH₃, and dust continuum emission (see Appendix A.14).

In summary, L492 has been confirmed as a definitive CCPR that is found for the first time outside the Taurus molecular cloud, and the others are possible candidates for CCPRs, for which further high-resolution mapping observations of carbon-chain molecules and NH₃, together with the dust continuum emission, are necessary for confirmation.

According to our results, overall detection rates of the observed molecular lines are 43% (17/40), 43% (17/40), 14% (5/35), and 62% (24/39) for CCS, HC₃N, HC₅N, and NH₃, respectively, which are slightly lower than those of Suzuki et al. (1992), 55%, 67%, 31%, and 80%, respectively, as listed in Table 6. Although we selectively observed the cores with low abundance of NH₃, based on the anticorrelation between carbon-chain molecules and NH₃ (Suzuki et al. 1992), the detection rates of carbon-chain molecules are not so high as expected. Most of the observed cores with weak NH₃ emission, except for five newly found candidates as mentioned in the previous section, are turned out to have "normal" NH₃/CCS ratios and hence, would be just low column density objects. As mentioned previously, none of the observed lines were detected toward some of our samples probably due to offsets from the real emission peak. Otherwise, they would have less densities so that the NH₃ lines along with those of carbon-chain molecules are hardly excited. These results mean that the CCPRs are intrinsically rare even in the Taurus region where four CCPRs have already been identified.

Table 6. Detection Rate for CCS, HC₃N, HC₅N, and NH₃

			CCS		HC3N		HC5N		NH3
Sample	Total	Star-forming Cores	Detect/Observed	Ratio	Detect/Observed	Ratio	Detect/Observed	Ratio	Detect/Observed	Ratio	$\frac{\mbox{Num. of CCS Cores}}{\mbox{Num. of NH}_{3}\mbox{ Cores}}$
Suzuki et al. (1992)a	49	22	27/49	0.55	33/49	0.67	14/49	0.29	39/49	0.80	0.69
Present study	40	9	17/40	0.43	17/40	0.43	5/35	0.14	24/39	0.62	0.71
Compiled ensemble	90	34	48/90	0.53	46/83	0.55	18/80	0.23	66/89	0.74	0.73
Taurus	29	9	18/29	0.62	19/29	0.66	12/29	0.41	20/29	0.69	0.90
Ophiuchus	24	9	6/24	0.25	5/21	0.24	0/22	0.00	18/24	0.75	0.33
Other than Taurus	61	25	30/61	0.49	27/54	0.50	6/51	0.12	37/60	0.62	0.81
Pipe Nebulab	46	1	13/46	0.28	...	...	4/46	0.09	29/46	0.63	0.45
Perseusc	193	>44	96/193	0.51	...	...	...	...	162/193	0.84	0.59

Notes. ^aTwo sources in the Orion region, L1641N and NGC2071N, are included, which are not listed in Table A.1. ^bRathborne et al. (2008). We identified the Pipe core 12 as the only site of star formation (Brooke et al. 2007). ^cRosolowsky et al. (2008). We identified the star-forming cores according to Enoch et al. (2006), which are labeled as "Bolocam" sources in Rosolowsky et al. (2008). We tentatively regard the sources other than "Bolocam" in Rosolowsky et al. (2008) as the starless cores. Therefore, the number of star-forming cores in the Perseus gives a lower limit.

Download table as:  ASCIITypeset image

As discussed in the previous section, we calculated the column densities of CCS, HC₃N, HC₅N, and NH₃ by the similar method described by Suzuki et al. (1992). This enables us to discuss statistically about the carbon-chain chemistry in dark cloud cores by combining our results with those of previous surveys (Suzuki et al. 1992; Benson & Myers 1989; Benson et al. 1998; Hirota et al. 2001) supplemented with several individual observations (e.g., Hirota et al. 2002, 2004), which were made by almost the same method of observations and data analysis. Details of our samples are discussed in Appendix B. We compiled a sample of nearby dark cloud cores consisting of 90 sources in total for the further statistical studies. The total detection rates are 53%, 55%, 23%, and 74% for CCS, HC₃N, HC₅N, and NH₃, respectively.

As noted in the Introduction, the CCPRs have so far been detected only in the Taurus molecular cloud (Suzuki et al. 1992). This result would suggest a difference in the overall chemical compositions from cloud to cloud. Hence, it is worthy to compare the detection rates of carbon-chain molecules between the Taurus and other regions in order to investigate whether such a chemical abundance variation occurs. Because the numbers of available samples are sufficient only for the Taurus and Ophiuchus cores (∼30 each), we here summarize the results for these two regions in Table 6. It is easily found that the detection rates of carbon-chain molecules are significantly lower in the Ophiuchus cores than in the Taurus cores. A ratio of the number of the CCS cores relative to that of the NH₃ cores in the Taurus region is three times higher than that of the Ophiuchus region, as listed in Table 6. Since our sample was not prepared in an unbiased way, these results are not statistical in a strict sense. Nevertheless, we can extract "statistical" trends from the results because the sample involves a number of sources with and without star-forming activities. Therefore, it is very likely that carbon-chain molecules are generally less abundant in the Ophiuchus cores than in the Taurus cores.

We also compared our results with those of recent complete survey of CCS and NH₃ in the Pipe Nebula (Rathborne et al. 2008) and the Perseus molecular cloud (Rosolowsky et al. 2008), as listed in Table 6. Detection rates of the observed molecules are different from region to region. In the Pipe Nebula, the detection rate of CCS is lower than that in the Taurus region and is comparable to that in the Ophiuchus region, while the detection rate of NH₃ is almost the same as those in the other regions. On the other hand, the detection rate of CCS in the Perseus region is rather close to that in the Taurus region, and the detection rate of NH₃ in the Perseus region is the highest among our four sampled regions. Although the methods of the surveys are different from ours in various ways, such as observed transitions, sensitivities, spectral and spatial resolutions, methods of data analysis, and definition of star formation activities (association of IRAS and/or Spitzer sources), the possible difference in the ratio of the number of the CCS cores relative to that of the NH₃ cores can be seen in Table 6; the ratio in the Taurus region (90%) is nearly 3, 2, and 1.5 times higher than that in the Ophiuchus region, the Pipe Nebula, and the Perseus region, respectively, which would be indicative of difference in an average chemical age of cores in each region, as discussed later (Suzuki et al. 1992).

In order to investigate a source-to-source variation in molecular abundances, we focus on the abundance ratio of NH₃/CCS, which is recognized as a useful probe of the chemical evolutionary stage of dark cloud cores (Suzuki et al. 1992). Following the discussion by Ohishi & Kaifu (1998), we investigate the relationship between the CCS abundance and the NH₃/CCS ratio in Figure 5. We plotted the sources where either the CCS or NH₃ line is detected (Table A.1). Because of the lack of a complete data set for the column densities of H₂ for all the cores, we plot the column densities of CCS instead of the fractional abundance of CCS relative to H₂. Therefore, a scatter in the column density of CCS is not only due to the variation in the fractional abundance of CCS relative to H₂, but is also attributed to the difference in the column density of H₂.

In Figure 5, we can find CCPRs (L1495B, L1521B, L1521E, the cyanopolyyne peak of TMC-1, and L492) at the upper left part, and the CCPR candidates (L1517D, L530D, L1147, and the (−40'', −40'') and (+40'', +40'') positions in L1172B) at the middle left part. We stress that there is no carbon-chain-rich cores in the Ophiuchus region as depicted in Figure 5(d), even though we have almost the same number of samples as in the Taurus region. Such a systematic abundance variation from cloud to cloud has been found in the deuterium fractionation ratios of DNC/HN¹³C (Hirota et al. 2001) and N₂D⁺/N₂H⁺ (Crapsi et al. 2005).

We again refer to the results of the Pipe Nebula (Rathborne et al. 2008) and the Perseus molecular cloud (Rosolowsky et al. 2008). It can be found that there is no carbon-chain-rich core in the Perseus region in spite of the completeness of their survey, as shown in Figure 6. On the other hand, we find several candidates for CCPRs in the Pipe nebula (Rathborne et al. 2008), where the NH₃/CCS ratios are less than 10. In the Pipe nebula, there exists a dense core (core 37 of Rathborne et al. 2008) where the CCS line is detected while NH₃ is not. Because of the lack of the NH₃ data, it is not plotted in Figure 6(b). The column density of CCS toward this core is derived to be 2.0 × 10¹² cm⁻² (Rathborne et al. 2008). Although this value is much lower than those of the known CCPRs, it would be a possible candidate of a CCPR. This result would suggest that the cores in the Pipe nebula are generally in the early stages of both chemical and dynamical evolution (Rathborne et al. 2008).

Zoom In Zoom Out Reset image size

Although the above discussions are not based on unbiased, complete, and uniform surveys, the apparent region-to-region variation of the chemical composition would be indicative of difference in chemical and/or physical properties of each region. In fact, chemical and dynamical model calculations demonstrate that molecular abundances and distributions could reflect the physical properties and their initial conditions of the cores such as the density, degree of depletion, and velocity structure of the cores (Aikawa et al. 2001, 2003, 2005). Observational evidences for the variations of physical properties are reported by Jijina et al. (1999) based on the database of a survey of the NH₃ lines. Such a variation of the core properties would affect not only chemistry but also star formation processes, which results in either isolated or clustered mode, and low-mass or high-mass star formation.

In the above sections, we recognize significant source-to-source chemical abundance variation, in particular between the Taurus and Ophiuchus regions. Here we discuss two possible scenarios which can explain qualitatively the systematic difference in the molecular abundances between the Taurus cores and Ophiuchus ones.

4.4.1. Difference in Ages of the Cores

4.4.2. Difference in Contraction Timescales of the Cores

Finally, we comment on Warm Carbon-Chain Chemistry (WCCC) recently proposed by Sakai et al. (2008a). The WCCC is first identified in L1527 and is characterized by high abundances of long carbon-chain molecules such as C_nH (n = 4, 5, 6) and C_nH₂ (n = 3, 4, 6), and intense spectra of higher excitation lines of carbon-chain molecules. According to these criteria, L483 is considered to be a possible candidate of the WCCC source. When it is compared with L1527, both are associated with the IRAS point source, and the NH₃/CCS ratios are relatively high due to high NH₃ column density. The abundances of HC₃N are remarkably high and are comparable to those of CCPRs such as L1495B, L1521B, L1521E, the cyanopolyyne peak of TMC-1, and L492. Furthermore, we stress that L483 and L1527 show stronger emission of longer carbon-chain molecule HC₅N than those in CCPRs. Except for relatively high abundance of CCS, chemistry in L483 seems to be quite similar to that in L1527. Observational studies of WCCC have just been started, and hence, the reason for such similarities and differences will be investigated in the near future (Sakai et al. 2008a, 2009). Interestingly, the regions where the WCCC sources are found, Taurus (L1527) and Aquila (L483), include CCPRs such as L1495B, L1521B, L1521E, the cyanopolyyne peak of TMC-1, and L492. On the other hand, no WCCC sources are found currently in the Ophiuchus and Perseus regions (Sakai et al. 2008a, 2009). Although the discussion is based on a limited number of samples at this moment, the WCCC might have a close connection to the CCPRs. Therefore, chemistry of carbon-chain molecules would be indicative of difference in the physical and chemical properties among molecular cloud complexes.

This core, taken from Benson & Myers (1989), is a starless core located in the Taurus molecular cloud at a distance of 137 pc (Torres et al. 2007). We first detected the CCS and HC₃N lines at the (+80'', −80'') position in right ascension and declination, respectively, with respect to the reference position of L1495C. Then we extended our maps to about 4' southeast from the reference position, and found that the peak positions of CCS and HC₃N are located within the NH₃ core of L1495N observed by Benson & Myers (1989). Its position offset is (−60'', + 42'') with respect to the reference position of L1495N. Therefore, the CCS and HC₃N cores detected in the present study should be referred to L1495N rather than L1495C. The integrated intensity maps for CCS and HC₃N are shown in Figure 4(a). The IRAS source and the NH₃ core are located at the southeast edge of our map. Because the CCS and HC₃N peaks correspond to the northwestern edge of the NH₃ core of L1495N (Benson & Myers 1989), the carbon-chain molecules seem to be depleted at the NH₃ peak of L1495N.

This core is one of the starless cores listed in Benson & Myers (1989), whose distance is 170 pc (Hilton & Lahulla 1995). We detected the CCS and HC₃N lines toward the (−80'', + 40'') northwest from the reference position. The integrated intensity maps for CCS and HC₃N are shown in Figure 4(b). The peak position of CCS may be shifted from that of HC₃N, although it is unclear due to the low signal-to-noise ratio and the coarse sampling. We also detected the NH₃ lines toward the (−80'', +40'') position. Because the NH₃ lines were not detected by Benson & Myers (1989), this is the first detection of the NH₃ lines toward L1400B. The line parameters are listed in Table 4.

A dense core associated with L1527 in the Taurus molecular cloud is a host of Class 0 protostar IRAS 04368+2557, which is listed in Benson & Myers (1989). The CCS line was detected in previous observations (Hirota et al. 2001). Recently, Sakai et al. (2008a) detected the lines of various carbon-chain molecules such as C_nH (n = 4, 5, 6) and C_nH₂ (n = 3, 4, 6) in this source. Because of their high excitation temperature (>10 K), Sakai et al. (2008a) proposed a different type of carbon-chain chemistry other than that in the cold dark cloud cores, and called it as WCCC. The column density of HC₅N listed in Table 5 is 3.5 times larger than that of Sakai et al. (2008a) due to a difference in the assumed excitation temperature (6.5 K); Sakai et al. (2008a) employed a higher excitation temperature of 12.3 K by assuming that the line mainly comes from a dense and warm part near the protostar. Since we observed a single transition for each molecule only toward the IRAS position, it is not possible from our data to discuss molecular distribution and their excitation condition in detail.

L1517D is a starless dense core in the Taurus molecular cloud (Benson & Myers 1989). We searched for the molecular lines around the reference position. The CCS line was detected only toward the (+80'', +80'') position, while the HC₃N and HC₅N lines were not detected toward all the observed positions, as shown in Figure 3(a). The NH₃ line was detected neither toward the reference position (Benson & Myers 1989) nor the CCS position above, suggesting that L1517D would be one of the new candidates for CCPRs with an extremely low NH₃/CCS ratio. On the other hand, it would be possible that the weak CCS emission and the low NH₃/CCS ratio are due to the offset from the core center, as in the case of L1498 and L1544 where the CCS lines tend to be intense at the outer part of the dense cores (Kuiper et al. 1996; Ohashi et al. 1999). In L1517D, the position where the CCS line was detected corresponds to the edge of the Spitzer 160 μm continuum emission image with the 40'' resolution (Kirk et al. 2007), although the dense core is not detected in the higher resolution observations with SCUBA. Further high-resolution mapping observations of both dust continuum emission and molecular lines are required to confirm whether L1517D is one of the CCPRs.

L1512 is one of the well-studied starless dense cores in the Taurus molecular cloud (Benson & Myers 1989). Since the CCS, HC₃N, and HC₅N lines were detected in the previous observations (Suzuki et al. 1992), we carried out mapping observations of the CCS and HC₃N lines in the present study. The results are shown in Figure 4(c). The derived column densities of CCS and HC₃N are larger than those reported by Suzuki et al. (1992) because the peak positions of CCS and HC₃N are significantly shifted from the reference position (Suzuki et al. 1992). The position of the dust continuum emission is plotted in Figure 4(c) (Kirk et al. 2005). We found that the integrated intensity maps of CCS and HC₃N do not show a central hole, which is consistent with the results by Lee et al. (2003). However, their peak positions are shifted from the dust continuum peak by (−24'', −17'') in right ascension and declination, respectively. In particular, the CCS emission traces the southwest edge of the dust continuum emission, possibly suggesting the depletion of CCS at the dust continuum peak.

L183 is a well-studied dark cloud, which is also known as L134N, and L183(N) is one of the ammonia cores in the filamentary dense molecular cloud L183 (Ungerechts et al. 1980; Benson & Myers 1989). The estimated distance of L183 is about 140 pc, although some other results are also reported (Hilton & Lahulla 1995). The CCS lines were observed previously by Benson et al. (1998). We observed only toward one position, which corresponds to the position for the DNC and HN¹³C observations (Hirota et al. 2001). The column density of CCS derived for this position agrees well with the previous result (Benson et al. 1998).

B68 is a dense starless core located in the Pipe Nebula at an estimated distance of 130-200 pc (Hilton & Lahulla 1995; Rathborne et al. 2008). We detected the CCS, HC₃N, and HC₅N lines toward the reference position, although we could not carry out the mapping observations due to the limited observing time. The CCS and NH₃ maps were presented by Lai et al. (2003). In our study, the column density is derived toward the NH₃ peak, which is offset by 45'' from the CCS peak (Lai et al. 2003), so that the derived column density of CCS is different from that of Lai et al. (2003), by a factor of 2–3.

L492 is a Bok globule located in the Aquila rift at a distance of 200 pc (Lee & Myers 1999). This core was not listed in the NH₃ survey by Benson & Myers (1989), while we included it in our survey because an intense CS line was detected toward this source (Lee et al. 2001). In the initial survey observations conducted in 2002, we found that the CCS, HC₃N, and HC₅N lines toward the reference position of L492 are extremely strong. Therefore, we carried out follow-up mapping observations of L492 in various molecular lines. Further details were reported in Hirota & Yamamoto (2006).

L429-1 is also a Bok globule located in the Aquila rift at a distance of 200 pc (Lee & Myers 1999). Similar to L492, this core was not included in the NH₃ survey by Benson & Myers (1989), while we observed it based on the results of the CS survey by Lee et al. (2001). Because intense CCS and HC₃N lines were detected toward the reference position of L429-1, we carried out mapping observations of the CCS and HC₃N lines. The results are shown in Figure 4(d). The overall structures seen in the integrated intensity maps of CCS and HC₃N are similar, while the HC₃N map shows relatively complex clumpy structure. There seem to be three cores in our map. According to the dust continuum observations (Crapsi et al. 2005), the CCS and HC₃N distributions are significantly shifted from the continuum peak position, although the mapped region of the dust continuum emission (Crapsi et al. 2005) is limited to the northwestern peak of the CCS and HC₃N cores. For much wider area, the mid-infrared absorption map is available. It shows an extended high-extinction region toward east of the dust continuum peak (Bacmann et al. 2000), which agrees well with the HC₃N and CCS maps. Another peak of HC₃N and the weak emission of the CCS lines can be seen toward 150'' east of the dust continuum peak. We observed the NH₃ line only toward the reference position of L429-1, and hence, the NH₃/CCS ratio at the eastern peak of HC₃N, is not calculated. In spite of the intense HC₃N emission, the HC₅N line was not detected in L429-1.

L483 is a star-forming dense core in the Aquila rift at a distance of 200 pc (Hilton & Lahulla 1995). An intense CCS line was detected by Benson et al. (1998). We derived the column density which is consistent with that of Benson et al. (1998). We also detected intense HC₃N and HC₅N lines toward the reference position of L483 as shown in Figure 2. Then, we carried out mapping observations in the HC₃N and CCS lines. The results are shown in Figure 4(e). According to the previous observations of molecular lines and submillimeter continuum emission (Fuller et al. 1995), L483 harbors a Class 0 protostar IRAS 18148−0440 associated with molecular outflows. The CCS and HC₅N lines are quite strong toward the central protostar, although we cannot completely rule out the possibility that these centrally peaked structure would be an artifact due to insufficient spatial resolution, as seen in Velusamy et al. (1995). Extremely high abundances of carbon-chain molecules toward the Class 0 protostar are apparently similar to the L1527 core (Sakai et al. 2008a), and are indicative of WCCC (Sakai et al. 2009). However, higher abundance of CCS than that in L1527 seems to be rather common characteristics for CCPRs such as L1495B, L1521B, L1521E, and the cyanopolyyne peak of TMC-1 (Hirota et al. 2002, 2004). Observations of higher excitation lines of carbon-chain molecules in L483 would be crucial to distinguish whether L483 is categorized as WCCC or CCPR.

L530D is a dense core which is included in the survey of NH₃ by Benson & Myers (1989). The estimated distance is 350 pc (Hilton & Lahulla 1995). We searched for the lines around the reference position of L530D, and found that the peak position of the CCS line is located at (+80'', +40''). However, the HC₃N and HC₅N lines were not detected toward this position, as shown in Figures 3(b). This position is offset by 85'' northeast of the dust continuum peak (Kirk et al. 2005). On the other hand, the NH₃ line was detected toward the dust continuum peak (Benson & Myers 1989; Kirk et al. 2005), while it was not detected toward the CCS peak in our observations. Although the NH₃/CCS ratio at the CCS peak in L530D is as low as those of the other CCPRs, lower column densities of CCS and other carbon-chain molecules, the offsets of the CCS peak from the dust continuum emission, and the detection of NH₃ at the dust continuum peak would suggest that the CCS line traces the outer part of this core, similar to the case of L1517D. In fact, the column density of NH₃ toward the dust continuum peak is derived to be 7.9×10¹³ cm⁻² from the data of Benson & Myers (1989), which results in the NH₃/CCS ratio toward the dust continuum peak 10 times larger than that toward the CCS peak. Molecular distributions in L530D seem to be similar to those in L1498 and L1544, where the NH₃ peak is coincident with the dust continuum and CCS is distributed in the surrounding region (Kuiper et al. 1996; Ohashi et al. 1999). However, we cannot totally rule out the possibility that L530D is a candidate for the CCPR because of a lack of complete mapping observations in the CCS and NH₃ lines.

L1147 is a dense core located in the Cepheus region at a distance of 325 pc (Hilton & Lahulla 1995). This source is included in the survey of the NH₃ lines (Benson & Myers 1989). We searched for the CCS, HC₃N, and HC₅N lines around the reference position of L1147, and detected the CCS and HC₃N lines only toward the (0'', −80'') position, whose spectra are shown in Figures 3(c). The HC₅N line was not detected toward all the observed position. The NH₃ line was detected toward both the reference position (Benson & Myers 1989) and the CCS peak. The position of the CCS peak is offset by only 20'' toward south of the dust continuum peak of L1148, which is 1' south of L1147 (Myers et al. 1983; Kirk et al. 2005), and hence, the dense core of L1147 found in the present survey can be regarded as L1148. The relatively low NH₃/CCS ratio toward the dust continuum peak of L1147 means that it could be a candidate for the CCPR.

L1172D was originally identified as a starless dense core in the survey of the NH₃ lines (Benson & Myers 1989), which is located in the Cepheus region at a distance of 288 pc (Hilton & Lahulla 1995). At first, the CCS and HC₃N lines were found to be intense toward the eastern part of the core of L1172D, and hence, the mapping region was extended toward 1'–2' east from the reference position. Then the CCS and HC₃N emission peaks were found to be located toward L1172A, another star-forming dense core identified by Benson & Myers (1989). There is a Class I protostar, IRAS 21017+6742, at the center of the core, which is known to be an engine of the molecular outflow (Myers et al. 1988). The integrated intensity maps for HC₃N and CCS are shown in Figure 4(f). The HC₃N peak corresponds to the IRAS position, while CCS is significantly depleted toward the center of the HC₃N core. The shape and size of the HC₃N core is quite similar to that of NH₃ (Benson & Myers 1989). These results suggest that CCS is depleted as in the case of B335 (Velusamy et al. 1995), while HC₃N is abundant toward the central part of the core.

L1172B is also identified as a starless dense core in the Cepheus region, which is observed in the survey of the NH₃ lines (Benson & Myers 1989). The observed spectra toward three positions in L1172B and the integrated intensity maps for HC₃N and CCS are shown in Figures 3(d)–(f) and 4(g), respectively. A dense core traced by the CCS and HC₃N lines shows an elongated structure extending from northeast to southwest, and consists of at least two clumps. The integrated intensity map of the NH₃ lines is also shown in Figure 4(g) in thick contour lines. The NH₃ lines are detected only toward the western edge of the elongated molecular cloud core traced by the CCS and HC₃N lines. The NH₃/CCS abundance ratios are relatively low in the northeast clump (+40'', +40''), and hence, this clump can be a candidate of CCPR. An IRAS point source, IRAS 21025+6801, is located at northwest of the core, although it is not clear whether this source is physically associated with the core of L1172B.

The results of the CCS observations have already been reported for these sources (Hirota et al. 2001). We re-analyzed the HC₃N data for these sources, while we did not make further observations of the HC₅N lines.

For the statistical study with uniform samples of nearby dark cloud cores, we excluded results of other surveys which were conducted in the different frequency bands and/or with interferometers (Scappini & Codella 1996; Lai & Crutcher 2000; de Gregorio-Monsalvo et al. 2006; Rathborne et al. 2008; Rosolowsky et al. 2008), while we referred to the results by Rathborne et al. (2008) and Rosolowsky et al. (2008) for comparison of source-to-source variation of molecular abundances. We only employed dark cloud cores and low-mass star-forming regions in our sample as uniformly as possible. As a result, some of the sources in the high-mass star-forming molecular clouds in the Orion region (e.g., Tatematsu et al. 2008) were not included in our samples. We also excluded the different kind of samples such as translucent clouds (Turner et al. 1998) and distant infrared dark clouds (Sakai et al. 2008b). For a similar reason, we did not include the results of individual sources (e.g., Velusamy et al. 1995; Kuiper et al. 1996; Ohashi et al. 1999; Lai et al. 2003; Lee et al. 2003), although these sources are included in the part of the survey observations. For the sources which are included in more than two literatures, we adopted one of the results as indicated in Table A.1 to avoid overlap of the samples. As a result, our sample consists of 90 nearby dark cloud cores as listed in Table A.1.

We compared the results of several sources which are observed both in our observations and previous ones, such as L1527, L1512, L1523, L183(N), L1719B, B68, L483, and L1172A (L1172D) (Suzuki et al. 1992; Benson et al. 1998; Hirota et al. 2001; Lai et al. 2003). As a result, we found that the difference in the derived column density is typically within 40% but is a factor of 3 in the worst case, which is mostly due to the difference in the observed position (e.g., in the case of L1512; Suzuki et al. 1992; Benson et al. 1998). These uncertainties are comparable to those attributed to the unknown excitation temperatures as already discussed in the main text.

For the NH₃ lines, substantial amount of the sources listed in Table A.1 were observed by Benson & Myers (1989). However, the column densities of NH₃ were reported for a part of the sources where the intense spectra were detected. Therefore, we calculated the column densities of NH₃ lines using the line parameters reported by Benson & Myers (1989). The method is described by Suzuki et al. (1992), in which we assumed the excitation temperature of 6.5 K. If the NH₃ lines were not detected, we calculated the upper limits for the column densities by assuming the line width of 0.5 km s⁻¹.

The data for the NH₃ lines are mostly taken from the systematic survey observations by Benson & Myers (1989) and Suzuki et al. (1992), while some individual observations are included (Ungerechts et al. 1980; Mundy et al. 1990; Bachiller et al. 1993; Ladd et al. 1994; Anglada et al. 1997; Hirota et al. 2002, 2004). Therefore, the uncertainties in the column densities of NH₃ might be affected by the difference in the observations and data analysis such as the telescope beam sizes and assumed and/or derived kinetic temperatures. However, we estimate that the uncertainties in the column densities of NH₃ is within a factor of 2 for most of the sources as discussed in Suzuki et al. (1992).